.\" Man page generated from reStructuredText.
.
.TH "LIBUV" "3" "May 24, 2020" "1.38.0" "libuv API documentation"
.SH NAME
libuv \- libuv documentation
.
.nr rst2man-indent-level 0
.
.de1 rstReportMargin
\\$1 \\n[an-margin]
level \\n[rst2man-indent-level]
level margin: \\n[rst2man-indent\\n[rst2man-indent-level]]
-
\\n[rst2man-indent0]
\\n[rst2man-indent1]
\\n[rst2man-indent2]
..
.de1 INDENT
.\" .rstReportMargin pre:
. RS \\$1
. nr rst2man-indent\\n[rst2man-indent-level] \\n[an-margin]
. nr rst2man-indent-level +1
.\" .rstReportMargin post:
..
.de UNINDENT
. RE
.\" indent \\n[an-margin]
.\" old: \\n[rst2man-indent\\n[rst2man-indent-level]]
.nr rst2man-indent-level -1
.\" new: \\n[rst2man-indent\\n[rst2man-indent-level]]
.in \\n[rst2man-indent\\n[rst2man-indent-level]]u
..
.SH OVERVIEW
.sp
libuv is a multi\-platform support library with a focus on asynchronous I/O. It
was primarily developed for use by \fI\%Node.js\fP, but it\(aqs also used by \fI\%Luvit\fP,
\fI\%Julia\fP, \fI\%pyuv\fP, and \fI\%others\fP\&.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
In case you find errors in this documentation you can help by sending
\fI\%pull requests\fP!
.UNINDENT
.UNINDENT
.SH FEATURES
.INDENT 0.0
.IP \(bu 2
Full\-featured event loop backed by epoll, kqueue, IOCP, event ports.
.IP \(bu 2
Asynchronous TCP and UDP sockets
.IP \(bu 2
Asynchronous DNS resolution
.IP \(bu 2
Asynchronous file and file system operations
.IP \(bu 2
File system events
.IP \(bu 2
ANSI escape code controlled TTY
.IP \(bu 2
IPC with socket sharing, using Unix domain sockets or named pipes (Windows)
.IP \(bu 2
Child processes
.IP \(bu 2
Thread pool
.IP \(bu 2
Signal handling
.IP \(bu 2
High resolution clock
.IP \(bu 2
Threading and synchronization primitives
.UNINDENT
.SH DOCUMENTATION
.SS Design overview
.sp
libuv is cross\-platform support library which was originally written for \fI\%Node.js\fP\&. It\(aqs designed
around the event\-driven asynchronous I/O model.
.sp
The library provides much more than a simple abstraction over different I/O polling mechanisms:
\(aqhandles\(aq and \(aqstreams\(aq provide a high level abstraction for sockets and other entities;
cross\-platform file I/O and threading functionality is also provided, amongst other things.
.sp
Here is a diagram illustrating the different parts that compose libuv and what subsystem they
relate to:
[image]
.SS Handles and requests
.sp
libuv provides users with 2 abstractions to work with, in combination with the event loop:
handles and requests.
.sp
Handles represent long\-lived objects capable of performing certain operations while active. Some examples:
.INDENT 0.0
.IP \(bu 2
A prepare handle gets its callback called once every loop iteration when active.
.IP \(bu 2
A TCP server handle that gets its connection callback called every time there is a new connection.
.UNINDENT
.sp
Requests represent (typically) short\-lived operations. These operations can be performed over a
handle: write requests are used to write data on a handle; or standalone: getaddrinfo requests
don\(aqt need a handle they run directly on the loop.
.SS The I/O loop
.sp
The I/O (or event) loop is the central part of libuv. It establishes the content for all I/O
operations, and it\(aqs meant to be tied to a single thread. One can run multiple event loops
as long as each runs in a different thread. The libuv event loop (or any other API involving
the loop or handles, for that matter) \fBis not thread\-safe\fP except where stated otherwise.
.sp
The event loop follows the rather usual single threaded asynchronous I/O approach: all (network)
I/O is performed on non\-blocking sockets which are polled using the best mechanism available
on the given platform: epoll on Linux, kqueue on OSX and other BSDs, event ports on SunOS and IOCP
on Windows. As part of a loop iteration the loop will block waiting for I/O activity on sockets
which have been added to the poller and callbacks will be fired indicating socket conditions
(readable, writable hangup) so handles can read, write or perform the desired I/O operation.
.sp
In order to better understand how the event loop operates, the following diagram illustrates all
stages of a loop iteration:
[image]
.INDENT 0.0
.IP 1. 4
The loop concept of \(aqnow\(aq is updated. The event loop caches the current time at the start of
the event loop tick in order to reduce the number of time\-related system calls.
.IP 2. 4
If the loop is \fIalive\fP an iteration is started, otherwise the loop will exit immediately. So,
when is a loop considered to be \fIalive\fP? If a loop has active and ref\(aqd handles, active
requests or closing handles it\(aqs considered to be \fIalive\fP\&.
.IP 3. 4
Due timers are run. All active timers scheduled for a time before the loop\(aqs concept of \fInow\fP
get their callbacks called.
.IP 4. 4
Pending callbacks are called. All I/O callbacks are called right after polling for I/O, for the
most part. There are cases, however, in which calling such a callback is deferred for the next
loop iteration. If the previous iteration deferred any I/O callback it will be run at this point.
.IP 5. 4
Idle handle callbacks are called. Despite the unfortunate name, idle handles are run on every
loop iteration, if they are active.
.IP 6. 4
Prepare handle callbacks are called. Prepare handles get their callbacks called right before
the loop will block for I/O.
.IP 7. 4
Poll timeout is calculated. Before blocking for I/O the loop calculates for how long it should
block. These are the rules when calculating the timeout:
.INDENT 4.0
.INDENT 3.5
.INDENT 0.0
.IP \(bu 2
If the loop was run with the \fBUV_RUN_NOWAIT\fP flag, the timeout is 0.
.IP \(bu 2
If the loop is going to be stopped (\fBuv_stop()\fP was called), the timeout is 0.
.IP \(bu 2
If there are no active handles or requests, the timeout is 0.
.IP \(bu 2
If there are any idle handles active, the timeout is 0.
.IP \(bu 2
If there are any handles pending to be closed, the timeout is 0.
.IP \(bu 2
If none of the above cases matches, the timeout of the closest timer is taken, or
if there are no active timers, infinity.
.UNINDENT
.UNINDENT
.UNINDENT
.IP 8. 4
The loop blocks for I/O. At this point the loop will block for I/O for the duration calculated
in the previous step. All I/O related handles that were monitoring a given file descriptor
for a read or write operation get their callbacks called at this point.
.IP 9. 4
Check handle callbacks are called. Check handles get their callbacks called right after the
loop has blocked for I/O. Check handles are essentially the counterpart of prepare handles.
.IP 10. 4
Close callbacks are called. If a handle was closed by calling \fBuv_close()\fP it will
get the close callback called.
.IP 11. 4
Special case in case the loop was run with \fBUV_RUN_ONCE\fP, as it implies forward progress.
It\(aqs possible that no I/O callbacks were fired after blocking for I/O, but some time has passed
so there might be timers which are due, those timers get their callbacks called.
.IP 12. 4
Iteration ends. If the loop was run with \fBUV_RUN_NOWAIT\fP or \fBUV_RUN_ONCE\fP modes the
iteration ends and \fBuv_run()\fP will return. If the loop was run with \fBUV_RUN_DEFAULT\fP
it will continue from the start if it\(aqs still \fIalive\fP, otherwise it will also end.
.UNINDENT
.sp
\fBIMPORTANT:\fP
.INDENT 0.0
.INDENT 3.5
libuv uses a thread pool to make asynchronous file I/O operations possible, but
network I/O is \fBalways\fP performed in a single thread, each loop\(aqs thread.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
While the polling mechanism is different, libuv makes the execution model consistent
across Unix systems and Windows.
.UNINDENT
.UNINDENT
.SS File I/O
.sp
Unlike network I/O, there are no platform\-specific file I/O primitives libuv could rely on,
so the current approach is to run blocking file I/O operations in a thread pool.
.sp
For a thorough explanation of the cross\-platform file I/O landscape, checkout
\fI\%this post\fP\&.
.sp
libuv currently uses a global thread pool on which all loops can queue work. 3 types of
operations are currently run on this pool:
.INDENT 0.0
.INDENT 3.5
.INDENT 0.0
.IP \(bu 2
File system operations
.IP \(bu 2
DNS functions (getaddrinfo and getnameinfo)
.IP \(bu 2
User specified code via \fBuv_queue_work()\fP
.UNINDENT
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
See the threadpool section for more details, but keep in mind the thread pool size
is quite limited.
.UNINDENT
.UNINDENT
.SS API documentation
.SS Error handling
.sp
In libuv errors are negative numbered constants. As a rule of thumb, whenever
there is a status parameter, or an API functions returns an integer, a negative
number will imply an error.
.sp
When a function which takes a callback returns an error, the callback will never
be called.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
Implementation detail: on Unix error codes are the negated \fIerrno\fP (or \fI\-errno\fP), while on
Windows they are defined by libuv to arbitrary negative numbers.
.UNINDENT
.UNINDENT
.SS Error constants
.INDENT 0.0
.TP
.B UV_E2BIG
argument list too long
.UNINDENT
.INDENT 0.0
.TP
.B UV_EACCES
permission denied
.UNINDENT
.INDENT 0.0
.TP
.B UV_EADDRINUSE
address already in use
.UNINDENT
.INDENT 0.0
.TP
.B UV_EADDRNOTAVAIL
address not available
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAFNOSUPPORT
address family not supported
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAGAIN
resource temporarily unavailable
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_ADDRFAMILY
address family not supported
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_AGAIN
temporary failure
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_BADFLAGS
bad ai_flags value
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_BADHINTS
invalid value for hints
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_CANCELED
request canceled
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_FAIL
permanent failure
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_FAMILY
ai_family not supported
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_MEMORY
out of memory
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_NODATA
no address
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_NONAME
unknown node or service
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_OVERFLOW
argument buffer overflow
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_PROTOCOL
resolved protocol is unknown
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_SERVICE
service not available for socket type
.UNINDENT
.INDENT 0.0
.TP
.B UV_EAI_SOCKTYPE
socket type not supported
.UNINDENT
.INDENT 0.0
.TP
.B UV_EALREADY
connection already in progress
.UNINDENT
.INDENT 0.0
.TP
.B UV_EBADF
bad file descriptor
.UNINDENT
.INDENT 0.0
.TP
.B UV_EBUSY
resource busy or locked
.UNINDENT
.INDENT 0.0
.TP
.B UV_ECANCELED
operation canceled
.UNINDENT
.INDENT 0.0
.TP
.B UV_ECHARSET
invalid Unicode character
.UNINDENT
.INDENT 0.0
.TP
.B UV_ECONNABORTED
software caused connection abort
.UNINDENT
.INDENT 0.0
.TP
.B UV_ECONNREFUSED
connection refused
.UNINDENT
.INDENT 0.0
.TP
.B UV_ECONNRESET
connection reset by peer
.UNINDENT
.INDENT 0.0
.TP
.B UV_EDESTADDRREQ
destination address required
.UNINDENT
.INDENT 0.0
.TP
.B UV_EEXIST
file already exists
.UNINDENT
.INDENT 0.0
.TP
.B UV_EFAULT
bad address in system call argument
.UNINDENT
.INDENT 0.0
.TP
.B UV_EFBIG
file too large
.UNINDENT
.INDENT 0.0
.TP
.B UV_EHOSTUNREACH
host is unreachable
.UNINDENT
.INDENT 0.0
.TP
.B UV_EINTR
interrupted system call
.UNINDENT
.INDENT 0.0
.TP
.B UV_EINVAL
invalid argument
.UNINDENT
.INDENT 0.0
.TP
.B UV_EIO
i/o error
.UNINDENT
.INDENT 0.0
.TP
.B UV_EISCONN
socket is already connected
.UNINDENT
.INDENT 0.0
.TP
.B UV_EISDIR
illegal operation on a directory
.UNINDENT
.INDENT 0.0
.TP
.B UV_ELOOP
too many symbolic links encountered
.UNINDENT
.INDENT 0.0
.TP
.B UV_EMFILE
too many open files
.UNINDENT
.INDENT 0.0
.TP
.B UV_EMSGSIZE
message too long
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENAMETOOLONG
name too long
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENETDOWN
network is down
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENETUNREACH
network is unreachable
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENFILE
file table overflow
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOBUFS
no buffer space available
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENODEV
no such device
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOENT
no such file or directory
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOMEM
not enough memory
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENONET
machine is not on the network
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOPROTOOPT
protocol not available
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOSPC
no space left on device
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOSYS
function not implemented
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOTCONN
socket is not connected
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOTDIR
not a directory
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOTEMPTY
directory not empty
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOTSOCK
socket operation on non\-socket
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENOTSUP
operation not supported on socket
.UNINDENT
.INDENT 0.0
.TP
.B UV_EPERM
operation not permitted
.UNINDENT
.INDENT 0.0
.TP
.B UV_EPIPE
broken pipe
.UNINDENT
.INDENT 0.0
.TP
.B UV_EPROTO
protocol error
.UNINDENT
.INDENT 0.0
.TP
.B UV_EPROTONOSUPPORT
protocol not supported
.UNINDENT
.INDENT 0.0
.TP
.B UV_EPROTOTYPE
protocol wrong type for socket
.UNINDENT
.INDENT 0.0
.TP
.B UV_ERANGE
result too large
.UNINDENT
.INDENT 0.0
.TP
.B UV_EROFS
read\-only file system
.UNINDENT
.INDENT 0.0
.TP
.B UV_ESHUTDOWN
cannot send after transport endpoint shutdown
.UNINDENT
.INDENT 0.0
.TP
.B UV_ESPIPE
invalid seek
.UNINDENT
.INDENT 0.0
.TP
.B UV_ESRCH
no such process
.UNINDENT
.INDENT 0.0
.TP
.B UV_ETIMEDOUT
connection timed out
.UNINDENT
.INDENT 0.0
.TP
.B UV_ETXTBSY
text file is busy
.UNINDENT
.INDENT 0.0
.TP
.B UV_EXDEV
cross\-device link not permitted
.UNINDENT
.INDENT 0.0
.TP
.B UV_UNKNOWN
unknown error
.UNINDENT
.INDENT 0.0
.TP
.B UV_EOF
end of file
.UNINDENT
.INDENT 0.0
.TP
.B UV_ENXIO
no such device or address
.UNINDENT
.INDENT 0.0
.TP
.B UV_EMLINK
too many links
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B UV_ERRNO_MAP(iter_macro)
Macro that expands to a series of invocations of \fIiter_macro\fP for
each of the error constants above. \fIiter_macro\fP is invoked with two
arguments: the name of the error constant without the \fIUV_\fP prefix,
and the error message string literal.
.UNINDENT
.INDENT 0.0
.TP
.B const char* uv_strerror(int\fI\ err\fP)
Returns the error message for the given error code. Leaks a few bytes
of memory when you call it with an unknown error code.
.UNINDENT
.INDENT 0.0
.TP
.B char* uv_strerror_r(int\fI\ err\fP, char*\fI\ buf\fP, size_t\fI\ buflen\fP)
Returns the error message for the given error code. The zero\-terminated
message is stored in the user\-supplied buffer \fIbuf\fP of at most \fIbuflen\fP bytes.
.sp
New in version 1.22.0.
.UNINDENT
.INDENT 0.0
.TP
.B const char* uv_err_name(int\fI\ err\fP)
Returns the error name for the given error code. Leaks a few bytes
of memory when you call it with an unknown error code.
.UNINDENT
.INDENT 0.0
.TP
.B char* uv_err_name_r(int\fI\ err\fP, char*\fI\ buf\fP, size_t\fI\ buflen\fP)
Returns the error name for the given error code. The zero\-terminated
name is stored in the user\-supplied buffer \fIbuf\fP of at most \fIbuflen\fP bytes.
.sp
New in version 1.22.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_translate_sys_error(int\fI\ sys_errno\fP)
Returns the libuv error code equivalent to the given platform dependent error
code: POSIX error codes on Unix (the ones stored in \fIerrno\fP), and Win32 error
codes on Windows (those returned by \fIGetLastError()\fP or \fIWSAGetLastError()\fP).
.sp
If \fIsys_errno\fP is already a libuv error, it is simply returned.
.sp
Changed in version 1.10.0: function declared public.
.UNINDENT
.SS Version\-checking macros and functions
.sp
Starting with version 1.0.0 libuv follows the \fI\%semantic versioning\fP
scheme. This means that new APIs can be introduced throughout the lifetime of
a major release. In this section you\(aqll find all macros and functions that
will allow you to write or compile code conditionally, in order to work with
multiple libuv versions.
.SS Macros
.INDENT 0.0
.TP
.B UV_VERSION_MAJOR
libuv version\(aqs major number.
.UNINDENT
.INDENT 0.0
.TP
.B UV_VERSION_MINOR
libuv version\(aqs minor number.
.UNINDENT
.INDENT 0.0
.TP
.B UV_VERSION_PATCH
libuv version\(aqs patch number.
.UNINDENT
.INDENT 0.0
.TP
.B UV_VERSION_IS_RELEASE
Set to 1 to indicate a release version of libuv, 0 for a development
snapshot.
.UNINDENT
.INDENT 0.0
.TP
.B UV_VERSION_SUFFIX
libuv version suffix. Certain development releases such as Release Candidates
might have a suffix such as "rc".
.UNINDENT
.INDENT 0.0
.TP
.B UV_VERSION_HEX
Returns the libuv version packed into a single integer. 8 bits are used for
each component, with the patch number stored in the 8 least significant
bits. E.g. for libuv 1.2.3 this would be 0x010203.
.sp
New in version 1.7.0.
.UNINDENT
.SS Functions
.INDENT 0.0
.TP
.B unsigned int uv_version(void)
Returns \fI\%UV_VERSION_HEX\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B const char* uv_version_string(void)
Returns the libuv version number as a string. For non\-release versions the
version suffix is included.
.UNINDENT
.SS \fI\%uv_loop_t\fP \-\-\- Event loop
.sp
The event loop is the central part of libuv\(aqs functionality. It takes care
of polling for i/o and scheduling callbacks to be run based on different sources
of events.
.SS Data types
.INDENT 0.0
.TP
.B uv_loop_t
Loop data type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_run_mode
Mode used to run the loop with \fI\%uv_run()\fP\&.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
UV_RUN_DEFAULT = 0,
UV_RUN_ONCE,
UV_RUN_NOWAIT
} uv_run_mode;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_walk_cb)(uv_handle_t*\fI\ handle\fP, void*\fI\ arg\fP)
Type definition for callback passed to \fI\%uv_walk()\fP\&.
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B void* uv_loop_t.data
Space for user\-defined arbitrary data. libuv does not use and does not
touch this field.
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_loop_init(uv_loop_t*\fI\ loop\fP)
Initializes the given \fIuv_loop_t\fP structure.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_loop_configure(uv_loop_t*\fI\ loop\fP, uv_loop_option\fI\ option\fP, \&...)
New in version 1.0.2.
.sp
Set additional loop options. You should normally call this before the
first call to \fI\%uv_run()\fP unless mentioned otherwise.
.sp
Returns 0 on success or a UV_E* error code on failure. Be prepared to
handle UV_ENOSYS; it means the loop option is not supported by the platform.
.sp
Supported options:
.INDENT 7.0
.IP \(bu 2
UV_LOOP_BLOCK_SIGNAL: Block a signal when polling for new events. The
second argument to \fI\%uv_loop_configure()\fP is the signal number.
.sp
This operation is currently only implemented for SIGPROF signals,
to suppress unnecessary wakeups when using a sampling profiler.
Requesting other signals will fail with UV_EINVAL.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_loop_close(uv_loop_t*\fI\ loop\fP)
Releases all internal loop resources. Call this function only when the loop
has finished executing and all open handles and requests have been closed,
or it will return UV_EBUSY. After this function returns, the user can free
the memory allocated for the loop.
.UNINDENT
.INDENT 0.0
.TP
.B uv_loop_t* uv_default_loop(void)
Returns the initialized default loop. It may return NULL in case of
allocation failure.
.sp
This function is just a convenient way for having a global loop throughout
an application, the default loop is in no way different than the ones
initialized with \fI\%uv_loop_init()\fP\&. As such, the default loop can (and
should) be closed with \fI\%uv_loop_close()\fP so the resources associated
with it are freed.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
This function is not thread safe.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_run(uv_loop_t*\fI\ loop\fP, uv_run_mode\fI\ mode\fP)
This function runs the event loop. It will act differently depending on the
specified mode:
.INDENT 7.0
.IP \(bu 2
UV_RUN_DEFAULT: Runs the event loop until there are no more active and
referenced handles or requests. Returns non\-zero if \fI\%uv_stop()\fP
was called and there are still active handles or requests. Returns
zero in all other cases.
.IP \(bu 2
UV_RUN_ONCE: Poll for i/o once. Note that this function blocks if
there are no pending callbacks. Returns zero when done (no active handles
or requests left), or non\-zero if more callbacks are expected (meaning
you should run the event loop again sometime in the future).
.IP \(bu 2
UV_RUN_NOWAIT: Poll for i/o once but don\(aqt block if there are no
pending callbacks. Returns zero if done (no active handles
or requests left), or non\-zero if more callbacks are expected (meaning
you should run the event loop again sometime in the future).
.UNINDENT
.sp
\fI\%uv_run()\fP is not reentrant. It must not be called from a callback.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_loop_alive(const uv_loop_t*\fI\ loop\fP)
Returns non\-zero if there are referenced active handles, active
requests or closing handles in the loop.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_stop(uv_loop_t*\fI\ loop\fP)
Stop the event loop, causing \fI\%uv_run()\fP to end as soon as
possible. This will happen not sooner than the next loop iteration.
If this function was called before blocking for i/o, the loop won\(aqt block
for i/o on this iteration.
.UNINDENT
.INDENT 0.0
.TP
.B size_t uv_loop_size(void)
Returns the size of the \fIuv_loop_t\fP structure. Useful for FFI binding
writers who don\(aqt want to know the structure layout.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_backend_fd(const uv_loop_t*\fI\ loop\fP)
Get backend file descriptor. Only kqueue, epoll and event ports are
supported.
.sp
This can be used in conjunction with \fIuv_run(loop, UV_RUN_NOWAIT)\fP to
poll in one thread and run the event loop\(aqs callbacks in another see
test/test\-embed.c for an example.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Embedding a kqueue fd in another kqueue pollset doesn\(aqt work on all platforms. It\(aqs not
an error to add the fd but it never generates events.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_backend_timeout(const uv_loop_t*\fI\ loop\fP)
Get the poll timeout. The return value is in milliseconds, or \-1 for no
timeout.
.UNINDENT
.INDENT 0.0
.TP
.B uint64_t uv_now(const uv_loop_t*\fI\ loop\fP)
Return the current timestamp in milliseconds. The timestamp is cached at
the start of the event loop tick, see \fI\%uv_update_time()\fP for details
and rationale.
.sp
The timestamp increases monotonically from some arbitrary point in time.
Don\(aqt make assumptions about the starting point, you will only get
disappointed.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Use \fBuv_hrtime()\fP if you need sub\-millisecond granularity.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void uv_update_time(uv_loop_t*\fI\ loop\fP)
Update the event loop\(aqs concept of "now". Libuv caches the current time
at the start of the event loop tick in order to reduce the number of
time\-related system calls.
.sp
You won\(aqt normally need to call this function unless you have callbacks
that block the event loop for longer periods of time, where "longer" is
somewhat subjective but probably on the order of a millisecond or more.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_walk(uv_loop_t*\fI\ loop\fP, uv_walk_cb\fI\ walk_cb\fP, void*\fI\ arg\fP)
Walk the list of handles: \fIwalk_cb\fP will be executed with the given \fIarg\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_loop_fork(uv_loop_t*\fI\ loop\fP)
New in version 1.12.0.
.sp
Reinitialize any kernel state necessary in the child process after
a \fI\%fork(2)\fP system call.
.sp
Previously started watchers will continue to be started in the
child process.
.sp
It is necessary to explicitly call this function on every event
loop created in the parent process that you plan to continue to
use in the child, including the default loop (even if you don\(aqt
continue to use it in the parent). This function must be called
before calling \fI\%uv_run()\fP or any other API function using
the loop in the child. Failure to do so will result in undefined
behaviour, possibly including duplicate events delivered to both
parent and child or aborting the child process.
.sp
When possible, it is preferred to create a new loop in the child
process instead of reusing a loop created in the parent. New loops
created in the child process after the fork should not use this
function.
.sp
This function is not implemented on Windows, where it returns \fBUV_ENOSYS\fP\&.
.sp
\fBCAUTION:\fP
.INDENT 7.0
.INDENT 3.5
This function is experimental. It may contain bugs, and is subject to
change or removal. API and ABI stability is not guaranteed.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Mac OS X, if directory FS event handles were in use in the
parent process \fIfor any event loop\fP, the child process will no
longer be able to use the most efficient FSEvent
implementation. Instead, uses of directory FS event handles in
the child will fall back to the same implementation used for
files and on other kqueue\-based systems.
.UNINDENT
.UNINDENT
.sp
\fBCAUTION:\fP
.INDENT 7.0
.INDENT 3.5
On AIX and SunOS, FS event handles that were already started in
the parent process at the time of forking will \fInot\fP deliver
events in the child process; they must be closed and restarted.
On all other platforms, they will continue to work normally
without any further intervention.
.UNINDENT
.UNINDENT
.sp
\fBCAUTION:\fP
.INDENT 7.0
.INDENT 3.5
Any previous value returned from \fI\%uv_backend_fd()\fP is now
invalid. That function must be called again to determine the
correct backend file descriptor.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_loop_get_data(const uv_loop_t*\fI\ loop\fP)
Returns \fIloop\->data\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_loop_set_data(uv_loop_t*\fI\ loop\fP, void*\fI\ data\fP)
Sets \fIloop\->data\fP to \fIdata\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.SS \fI\%uv_handle_t\fP \-\-\- Base handle
.sp
\fIuv_handle_t\fP is the base type for all libuv handle types.
.sp
Structures are aligned so that any libuv handle can be cast to \fIuv_handle_t\fP\&.
All API functions defined here work with any handle type.
.sp
Libuv handles are not movable. Pointers to handle structures passed to
functions must remain valid for the duration of the requested operation. Take
care when using stack allocated handles.
.SS Data types
.INDENT 0.0
.TP
.B uv_handle_t
The base libuv handle type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_handle_type
The kind of the libuv handle.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
UV_UNKNOWN_HANDLE = 0,
UV_ASYNC,
UV_CHECK,
UV_FS_EVENT,
UV_FS_POLL,
UV_HANDLE,
UV_IDLE,
UV_NAMED_PIPE,
UV_POLL,
UV_PREPARE,
UV_PROCESS,
UV_STREAM,
UV_TCP,
UV_TIMER,
UV_TTY,
UV_UDP,
UV_SIGNAL,
UV_FILE,
UV_HANDLE_TYPE_MAX
} uv_handle_type;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_any_handle
Union of all handle types.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_alloc_cb)(uv_handle_t*\fI\ handle\fP, size_t\fI\ suggested_size\fP, uv_buf_t*\fI\ buf\fP)
Type definition for callback passed to \fBuv_read_start()\fP and
\fBuv_udp_recv_start()\fP\&. The user must allocate memory and fill the supplied
\fBuv_buf_t\fP structure. If NULL is assigned as the buffer\(aqs base or 0 as its length,
a \fBUV_ENOBUFS\fP error will be triggered in the \fBuv_udp_recv_cb\fP or the
\fBuv_read_cb\fP callback.
.sp
Each buffer is used only once and the user is responsible for freeing it in the
\fBuv_udp_recv_cb\fP or the \fBuv_read_cb\fP callback.
.sp
A suggested size (65536 at the moment in most cases) is provided, but it\(aqs just an indication,
not related in any way to the pending data to be read. The user is free to allocate the amount
of memory they decide.
.sp
As an example, applications with custom allocation schemes such as using freelists, allocation
pools or slab based allocators may decide to use a different size which matches the memory
chunks they already have.
.sp
Example:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
static void my_alloc_cb(uv_handle_t* handle, size_t suggested_size, uv_buf_t* buf) {
buf\->base = malloc(suggested_size);
buf\->len = suggested_size;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_close_cb)(uv_handle_t*\fI\ handle\fP)
Type definition for callback passed to \fI\%uv_close()\fP\&.
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B uv_loop_t* uv_handle_t.loop
Pointer to the \fBuv_loop_t\fP the handle is running on. Readonly.
.UNINDENT
.INDENT 0.0
.TP
.B uv_handle_type uv_handle_t.type
The \fI\%uv_handle_type\fP, indicating the type of the underlying handle. Readonly.
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_handle_t.data
Space for user\-defined arbitrary data. libuv does not use this field.
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B UV_HANDLE_TYPE_MAP(iter_macro)
Macro that expands to a series of invocations of \fIiter_macro\fP for
each of the handle types. \fIiter_macro\fP is invoked with two
arguments: the name of the \fIuv_handle_type\fP element without the
\fIUV_\fP prefix, and the name of the corresponding structure type
without the \fIuv_\fP prefix and \fI_t\fP suffix.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_is_active(const uv_handle_t*\fI\ handle\fP)
Returns non\-zero if the handle is active, zero if it\(aqs inactive. What
"active" means depends on the type of handle:
.INDENT 7.0
.IP \(bu 2
A uv_async_t handle is always active and cannot be deactivated, except
by closing it with uv_close().
.IP \(bu 2
A uv_pipe_t, uv_tcp_t, uv_udp_t, etc. handle \- basically any handle that
deals with i/o \- is active when it is doing something that involves i/o,
like reading, writing, connecting, accepting new connections, etc.
.IP \(bu 2
A uv_check_t, uv_idle_t, uv_timer_t, etc. handle is active when it has
been started with a call to uv_check_start(), uv_idle_start(), etc.
.UNINDENT
.sp
Rule of thumb: if a handle of type \fIuv_foo_t\fP has a \fIuv_foo_start()\fP
function, then it\(aqs active from the moment that function is called.
Likewise, \fIuv_foo_stop()\fP deactivates the handle again.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_is_closing(const uv_handle_t*\fI\ handle\fP)
Returns non\-zero if the handle is closing or closed, zero otherwise.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
This function should only be used between the initialization of the handle and the
arrival of the close callback.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void uv_close(uv_handle_t*\fI\ handle\fP, uv_close_cb\fI\ close_cb\fP)
Request handle to be closed. \fIclose_cb\fP will be called asynchronously after
this call. This MUST be called on each handle before memory is released.
Moreover, the memory can only be released in \fIclose_cb\fP or after it has
returned.
.sp
Handles that wrap file descriptors are closed immediately but
\fIclose_cb\fP will still be deferred to the next iteration of the event loop.
It gives you a chance to free up any resources associated with the handle.
.sp
In\-progress requests, like uv_connect_t or uv_write_t, are cancelled and
have their callbacks called asynchronously with status=UV_ECANCELED.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_ref(uv_handle_t*\fI\ handle\fP)
Reference the given handle. References are idempotent, that is, if a handle
is already referenced calling this function again will have no effect.
.sp
See \fI\%Reference counting\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_unref(uv_handle_t*\fI\ handle\fP)
Un\-reference the given handle. References are idempotent, that is, if a handle
is not referenced calling this function again will have no effect.
.sp
See \fI\%Reference counting\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_has_ref(const uv_handle_t*\fI\ handle\fP)
Returns non\-zero if the handle referenced, zero otherwise.
.sp
See \fI\%Reference counting\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B size_t uv_handle_size(uv_handle_type\fI\ type\fP)
Returns the size of the given handle type. Useful for FFI binding writers
who don\(aqt want to know the structure layout.
.UNINDENT
.SS Miscellaneous API functions
.sp
The following API functions take a \fI\%uv_handle_t\fP argument but they work
just for some handle types.
.INDENT 0.0
.TP
.B int uv_send_buffer_size(uv_handle_t*\fI\ handle\fP, int*\fI\ value\fP)
Gets or sets the size of the send buffer that the operating
system uses for the socket.
.sp
If \fI*value\fP == 0, then it will set \fI*value\fP to the current send buffer size.
If \fI*value\fP > 0 then it will use \fI*value\fP to set the new send buffer size.
.sp
On success, zero is returned. On error, a negative result is
returned.
.sp
This function works for TCP, pipe and UDP handles on Unix and for TCP and
UDP handles on Windows.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Linux will set double the size and return double the size of the original set value.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_recv_buffer_size(uv_handle_t*\fI\ handle\fP, int*\fI\ value\fP)
Gets or sets the size of the receive buffer that the operating
system uses for the socket.
.sp
If \fI*value\fP == 0, then it will set \fI*value\fP to the current receive buffer size.
If \fI*value\fP > 0 then it will use \fI*value\fP to set the new receive buffer size.
.sp
On success, zero is returned. On error, a negative result is
returned.
.sp
This function works for TCP, pipe and UDP handles on Unix and for TCP and
UDP handles on Windows.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Linux will set double the size and return double the size of the original set value.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fileno(const uv_handle_t*\fI\ handle\fP, uv_os_fd_t*\fI\ fd\fP)
Gets the platform dependent file descriptor equivalent.
.sp
The following handles are supported: TCP, pipes, TTY, UDP and poll. Passing
any other handle type will fail with \fIUV_EINVAL\fP\&.
.sp
If a handle doesn\(aqt have an attached file descriptor yet or the handle
itself has been closed, this function will return \fIUV_EBADF\fP\&.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
Be very careful when using this function. libuv assumes it\(aqs in control of the file
descriptor so any change to it may lead to malfunction.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_loop_t* uv_handle_get_loop(const uv_handle_t*\fI\ handle\fP)
Returns \fIhandle\->loop\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_handle_get_data(const uv_handle_t*\fI\ handle\fP)
Returns \fIhandle\->data\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_handle_set_data(uv_handle_t*\fI\ handle\fP, void*\fI\ data\fP)
Sets \fIhandle\->data\fP to \fIdata\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B uv_handle_type uv_handle_get_type(const uv_handle_t*\fI\ handle\fP)
Returns \fIhandle\->type\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B const char* uv_handle_type_name(uv_handle_type\fI\ type\fP)
Returns the name for the equivalent struct for a given handle type,
e.g. \fI"pipe"\fP (as in \fBuv_pipe_t\fP) for \fIUV_NAMED_PIPE\fP\&.
.sp
If no such handle type exists, this returns \fINULL\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.SS Reference counting
.sp
The libuv event loop (if run in the default mode) will run until there are no
active \fIand\fP referenced handles left. The user can force the loop to exit early
by unreferencing handles which are active, for example by calling \fI\%uv_unref()\fP
after calling \fBuv_timer_start()\fP\&.
.sp
A handle can be referenced or unreferenced, the refcounting scheme doesn\(aqt use
a counter, so both operations are idempotent.
.sp
All handles are referenced when active by default, see \fI\%uv_is_active()\fP
for a more detailed explanation on what being \fIactive\fP involves.
.SS \fI\%uv_req_t\fP \-\-\- Base request
.sp
\fIuv_req_t\fP is the base type for all libuv request types.
.sp
Structures are aligned so that any libuv request can be cast to \fIuv_req_t\fP\&.
All API functions defined here work with any request type.
.SS Data types
.INDENT 0.0
.TP
.B uv_req_t
The base libuv request structure.
.UNINDENT
.INDENT 0.0
.TP
.B uv_any_req
Union of all request types.
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B void* uv_req_t.data
Space for user\-defined arbitrary data. libuv does not use this field.
.UNINDENT
.INDENT 0.0
.TP
.B uv_req_type uv_req_t.type
Indicated the type of request. Readonly.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
UV_UNKNOWN_REQ = 0,
UV_REQ,
UV_CONNECT,
UV_WRITE,
UV_SHUTDOWN,
UV_UDP_SEND,
UV_FS,
UV_WORK,
UV_GETADDRINFO,
UV_GETNAMEINFO,
UV_REQ_TYPE_MAX,
} uv_req_type;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B UV_REQ_TYPE_MAP(iter_macro)
Macro that expands to a series of invocations of \fIiter_macro\fP for
each of the request types. \fIiter_macro\fP is invoked with two
arguments: the name of the \fIuv_req_type\fP element without the \fIUV_\fP
prefix, and the name of the corresponding structure type without the
\fIuv_\fP prefix and \fI_t\fP suffix.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_cancel(uv_req_t*\fI\ req\fP)
Cancel a pending request. Fails if the request is executing or has finished
executing.
.sp
Returns 0 on success, or an error code < 0 on failure.
.sp
Only cancellation of \fBuv_fs_t\fP, \fBuv_getaddrinfo_t\fP,
\fBuv_getnameinfo_t\fP, \fBuv_random_t\fP and \fBuv_work_t\fP
requests is currently supported.
.sp
Cancelled requests have their callbacks invoked some time in the future.
It\(aqs \fBnot\fP safe to free the memory associated with the request until the
callback is called.
.sp
Here is how cancellation is reported to the callback:
.INDENT 7.0
.IP \(bu 2
A \fBuv_fs_t\fP request has its req\->result field set to \fIUV_ECANCELED\fP\&.
.IP \(bu 2
A \fBuv_work_t\fP, \fBuv_getaddrinfo_t\fP,
\fBuv_getnameinfo_t\fP or \fBuv_random_t\fP request has its
callback invoked with status == \fIUV_ECANCELED\fP\&.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B size_t uv_req_size(uv_req_type\fI\ type\fP)
Returns the size of the given request type. Useful for FFI binding writers
who don\(aqt want to know the structure layout.
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_req_get_data(const uv_req_t*\fI\ req\fP)
Returns \fIreq\->data\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_req_set_data(uv_req_t*\fI\ req\fP, void*\fI\ data\fP)
Sets \fIreq\->data\fP to \fIdata\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B uv_req_type uv_req_get_type(const uv_req_t*\fI\ req\fP)
Returns \fIreq\->type\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B const char* uv_req_type_name(uv_req_type\fI\ type\fP)
Returns the name for the equivalent struct for a given request type,
e.g. \fI"connect"\fP (as in \fBuv_connect_t\fP) for \fIUV_CONNECT\fP\&.
.sp
If no such request type exists, this returns \fINULL\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.SS \fI\%uv_timer_t\fP \-\-\- Timer handle
.sp
Timer handles are used to schedule callbacks to be called in the future.
.SS Data types
.INDENT 0.0
.TP
.B uv_timer_t
Timer handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_timer_cb)(uv_timer_t*\fI\ handle\fP)
Type definition for callback passed to \fI\%uv_timer_start()\fP\&.
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_timer_init(uv_loop_t*\fI\ loop\fP, uv_timer_t*\fI\ handle\fP)
Initialize the handle.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_timer_start(uv_timer_t*\fI\ handle\fP, uv_timer_cb\fI\ cb\fP, uint64_t\fI\ timeout\fP, uint64_t\fI\ repeat\fP)
Start the timer. \fItimeout\fP and \fIrepeat\fP are in milliseconds.
.sp
If \fItimeout\fP is zero, the callback fires on the next event loop iteration.
If \fIrepeat\fP is non\-zero, the callback fires first after \fItimeout\fP
milliseconds and then repeatedly after \fIrepeat\fP milliseconds.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Does not update the event loop\(aqs concept of "now". See \fBuv_update_time()\fP for more information.
.sp
If the timer is already active, it is simply updated.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_timer_stop(uv_timer_t*\fI\ handle\fP)
Stop the timer, the callback will not be called anymore.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_timer_again(uv_timer_t*\fI\ handle\fP)
Stop the timer, and if it is repeating restart it using the repeat value
as the timeout. If the timer has never been started before it returns
UV_EINVAL.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_timer_set_repeat(uv_timer_t*\fI\ handle\fP, uint64_t\fI\ repeat\fP)
Set the repeat interval value in milliseconds. The timer will be scheduled
to run on the given interval, regardless of the callback execution
duration, and will follow normal timer semantics in the case of a
time\-slice overrun.
.sp
For example, if a 50ms repeating timer first runs for 17ms, it will be
scheduled to run again 33ms later. If other tasks consume more than the
33ms following the first timer callback, then the callback will run as soon
as possible.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
If the repeat value is set from a timer callback it does not immediately take effect.
If the timer was non\-repeating before, it will have been stopped. If it was repeating,
then the old repeat value will have been used to schedule the next timeout.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uint64_t uv_timer_get_repeat(const uv_timer_t*\fI\ handle\fP)
Get the timer repeat value.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_prepare_t\fP \-\-\- Prepare handle
.sp
Prepare handles will run the given callback once per loop iteration, right
before polling for i/o.
.SS Data types
.INDENT 0.0
.TP
.B uv_prepare_t
Prepare handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_prepare_cb)(uv_prepare_t*\fI\ handle\fP)
Type definition for callback passed to \fI\%uv_prepare_start()\fP\&.
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_prepare_init(uv_loop_t*\fI\ loop\fP, uv_prepare_t*\fI\ prepare\fP)
Initialize the handle. This function always succeeds.
.INDENT 7.0
.TP
.B Returns
0
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_prepare_start(uv_prepare_t*\fI\ prepare\fP, uv_prepare_cb\fI\ cb\fP)
Start the handle with the given callback. This function always succeeds,
except when \fIcb\fP is \fINULL\fP\&.
.INDENT 7.0
.TP
.B Returns
0 on success, or \fIUV_EINVAL\fP when \fIcb == NULL\fP\&.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_prepare_stop(uv_prepare_t*\fI\ prepare\fP)
Stop the handle, the callback will no longer be called.
This function always succeeds.
.INDENT 7.0
.TP
.B Returns
0
.UNINDENT
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_check_t\fP \-\-\- Check handle
.sp
Check handles will run the given callback once per loop iteration, right
after polling for i/o.
.SS Data types
.INDENT 0.0
.TP
.B uv_check_t
Check handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_check_cb)(uv_check_t*\fI\ handle\fP)
Type definition for callback passed to \fI\%uv_check_start()\fP\&.
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_check_init(uv_loop_t*\fI\ loop\fP, uv_check_t*\fI\ check\fP)
Initialize the handle. This function always succeeds.
.INDENT 7.0
.TP
.B Returns
0
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_check_start(uv_check_t*\fI\ check\fP, uv_check_cb\fI\ cb\fP)
Start the handle with the given callback. This function always succeeds,
except when \fIcb\fP is \fINULL\fP\&.
.INDENT 7.0
.TP
.B Returns
0 on success, or \fIUV_EINVAL\fP when \fIcb == NULL\fP\&.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_check_stop(uv_check_t*\fI\ check\fP)
Stop the handle, the callback will no longer be called.
This function always succeeds.
.INDENT 7.0
.TP
.B Returns
0
.UNINDENT
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_idle_t\fP \-\-\- Idle handle
.sp
Idle handles will run the given callback once per loop iteration, right
before the \fBuv_prepare_t\fP handles.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
The notable difference with prepare handles is that when there are active idle handles,
the loop will perform a zero timeout poll instead of blocking for i/o.
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
Despite the name, idle handles will get their callbacks called on every loop iteration,
not when the loop is actually "idle".
.UNINDENT
.UNINDENT
.SS Data types
.INDENT 0.0
.TP
.B uv_idle_t
Idle handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_idle_cb)(uv_idle_t*\fI\ handle\fP)
Type definition for callback passed to \fI\%uv_idle_start()\fP\&.
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_idle_init(uv_loop_t*\fI\ loop\fP, uv_idle_t*\fI\ idle\fP)
Initialize the handle. This function always succeeds.
.INDENT 7.0
.TP
.B Returns
0
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_idle_start(uv_idle_t*\fI\ idle\fP, uv_idle_cb\fI\ cb\fP)
Start the handle with the given callback. This function always succeeds,
except when \fIcb\fP is \fINULL\fP\&.
.INDENT 7.0
.TP
.B Returns
0 on success, or \fIUV_EINVAL\fP when \fIcb == NULL\fP\&.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_idle_stop(uv_idle_t*\fI\ idle\fP)
Stop the handle, the callback will no longer be called.
This function always succeeds.
.INDENT 7.0
.TP
.B Returns
0
.UNINDENT
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_async_t\fP \-\-\- Async handle
.sp
Async handles allow the user to "wakeup" the event loop and get a callback
called from another thread.
.SS Data types
.INDENT 0.0
.TP
.B uv_async_t
Async handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_async_cb)(uv_async_t*\fI\ handle\fP)
Type definition for callback passed to \fI\%uv_async_init()\fP\&.
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_async_init(uv_loop_t*\fI\ loop\fP, uv_async_t*\fI\ async\fP, uv_async_cb\fI\ async_cb\fP)
Initialize the handle. A NULL callback is allowed.
.INDENT 7.0
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Unlike other handle initialization functions, it immediately starts the handle.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_async_send(uv_async_t*\fI\ async\fP)
Wake up the event loop and call the async handle\(aqs callback.
.INDENT 7.0
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
It\(aqs safe to call this function from any thread. The callback will be called on the
loop thread.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fI\%uv_async_send()\fP is \fI\%async\-signal\-safe\fP\&.
It\(aqs safe to call this function from a signal handler.
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
libuv will coalesce calls to \fI\%uv_async_send()\fP, that is, not every call to it will
yield an execution of the callback. For example: if \fI\%uv_async_send()\fP is called 5
times in a row before the callback is called, the callback will only be called once. If
\fI\%uv_async_send()\fP is called again after the callback was called, it will be called
again.
.UNINDENT
.UNINDENT
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_poll_t\fP \-\-\- Poll handle
.sp
Poll handles are used to watch file descriptors for readability,
writability and disconnection similar to the purpose of \fI\%poll(2)\fP\&.
.sp
The purpose of poll handles is to enable integrating external libraries that
rely on the event loop to signal it about the socket status changes, like
c\-ares or libssh2. Using uv_poll_t for any other purpose is not recommended;
\fBuv_tcp_t\fP, \fBuv_udp_t\fP, etc. provide an implementation that is faster and
more scalable than what can be achieved with \fI\%uv_poll_t\fP, especially on
Windows.
.sp
It is possible that poll handles occasionally signal that a file descriptor is
readable or writable even when it isn\(aqt. The user should therefore always
be prepared to handle EAGAIN or equivalent when it attempts to read from or
write to the fd.
.sp
It is not okay to have multiple active poll handles for the same socket, this
can cause libuv to busyloop or otherwise malfunction.
.sp
The user should not close a file descriptor while it is being polled by an
active poll handle. This can cause the handle to report an error,
but it might also start polling another socket. However the fd can be safely
closed immediately after a call to \fI\%uv_poll_stop()\fP or \fBuv_close()\fP\&.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
On windows only sockets can be polled with poll handles. On Unix any file
descriptor that would be accepted by \fI\%poll(2)\fP can be used.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
On AIX, watching for disconnection is not supported.
.UNINDENT
.UNINDENT
.SS Data types
.INDENT 0.0
.TP
.B uv_poll_t
Poll handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_poll_cb)(uv_poll_t*\fI\ handle\fP, int\fI\ status\fP, int\fI\ events\fP)
Type definition for callback passed to \fI\%uv_poll_start()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B uv_poll_event
Poll event types
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
enum uv_poll_event {
UV_READABLE = 1,
UV_WRITABLE = 2,
UV_DISCONNECT = 4,
UV_PRIORITIZED = 8
};
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_poll_init(uv_loop_t*\fI\ loop\fP, uv_poll_t*\fI\ handle\fP, int\fI\ fd\fP)
Initialize the handle using a file descriptor.
.sp
Changed in version 1.2.2: the file descriptor is set to non\-blocking mode.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_poll_init_socket(uv_loop_t*\fI\ loop\fP, uv_poll_t*\fI\ handle\fP, uv_os_sock_t\fI\ socket\fP)
Initialize the handle using a socket descriptor. On Unix this is identical
to \fI\%uv_poll_init()\fP\&. On windows it takes a SOCKET handle.
.sp
Changed in version 1.2.2: the socket is set to non\-blocking mode.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_poll_start(uv_poll_t*\fI\ handle\fP, int\fI\ events\fP, uv_poll_cb\fI\ cb\fP)
Starts polling the file descriptor. \fIevents\fP is a bitmask made up of
UV_READABLE, UV_WRITABLE, UV_PRIORITIZED and UV_DISCONNECT. As soon as an
event is detected the callback will be called with \fIstatus\fP set to 0, and the
detected events set on the \fIevents\fP field.
.sp
The UV_PRIORITIZED event is used to watch for sysfs interrupts or TCP out\-of\-band
messages.
.sp
The UV_DISCONNECT event is optional in the sense that it may not be
reported and the user is free to ignore it, but it can help optimize the shutdown
path because an extra read or write call might be avoided.
.sp
If an error happens while polling, \fIstatus\fP will be < 0 and corresponds
with one of the UV_E* error codes (see errors). The user should
not close the socket while the handle is active. If the user does that
anyway, the callback \fImay\fP be called reporting an error status, but this
is \fBnot\fP guaranteed.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Calling \fI\%uv_poll_start()\fP on a handle that is already active is fine. Doing so
will update the events mask that is being watched for.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Though UV_DISCONNECT can be set, it is unsupported on AIX and as such will not be set
on the \fIevents\fP field in the callback.
.UNINDENT
.UNINDENT
.sp
Changed in version 1.9.0: Added the UV_DISCONNECT event.
.sp
Changed in version 1.14.0: Added the UV_PRIORITIZED event.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_poll_stop(uv_poll_t*\fI\ poll\fP)
Stop polling the file descriptor, the callback will no longer be called.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_signal_t\fP \-\-\- Signal handle
.sp
Signal handles implement Unix style signal handling on a per\-event loop bases.
.SS Windows notes
.sp
Reception of some signals is emulated:
.INDENT 0.0
.IP \(bu 2
SIGINT is normally delivered when the user presses CTRL+C. However, like
on Unix, it is not generated when terminal raw mode is enabled.
.IP \(bu 2
SIGBREAK is delivered when the user pressed CTRL + BREAK.
.IP \(bu 2
SIGHUP is generated when the user closes the console window. On SIGHUP the
program is given approximately 10 seconds to perform cleanup. After that
Windows will unconditionally terminate it.
.IP \(bu 2
SIGWINCH is raised whenever libuv detects that the console has been
resized. When a libuv app is running under a console emulator, or when a
32\-bit libuv app is running on 64\-bit system, SIGWINCH will be emulated. In
such cases SIGWINCH signals may not always be delivered in a timely manner.
For a writable \fBuv_tty_t\fP handle libuv will only detect size changes
when the cursor is moved. When a readable \fBuv_tty_t\fP handle is used,
resizing of the console buffer will be detected only if the handle is in raw
mode and is being read.
.IP \(bu 2
Watchers for other signals can be successfully created, but these signals
are never received. These signals are: \fISIGILL\fP, \fISIGABRT\fP, \fISIGFPE\fP, \fISIGSEGV\fP,
\fISIGTERM\fP and \fISIGKILL.\fP
.IP \(bu 2
Calls to raise() or abort() to programmatically raise a signal are
not detected by libuv; these will not trigger a signal watcher.
.UNINDENT
.sp
Changed in version 1.15.0: SIGWINCH support on Windows was improved.
.sp
Changed in version 1.31.0: 32\-bit libuv SIGWINCH support on 64\-bit Windows was
rolled back to old implementation.
.SS Unix notes
.INDENT 0.0
.IP \(bu 2
SIGKILL and SIGSTOP are impossible to catch.
.IP \(bu 2
Handling SIGBUS, SIGFPE, SIGILL or SIGSEGV via libuv results into undefined behavior.
.IP \(bu 2
SIGABRT will not be caught by libuv if generated by \fIabort()\fP, e.g. through \fIassert()\fP\&.
.IP \(bu 2
On Linux SIGRT0 and SIGRT1 (signals 32 and 33) are used by the NPTL pthreads library to
manage threads. Installing watchers for those signals will lead to unpredictable behavior
and is strongly discouraged. Future versions of libuv may simply reject them.
.UNINDENT
.SS Data types
.INDENT 0.0
.TP
.B uv_signal_t
Signal handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_signal_cb)(uv_signal_t*\fI\ handle\fP, int\fI\ signum\fP)
Type definition for callback passed to \fI\%uv_signal_start()\fP\&.
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B int uv_signal_t.signum
Signal being monitored by this handle. Readonly.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_signal_init(uv_loop_t*\fI\ loop\fP, uv_signal_t*\fI\ signal\fP)
Initialize the handle.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_signal_start(uv_signal_t*\fI\ signal\fP, uv_signal_cb\fI\ cb\fP, int\fI\ signum\fP)
Start the handle with the given callback, watching for the given signal.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_signal_start_oneshot(uv_signal_t*\fI\ signal\fP, uv_signal_cb\fI\ cb\fP, int\fI\ signum\fP)
New in version 1.12.0.
.sp
Same functionality as \fI\%uv_signal_start()\fP but the signal handler is reset the moment
the signal is received.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_signal_stop(uv_signal_t*\fI\ signal\fP)
Stop the handle, the callback will no longer be called.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_process_t\fP \-\-\- Process handle
.sp
Process handles will spawn a new process and allow the user to control it and
establish communication channels with it using streams.
.SS Data types
.INDENT 0.0
.TP
.B uv_process_t
Process handle type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t
Options for spawning the process (passed to \fI\%uv_spawn()\fP\&.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_process_options_s {
uv_exit_cb exit_cb;
const char* file;
char** args;
char** env;
const char* cwd;
unsigned int flags;
int stdio_count;
uv_stdio_container_t* stdio;
uv_uid_t uid;
uv_gid_t gid;
} uv_process_options_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_exit_cb)(uv_process_t*, int64_t\fI\ exit_status\fP, int\fI\ term_signal\fP)
Type definition for callback passed in \fI\%uv_process_options_t\fP which
will indicate the exit status and the signal that caused the process to
terminate, if any.
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_flags
Flags to be set on the flags field of \fI\%uv_process_options_t\fP\&.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
enum uv_process_flags {
/*
* Set the child process\(aq user id.
*/
UV_PROCESS_SETUID = (1 << 0),
/*
* Set the child process\(aq group id.
*/
UV_PROCESS_SETGID = (1 << 1),
/*
* Do not wrap any arguments in quotes, or perform any other escaping, when
* converting the argument list into a command line string. This option is
* only meaningful on Windows systems. On Unix it is silently ignored.
*/
UV_PROCESS_WINDOWS_VERBATIM_ARGUMENTS = (1 << 2),
/*
* Spawn the child process in a detached state \- this will make it a process
* group leader, and will effectively enable the child to keep running after
* the parent exits. Note that the child process will still keep the
* parent\(aqs event loop alive unless the parent process calls uv_unref() on
* the child\(aqs process handle.
*/
UV_PROCESS_DETACHED = (1 << 3),
/*
* Hide the subprocess window that would normally be created. This option is
* only meaningful on Windows systems. On Unix it is silently ignored.
*/
UV_PROCESS_WINDOWS_HIDE = (1 << 4),
/*
* Hide the subprocess console window that would normally be created. This
* option is only meaningful on Windows systems. On Unix it is silently
* ignored.
*/
UV_PROCESS_WINDOWS_HIDE_CONSOLE = (1 << 5),
/*
* Hide the subprocess GUI window that would normally be created. This
* option is only meaningful on Windows systems. On Unix it is silently
* ignored.
*/
UV_PROCESS_WINDOWS_HIDE_GUI = (1 << 6)
};
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_stdio_container_t
Container for each stdio handle or fd passed to a child process.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_stdio_container_s {
uv_stdio_flags flags;
union {
uv_stream_t* stream;
int fd;
} data;
} uv_stdio_container_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_stdio_flags
Flags specifying how a stdio should be transmitted to the child process.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
UV_IGNORE = 0x00,
UV_CREATE_PIPE = 0x01,
UV_INHERIT_FD = 0x02,
UV_INHERIT_STREAM = 0x04,
/*
* When UV_CREATE_PIPE is specified, UV_READABLE_PIPE and UV_WRITABLE_PIPE
* determine the direction of flow, from the child process\(aq perspective. Both
* flags may be specified to create a duplex data stream.
*/
UV_READABLE_PIPE = 0x10,
UV_WRITABLE_PIPE = 0x20
/*
* Open the child pipe handle in overlapped mode on Windows.
* On Unix it is silently ignored.
*/
UV_OVERLAPPED_PIPE = 0x40
} uv_stdio_flags;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B uv_process_t.pid
The PID of the spawned process. It\(aqs set after calling \fI\%uv_spawn()\fP\&.
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.exit_cb
Callback called after the process exits.
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.file
Path pointing to the program to be executed.
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.args
Command line arguments. args[0] should be the path to the program. On
Windows this uses \fICreateProcess\fP which concatenates the arguments into a
string this can cause some strange errors. See the
\fBUV_PROCESS_WINDOWS_VERBATIM_ARGUMENTS\fP flag on \fI\%uv_process_flags\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.env
Environment for the new process. If NULL the parents environment is used.
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.cwd
Current working directory for the subprocess.
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.flags
Various flags that control how \fI\%uv_spawn()\fP behaves. See
\fI\%uv_process_flags\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.stdio_count
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.stdio
The \fIstdio\fP field points to an array of \fI\%uv_stdio_container_t\fP
structs that describe the file descriptors that will be made available to
the child process. The convention is that stdio[0] points to stdin,
fd 1 is used for stdout, and fd 2 is stderr.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Windows file descriptors greater than 2 are available to the child process only if
the child processes uses the MSVCRT runtime.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.uid
.UNINDENT
.INDENT 0.0
.TP
.B uv_process_options_t.gid
Libuv can change the child process\(aq user/group id. This happens only when
the appropriate bits are set in the flags fields.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
This is not supported on Windows, \fI\%uv_spawn()\fP will fail and set the error
to \fBUV_ENOTSUP\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_stdio_container_t.flags
Flags specifying how the stdio container should be passed to the child. See
\fI\%uv_stdio_flags\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B uv_stdio_container_t.data
Union containing either the stream or fd to be passed on to the child
process.
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B void uv_disable_stdio_inheritance(void)
Disables inheritance for file descriptors / handles that this process
inherited from its parent. The effect is that child processes spawned by
this process don\(aqt accidentally inherit these handles.
.sp
It is recommended to call this function as early in your program as possible,
before the inherited file descriptors can be closed or duplicated.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
This function works on a best\-effort basis: there is no guarantee that libuv can discover
all file descriptors that were inherited. In general it does a better job on Windows than
it does on Unix.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_spawn(uv_loop_t*\fI\ loop\fP, uv_process_t*\fI\ handle\fP, const uv_process_options_t*\fI\ options\fP)
Initializes the process handle and starts the process. If the process is
successfully spawned, this function will return 0. Otherwise, the
negative error code corresponding to the reason it couldn\(aqt spawn is
returned.
.sp
Possible reasons for failing to spawn would include (but not be limited to)
the file to execute not existing, not having permissions to use the setuid or
setgid specified, or not having enough memory to allocate for the new
process.
.sp
Changed in version 1.24.0: Added \fIUV_PROCESS_WINDOWS_HIDE_CONSOLE\fP and
\fIUV_PROCESS_WINDOWS_HIDE_GUI\fP flags.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_process_kill(uv_process_t*\fI\ handle\fP, int\fI\ signum\fP)
Sends the specified signal to the given process handle. Check the documentation
on signal for signal support, specially on Windows.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_kill(int\fI\ pid\fP, int\fI\ signum\fP)
Sends the specified signal to the given PID. Check the documentation
on signal for signal support, specially on Windows.
.UNINDENT
.INDENT 0.0
.TP
.B uv_pid_t uv_process_get_pid(const uv_process_t*\fI\ handle\fP)
Returns \fIhandle\->pid\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_stream_t\fP \-\-\- Stream handle
.sp
Stream handles provide an abstraction of a duplex communication channel.
\fI\%uv_stream_t\fP is an abstract type, libuv provides 3 stream implementations
in the form of \fBuv_tcp_t\fP, \fBuv_pipe_t\fP and \fBuv_tty_t\fP\&.
.SS Data types
.INDENT 0.0
.TP
.B uv_stream_t
Stream handle type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_connect_t
Connect request type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_shutdown_t
Shutdown request type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_write_t
Write request type. Careful attention must be paid when reusing objects of
this type. When a stream is in non\-blocking mode, write requests sent
with \fBuv_write\fP will be queued. Reusing objects at this point is undefined
behaviour. It is safe to reuse the \fBuv_write_t\fP object only after the
callback passed to \fBuv_write\fP is fired.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_read_cb)(uv_stream_t*\fI\ stream\fP, ssize_t\fI\ nread\fP, const uv_buf_t*\fI\ buf\fP)
Callback called when data was read on a stream.
.sp
\fInread\fP is > 0 if there is data available or < 0 on error. When we\(aqve
reached EOF, \fInread\fP will be set to \fBUV_EOF\fP\&. When \fInread\fP < 0,
the \fIbuf\fP parameter might not point to a valid buffer; in that case
\fIbuf.len\fP and \fIbuf.base\fP are both set to 0.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fInread\fP might be 0, which does \fInot\fP indicate an error or EOF. This
is equivalent to \fBEAGAIN\fP or \fBEWOULDBLOCK\fP under \fBread(2)\fP\&.
.UNINDENT
.UNINDENT
.sp
The callee is responsible for stopping/closing the stream when an error happens
by calling \fI\%uv_read_stop()\fP or \fBuv_close()\fP\&. Trying to read
from the stream again is undefined.
.sp
The callee is responsible for freeing the buffer, libuv does not reuse it.
The buffer may be a null buffer (where \fIbuf\->base\fP == NULL and \fIbuf\->len\fP == 0)
on error.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_write_cb)(uv_write_t*\fI\ req\fP, int\fI\ status\fP)
Callback called after data was written on a stream. \fIstatus\fP will be 0 in
case of success, < 0 otherwise.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_connect_cb)(uv_connect_t*\fI\ req\fP, int\fI\ status\fP)
Callback called after a connection started by \fBuv_connect()\fP is done.
\fIstatus\fP will be 0 in case of success, < 0 otherwise.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_shutdown_cb)(uv_shutdown_t*\fI\ req\fP, int\fI\ status\fP)
Callback called after a shutdown request has been completed. \fIstatus\fP will
be 0 in case of success, < 0 otherwise.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_connection_cb)(uv_stream_t*\fI\ server\fP, int\fI\ status\fP)
Callback called when a stream server has received an incoming connection.
The user can accept the connection by calling \fI\%uv_accept()\fP\&.
\fIstatus\fP will be 0 in case of success, < 0 otherwise.
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B size_t uv_stream_t.write_queue_size
Contains the amount of queued bytes waiting to be sent. Readonly.
.UNINDENT
.INDENT 0.0
.TP
.B uv_stream_t* uv_connect_t.handle
Pointer to the stream where this connection request is running.
.UNINDENT
.INDENT 0.0
.TP
.B uv_stream_t* uv_shutdown_t.handle
Pointer to the stream where this shutdown request is running.
.UNINDENT
.INDENT 0.0
.TP
.B uv_stream_t* uv_write_t.handle
Pointer to the stream where this write request is running.
.UNINDENT
.INDENT 0.0
.TP
.B uv_stream_t* uv_write_t.send_handle
Pointer to the stream being sent using this write request.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_shutdown(uv_shutdown_t*\fI\ req\fP, uv_stream_t*\fI\ handle\fP, uv_shutdown_cb\fI\ cb\fP)
Shutdown the outgoing (write) side of a duplex stream. It waits for pending
write requests to complete. The \fIhandle\fP should refer to a initialized stream.
\fIreq\fP should be an uninitialized shutdown request struct. The \fIcb\fP is called
after shutdown is complete.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_listen(uv_stream_t*\fI\ stream\fP, int\fI\ backlog\fP, uv_connection_cb\fI\ cb\fP)
Start listening for incoming connections. \fIbacklog\fP indicates the number of
connections the kernel might queue, same as \fI\%listen(2)\fP\&. When a new
incoming connection is received the \fI\%uv_connection_cb\fP callback is
called.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_accept(uv_stream_t*\fI\ server\fP, uv_stream_t*\fI\ client\fP)
This call is used in conjunction with \fI\%uv_listen()\fP to accept incoming
connections. Call this function after receiving a \fI\%uv_connection_cb\fP
to accept the connection. Before calling this function the client handle must
be initialized. < 0 return value indicates an error.
.sp
When the \fI\%uv_connection_cb\fP callback is called it is guaranteed that
this function will complete successfully the first time. If you attempt to use
it more than once, it may fail. It is suggested to only call this function once
per \fI\%uv_connection_cb\fP call.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIserver\fP and \fIclient\fP must be handles running on the same loop.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_read_start(uv_stream_t*\fI\ stream\fP, uv_alloc_cb\fI\ alloc_cb\fP, uv_read_cb\fI\ read_cb\fP)
Read data from an incoming stream. The \fI\%uv_read_cb\fP callback will
be made several times until there is no more data to read or
\fI\%uv_read_stop()\fP is called.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_read_stop(uv_stream_t*)
Stop reading data from the stream. The \fI\%uv_read_cb\fP callback will
no longer be called.
.sp
This function is idempotent and may be safely called on a stopped stream.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_write(uv_write_t*\fI\ req\fP, uv_stream_t*\fI\ handle\fP, const uv_buf_t\fI\ bufs[]\fP, unsigned int\fI\ nbufs\fP, uv_write_cb\fI\ cb\fP)
Write data to stream. Buffers are written in order. Example:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
void cb(uv_write_t* req, int status) {
/* Logic which handles the write result */
}
uv_buf_t a[] = {
{ .base = "1", .len = 1 },
{ .base = "2", .len = 1 }
};
uv_buf_t b[] = {
{ .base = "3", .len = 1 },
{ .base = "4", .len = 1 }
};
uv_write_t req1;
uv_write_t req2;
/* writes "1234" */
uv_write(&req1, stream, a, 2, cb);
uv_write(&req2, stream, b, 2, cb);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
The memory pointed to by the buffers must remain valid until the callback gets called.
This also holds for \fI\%uv_write2()\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_write2(uv_write_t*\fI\ req\fP, uv_stream_t*\fI\ handle\fP, const uv_buf_t\fI\ bufs[]\fP, unsigned int\fI\ nbufs\fP, uv_stream_t*\fI\ send_handle\fP, uv_write_cb\fI\ cb\fP)
Extended write function for sending handles over a pipe. The pipe must be
initialized with \fIipc\fP == 1.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIsend_handle\fP must be a TCP socket or pipe, which is a server or a connection (listening
or connected state). Bound sockets or pipes will be assumed to be servers.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_try_write(uv_stream_t*\fI\ handle\fP, const uv_buf_t\fI\ bufs[]\fP, unsigned int\fI\ nbufs\fP)
Same as \fI\%uv_write()\fP, but won\(aqt queue a write request if it can\(aqt be
completed immediately.
.sp
Will return either:
.INDENT 7.0
.IP \(bu 2
> 0: number of bytes written (can be less than the supplied buffer size).
.IP \(bu 2
< 0: negative error code (\fBUV_EAGAIN\fP is returned if no data can be sent
immediately).
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_is_readable(const uv_stream_t*\fI\ handle\fP)
Returns 1 if the stream is readable, 0 otherwise.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_is_writable(const uv_stream_t*\fI\ handle\fP)
Returns 1 if the stream is writable, 0 otherwise.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_stream_set_blocking(uv_stream_t*\fI\ handle\fP, int\fI\ blocking\fP)
Enable or disable blocking mode for a stream.
.sp
When blocking mode is enabled all writes complete synchronously. The
interface remains unchanged otherwise, e.g. completion or failure of the
operation will still be reported through a callback which is made
asynchronously.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
Relying too much on this API is not recommended. It is likely to change
significantly in the future.
.sp
Currently only works on Windows for \fBuv_pipe_t\fP handles.
On UNIX platforms, all \fI\%uv_stream_t\fP handles are supported.
.sp
Also libuv currently makes no ordering guarantee when the blocking mode
is changed after write requests have already been submitted. Therefore it is
recommended to set the blocking mode immediately after opening or creating
the stream.
.UNINDENT
.UNINDENT
.sp
Changed in version 1.4.0: UNIX implementation added.
.UNINDENT
.INDENT 0.0
.TP
.B size_t uv_stream_get_write_queue_size(const uv_stream_t*\fI\ stream\fP)
Returns \fIstream\->write_queue_size\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_tcp_t\fP \-\-\- TCP handle
.sp
TCP handles are used to represent both TCP streams and servers.
.sp
\fI\%uv_tcp_t\fP is a \(aqsubclass\(aq of \fBuv_stream_t\fP\&.
.SS Data types
.INDENT 0.0
.TP
.B uv_tcp_t
TCP handle type.
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_stream_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_tcp_init(uv_loop_t*\fI\ loop\fP, uv_tcp_t*\fI\ handle\fP)
Initialize the handle. No socket is created as of yet.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_init_ex(uv_loop_t*\fI\ loop\fP, uv_tcp_t*\fI\ handle\fP, unsigned int\fI\ flags\fP)
Initialize the handle with the specified flags. At the moment only the lower 8 bits
of the \fIflags\fP parameter are used as the socket domain. A socket will be created
for the given domain. If the specified domain is \fBAF_UNSPEC\fP no socket is created,
just like \fI\%uv_tcp_init()\fP\&.
.sp
New in version 1.7.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_open(uv_tcp_t*\fI\ handle\fP, uv_os_sock_t\fI\ sock\fP)
Open an existing file descriptor or SOCKET as a TCP handle.
.sp
Changed in version 1.2.1: the file descriptor is set to non\-blocking mode.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
The passed file descriptor or SOCKET is not checked for its type, but
it\(aqs required that it represents a valid stream socket.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_nodelay(uv_tcp_t*\fI\ handle\fP, int\fI\ enable\fP)
Enable \fITCP_NODELAY\fP, which disables Nagle\(aqs algorithm.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_keepalive(uv_tcp_t*\fI\ handle\fP, int\fI\ enable\fP, unsigned int\fI\ delay\fP)
Enable / disable TCP keep\-alive. \fIdelay\fP is the initial delay in seconds,
ignored when \fIenable\fP is zero.
.sp
After \fIdelay\fP has been reached, 10 successive probes, each spaced 1 second
from the previous one, will still happen. If the connection is still lost
at the end of this procedure, then the handle is destroyed with a
\fBUV_ETIMEDOUT\fP error passed to the corresponding callback.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_simultaneous_accepts(uv_tcp_t*\fI\ handle\fP, int\fI\ enable\fP)
Enable / disable simultaneous asynchronous accept requests that are
queued by the operating system when listening for new TCP connections.
.sp
This setting is used to tune a TCP server for the desired performance.
Having simultaneous accepts can significantly improve the rate of accepting
connections (which is why it is enabled by default) but may lead to uneven
load distribution in multi\-process setups.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_bind(uv_tcp_t*\fI\ handle\fP, const struct sockaddr*\fI\ addr\fP, unsigned int\fI\ flags\fP)
Bind the handle to an address and port. \fIaddr\fP should point to an
initialized \fBstruct sockaddr_in\fP or \fBstruct sockaddr_in6\fP\&.
.sp
When the port is already taken, you can expect to see an \fBUV_EADDRINUSE\fP
error from either \fI\%uv_tcp_bind()\fP, \fBuv_listen()\fP or
\fI\%uv_tcp_connect()\fP\&. That is, a successful call to this function does
not guarantee that the call to \fBuv_listen()\fP or \fI\%uv_tcp_connect()\fP
will succeed as well.
.sp
\fIflags\fP can contain \fBUV_TCP_IPV6ONLY\fP, in which case dual\-stack support
is disabled and only IPv6 is used.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_getsockname(const uv_tcp_t*\fI\ handle\fP, struct sockaddr*\fI\ name\fP, int*\fI\ namelen\fP)
Get the current address to which the handle is bound. \fIname\fP must point to
a valid and big enough chunk of memory, \fBstruct sockaddr_storage\fP is
recommended for IPv4 and IPv6 support.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_getpeername(const uv_tcp_t*\fI\ handle\fP, struct sockaddr*\fI\ name\fP, int*\fI\ namelen\fP)
Get the address of the peer connected to the handle. \fIname\fP must point to
a valid and big enough chunk of memory, \fBstruct sockaddr_storage\fP is
recommended for IPv4 and IPv6 support.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_connect(uv_connect_t*\fI\ req\fP, uv_tcp_t*\fI\ handle\fP, const struct sockaddr*\fI\ addr\fP, uv_connect_cb\fI\ cb\fP)
Establish an IPv4 or IPv6 TCP connection. Provide an initialized TCP handle
and an uninitialized \fBuv_connect_t\fP\&. \fIaddr\fP should point to an
initialized \fBstruct sockaddr_in\fP or \fBstruct sockaddr_in6\fP\&.
.sp
On Windows if the \fIaddr\fP is initialized to point to an unspecified address
(\fB0.0.0.0\fP or \fB::\fP) it will be changed to point to \fBlocalhost\fP\&.
This is done to match the behavior of Linux systems.
.sp
The callback is made when the connection has been established or when a
connection error happened.
.sp
Changed in version 1.19.0: added \fB0.0.0.0\fP and \fB::\fP to \fBlocalhost\fP
mapping
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_stream_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tcp_close_reset(uv_tcp_t*\fI\ handle\fP, uv_close_cb\fI\ close_cb\fP)
Resets a TCP connection by sending a RST packet. This is accomplished by
setting the \fISO_LINGER\fP socket option with a linger interval of zero and
then calling \fBuv_close()\fP\&.
Due to some platform inconsistencies, mixing of \fBuv_shutdown()\fP and
\fI\%uv_tcp_close_reset()\fP calls is not allowed.
.sp
New in version 1.32.0.
.UNINDENT
.SS \fI\%uv_pipe_t\fP \-\-\- Pipe handle
.sp
Pipe handles provide an abstraction over streaming files on Unix (including
local domain sockets, pipes, and FIFOs) and named pipes on Windows.
.sp
\fI\%uv_pipe_t\fP is a \(aqsubclass\(aq of \fBuv_stream_t\fP\&.
.SS Data types
.INDENT 0.0
.TP
.B uv_pipe_t
Pipe handle type.
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B int uv_pipe_t.ipc
Whether this pipe is suitable for handle passing between processes.
Only a connected pipe that will be passing the handles should have this flag
set, not the listening pipe that uv_accept is called on.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_stream_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_pipe_init(uv_loop_t*\fI\ loop\fP, uv_pipe_t*\fI\ handle\fP, int\fI\ ipc\fP)
Initialize a pipe handle. The \fIipc\fP argument is a boolean to indicate if
this pipe will be used for handle passing between processes (which may
change the bytes on the wire). Only a connected pipe that will be
passing the handles should have this flag set, not the listening pipe
that uv_accept is called on.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_pipe_open(uv_pipe_t*\fI\ handle\fP, uv_file\fI\ file\fP)
Open an existing file descriptor or HANDLE as a pipe.
.sp
Changed in version 1.2.1: the file descriptor is set to non\-blocking mode.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
The passed file descriptor or HANDLE is not checked for its type, but
it\(aqs required that it represents a valid pipe.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_pipe_bind(uv_pipe_t*\fI\ handle\fP, const char*\fI\ name\fP)
Bind the pipe to a file path (Unix) or a name (Windows).
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Paths on Unix get truncated to \fBsizeof(sockaddr_un.sun_path)\fP bytes, typically between
92 and 108 bytes.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void uv_pipe_connect(uv_connect_t*\fI\ req\fP, uv_pipe_t*\fI\ handle\fP, const char*\fI\ name\fP, uv_connect_cb\fI\ cb\fP)
Connect to the Unix domain socket or the named pipe.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Paths on Unix get truncated to \fBsizeof(sockaddr_un.sun_path)\fP bytes, typically between
92 and 108 bytes.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_pipe_getsockname(const uv_pipe_t*\fI\ handle\fP, char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Get the name of the Unix domain socket or the named pipe.
.sp
A preallocated buffer must be provided. The size parameter holds the length
of the buffer and it\(aqs set to the number of bytes written to the buffer on
output. If the buffer is not big enough \fBUV_ENOBUFS\fP will be returned and
len will contain the required size.
.sp
Changed in version 1.3.0: the returned length no longer includes the terminating null byte,
and the buffer is not null terminated.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_pipe_getpeername(const uv_pipe_t*\fI\ handle\fP, char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Get the name of the Unix domain socket or the named pipe to which the handle
is connected.
.sp
A preallocated buffer must be provided. The size parameter holds the length
of the buffer and it\(aqs set to the number of bytes written to the buffer on
output. If the buffer is not big enough \fBUV_ENOBUFS\fP will be returned and
len will contain the required size.
.sp
New in version 1.3.0.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_pipe_pending_instances(uv_pipe_t*\fI\ handle\fP, int\fI\ count\fP)
Set the number of pending pipe instance handles when the pipe server is
waiting for connections.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
This setting applies to Windows only.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_pipe_pending_count(uv_pipe_t*\fI\ handle\fP)
.UNINDENT
.INDENT 0.0
.TP
.B uv_handle_type uv_pipe_pending_type(uv_pipe_t*\fI\ handle\fP)
Used to receive handles over IPC pipes.
.sp
First \- call \fI\%uv_pipe_pending_count()\fP, if it\(aqs > 0 then initialize
a handle of the given \fItype\fP, returned by \fI\%uv_pipe_pending_type()\fP
and call \fBuv_accept(pipe, handle)\fP\&.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_stream_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_pipe_chmod(uv_pipe_t*\fI\ handle\fP, int\fI\ flags\fP)
Alters pipe permissions, allowing it to be accessed from processes run by
different users. Makes the pipe writable or readable by all users. Mode can
be \fBUV_WRITABLE\fP, \fBUV_READABLE\fP or \fBUV_WRITABLE | UV_READABLE\fP\&. This
function is blocking.
.sp
New in version 1.16.0.
.UNINDENT
.SS \fI\%uv_tty_t\fP \-\-\- TTY handle
.sp
TTY handles represent a stream for the console.
.sp
\fI\%uv_tty_t\fP is a \(aqsubclass\(aq of \fBuv_stream_t\fP\&.
.SS Data types
.INDENT 0.0
.TP
.B uv_tty_t
TTY handle type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_tty_mode_t
New in version 1.2.0.
.sp
TTY mode type:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
/* Initial/normal terminal mode */
UV_TTY_MODE_NORMAL,
/* Raw input mode (On Windows, ENABLE_WINDOW_INPUT is also enabled) */
UV_TTY_MODE_RAW,
/* Binary\-safe I/O mode for IPC (Unix\-only) */
UV_TTY_MODE_IO
} uv_tty_mode_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_tty_vtermstate_t
.TP
.B Console virtual terminal mode type:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
/*
* The console supports handling of virtual terminal sequences
* (Windows10 new console, ConEmu)
*/
UV_TTY_SUPPORTED,
/* The console cannot process virtual terminal sequences. (Legacy
* console)
*/
UV_TTY_UNSUPPORTED
} uv_tty_vtermstate_t
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_stream_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_tty_init(uv_loop_t*\fI\ loop\fP, uv_tty_t*\fI\ handle\fP, uv_file\fI\ fd\fP, int\fI\ unused\fP)
Initialize a new TTY stream with the given file descriptor. Usually the
file descriptor will be:
.INDENT 7.0
.IP \(bu 2
0 = stdin
.IP \(bu 2
1 = stdout
.IP \(bu 2
2 = stderr
.UNINDENT
.sp
On Unix this function will determine the path of the fd of the terminal
using \fI\%ttyname_r(3)\fP, open it, and use it if the passed file descriptor
refers to a TTY. This lets libuv put the tty in non\-blocking mode without
affecting other processes that share the tty.
.sp
This function is not thread safe on systems that don\(aqt support
ioctl TIOCGPTN or TIOCPTYGNAME, for instance OpenBSD and Solaris.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
If reopening the TTY fails, libuv falls back to blocking writes.
.UNINDENT
.UNINDENT
.sp
Changed in version 1.23.1:: the \fIreadable\fP parameter is now unused and ignored.
The correct value will now be auto\-detected from the kernel.
.sp
Changed in version 1.9.0:: the path of the TTY is determined by
\fI\%ttyname_r(3)\fP\&. In earlier versions libuv opened
\fI/dev/tty\fP instead.
.sp
Changed in version 1.5.0:: trying to initialize a TTY stream with a file
descriptor that refers to a file returns \fIUV_EINVAL\fP
on UNIX.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tty_set_mode(uv_tty_t*\fI\ handle\fP, uv_tty_mode_t\fI\ mode\fP)
Changed in version 1.2.0:: the mode is specified as a
\fI\%uv_tty_mode_t\fP value.
.sp
Set the TTY using the specified terminal mode.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tty_reset_mode(void)
To be called when the program exits. Resets TTY settings to default
values for the next process to take over.
.sp
This function is async signal\-safe on Unix platforms but can fail with error
code \fBUV_EBUSY\fP if you call it when execution is inside
\fI\%uv_tty_set_mode()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tty_get_winsize(uv_tty_t*\fI\ handle\fP, int*\fI\ width\fP, int*\fI\ height\fP)
Gets the current Window size. On success it returns 0.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_stream_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void uv_tty_set_vterm_state(uv_tty_vtermstate_t\fI\ state\fP)
Controls whether console virtual terminal sequences are processed by libuv
or console.
Useful in particular for enabling ConEmu support of ANSI X3.64 and Xterm
256 colors. Otherwise Windows10 consoles are usually detected automatically.
.sp
This function is only meaningful on Windows systems. On Unix it is silently
ignored.
.sp
New in version 1.33.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_tty_get_vterm_state(uv_tty_vtermstate_t*\fI\ state\fP)
Get the current state of whether console virtual terminal sequences are
handled by libuv or the console.
.sp
This function is not implemented on Unix, where it returns \fBUV_ENOTSUP\fP\&.
.sp
New in version 1.33.0.
.UNINDENT
.SS \fI\%uv_udp_t\fP \-\-\- UDP handle
.sp
UDP handles encapsulate UDP communication for both clients and servers.
.SS Data types
.INDENT 0.0
.TP
.B uv_udp_t
UDP handle type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_udp_send_t
UDP send request type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_udp_flags
Flags used in \fI\%uv_udp_bind()\fP and \fI\%uv_udp_recv_cb\fP\&..
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
enum uv_udp_flags {
/* Disables dual stack mode. */
UV_UDP_IPV6ONLY = 1,
/*
* Indicates message was truncated because read buffer was too small. The
* remainder was discarded by the OS. Used in uv_udp_recv_cb.
*/
UV_UDP_PARTIAL = 2,
/*
* Indicates if SO_REUSEADDR will be set when binding the handle in
* uv_udp_bind.
* This sets the SO_REUSEPORT socket flag on the BSDs and OS X. On other
* Unix platforms, it sets the SO_REUSEADDR flag. What that means is that
* multiple threads or processes can bind to the same address without error
* (provided they all set the flag) but only the last one to bind will receive
* any traffic, in effect "stealing" the port from the previous listener.
*/
UV_UDP_REUSEADDR = 4,
/*
* Indicates that the message was received by recvmmsg, so the buffer provided
* must not be freed by the recv_cb callback.
*/
UV_UDP_MMSG_CHUNK = 8,
/*
* Indicates that recvmmsg should be used, if available.
*/
UV_UDP_RECVMMSG = 256
};
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_udp_send_cb)(uv_udp_send_t*\fI\ req\fP, int\fI\ status\fP)
Type definition for callback passed to \fI\%uv_udp_send()\fP, which is
called after the data was sent.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_udp_recv_cb)(uv_udp_t*\fI\ handle\fP, ssize_t\fI\ nread\fP, const uv_buf_t*\fI\ buf\fP, const struct sockaddr*\fI\ addr\fP, unsigned\fI\ flags\fP)
Type definition for callback passed to \fI\%uv_udp_recv_start()\fP, which
is called when the endpoint receives data.
.INDENT 7.0
.IP \(bu 2
\fIhandle\fP: UDP handle
.IP \(bu 2
\fInread\fP: Number of bytes that have been received.
0 if there is no more data to read. Note that 0 may also mean that an
empty datagram was received (in this case \fIaddr\fP is not NULL). < 0 if
a transmission error was detected.
.IP \(bu 2
\fIbuf\fP: \fBuv_buf_t\fP with the received data.
.IP \(bu 2
\fIaddr\fP: \fBstruct sockaddr*\fP containing the address of the sender.
Can be NULL. Valid for the duration of the callback only.
.IP \(bu 2
\fIflags\fP: One or more or\(aqed UV_UDP_* constants.
.UNINDENT
.sp
The callee is responsible for freeing the buffer, libuv does not reuse it.
The buffer may be a null buffer (where \fIbuf\->base\fP == NULL and \fIbuf\->len\fP == 0)
on error.
.sp
When using \fI\%recvmmsg(2)\fP, chunks will have the \fIUV_UDP_MMSG_CHUNK\fP flag set,
those must not be freed. There will be a final callback with \fInread\fP set to 0,
\fIaddr\fP set to NULL and the buffer pointing at the initially allocated data with
the \fIUV_UDP_MMSG_CHUNK\fP flag cleared. This is a good chance for the callee to
free the provided buffer.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
The receive callback will be called with \fInread\fP == 0 and \fIaddr\fP == NULL when there is
nothing to read, and with \fInread\fP == 0 and \fIaddr\fP != NULL when an empty UDP packet is
received.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_membership
Membership type for a multicast address.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
UV_LEAVE_GROUP = 0,
UV_JOIN_GROUP
} uv_membership;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B size_t uv_udp_t.send_queue_size
Number of bytes queued for sending. This field strictly shows how much
information is currently queued.
.UNINDENT
.INDENT 0.0
.TP
.B size_t uv_udp_t.send_queue_count
Number of send requests currently in the queue awaiting to be processed.
.UNINDENT
.INDENT 0.0
.TP
.B uv_udp_t* uv_udp_send_t.handle
UDP handle where this send request is taking place.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_udp_init(uv_loop_t*\fI\ loop\fP, uv_udp_t*\fI\ handle\fP)
Initialize a new UDP handle. The actual socket is created lazily.
Returns 0 on success.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_init_ex(uv_loop_t*\fI\ loop\fP, uv_udp_t*\fI\ handle\fP, unsigned int\fI\ flags\fP)
Initialize the handle with the specified flags. The lower 8 bits of the \fIflags\fP
parameter are used as the socket domain. A socket will be created for the given domain.
If the specified domain is \fBAF_UNSPEC\fP no socket is created, just like \fI\%uv_udp_init()\fP\&.
.sp
The remaining bits can be used to set one of these flags:
.INDENT 7.0
.IP \(bu 2
\fIUV_UDP_RECVMMSG\fP: if set, and the platform supports it, \fI\%recvmmsg(2)\fP will
be used.
.UNINDENT
.sp
New in version 1.7.0.
.sp
Changed in version 1.37.0: added the \fIUV_UDP_RECVMMSG\fP flag.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_open(uv_udp_t*\fI\ handle\fP, uv_os_sock_t\fI\ sock\fP)
Opens an existing file descriptor or Windows SOCKET as a UDP handle.
.sp
Unix only:
The only requirement of the \fIsock\fP argument is that it follows the datagram
contract (works in unconnected mode, supports sendmsg()/recvmsg(), etc).
In other words, other datagram\-type sockets like raw sockets or netlink
sockets can also be passed to this function.
.sp
Changed in version 1.2.1: the file descriptor is set to non\-blocking mode.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
The passed file descriptor or SOCKET is not checked for its type, but
it\(aqs required that it represents a valid datagram socket.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_bind(uv_udp_t*\fI\ handle\fP, const struct sockaddr*\fI\ addr\fP, unsigned int\fI\ flags\fP)
Bind the UDP handle to an IP address and port.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBaddr\fP \-\- \fIstruct sockaddr_in\fP or \fIstruct sockaddr_in6\fP
with the address and port to bind to.
.IP \(bu 2
\fBflags\fP \-\- Indicate how the socket will be bound,
\fBUV_UDP_IPV6ONLY\fP and \fBUV_UDP_REUSEADDR\fP are supported.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_connect(uv_udp_t*\fI\ handle\fP, const struct sockaddr*\fI\ addr\fP)
Associate the UDP handle to a remote address and port, so every
message sent by this handle is automatically sent to that destination.
Calling this function with a \fINULL\fP \fIaddr\fP disconnects the handle.
Trying to call \fIuv_udp_connect()\fP on an already connected handle will result
in an \fIUV_EISCONN\fP error. Trying to disconnect a handle that is not
connected will return an \fIUV_ENOTCONN\fP error.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBaddr\fP \-\- \fIstruct sockaddr_in\fP or \fIstruct sockaddr_in6\fP
with the address and port to associate to.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.sp
New in version 1.27.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_getpeername(const uv_udp_t*\fI\ handle\fP, struct sockaddr*\fI\ name\fP, int*\fI\ namelen\fP)
Get the remote IP and port of the UDP handle on connected UDP handles.
On unconnected handles, it returns \fIUV_ENOTCONN\fP\&.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP and bound.
.IP \(bu 2
\fBname\fP \-\- Pointer to the structure to be filled with the address data.
In order to support IPv4 and IPv6 \fIstruct sockaddr_storage\fP should be
used.
.IP \(bu 2
\fBnamelen\fP \-\- On input it indicates the data of the \fIname\fP field. On
output it indicates how much of it was filled.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure
.UNINDENT
.sp
New in version 1.27.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_getsockname(const uv_udp_t*\fI\ handle\fP, struct sockaddr*\fI\ name\fP, int*\fI\ namelen\fP)
Get the local IP and port of the UDP handle.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP and bound.
.IP \(bu 2
\fBname\fP \-\- Pointer to the structure to be filled with the address data.
In order to support IPv4 and IPv6 \fIstruct sockaddr_storage\fP should be
used.
.IP \(bu 2
\fBnamelen\fP \-\- On input it indicates the data of the \fIname\fP field. On
output it indicates how much of it was filled.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_set_membership(uv_udp_t*\fI\ handle\fP, const char*\fI\ multicast_addr\fP, const char*\fI\ interface_addr\fP, uv_membership\fI\ membership\fP)
Set membership for a multicast address
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBmulticast_addr\fP \-\- Multicast address to set membership for.
.IP \(bu 2
\fBinterface_addr\fP \-\- Interface address.
.IP \(bu 2
\fBmembership\fP \-\- Should be \fBUV_JOIN_GROUP\fP or \fBUV_LEAVE_GROUP\fP\&.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_set_source_membership(uv_udp_t*\fI\ handle\fP, const char*\fI\ multicast_addr\fP, const char*\fI\ interface_addr\fP, const char*\fI\ source_addr\fP, uv_membership\fI\ membership\fP)
Set membership for a source\-specific multicast group.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBmulticast_addr\fP \-\- Multicast address to set membership for.
.IP \(bu 2
\fBinterface_addr\fP \-\- Interface address.
.IP \(bu 2
\fBsource_addr\fP \-\- Source address.
.IP \(bu 2
\fBmembership\fP \-\- Should be \fBUV_JOIN_GROUP\fP or \fBUV_LEAVE_GROUP\fP\&.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.sp
New in version 1.32.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_set_multicast_loop(uv_udp_t*\fI\ handle\fP, int\fI\ on\fP)
Set IP multicast loop flag. Makes multicast packets loop back to
local sockets.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBon\fP \-\- 1 for on, 0 for off.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_set_multicast_ttl(uv_udp_t*\fI\ handle\fP, int\fI\ ttl\fP)
Set the multicast ttl.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBttl\fP \-\- 1 through 255.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_set_multicast_interface(uv_udp_t*\fI\ handle\fP, const char*\fI\ interface_addr\fP)
Set the multicast interface to send or receive data on.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBinterface_addr\fP \-\- interface address.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_set_broadcast(uv_udp_t*\fI\ handle\fP, int\fI\ on\fP)
Set broadcast on or off.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBon\fP \-\- 1 for on, 0 for off.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_set_ttl(uv_udp_t*\fI\ handle\fP, int\fI\ ttl\fP)
Set the time to live.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBttl\fP \-\- 1 through 255.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_send(uv_udp_send_t*\fI\ req\fP, uv_udp_t*\fI\ handle\fP, const uv_buf_t\fI\ bufs[]\fP, unsigned int\fI\ nbufs\fP, const struct sockaddr*\fI\ addr\fP, uv_udp_send_cb\fI\ send_cb\fP)
Send data over the UDP socket. If the socket has not previously been bound
with \fI\%uv_udp_bind()\fP it will be bound to 0.0.0.0
(the "all interfaces" IPv4 address) and a random port number.
.sp
On Windows if the \fIaddr\fP is initialized to point to an unspecified address
(\fB0.0.0.0\fP or \fB::\fP) it will be changed to point to \fBlocalhost\fP\&.
This is done to match the behavior of Linux systems.
.sp
For connected UDP handles, \fIaddr\fP must be set to \fINULL\fP, otherwise it will
return \fIUV_EISCONN\fP error.
.sp
For connectionless UDP handles, \fIaddr\fP cannot be \fINULL\fP, otherwise it will
return \fIUV_EDESTADDRREQ\fP error.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBreq\fP \-\- UDP request handle. Need not be initialized.
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBbufs\fP \-\- List of buffers to send.
.IP \(bu 2
\fBnbufs\fP \-\- Number of buffers in \fIbufs\fP\&.
.IP \(bu 2
\fBaddr\fP \-\- \fIstruct sockaddr_in\fP or \fIstruct sockaddr_in6\fP with the
address and port of the remote peer.
.IP \(bu 2
\fBsend_cb\fP \-\- Callback to invoke when the data has been sent out.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.sp
Changed in version 1.19.0: added \fB0.0.0.0\fP and \fB::\fP to \fBlocalhost\fP
mapping
.sp
Changed in version 1.27.0: added support for connected sockets
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_try_send(uv_udp_t*\fI\ handle\fP, const uv_buf_t\fI\ bufs[]\fP, unsigned int\fI\ nbufs\fP, const struct sockaddr*\fI\ addr\fP)
Same as \fI\%uv_udp_send()\fP, but won\(aqt queue a send request if it can\(aqt
be completed immediately.
.sp
For connected UDP handles, \fIaddr\fP must be set to \fINULL\fP, otherwise it will
return \fIUV_EISCONN\fP error.
.sp
For connectionless UDP handles, \fIaddr\fP cannot be \fINULL\fP, otherwise it will
return \fIUV_EDESTADDRREQ\fP error.
.INDENT 7.0
.TP
.B Returns
>= 0: number of bytes sent (it matches the given buffer size).
< 0: negative error code (\fBUV_EAGAIN\fP is returned when the message
can\(aqt be sent immediately).
.UNINDENT
.sp
Changed in version 1.27.0: added support for connected sockets
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_recv_start(uv_udp_t*\fI\ handle\fP, uv_alloc_cb\fI\ alloc_cb\fP, uv_udp_recv_cb\fI\ recv_cb\fP)
Prepare for receiving data. If the socket has not previously been bound
with \fI\%uv_udp_bind()\fP it is bound to 0.0.0.0 (the "all interfaces"
IPv4 address) and a random port number.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.IP \(bu 2
\fBalloc_cb\fP \-\- Callback to invoke when temporary storage is needed.
.IP \(bu 2
\fBrecv_cb\fP \-\- Callback to invoke with received data.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.sp
Changed in version 1.35.0: added support for \fI\%recvmmsg(2)\fP on supported platforms).
The use of this feature requires a buffer larger than
2 * 64KB to be passed to \fIalloc_cb\fP\&.
.sp
Changed in version 1.37.0: \fI\%recvmmsg(2)\fP support is no longer enabled implicitly,
it must be explicitly requested by passing the \fIUV_UDP_RECVMMSG\fP flag to
\fI\%uv_udp_init_ex()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_udp_recv_stop(uv_udp_t*\fI\ handle\fP)
Stop listening for incoming datagrams.
.INDENT 7.0
.TP
.B Parameters
.INDENT 7.0
.IP \(bu 2
\fBhandle\fP \-\- UDP handle. Should have been initialized with
\fI\%uv_udp_init()\fP\&.
.UNINDENT
.TP
.B Returns
0 on success, or an error code < 0 on failure.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B size_t uv_udp_get_send_queue_size(const uv_udp_t*\fI\ handle\fP)
Returns \fIhandle\->send_queue_size\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B size_t uv_udp_get_send_queue_count(const uv_udp_t*\fI\ handle\fP)
Returns \fIhandle\->send_queue_count\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_fs_event_t\fP \-\-\- FS Event handle
.sp
FS Event handles allow the user to monitor a given path for changes, for example,
if the file was renamed or there was a generic change in it. This handle uses
the best backend for the job on each platform.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
For AIX, the non default IBM bos.ahafs package has to be installed.
The AIX Event Infrastructure file system (ahafs) has some limitations:
.INDENT 0.0
.INDENT 3.5
.INDENT 0.0
.IP \(bu 2
ahafs tracks monitoring per process and is not thread safe. A separate process
must be spawned for each monitor for the same event.
.IP \(bu 2
Events for file modification (writing to a file) are not received if only the
containing folder is watched.
.UNINDENT
.UNINDENT
.UNINDENT
.sp
See \fI\%documentation\fP for more details.
.sp
The z/OS file system events monitoring infrastructure does not notify of file
creation/deletion within a directory that is being monitored.
See the \fI\%IBM Knowledge centre\fP for more details.
.UNINDENT
.UNINDENT
.SS Data types
.INDENT 0.0
.TP
.B uv_fs_event_t
FS Event handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_fs_event_cb)(uv_fs_event_t*\fI\ handle\fP, const char*\fI\ filename\fP, int\fI\ events\fP, int\fI\ status\fP)
Callback passed to \fI\%uv_fs_event_start()\fP which will be called repeatedly
after the handle is started. If the handle was started with a directory the
\fIfilename\fP parameter will be a relative path to a file contained in the directory.
The \fIevents\fP parameter is an ORed mask of \fI\%uv_fs_event\fP elements.
.UNINDENT
.INDENT 0.0
.TP
.B uv_fs_event
Event types that \fI\%uv_fs_event_t\fP handles monitor.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
enum uv_fs_event {
UV_RENAME = 1,
UV_CHANGE = 2
};
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_fs_event_flags
Flags that can be passed to \fI\%uv_fs_event_start()\fP to control its
behavior.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
enum uv_fs_event_flags {
/*
* By default, if the fs event watcher is given a directory name, we will
* watch for all events in that directory. This flags overrides this behavior
* and makes fs_event report only changes to the directory entry itself. This
* flag does not affect individual files watched.
* This flag is currently not implemented yet on any backend.
*/
UV_FS_EVENT_WATCH_ENTRY = 1,
/*
* By default uv_fs_event will try to use a kernel interface such as inotify
* or kqueue to detect events. This may not work on remote file systems such
* as NFS mounts. This flag makes fs_event fall back to calling stat() on a
* regular interval.
* This flag is currently not implemented yet on any backend.
*/
UV_FS_EVENT_STAT = 2,
/*
* By default, event watcher, when watching directory, is not registering
* (is ignoring) changes in its subdirectories.
* This flag will override this behaviour on platforms that support it.
*/
UV_FS_EVENT_RECURSIVE = 4
};
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_fs_event_init(uv_loop_t*\fI\ loop\fP, uv_fs_event_t*\fI\ handle\fP)
Initialize the handle.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_event_start(uv_fs_event_t*\fI\ handle\fP, uv_fs_event_cb\fI\ cb\fP, const char*\fI\ path\fP, unsigned int\fI\ flags\fP)
Start the handle with the given callback, which will watch the specified
\fIpath\fP for changes. \fIflags\fP can be an ORed mask of \fI\%uv_fs_event_flags\fP\&.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Currently the only supported flag is \fBUV_FS_EVENT_RECURSIVE\fP and
only on OSX and Windows.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_event_stop(uv_fs_event_t*\fI\ handle\fP)
Stop the handle, the callback will no longer be called.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_event_getpath(uv_fs_event_t*\fI\ handle\fP, char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Get the path being monitored by the handle. The buffer must be preallocated
by the user. Returns 0 on success or an error code < 0 in case of failure.
On success, \fIbuffer\fP will contain the path and \fIsize\fP its length. If the buffer
is not big enough \fIUV_ENOBUFS\fP will be returned and \fIsize\fP will be set to
the required size, including the null terminator.
.sp
Changed in version 1.3.0: the returned length no longer includes the terminating null byte,
and the buffer is not null terminated.
.sp
Changed in version 1.9.0: the returned length includes the terminating null
byte on \fIUV_ENOBUFS\fP, and the buffer is null terminated
on success.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS \fI\%uv_fs_poll_t\fP \-\-\- FS Poll handle
.sp
FS Poll handles allow the user to monitor a given path for changes. Unlike
\fBuv_fs_event_t\fP, fs poll handles use \fIstat\fP to detect when a file has
changed so they can work on file systems where fs event handles can\(aqt.
.SS Data types
.INDENT 0.0
.TP
.B uv_fs_poll_t
FS Poll handle type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_fs_poll_cb)(uv_fs_poll_t*\fI\ handle\fP, int\fI\ status\fP, const uv_stat_t*\fI\ prev\fP, const uv_stat_t*\fI\ curr\fP)
Callback passed to \fI\%uv_fs_poll_start()\fP which will be called repeatedly
after the handle is started, when any change happens to the monitored path.
.sp
The callback is invoked with \fIstatus < 0\fP if \fIpath\fP does not exist
or is inaccessible. The watcher is \fInot\fP stopped but your callback is
not called again until something changes (e.g. when the file is created
or the error reason changes).
.sp
When \fIstatus == 0\fP, the callback receives pointers to the old and new
\fBuv_stat_t\fP structs. They are valid for the duration of the
callback only.
.UNINDENT
.SS Public members
.sp
N/A
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_fs_poll_init(uv_loop_t*\fI\ loop\fP, uv_fs_poll_t*\fI\ handle\fP)
Initialize the handle.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_poll_start(uv_fs_poll_t*\fI\ handle\fP, uv_fs_poll_cb\fI\ poll_cb\fP, const char*\fI\ path\fP, unsigned int\fI\ interval\fP)
Check the file at \fIpath\fP for changes every \fIinterval\fP milliseconds.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
For maximum portability, use multi\-second intervals. Sub\-second intervals will not detect
all changes on many file systems.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_poll_stop(uv_fs_poll_t*\fI\ handle\fP)
Stop the handle, the callback will no longer be called.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_poll_getpath(uv_fs_poll_t*\fI\ handle\fP, char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Get the path being monitored by the handle. The buffer must be preallocated
by the user. Returns 0 on success or an error code < 0 in case of failure.
On success, \fIbuffer\fP will contain the path and \fIsize\fP its length. If the buffer
is not big enough \fIUV_ENOBUFS\fP will be returned and \fIsize\fP will be set to
the required size.
.sp
Changed in version 1.3.0: the returned length no longer includes the terminating null byte,
and the buffer is not null terminated.
.sp
Changed in version 1.9.0: the returned length includes the terminating null
byte on \fIUV_ENOBUFS\fP, and the buffer is null terminated
on success.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_handle_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS File system operations
.sp
libuv provides a wide variety of cross\-platform sync and async file system
operations. All functions defined in this document take a callback, which is
allowed to be NULL. If the callback is NULL the request is completed synchronously,
otherwise it will be performed asynchronously.
.sp
All file operations are run on the threadpool. See threadpool for information
on the threadpool size.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
On Windows \fIuv_fs_*\fP functions use utf\-8 encoding.
.UNINDENT
.UNINDENT
.SS Data types
.INDENT 0.0
.TP
.B uv_fs_t
File system request type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_timespec_t
Portable equivalent of \fBstruct timespec\fP\&.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct {
long tv_sec;
long tv_nsec;
} uv_timespec_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_stat_t
Portable equivalent of \fBstruct stat\fP\&.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct {
uint64_t st_dev;
uint64_t st_mode;
uint64_t st_nlink;
uint64_t st_uid;
uint64_t st_gid;
uint64_t st_rdev;
uint64_t st_ino;
uint64_t st_size;
uint64_t st_blksize;
uint64_t st_blocks;
uint64_t st_flags;
uint64_t st_gen;
uv_timespec_t st_atim;
uv_timespec_t st_mtim;
uv_timespec_t st_ctim;
uv_timespec_t st_birthtim;
} uv_stat_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_fs_type
File system request type.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
UV_FS_UNKNOWN = \-1,
UV_FS_CUSTOM,
UV_FS_OPEN,
UV_FS_CLOSE,
UV_FS_READ,
UV_FS_WRITE,
UV_FS_SENDFILE,
UV_FS_STAT,
UV_FS_LSTAT,
UV_FS_FSTAT,
UV_FS_FTRUNCATE,
UV_FS_UTIME,
UV_FS_FUTIME,
UV_FS_ACCESS,
UV_FS_CHMOD,
UV_FS_FCHMOD,
UV_FS_FSYNC,
UV_FS_FDATASYNC,
UV_FS_UNLINK,
UV_FS_RMDIR,
UV_FS_MKDIR,
UV_FS_MKDTEMP,
UV_FS_RENAME,
UV_FS_SCANDIR,
UV_FS_LINK,
UV_FS_SYMLINK,
UV_FS_READLINK,
UV_FS_CHOWN,
UV_FS_FCHOWN,
UV_FS_REALPATH,
UV_FS_COPYFILE,
UV_FS_LCHOWN,
UV_FS_OPENDIR,
UV_FS_READDIR,
UV_FS_CLOSEDIR,
UV_FS_MKSTEMP,
UV_FS_LUTIME
} uv_fs_type;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_statfs_t
Reduced cross platform equivalent of \fBstruct statfs\fP\&.
Used in \fI\%uv_fs_statfs()\fP\&.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_statfs_s {
uint64_t f_type;
uint64_t f_bsize;
uint64_t f_blocks;
uint64_t f_bfree;
uint64_t f_bavail;
uint64_t f_files;
uint64_t f_ffree;
uint64_t f_spare[4];
} uv_statfs_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_dirent_t
Cross platform (reduced) equivalent of \fBstruct dirent\fP\&.
Used in \fI\%uv_fs_scandir_next()\fP\&.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef enum {
UV_DIRENT_UNKNOWN,
UV_DIRENT_FILE,
UV_DIRENT_DIR,
UV_DIRENT_LINK,
UV_DIRENT_FIFO,
UV_DIRENT_SOCKET,
UV_DIRENT_CHAR,
UV_DIRENT_BLOCK
} uv_dirent_type_t;
typedef struct uv_dirent_s {
const char* name;
uv_dirent_type_t type;
} uv_dirent_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_dir_t
Data type used for streaming directory iteration.
Used by \fI\%uv_fs_opendir()\fP, \fI\%uv_fs_readdir()\fP, and
\fI\%uv_fs_closedir()\fP\&. \fIdirents\fP represents a user provided array of
\fIuv_dirent_t\(gas used to hold results. \(ganentries\fP is the user provided maximum
array size of \fIdirents\fP\&.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_dir_s {
uv_dirent_t* dirents;
size_t nentries;
} uv_dir_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B uv_loop_t* uv_fs_t.loop
Loop that started this request and where completion will be reported.
Readonly.
.UNINDENT
.INDENT 0.0
.TP
.B uv_fs_type uv_fs_t.fs_type
FS request type.
.UNINDENT
.INDENT 0.0
.TP
.B const char* uv_fs_t.path
Path affecting the request.
.UNINDENT
.INDENT 0.0
.TP
.B ssize_t uv_fs_t.result
Result of the request. < 0 means error, success otherwise. On requests such
as \fI\%uv_fs_read()\fP or \fI\%uv_fs_write()\fP it indicates the amount of
data that was read or written, respectively.
.UNINDENT
.INDENT 0.0
.TP
.B uv_stat_t uv_fs_t.statbuf
Stores the result of \fI\%uv_fs_stat()\fP and other stat requests.
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_fs_t.ptr
Stores the result of \fI\%uv_fs_readlink()\fP and
\fI\%uv_fs_realpath()\fP and serves as an alias to \fIstatbuf\fP\&.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_req_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B void uv_fs_req_cleanup(uv_fs_t*\fI\ req\fP)
Cleanup request. Must be called after a request is finished to deallocate
any memory libuv might have allocated.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_close(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%close(2)\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_open(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, int\fI\ flags\fP, int\fI\ mode\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%open(2)\fP\&.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Windows libuv uses \fICreateFileW\fP and thus the file is always opened
in binary mode. Because of this the O_BINARY and O_TEXT flags are not
supported.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_read(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, const uv_buf_t\fI\ bufs[]\fP, unsigned int\fI\ nbufs\fP, int64_t\fI\ offset\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%preadv(2)\fP\&.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
On Windows, under non\-MSVC environments (e.g. when GCC or Clang is used
to build libuv), files opened using \fBUV_FS_O_FILEMAP\fP may cause a fatal
crash if the memory mapped read operation fails.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_unlink(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%unlink(2)\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_write(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, const uv_buf_t\fI\ bufs[]\fP, unsigned int\fI\ nbufs\fP, int64_t\fI\ offset\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%pwritev(2)\fP\&.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
On Windows, under non\-MSVC environments (e.g. when GCC or Clang is used
to build libuv), files opened using \fBUV_FS_O_FILEMAP\fP may cause a fatal
crash if the memory mapped write operation fails.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_mkdir(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, int\fI\ mode\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%mkdir(2)\fP\&.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fImode\fP is currently not implemented on Windows.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_mkdtemp(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ tpl\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%mkdtemp(3)\fP\&. The result can be found as a null terminated string at \fIreq\->path\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_mkstemp(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ tpl\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%mkstemp(3)\fP\&. The created file path can be found as a null terminated string at \fIreq\->path\fP\&.
The file descriptor can be found as an integer at \fIreq\->result\fP\&.
.sp
New in version 1.34.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_rmdir(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%rmdir(2)\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_opendir(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_fs_cb\fI\ cb\fP)
Opens \fIpath\fP as a directory stream. On success, a \fIuv_dir_t\fP is allocated
and returned via \fIreq\->ptr\fP\&. This memory is not freed by
\fIuv_fs_req_cleanup()\fP, although \fIreq\->ptr\fP is set to \fINULL\fP\&. The allocated
memory must be freed by calling \fIuv_fs_closedir()\fP\&. On failure, no memory
is allocated.
.sp
The contents of the directory can be iterated over by passing the resulting
\fIuv_dir_t\fP to \fIuv_fs_readdir()\fP\&.
.sp
New in version 1.28.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_closedir(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_dir_t*\fI\ dir\fP, uv_fs_cb\fI\ cb\fP)
Closes the directory stream represented by \fIdir\fP and frees the memory
allocated by \fIuv_fs_opendir()\fP\&.
.sp
New in version 1.28.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_readdir(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_dir_t*\fI\ dir\fP, uv_fs_cb\fI\ cb\fP)
Iterates over the directory stream, \fIdir\fP, returned by a successful
\fIuv_fs_opendir()\fP call. Prior to invoking \fIuv_fs_readdir()\fP, the caller
must set \fIdir\->dirents\fP and \fIdir\->nentries\fP, representing the array of
\fI\%uv_dirent_t\fP elements used to hold the read directory entries and
its size.
.sp
On success, the result is an integer >= 0 representing the number of entries
read from the stream.
.sp
New in version 1.28.0.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
\fIuv_fs_readdir()\fP is not thread safe.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
This function does not return the "." and ".." entries.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On success this function allocates memory that must be freed using
\fIuv_fs_req_cleanup()\fP\&. \fIuv_fs_req_cleanup()\fP must be called before
closing the directory with \fIuv_fs_closedir()\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_scandir(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, int\fI\ flags\fP, uv_fs_cb\fI\ cb\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_scandir_next(uv_fs_t*\fI\ req\fP, uv_dirent_t*\fI\ ent\fP)
Equivalent to \fI\%scandir(3)\fP, with a slightly different API. Once the callback
for the request is called, the user can use \fI\%uv_fs_scandir_next()\fP to
get \fIent\fP populated with the next directory entry data. When there are no
more entries \fBUV_EOF\fP will be returned.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Unlike \fIscandir(3)\fP, this function does not return the "." and ".." entries.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Linux, getting the type of an entry is only supported by some file systems (btrfs, ext2,
ext3 and ext4 at the time of this writing), check the \fI\%getdents(2)\fP man page.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_stat(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_fs_cb\fI\ cb\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_fstat(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, uv_fs_cb\fI\ cb\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_lstat(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%stat(2)\fP, \fI\%fstat(2)\fP and \fI\%lstat(2)\fP respectively.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_statfs(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%statfs(2)\fP\&. On success, a \fIuv_statfs_t\fP is allocated
and returned via \fIreq\->ptr\fP\&. This memory is freed by \fIuv_fs_req_cleanup()\fP\&.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Any fields in the resulting \fIuv_statfs_t\fP that are not supported by the
underlying operating system are set to zero.
.UNINDENT
.UNINDENT
.sp
New in version 1.31.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_rename(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, const char*\fI\ new_path\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%rename(2)\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_fsync(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%fsync(2)\fP\&.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
For AIX, \fIuv_fs_fsync\fP returns \fIUV_EBADF\fP on file descriptors referencing
non regular files.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_fdatasync(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%fdatasync(2)\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_ftruncate(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, int64_t\fI\ offset\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%ftruncate(2)\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_copyfile(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, const char*\fI\ new_path\fP, int\fI\ flags\fP, uv_fs_cb\fI\ cb\fP)
Copies a file from \fIpath\fP to \fInew_path\fP\&. Supported \fIflags\fP are described below.
.INDENT 7.0
.IP \(bu 2
\fIUV_FS_COPYFILE_EXCL\fP: If present, \fIuv_fs_copyfile()\fP will fail with
\fIUV_EEXIST\fP if the destination path already exists. The default behavior
is to overwrite the destination if it exists.
.IP \(bu 2
\fIUV_FS_COPYFILE_FICLONE\fP: If present, \fIuv_fs_copyfile()\fP will attempt to
create a copy\-on\-write reflink. If the underlying platform does not
support copy\-on\-write, or an error occurs while attempting to use
copy\-on\-write, a fallback copy mechanism based on
\fI\%uv_fs_sendfile()\fP is used.
.IP \(bu 2
\fIUV_FS_COPYFILE_FICLONE_FORCE\fP: If present, \fIuv_fs_copyfile()\fP will
attempt to create a copy\-on\-write reflink. If the underlying platform does
not support copy\-on\-write, or an error occurs while attempting to use
copy\-on\-write, then an error is returned.
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
If the destination path is created, but an error occurs while copying
the data, then the destination path is removed. There is a brief window
of time between closing and removing the file where another process
could access the file.
.UNINDENT
.UNINDENT
.sp
New in version 1.14.0.
.sp
Changed in version 1.20.0: \fIUV_FS_COPYFILE_FICLONE\fP and
\fIUV_FS_COPYFILE_FICLONE_FORCE\fP are supported.
.sp
Changed in version 1.33.0: If an error occurs while using
\fIUV_FS_COPYFILE_FICLONE_FORCE\fP, that error is returned. Previously,
all errors were mapped to \fIUV_ENOTSUP\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_sendfile(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ out_fd\fP, uv_file\fI\ in_fd\fP, int64_t\fI\ in_offset\fP, size_t\fI\ length\fP, uv_fs_cb\fI\ cb\fP)
Limited equivalent to \fI\%sendfile(2)\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_access(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, int\fI\ mode\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%access(2)\fP on Unix. Windows uses \fBGetFileAttributesW()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_chmod(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, int\fI\ mode\fP, uv_fs_cb\fI\ cb\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_fchmod(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, int\fI\ mode\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%chmod(2)\fP and \fI\%fchmod(2)\fP respectively.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_utime(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, double\fI\ atime\fP, double\fI\ mtime\fP, uv_fs_cb\fI\ cb\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_futime(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, double\fI\ atime\fP, double\fI\ mtime\fP, uv_fs_cb\fI\ cb\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_lutime(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, double\fI\ atime\fP, double\fI\ mtime\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%utime(2)\fP, \fI\%futimes(3)\fP and \fI\%lutimes(3)\fP respectively.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
z/OS: \fIuv_fs_lutime()\fP is not implemented for z/OS. It can still be called but will return
\fBUV_ENOSYS\fP\&.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
AIX: \fIuv_fs_futime()\fP and \fIuv_fs_lutime()\fP functions only work for AIX 7.1 and newer.
They can still be called on older versions but will return \fBUV_ENOSYS\fP\&.
.UNINDENT
.UNINDENT
.sp
Changed in version 1.10.0: sub\-second precission is supported on Windows
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_link(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, const char*\fI\ new_path\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%link(2)\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_symlink(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, const char*\fI\ new_path\fP, int\fI\ flags\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%symlink(2)\fP\&.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Windows the \fIflags\fP parameter can be specified to control how the symlink will
be created:
.INDENT 0.0
.INDENT 3.5
.INDENT 0.0
.IP \(bu 2
\fBUV_FS_SYMLINK_DIR\fP: indicates that \fIpath\fP points to a directory.
.IP \(bu 2
\fBUV_FS_SYMLINK_JUNCTION\fP: request that the symlink is created
using junction points.
.UNINDENT
.UNINDENT
.UNINDENT
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_readlink(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%readlink(2)\fP\&.
The resulting string is stored in \fIreq\->ptr\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_realpath(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%realpath(3)\fP on Unix. Windows uses \fI\%GetFinalPathNameByHandle\fP\&.
The resulting string is stored in \fIreq\->ptr\fP\&.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
This function has certain platform\-specific caveats that were discovered when used in Node.
.INDENT 0.0
.IP \(bu 2
macOS and other BSDs: this function will fail with UV_ELOOP if more than 32 symlinks are
found while resolving the given path. This limit is hardcoded and cannot be sidestepped.
.IP \(bu 2
Windows: while this function works in the common case, there are a number of corner cases
where it doesn\(aqt:
.INDENT 2.0
.IP \(bu 2
Paths in ramdisk volumes created by tools which sidestep the Volume Manager (such as ImDisk)
cannot be resolved.
.IP \(bu 2
Inconsistent casing when using drive letters.
.IP \(bu 2
Resolved path bypasses subst\(aqd drives.
.UNINDENT
.UNINDENT
.sp
While this function can still be used, it\(aqs not recommended if scenarios such as the
above need to be supported.
.sp
The background story and some more details on these issues can be checked
\fI\%here\fP\&.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
This function is not implemented on Windows XP and Windows Server 2003.
On these systems, UV_ENOSYS is returned.
.UNINDENT
.UNINDENT
.sp
New in version 1.8.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_chown(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_uid_t\fI\ uid\fP, uv_gid_t\fI\ gid\fP, uv_fs_cb\fI\ cb\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_fchown(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, uv_file\fI\ file\fP, uv_uid_t\fI\ uid\fP, uv_gid_t\fI\ gid\fP, uv_fs_cb\fI\ cb\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_lchown(uv_loop_t*\fI\ loop\fP, uv_fs_t*\fI\ req\fP, const char*\fI\ path\fP, uv_uid_t\fI\ uid\fP, uv_gid_t\fI\ gid\fP, uv_fs_cb\fI\ cb\fP)
Equivalent to \fI\%chown(2)\fP, \fI\%fchown(2)\fP and \fI\%lchown(2)\fP respectively.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
These functions are not implemented on Windows.
.UNINDENT
.UNINDENT
.sp
Changed in version 1.21.0: implemented uv_fs_lchown
.UNINDENT
.INDENT 0.0
.TP
.B uv_fs_type uv_fs_get_type(const uv_fs_t*\fI\ req\fP)
Returns \fIreq\->fs_type\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B ssize_t uv_fs_get_result(const uv_fs_t*\fI\ req\fP)
Returns \fIreq\->result\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_fs_get_system_error(const uv_fs_t*\fI\ req\fP)
Returns the platform specific error code \- \fIGetLastError()\fP value on Windows
and \fI\-(req\->result)\fP on other platforms.
.sp
New in version 1.38.0.
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_fs_get_ptr(const uv_fs_t*\fI\ req\fP)
Returns \fIreq\->ptr\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B const char* uv_fs_get_path(const uv_fs_t*\fI\ req\fP)
Returns \fIreq\->path\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.INDENT 0.0
.TP
.B uv_stat_t* uv_fs_get_statbuf(uv_fs_t*\fI\ req\fP)
Returns \fI&req\->statbuf\fP\&.
.sp
New in version 1.19.0.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_req_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS Helper functions
.INDENT 0.0
.TP
.B uv_os_fd_t uv_get_osfhandle(int\fI\ fd\fP)
For a file descriptor in the C runtime, get the OS\-dependent handle.
On UNIX, returns the \fBfd\fP intact. On Windows, this calls \fI\%_get_osfhandle\fP\&.
Note that the return value is still owned by the C runtime,
any attempts to close it or to use it after closing the fd may lead to malfunction.
.INDENT 7.0
.INDENT 3.5
New in version 1.12.0.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_open_osfhandle(uv_os_fd_t\fI\ os_fd\fP)
For a OS\-dependent handle, get the file descriptor in the C runtime.
On UNIX, returns the \fBos_fd\fP intact. On Windows, this calls \fI\%_open_osfhandle\fP\&.
Note that the return value is still owned by the CRT,
any attempts to close it or to use it after closing the handle may lead to malfunction.
.INDENT 7.0
.INDENT 3.5
New in version 1.23.0.
.UNINDENT
.UNINDENT
.UNINDENT
.SS File open constants
.INDENT 0.0
.TP
.B UV_FS_O_APPEND
The file is opened in append mode. Before each write, the file offset is
positioned at the end of the file.
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_CREAT
The file is created if it does not already exist.
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_DIRECT
File I/O is done directly to and from user\-space buffers, which must be
aligned. Buffer size and address should be a multiple of the physical sector
size of the block device.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_DIRECT\fP is supported on Linux, and on Windows via
\fI\%FILE_FLAG_NO_BUFFERING\fP\&.
\fIUV_FS_O_DIRECT\fP is not supported on macOS.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_DIRECTORY
If the path is not a directory, fail the open.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_DIRECTORY\fP is not supported on Windows.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_DSYNC
The file is opened for synchronous I/O. Write operations will complete once
all data and a minimum of metadata are flushed to disk.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_DSYNC\fP is supported on Windows via
\fI\%FILE_FLAG_WRITE_THROUGH\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_EXCL
If the \fIO_CREAT\fP flag is set and the file already exists, fail the open.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
In general, the behavior of \fIO_EXCL\fP is undefined if it is used without
\fIO_CREAT\fP\&. There is one exception: on Linux 2.6 and later, \fIO_EXCL\fP can
be used without \fIO_CREAT\fP if pathname refers to a block device. If the
block device is in use by the system (e.g., mounted), the open will fail
with the error \fIEBUSY\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_EXLOCK
Atomically obtain an exclusive lock.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_EXLOCK\fP is only supported on macOS and Windows.
.UNINDENT
.UNINDENT
.sp
Changed in version 1.17.0: support is added for Windows.
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_FILEMAP
Use a memory file mapping to access the file. When using this flag, the
file cannot be open multiple times concurrently.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_FILEMAP\fP is only supported on Windows.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_NOATIME
Do not update the file access time when the file is read.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_NOATIME\fP is not supported on Windows.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_NOCTTY
If the path identifies a terminal device, opening the path will not cause
that terminal to become the controlling terminal for the process (if the
process does not already have one).
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_NOCTTY\fP is not supported on Windows.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_NOFOLLOW
If the path is a symbolic link, fail the open.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_NOFOLLOW\fP is not supported on Windows.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_NONBLOCK
Open the file in nonblocking mode if possible.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_NONBLOCK\fP is not supported on Windows.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_RANDOM
Access is intended to be random. The system can use this as a hint to
optimize file caching.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_RANDOM\fP is only supported on Windows via
\fI\%FILE_FLAG_RANDOM_ACCESS\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_RDONLY
Open the file for read\-only access.
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_RDWR
Open the file for read\-write access.
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_SEQUENTIAL
Access is intended to be sequential from beginning to end. The system can
use this as a hint to optimize file caching.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_SEQUENTIAL\fP is only supported on Windows via
\fI\%FILE_FLAG_SEQUENTIAL_SCAN\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_SHORT_LIVED
The file is temporary and should not be flushed to disk if possible.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_SHORT_LIVED\fP is only supported on Windows via
\fI\%FILE_ATTRIBUTE_TEMPORARY\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_SYMLINK
Open the symbolic link itself rather than the resource it points to.
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_SYNC
The file is opened for synchronous I/O. Write operations will complete once
all data and all metadata are flushed to disk.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_SYNC\fP is supported on Windows via
\fI\%FILE_FLAG_WRITE_THROUGH\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_TEMPORARY
The file is temporary and should not be flushed to disk if possible.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
\fIUV_FS_O_TEMPORARY\fP is only supported on Windows via
\fI\%FILE_ATTRIBUTE_TEMPORARY\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_TRUNC
If the file exists and is a regular file, and the file is opened
successfully for write access, its length shall be truncated to zero.
.UNINDENT
.INDENT 0.0
.TP
.B UV_FS_O_WRONLY
Open the file for write\-only access.
.UNINDENT
.SS Thread pool work scheduling
.sp
libuv provides a threadpool which can be used to run user code and get notified
in the loop thread. This thread pool is internally used to run all file system
operations, as well as getaddrinfo and getnameinfo requests.
.sp
Its default size is 4, but it can be changed at startup time by setting the
\fBUV_THREADPOOL_SIZE\fP environment variable to any value (the absolute maximum
is 1024).
.sp
Changed in version 1.30.0: the maximum UV_THREADPOOL_SIZE allowed was increased from 128 to 1024.
.sp
The threadpool is global and shared across all event loops. When a particular
function makes use of the threadpool (i.e. when using \fI\%uv_queue_work()\fP)
libuv preallocates and initializes the maximum number of threads allowed by
\fBUV_THREADPOOL_SIZE\fP\&. This causes a relatively minor memory overhead
(~1MB for 128 threads) but increases the performance of threading at runtime.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
Note that even though a global thread pool which is shared across all events
loops is used, the functions are not thread safe.
.UNINDENT
.UNINDENT
.SS Data types
.INDENT 0.0
.TP
.B uv_work_t
Work request type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_work_cb)(uv_work_t*\fI\ req\fP)
Callback passed to \fI\%uv_queue_work()\fP which will be run on the thread
pool.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_after_work_cb)(uv_work_t*\fI\ req\fP, int\fI\ status\fP)
Callback passed to \fI\%uv_queue_work()\fP which will be called on the loop
thread after the work on the threadpool has been completed. If the work
was cancelled using \fBuv_cancel()\fP \fIstatus\fP will be \fBUV_ECANCELED\fP\&.
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B uv_loop_t* uv_work_t.loop
Loop that started this request and where completion will be reported.
Readonly.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_req_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_queue_work(uv_loop_t*\fI\ loop\fP, uv_work_t*\fI\ req\fP, uv_work_cb\fI\ work_cb\fP, uv_after_work_cb\fI\ after_work_cb\fP)
Initializes a work request which will run the given \fIwork_cb\fP in a thread
from the threadpool. Once \fIwork_cb\fP is completed, \fIafter_work_cb\fP will be
called on the loop thread.
.sp
This request can be cancelled with \fBuv_cancel()\fP\&.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_req_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS DNS utility functions
.sp
libuv provides asynchronous variants of \fIgetaddrinfo\fP and \fIgetnameinfo\fP\&.
.SS Data types
.INDENT 0.0
.TP
.B uv_getaddrinfo_t
\fIgetaddrinfo\fP request type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_getaddrinfo_cb)(uv_getaddrinfo_t*\fI\ req\fP, int\fI\ status\fP, struct addrinfo*\fI\ res\fP)
Callback which will be called with the getaddrinfo request result once
complete. In case it was cancelled, \fIstatus\fP will have a value of
\fBUV_ECANCELED\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B uv_getnameinfo_t
\fIgetnameinfo\fP request type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_getnameinfo_cb)(uv_getnameinfo_t*\fI\ req\fP, int\fI\ status\fP, const char*\fI\ hostname\fP, const char*\fI\ service\fP)
Callback which will be called with the getnameinfo request result once
complete. In case it was cancelled, \fIstatus\fP will have a value of
\fBUV_ECANCELED\fP\&.
.UNINDENT
.SS Public members
.INDENT 0.0
.TP
.B uv_loop_t* uv_getaddrinfo_t.loop
Loop that started this getaddrinfo request and where completion will be
reported. Readonly.
.UNINDENT
.INDENT 0.0
.TP
.B struct addrinfo* uv_getaddrinfo_t.addrinfo
Pointer to a \fIstruct addrinfo\fP containing the result. Must be freed by the user
with \fI\%uv_freeaddrinfo()\fP\&.
.sp
Changed in version 1.3.0: the field is declared as public.
.UNINDENT
.INDENT 0.0
.TP
.B uv_loop_t* uv_getnameinfo_t.loop
Loop that started this getnameinfo request and where completion will be
reported. Readonly.
.UNINDENT
.INDENT 0.0
.TP
.B char[NI_MAXHOST] uv_getnameinfo_t.host
Char array containing the resulting host. It\(aqs null terminated.
.sp
Changed in version 1.3.0: the field is declared as public.
.UNINDENT
.INDENT 0.0
.TP
.B char[NI_MAXSERV] uv_getnameinfo_t.service
Char array containing the resulting service. It\(aqs null terminated.
.sp
Changed in version 1.3.0: the field is declared as public.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_req_t\fP members also apply.
.UNINDENT
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B int uv_getaddrinfo(uv_loop_t*\fI\ loop\fP, uv_getaddrinfo_t*\fI\ req\fP, uv_getaddrinfo_cb\fI\ getaddrinfo_cb\fP, const char*\fI\ node\fP, const char*\fI\ service\fP, const struct addrinfo*\fI\ hints\fP)
Asynchronous \fI\%getaddrinfo(3)\fP\&.
.sp
Either node or service may be NULL but not both.
.sp
\fIhints\fP is a pointer to a struct addrinfo with additional address type
constraints, or NULL. Consult \fIman \-s 3 getaddrinfo\fP for more details.
.sp
Returns 0 on success or an error code < 0 on failure. If successful, the
callback will get called sometime in the future with the lookup result,
which is either:
.INDENT 7.0
.IP \(bu 2
status == 0, the res argument points to a valid \fIstruct addrinfo\fP, or
.IP \(bu 2
status < 0, the res argument is NULL. See the UV_EAI_* constants.
.UNINDENT
.sp
Call \fI\%uv_freeaddrinfo()\fP to free the addrinfo structure.
.sp
Changed in version 1.3.0: the callback parameter is now allowed to be NULL,
in which case the request will run \fBsynchronously\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_freeaddrinfo(struct addrinfo*\fI\ ai\fP)
Free the struct addrinfo. Passing NULL is allowed and is a no\-op.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_getnameinfo(uv_loop_t*\fI\ loop\fP, uv_getnameinfo_t*\fI\ req\fP, uv_getnameinfo_cb\fI\ getnameinfo_cb\fP, const struct sockaddr*\fI\ addr\fP, int\fI\ flags\fP)
Asynchronous \fI\%getnameinfo(3)\fP\&.
.sp
Returns 0 on success or an error code < 0 on failure. If successful, the
callback will get called sometime in the future with the lookup result.
Consult \fIman \-s 3 getnameinfo\fP for more details.
.sp
Changed in version 1.3.0: the callback parameter is now allowed to be NULL,
in which case the request will run \fBsynchronously\fP\&.
.UNINDENT
.sp
\fBSEE ALSO:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_req_t\fP API functions also apply.
.UNINDENT
.UNINDENT
.SS Shared library handling
.sp
libuv provides cross platform utilities for loading shared libraries and
retrieving symbols from them, using the following API.
.SS Data types
.INDENT 0.0
.TP
.B uv_lib_t
Shared library data type.
.UNINDENT
.SS Public members
.sp
N/A
.SS API
.INDENT 0.0
.TP
.B int uv_dlopen(const char*\fI\ filename\fP, uv_lib_t*\fI\ lib\fP)
Opens a shared library. The filename is in utf\-8. Returns 0 on success and
\-1 on error. Call \fI\%uv_dlerror()\fP to get the error message.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_dlclose(uv_lib_t*\fI\ lib\fP)
Close the shared library.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_dlsym(uv_lib_t*\fI\ lib\fP, const char*\fI\ name\fP, void**\fI\ ptr\fP)
Retrieves a data pointer from a dynamic library. It is legal for a symbol
to map to NULL. Returns 0 on success and \-1 if the symbol was not found.
.UNINDENT
.INDENT 0.0
.TP
.B const char* uv_dlerror(const uv_lib_t*\fI\ lib\fP)
Returns the last uv_dlopen() or uv_dlsym() error message.
.UNINDENT
.SS Threading and synchronization utilities
.sp
libuv provides cross\-platform implementations for multiple threading and
synchronization primitives. The API largely follows the pthreads API.
.SS Data types
.INDENT 0.0
.TP
.B uv_thread_t
Thread data type.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_thread_cb)(void*\fI\ arg\fP)
Callback that is invoked to initialize thread execution. \fIarg\fP is the same
value that was passed to \fI\%uv_thread_create()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B uv_key_t
Thread\-local key data type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_once_t
Once\-only initializer data type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_mutex_t
Mutex data type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_rwlock_t
Read\-write lock data type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_sem_t
Semaphore data type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_cond_t
Condition data type.
.UNINDENT
.INDENT 0.0
.TP
.B uv_barrier_t
Barrier data type.
.UNINDENT
.SS API
.SS Threads
.INDENT 0.0
.TP
.B uv_thread_options_t
Options for spawning a new thread (passed to \fI\%uv_thread_create_ex()\fP).
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_thread_options_s {
enum {
UV_THREAD_NO_FLAGS = 0x00,
UV_THREAD_HAS_STACK_SIZE = 0x01
} flags;
size_t stack_size;
} uv_thread_options_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
More fields may be added to this struct at any time, so its exact
layout and size should not be relied upon.
.sp
New in version 1.26.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_thread_create(uv_thread_t*\fI\ tid\fP, uv_thread_cb\fI\ entry\fP, void*\fI\ arg\fP)
Changed in version 1.4.1: returns a UV_E* error code on failure
.UNINDENT
.INDENT 0.0
.TP
.B int uv_thread_create_ex(uv_thread_t*\fI\ tid\fP, const uv_thread_options_t*\fI\ params\fP, uv_thread_cb\fI\ entry\fP, void*\fI\ arg\fP)
Like \fI\%uv_thread_create()\fP, but additionally specifies options for creating a new thread.
.sp
If \fIUV_THREAD_HAS_STACK_SIZE\fP is set, \fIstack_size\fP specifies a stack size for the new thread.
\fI0\fP indicates that the default value should be used, i.e. behaves as if the flag was not set.
Other values will be rounded up to the nearest page boundary.
.sp
New in version 1.26.0.
.UNINDENT
.INDENT 0.0
.TP
.B uv_thread_t uv_thread_self(void)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_thread_join(uv_thread_t\fI\ *tid\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_thread_equal(const uv_thread_t*\fI\ t1\fP, const uv_thread_t*\fI\ t2\fP)
.UNINDENT
.SS Thread\-local storage
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
The total thread\-local storage size may be limited. That is, it may not be possible to
create many TLS keys.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_key_create(uv_key_t*\fI\ key\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_key_delete(uv_key_t*\fI\ key\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void* uv_key_get(uv_key_t*\fI\ key\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_key_set(uv_key_t*\fI\ key\fP, void*\fI\ value\fP)
.UNINDENT
.SS Once\-only initialization
.sp
Runs a function once and only once. Concurrent calls to \fI\%uv_once()\fP with the
same guard will block all callers except one (it\(aqs unspecified which one).
The guard should be initialized statically with the UV_ONCE_INIT macro.
.INDENT 0.0
.TP
.B void uv_once(uv_once_t*\fI\ guard\fP, void (\fI*callback\fP)(void))
.UNINDENT
.SS Mutex locks
.sp
Functions return 0 on success or an error code < 0 (unless the
return type is void, of course).
.INDENT 0.0
.TP
.B int uv_mutex_init(uv_mutex_t*\fI\ handle\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_mutex_init_recursive(uv_mutex_t*\fI\ handle\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_mutex_destroy(uv_mutex_t*\fI\ handle\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_mutex_lock(uv_mutex_t*\fI\ handle\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_mutex_trylock(uv_mutex_t*\fI\ handle\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_mutex_unlock(uv_mutex_t*\fI\ handle\fP)
.UNINDENT
.SS Read\-write locks
.sp
Functions return 0 on success or an error code < 0 (unless the
return type is void, of course).
.INDENT 0.0
.TP
.B int uv_rwlock_init(uv_rwlock_t*\fI\ rwlock\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_rwlock_destroy(uv_rwlock_t*\fI\ rwlock\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_rwlock_rdlock(uv_rwlock_t*\fI\ rwlock\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_rwlock_tryrdlock(uv_rwlock_t*\fI\ rwlock\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_rwlock_rdunlock(uv_rwlock_t*\fI\ rwlock\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_rwlock_wrlock(uv_rwlock_t*\fI\ rwlock\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_rwlock_trywrlock(uv_rwlock_t*\fI\ rwlock\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_rwlock_wrunlock(uv_rwlock_t*\fI\ rwlock\fP)
.UNINDENT
.SS Semaphores
.sp
Functions return 0 on success or an error code < 0 (unless the
return type is void, of course).
.INDENT 0.0
.TP
.B int uv_sem_init(uv_sem_t*\fI\ sem\fP, unsigned int\fI\ value\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_sem_destroy(uv_sem_t*\fI\ sem\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_sem_post(uv_sem_t*\fI\ sem\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_sem_wait(uv_sem_t*\fI\ sem\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_sem_trywait(uv_sem_t*\fI\ sem\fP)
.UNINDENT
.SS Conditions
.sp
Functions return 0 on success or an error code < 0 (unless the
return type is void, of course).
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
.INDENT 0.0
.IP 1. 3
Callers should be prepared to deal with spurious wakeups on \fI\%uv_cond_wait()\fP
and \fI\%uv_cond_timedwait()\fP\&.
.IP 2. 3
The timeout parameter for \fI\%uv_cond_timedwait()\fP is relative to the time
at which function is called.
.IP 3. 3
On z/OS, the timeout parameter for \fI\%uv_cond_timedwait()\fP is converted to an
absolute system time at which the wait expires. If the current system clock time
passes the absolute time calculated before the condition is signaled, an ETIMEDOUT
error results. After the wait begins, the wait time is not affected by changes
to the system clock.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_cond_init(uv_cond_t*\fI\ cond\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_cond_destroy(uv_cond_t*\fI\ cond\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_cond_signal(uv_cond_t*\fI\ cond\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_cond_broadcast(uv_cond_t*\fI\ cond\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_cond_wait(uv_cond_t*\fI\ cond\fP, uv_mutex_t*\fI\ mutex\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_cond_timedwait(uv_cond_t*\fI\ cond\fP, uv_mutex_t*\fI\ mutex\fP, uint64_t\fI\ timeout\fP)
.UNINDENT
.SS Barriers
.sp
Functions return 0 on success or an error code < 0 (unless the
return type is void, of course).
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
\fI\%uv_barrier_wait()\fP returns a value > 0 to an arbitrarily chosen "serializer" thread
to facilitate cleanup, i.e.
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
if (uv_barrier_wait(&barrier) > 0)
uv_barrier_destroy(&barrier);
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_barrier_init(uv_barrier_t*\fI\ barrier\fP, unsigned int\fI\ count\fP)
.UNINDENT
.INDENT 0.0
.TP
.B void uv_barrier_destroy(uv_barrier_t*\fI\ barrier\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_barrier_wait(uv_barrier_t*\fI\ barrier\fP)
.UNINDENT
.SS Miscellaneous utilities
.sp
This section contains miscellaneous functions that don\(aqt really belong in any
other section.
.SS Data types
.INDENT 0.0
.TP
.B uv_buf_t
Buffer data type.
.INDENT 7.0
.TP
.B char* uv_buf_t.base
Pointer to the base of the buffer.
.UNINDENT
.INDENT 7.0
.TP
.B size_t uv_buf_t.len
Total bytes in the buffer.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Windows this field is ULONG.
.UNINDENT
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void* (*uv_malloc_func)(size_t\fI\ size\fP)
Replacement function for \fI\%malloc(3)\fP\&.
See \fI\%uv_replace_allocator()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void* (*uv_realloc_func)(void*\fI\ ptr\fP, size_t\fI\ size\fP)
Replacement function for \fI\%realloc(3)\fP\&.
See \fI\%uv_replace_allocator()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void* (*uv_calloc_func)(size_t\fI\ count\fP, size_t\fI\ size\fP)
Replacement function for \fI\%calloc(3)\fP\&.
See \fI\%uv_replace_allocator()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_free_func)(void*\fI\ ptr\fP)
Replacement function for \fI\%free(3)\fP\&.
See \fI\%uv_replace_allocator()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void (*uv_random_cb)(uv_random_t*\fI\ req\fP, int\fI\ status\fP, void*\fI\ buf\fP, size_t\fI\ buflen\fP)
Callback passed to \fI\%uv_random()\fP\&. \fIstatus\fP is non\-zero in case of
error. The \fIbuf\fP pointer is the same pointer that was passed to
\fI\%uv_random()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B uv_file
Cross platform representation of a file handle.
.UNINDENT
.INDENT 0.0
.TP
.B uv_os_sock_t
Cross platform representation of a socket handle.
.UNINDENT
.INDENT 0.0
.TP
.B uv_os_fd_t
Abstract representation of a file descriptor. On Unix systems this is a
\fItypedef\fP of \fIint\fP and on Windows a \fIHANDLE\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B uv_pid_t
Cross platform representation of a \fIpid_t\fP\&.
.sp
New in version 1.16.0.
.UNINDENT
.INDENT 0.0
.TP
.B uv_timeval_t
Data type for storing times.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct {
long tv_sec;
long tv_usec;
} uv_timeval_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_timeval64_t
Alternative data type for storing times.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct {
int64_t tv_sec;
int32_t tv_usec;
} uv_timeval64_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_rusage_t
Data type for resource usage results.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct {
uv_timeval_t ru_utime; /* user CPU time used */
uv_timeval_t ru_stime; /* system CPU time used */
uint64_t ru_maxrss; /* maximum resident set size */
uint64_t ru_ixrss; /* integral shared memory size (X) */
uint64_t ru_idrss; /* integral unshared data size (X) */
uint64_t ru_isrss; /* integral unshared stack size (X) */
uint64_t ru_minflt; /* page reclaims (soft page faults) (X) */
uint64_t ru_majflt; /* page faults (hard page faults) */
uint64_t ru_nswap; /* swaps (X) */
uint64_t ru_inblock; /* block input operations */
uint64_t ru_oublock; /* block output operations */
uint64_t ru_msgsnd; /* IPC messages sent (X) */
uint64_t ru_msgrcv; /* IPC messages received (X) */
uint64_t ru_nsignals; /* signals received (X) */
uint64_t ru_nvcsw; /* voluntary context switches (X) */
uint64_t ru_nivcsw; /* involuntary context switches (X) */
} uv_rusage_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Members marked with \fI(X)\fP are unsupported on Windows.
See \fI\%getrusage(2)\fP for supported fields on Unix
.UNINDENT
.INDENT 0.0
.TP
.B uv_cpu_info_t
Data type for CPU information.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_cpu_info_s {
char* model;
int speed;
struct uv_cpu_times_s {
uint64_t user; /* milliseconds */
uint64_t nice; /* milliseconds */
uint64_t sys; /* milliseconds */
uint64_t idle; /* milliseconds */
uint64_t irq; /* milliseconds */
} cpu_times;
} uv_cpu_info_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_interface_address_t
Data type for interface addresses.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_interface_address_s {
char* name;
char phys_addr[6];
int is_internal;
union {
struct sockaddr_in address4;
struct sockaddr_in6 address6;
} address;
union {
struct sockaddr_in netmask4;
struct sockaddr_in6 netmask6;
} netmask;
} uv_interface_address_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_passwd_t
Data type for password file information.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_passwd_s {
char* username;
long uid;
long gid;
char* shell;
char* homedir;
} uv_passwd_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_utsname_t
Data type for operating system name and version information.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_utsname_s {
char sysname[256];
char release[256];
char version[256];
char machine[256];
} uv_utsname_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_env_item_t
Data type for environment variable storage.
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_env_item_s {
char* name;
char* value;
} uv_env_item_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_random_t
Random data request type.
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B uv_handle_type uv_guess_handle(uv_file\fI\ file\fP)
Used to detect what type of stream should be used with a given file
descriptor. Usually this will be used during initialization to guess the
type of the stdio streams.
.sp
For \fI\%isatty(3)\fP equivalent functionality use this function and test
for \fBUV_TTY\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_replace_allocator(uv_malloc_func\fI\ malloc_func\fP, uv_realloc_func\fI\ realloc_func\fP, uv_calloc_func\fI\ calloc_func\fP, uv_free_func\fI\ free_func\fP)
New in version 1.6.0.
.sp
Override the use of the standard library\(aqs \fI\%malloc(3)\fP,
\fI\%calloc(3)\fP, \fI\%realloc(3)\fP, \fI\%free(3)\fP, memory allocation
functions.
.sp
This function must be called before any other libuv function is called or
after all resources have been freed and thus libuv doesn\(aqt reference
any allocated memory chunk.
.sp
On success, it returns 0, if any of the function pointers is NULL it
returns UV_EINVAL.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
There is no protection against changing the allocator multiple
times. If the user changes it they are responsible for making
sure the allocator is changed while no memory was allocated with
the previous allocator, or that they are compatible.
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
Allocator must be thread\-safe.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void uv_library_shutdown(void);
New in version 1.38.0.
.sp
Release any global state that libuv is holding onto. Libuv will normally
do so automatically when it is unloaded but it can be instructed to perform
cleanup manually.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
Only call \fBuv_library_shutdown()\fP once.
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
Don\(aqt call \fBuv_library_shutdown()\fP when there are
still event loops or I/O requests active.
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
Don\(aqt call libuv functions after calling
\fBuv_library_shutdown()\fP\&.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_buf_t uv_buf_init(char*\fI\ base\fP, unsigned int\fI\ len\fP)
Constructor for \fI\%uv_buf_t\fP\&.
.sp
Due to platform differences the user cannot rely on the ordering of the
\fIbase\fP and \fIlen\fP members of the uv_buf_t struct. The user is responsible for
freeing \fIbase\fP after the uv_buf_t is done. Return struct passed by value.
.UNINDENT
.INDENT 0.0
.TP
.B char** uv_setup_args(int\fI\ argc\fP, char**\fI\ argv\fP)
Store the program arguments. Required for getting / setting the process title.
Libuv may take ownership of the memory that \fIargv\fP points to. This function
should be called exactly once, at program start\-up.
.sp
Example:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
argv = uv_setup_args(argc, argv); /* May return a copy of argv. */
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_get_process_title(char*\fI\ buffer\fP, size_t\fI\ size\fP)
Gets the title of the current process. You \fImust\fP call \fIuv_setup_args\fP
before calling this function. If \fIbuffer\fP is \fINULL\fP or \fIsize\fP is zero,
\fIUV_EINVAL\fP is returned. If \fIsize\fP cannot accommodate the process title and
terminating \fINULL\fP character, the function returns \fIUV_ENOBUFS\fP\&.
.sp
Changed in version 1.18.1: now thread\-safe on all supported platforms.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_set_process_title(const char*\fI\ title\fP)
Sets the current process title. You \fImust\fP call \fIuv_setup_args\fP before
calling this function. On platforms with a fixed size buffer for the process
title the contents of \fItitle\fP will be copied to the buffer and truncated if
larger than the available space. Other platforms will return \fIUV_ENOMEM\fP if
they cannot allocate enough space to duplicate the contents of \fItitle\fP\&.
.sp
Changed in version 1.18.1: now thread\-safe on all supported platforms.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_resident_set_memory(size_t*\fI\ rss\fP)
Gets the resident set size (RSS) for the current process.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_uptime(double*\fI\ uptime\fP)
Gets the current system uptime.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_getrusage(uv_rusage_t*\fI\ rusage\fP)
Gets the resource usage measures for the current process.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Windows not all fields are set, the unsupported fields are filled with zeroes.
See \fI\%uv_rusage_t\fP for more details.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B uv_pid_t uv_os_getpid(void)
Returns the current process ID.
.sp
New in version 1.18.0.
.UNINDENT
.INDENT 0.0
.TP
.B uv_pid_t uv_os_getppid(void)
Returns the parent process ID.
.sp
New in version 1.16.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_cpu_info(uv_cpu_info_t**\fI\ cpu_infos\fP, int*\fI\ count\fP)
Gets information about the CPUs on the system. The \fIcpu_infos\fP array will
have \fIcount\fP elements and needs to be freed with \fI\%uv_free_cpu_info()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_free_cpu_info(uv_cpu_info_t*\fI\ cpu_infos\fP, int\fI\ count\fP)
Frees the \fIcpu_infos\fP array previously allocated with \fI\%uv_cpu_info()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_interface_addresses(uv_interface_address_t**\fI\ addresses\fP, int*\fI\ count\fP)
Gets address information about the network interfaces on the system. An
array of \fIcount\fP elements is allocated and returned in \fIaddresses\fP\&. It must
be freed by the user, calling \fI\%uv_free_interface_addresses()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_free_interface_addresses(uv_interface_address_t*\fI\ addresses\fP, int\fI\ count\fP)
Free an array of \fI\%uv_interface_address_t\fP which was returned by
\fI\%uv_interface_addresses()\fP\&.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_loadavg(double\fI\ avg[3]\fP)
Gets the load average. See: \fI\%https://en.wikipedia.org/wiki/Load_(computing)\fP
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Returns [0,0,0] on Windows (i.e., it\(aqs not implemented).
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B int uv_ip4_addr(const char*\fI\ ip\fP, int\fI\ port\fP, struct sockaddr_in*\fI\ addr\fP)
Convert a string containing an IPv4 addresses to a binary structure.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_ip6_addr(const char*\fI\ ip\fP, int\fI\ port\fP, struct sockaddr_in6*\fI\ addr\fP)
Convert a string containing an IPv6 addresses to a binary structure.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_ip4_name(const struct sockaddr_in*\fI\ src\fP, char*\fI\ dst\fP, size_t\fI\ size\fP)
Convert a binary structure containing an IPv4 address to a string.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_ip6_name(const struct sockaddr_in6*\fI\ src\fP, char*\fI\ dst\fP, size_t\fI\ size\fP)
Convert a binary structure containing an IPv6 address to a string.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_inet_ntop(int\fI\ af\fP, const void*\fI\ src\fP, char*\fI\ dst\fP, size_t\fI\ size\fP)
.UNINDENT
.INDENT 0.0
.TP
.B int uv_inet_pton(int\fI\ af\fP, const char*\fI\ src\fP, void*\fI\ dst\fP)
Cross\-platform IPv6\-capable implementation of \fI\%inet_ntop(3)\fP
and \fI\%inet_pton(3)\fP\&. On success they return 0. In case of error
the target \fIdst\fP pointer is unmodified.
.UNINDENT
.INDENT 0.0
.TP
.B UV_IF_NAMESIZE
Maximum IPv6 interface identifier name length. Defined as
\fIIFNAMSIZ\fP on Unix and \fIIF_NAMESIZE\fP on Linux and Windows.
.sp
New in version 1.16.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_if_indextoname(unsigned int\fI\ ifindex\fP, char*\fI\ buffer\fP, size_t*\fI\ size\fP)
IPv6\-capable implementation of \fI\%if_indextoname(3)\fP\&. When called,
\fI*size\fP indicates the length of the \fIbuffer\fP, which is used to store the
result.
On success, zero is returned, \fIbuffer\fP contains the interface name, and
\fI*size\fP represents the string length of the \fIbuffer\fP, excluding the NUL
terminator byte from \fI*size\fP\&. On error, a negative result is
returned. If \fIbuffer\fP is not large enough to hold the result,
\fIUV_ENOBUFS\fP is returned, and \fI*size\fP represents the necessary size in
bytes, including the NUL terminator byte into the \fI*size\fP\&.
.sp
On Unix, the returned interface name can be used directly as an
interface identifier in scoped IPv6 addresses, e.g.
\fIfe80::abc:def1:2345%en0\fP\&.
.sp
On Windows, the returned interface cannot be used as an interface
identifier, as Windows uses numerical interface identifiers, e.g.
\fIfe80::abc:def1:2345%5\fP\&.
.sp
To get an interface identifier in a cross\-platform compatible way,
use \fIuv_if_indextoiid()\fP\&.
.sp
Example:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
char ifname[UV_IF_NAMESIZE];
size_t size = sizeof(ifname);
uv_if_indextoname(sin6\->sin6_scope_id, ifname, &size);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
New in version 1.16.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_if_indextoiid(unsigned int\fI\ ifindex\fP, char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Retrieves a network interface identifier suitable for use in an IPv6 scoped
address. On Windows, returns the numeric \fIifindex\fP as a string. On all other
platforms, \fIuv_if_indextoname()\fP is called. The result is written to
\fIbuffer\fP, with \fI*size\fP indicating the length of \fIbuffer\fP\&. If \fIbuffer\fP is not
large enough to hold the result, then \fIUV_ENOBUFS\fP is returned, and \fI*size\fP
represents the size, including the NUL byte, required to hold the
result.
.sp
See \fIuv_if_indextoname\fP for further details.
.sp
New in version 1.16.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_exepath(char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Gets the executable path.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_cwd(char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Gets the current working directory, and stores it in \fIbuffer\fP\&. If the
current working directory is too large to fit in \fIbuffer\fP, this function
returns \fIUV_ENOBUFS\fP, and sets \fIsize\fP to the required length, including the
null terminator.
.sp
Changed in version 1.1.0: On Unix the path no longer ends in a slash.
.sp
Changed in version 1.9.0: the returned length includes the terminating null
byte on \fIUV_ENOBUFS\fP, and the buffer is null terminated
on success.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_chdir(const char*\fI\ dir\fP)
Changes the current working directory.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_homedir(char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Gets the current user\(aqs home directory. On Windows, \fIuv_os_homedir()\fP first
checks the \fIUSERPROFILE\fP environment variable using
\fIGetEnvironmentVariableW()\fP\&. If \fIUSERPROFILE\fP is not set,
\fIGetUserProfileDirectoryW()\fP is called. On all other operating systems,
\fIuv_os_homedir()\fP first checks the \fIHOME\fP environment variable using
\fI\%getenv(3)\fP\&. If \fIHOME\fP is not set, \fI\%getpwuid_r(3)\fP is called. The
user\(aqs home directory is stored in \fIbuffer\fP\&. When \fIuv_os_homedir()\fP is
called, \fIsize\fP indicates the maximum size of \fIbuffer\fP\&. On success \fIsize\fP is set
to the string length of \fIbuffer\fP\&. On \fIUV_ENOBUFS\fP failure \fIsize\fP is set to the
required length for \fIbuffer\fP, including the null byte.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
\fIuv_os_homedir()\fP is not thread safe.
.UNINDENT
.UNINDENT
.sp
New in version 1.6.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_tmpdir(char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Gets the temp directory. On Windows, \fIuv_os_tmpdir()\fP uses \fIGetTempPathW()\fP\&.
On all other operating systems, \fIuv_os_tmpdir()\fP uses the first environment
variable found in the ordered list \fITMPDIR\fP, \fITMP\fP, \fITEMP\fP, and \fITEMPDIR\fP\&.
If none of these are found, the path \fI"/tmp"\fP is used, or, on Android,
\fI"/data/local/tmp"\fP is used. The temp directory is stored in \fIbuffer\fP\&. When
\fIuv_os_tmpdir()\fP is called, \fIsize\fP indicates the maximum size of \fIbuffer\fP\&.
On success \fIsize\fP is set to the string length of \fIbuffer\fP (which does not
include the terminating null). On \fIUV_ENOBUFS\fP failure \fIsize\fP is set to the
required length for \fIbuffer\fP, including the null byte.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
\fIuv_os_tmpdir()\fP is not thread safe.
.UNINDENT
.UNINDENT
.sp
New in version 1.9.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_get_passwd(uv_passwd_t*\fI\ pwd\fP)
Gets a subset of the password file entry for the current effective uid (not
the real uid). The populated data includes the username, euid, gid, shell,
and home directory. On non\-Windows systems, all data comes from
\fI\%getpwuid_r(3)\fP\&. On Windows, uid and gid are set to \-1 and have no
meaning, and shell is \fINULL\fP\&. After successfully calling this function, the
memory allocated to \fIpwd\fP needs to be freed with
\fI\%uv_os_free_passwd()\fP\&.
.sp
New in version 1.9.0.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_os_free_passwd(uv_passwd_t*\fI\ pwd\fP)
Frees the \fIpwd\fP memory previously allocated with \fI\%uv_os_get_passwd()\fP\&.
.sp
New in version 1.9.0.
.UNINDENT
.INDENT 0.0
.TP
.B uint64_t uv_get_free_memory(void)
Gets memory information (in bytes).
.UNINDENT
.INDENT 0.0
.TP
.B uint64_t uv_get_total_memory(void)
Gets memory information (in bytes).
.UNINDENT
.INDENT 0.0
.TP
.B uint64_t uv_get_constrained_memory(void)
Gets the amount of memory available to the process (in bytes) based on
limits imposed by the OS. If there is no such constraint, or the constraint
is unknown, \fI0\fP is returned. Note that it is not unusual for this value to
be less than or greater than \fI\%uv_get_total_memory()\fP\&.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
This function currently only returns a non\-zero value on Linux, based
on cgroups if it is present.
.UNINDENT
.UNINDENT
.sp
New in version 1.29.0.
.UNINDENT
.INDENT 0.0
.TP
.B uint64_t uv_hrtime(void)
Returns the current high\-resolution real time. This is expressed in
nanoseconds. It is relative to an arbitrary time in the past. It is not
related to the time of day and therefore not subject to clock drift. The
primary use is for measuring performance between intervals.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
Not every platform can support nanosecond resolution; however, this value will always
be in nanoseconds.
.UNINDENT
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B void uv_print_all_handles(uv_loop_t*\fI\ loop\fP, FILE*\fI\ stream\fP)
Prints all handles associated with the given \fIloop\fP to the given \fIstream\fP\&.
.sp
Example:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
uv_print_all_handles(uv_default_loop(), stderr);
/*
[\-\-I] signal 0x1a25ea8
[\-AI] async 0x1a25cf0
[R\-\-] idle 0x1a7a8c8
*/
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The format is \fI[flags] handle\-type handle\-address\fP\&. For \fIflags\fP:
.INDENT 7.0
.IP \(bu 2
\fIR\fP is printed for a handle that is referenced
.IP \(bu 2
\fIA\fP is printed for a handle that is active
.IP \(bu 2
\fII\fP is printed for a handle that is internal
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
This function is meant for ad hoc debugging, there is no API/ABI
stability guarantees.
.UNINDENT
.UNINDENT
.sp
New in version 1.8.0.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_print_active_handles(uv_loop_t*\fI\ loop\fP, FILE*\fI\ stream\fP)
This is the same as \fI\%uv_print_all_handles()\fP except only active handles
are printed.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
This function is meant for ad hoc debugging, there is no API/ABI
stability guarantees.
.UNINDENT
.UNINDENT
.sp
New in version 1.8.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_environ(uv_env_item_t**\fI\ envitems\fP, int*\fI\ count\fP)
Retrieves all environment variables. This function will allocate memory
which must be freed by calling \fBuv_os_free_environ()\fP\&.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
This function is not thread safe.
.UNINDENT
.UNINDENT
.sp
New in version 1.31.0.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_os_free_environ(uv_env_item_t* envitems, int count);
Frees the memory allocated for the environment variables by
\fI\%uv_os_environ()\fP\&.
.sp
New in version 1.31.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_getenv(const char*\fI\ name\fP, char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Retrieves the environment variable specified by \fIname\fP, copies its value
into \fIbuffer\fP, and sets \fIsize\fP to the string length of the value. When
calling this function, \fIsize\fP must be set to the amount of storage available
in \fIbuffer\fP, including the null terminator. If the environment variable
exceeds the storage available in \fIbuffer\fP, \fIUV_ENOBUFS\fP is returned, and
\fIsize\fP is set to the amount of storage required to hold the value. If no
matching environment variable exists, \fIUV_ENOENT\fP is returned.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
This function is not thread safe.
.UNINDENT
.UNINDENT
.sp
New in version 1.12.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_setenv(const char*\fI\ name\fP, const char*\fI\ value\fP)
Creates or updates the environment variable specified by \fIname\fP with
\fIvalue\fP\&.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
This function is not thread safe.
.UNINDENT
.UNINDENT
.sp
New in version 1.12.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_unsetenv(const char*\fI\ name\fP)
Deletes the environment variable specified by \fIname\fP\&. If no such environment
variable exists, this function returns successfully.
.sp
\fBWARNING:\fP
.INDENT 7.0
.INDENT 3.5
This function is not thread safe.
.UNINDENT
.UNINDENT
.sp
New in version 1.12.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_gethostname(char*\fI\ buffer\fP, size_t*\fI\ size\fP)
Returns the hostname as a null\-terminated string in \fIbuffer\fP, and sets
\fIsize\fP to the string length of the hostname. When calling this function,
\fIsize\fP must be set to the amount of storage available in \fIbuffer\fP, including
the null terminator. If the hostname exceeds the storage available in
\fIbuffer\fP, \fIUV_ENOBUFS\fP is returned, and \fIsize\fP is set to the amount of
storage required to hold the value.
.sp
New in version 1.12.0.
.sp
Changed in version 1.26.0: \fIUV_MAXHOSTNAMESIZE\fP is available and represents
the maximum \fIbuffer\fP size required to store a
hostname and terminating \fInul\fP character.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_getpriority(uv_pid_t\fI\ pid\fP, int*\fI\ priority\fP)
Retrieves the scheduling priority of the process specified by \fIpid\fP\&. The
returned value of \fIpriority\fP is between \-20 (high priority) and 19 (low
priority).
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Windows, the returned priority will equal one of the \fIUV_PRIORITY\fP
constants.
.UNINDENT
.UNINDENT
.sp
New in version 1.23.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_setpriority(uv_pid_t\fI\ pid\fP, int\fI\ priority\fP)
Sets the scheduling priority of the process specified by \fIpid\fP\&. The
\fIpriority\fP value range is between \-20 (high priority) and 19 (low priority).
The constants \fIUV_PRIORITY_LOW\fP, \fIUV_PRIORITY_BELOW_NORMAL\fP,
\fIUV_PRIORITY_NORMAL\fP, \fIUV_PRIORITY_ABOVE_NORMAL\fP, \fIUV_PRIORITY_HIGH\fP, and
\fIUV_PRIORITY_HIGHEST\fP are also provided for convenience.
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Windows, this function utilizes \fISetPriorityClass()\fP\&. The \fIpriority\fP
argument is mapped to a Windows priority class. When retrieving the
process priority, the result will equal one of the \fIUV_PRIORITY\fP
constants, and not necessarily the exact value of \fIpriority\fP\&.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On Windows, setting \fIPRIORITY_HIGHEST\fP will only work for elevated user,
for others it will be silently reduced to \fIPRIORITY_HIGH\fP\&.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On IBM i PASE, the highest process priority is \-10. The constant
\fIUV_PRIORITY_HIGHEST\fP is \-10, \fIUV_PRIORITY_HIGH\fP is \-7,
\fIUV_PRIORITY_ABOVE_NORMAL\fP is \-4, \fIUV_PRIORITY_NORMAL\fP is 0,
\fIUV_PRIORITY_BELOW_NORMAL\fP is 15 and \fIUV_PRIORITY_LOW\fP is 39.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
On IBM i PASE, you are not allowed to change your priority unless you
have the *JOBCTL special authority (even to lower it).
.UNINDENT
.UNINDENT
.sp
New in version 1.23.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_os_uname(uv_utsname_t*\fI\ buffer\fP)
Retrieves system information in \fIbuffer\fP\&. The populated data includes the
operating system name, release, version, and machine. On non\-Windows
systems, \fIuv_os_uname()\fP is a thin wrapper around \fI\%uname(2)\fP\&. Returns
zero on success, and a non\-zero error value otherwise.
.sp
New in version 1.25.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_gettimeofday(uv_timeval64_t*\fI\ tv\fP)
Cross\-platform implementation of \fI\%gettimeofday(2)\fP\&. The timezone
argument to \fIgettimeofday()\fP is not supported, as it is considered obsolete.
.sp
New in version 1.28.0.
.UNINDENT
.INDENT 0.0
.TP
.B int uv_random(uv_loop_t*\fI\ loop\fP, uv_random_t*\fI\ req\fP, void*\fI\ buf\fP, size_t\fI\ buflen\fP, unsigned int\fI\ flags\fP, uv_random_cb\fI\ cb\fP)
Fill \fIbuf\fP with exactly \fIbuflen\fP cryptographically strong random bytes
acquired from the system CSPRNG. \fIflags\fP is reserved for future extension
and must currently be 0.
.sp
Short reads are not possible. When less than \fIbuflen\fP random bytes are
available, a non\-zero error value is returned or passed to the callback.
.sp
The synchronous version may block indefinitely when not enough entropy
is available. The asynchronous version may not ever finish when the system
is low on entropy.
.sp
Sources of entropy:
.INDENT 7.0
.IP \(bu 2
Windows: \fIRtlGenRandom <https://docs.microsoft.com/en\-us/windows/desktop/api/ntsecapi/nf\-ntsecapi\-rtlgenrandom>_\fP\&.
.IP \(bu 2
Linux, Android: \fI\%getrandom(2)\fP if available, or \fI\%urandom(4)\fP
after reading from \fI/dev/random\fP once, or the \fIKERN_RANDOM\fP
\fI\%sysctl(2)\fP\&.
.IP \(bu 2
FreeBSD: \fIgetrandom(2) <https://www.freebsd.org/cgi/man.cgi?query=getrandom&sektion=2>_\fP,
or \fI/dev/urandom\fP after reading from \fI/dev/random\fP once.
.IP \(bu 2
NetBSD: \fIKERN_ARND\fP \fIsysctl(3) <https://netbsd.gw.com/cgi\-bin/man\-cgi?sysctl+3+NetBSD\-current>_\fP
.IP \(bu 2
macOS, OpenBSD: \fIgetentropy(2) <https://man.openbsd.org/getentropy.2>_\fP
if available, or \fI/dev/urandom\fP after reading from \fI/dev/random\fP once.
.IP \(bu 2
AIX: \fI/dev/random\fP\&.
.IP \(bu 2
IBM i: \fI/dev/urandom\fP\&.
.IP \(bu 2
Other UNIX: \fI/dev/urandom\fP after reading from \fI/dev/random\fP once.
.UNINDENT
.INDENT 7.0
.TP
.B Returns
0 on success, or an error code < 0 on failure. The contents of
\fIbuf\fP is undefined after an error.
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 7.0
.INDENT 3.5
When using the synchronous version, both \fIloop\fP and \fIreq\fP parameters
are not used and can be set to \fINULL\fP\&.
.UNINDENT
.UNINDENT
.sp
New in version 1.33.0.
.UNINDENT
.INDENT 0.0
.TP
.B void uv_sleep(unsigned int\fI\ msec\fP)
Causes the calling thread to sleep for \fImsec\fP milliseconds.
.sp
New in version 1.34.0.
.UNINDENT
.SS User guide
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
The contents of this guide have been recently incorporated into the libuv documentation
and it hasn\(aqt gone through thorough review yet. If you spot a mistake please file an
issue, or better yet, open a pull request!
.UNINDENT
.UNINDENT
.SS Introduction
.sp
This \(aqbook\(aq is a small set of tutorials about using \fI\%libuv\fP as
a high performance evented I/O library which offers the same API on Windows and Unix.
.sp
It is meant to cover the main areas of libuv, but is not a comprehensive
reference discussing every function and data structure. The \fI\%official libuv
documentation\fP may be consulted for full details.
.sp
This book is still a work in progress, so sections may be incomplete, but
I hope you will enjoy it as it grows.
.SS Who this book is for
.sp
If you are reading this book, you are either:
.INDENT 0.0
.IP 1. 3
a systems programmer, creating low\-level programs such as daemons or network
services and clients. You have found that the event loop approach is well
suited for your application and decided to use libuv.
.IP 2. 3
a node.js module writer, who wants to wrap platform APIs
written in C or C++ with a set of (a)synchronous APIs that are exposed to
JavaScript. You will use libuv purely in the context of node.js. For
this you will require some other resources as the book does not cover parts
specific to v8/node.js.
.UNINDENT
.sp
This book assumes that you are comfortable with the C programming language.
.SS Background
.sp
The \fI\%node.js\fP project began in 2009 as a JavaScript environment decoupled
from the browser. Using Google\(aqs \fI\%V8\fP and Marc Lehmann\(aqs \fI\%libev\fP, node.js
combined a model of I/O \-\- evented \-\- with a language that was well suited to
the style of programming; due to the way it had been shaped by browsers. As
node.js grew in popularity, it was important to make it work on Windows, but
libev ran only on Unix. The Windows equivalent of kernel event notification
mechanisms like kqueue or (e)poll is IOCP. libuv was an abstraction around libev
or IOCP depending on the platform, providing users an API based on libev.
In the node\-v0.9.0 version of libuv \fI\%libev was removed\fP\&.
.sp
Since then libuv has continued to mature and become a high quality standalone
library for system programming. Users outside of node.js include Mozilla\(aqs
\fI\%Rust\fP programming language, and a \fI\%variety\fP of language bindings.
.sp
This book and the code is based on libuv version \fI\%v1.3.0\fP\&.
.SS Code
.sp
All the code from this book is included as part of the source of the book on
Github. \fI\%Clone\fP/\fI\%Download\fP the book, then build libuv:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
cd libuv
\&./autogen.sh
\&./configure
make
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
There is no need to \fBmake install\fP\&. To build the examples run \fBmake\fP in the
\fBcode/\fP directory.
.SS Basics of libuv
.sp
libuv enforces an \fBasynchronous\fP, \fBevent\-driven\fP style of programming. Its
core job is to provide an event loop and callback based notifications of I/O
and other activities. libuv offers core utilities like timers, non\-blocking
networking support, asynchronous file system access, child processes and more.
.SS Event loops
.sp
In event\-driven programming, an application expresses interest in certain events
and respond to them when they occur. The responsibility of gathering events
from the operating system or monitoring other sources of events is handled by
libuv, and the user can register callbacks to be invoked when an event occurs.
The event\-loop usually keeps running \fIforever\fP\&. In pseudocode:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
while there are still events to process:
e = get the next event
if there is a callback associated with e:
call the callback
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Some examples of events are:
.INDENT 0.0
.IP \(bu 2
File is ready for writing
.IP \(bu 2
A socket has data ready to be read
.IP \(bu 2
A timer has timed out
.UNINDENT
.sp
This event loop is encapsulated by \fBuv_run()\fP \-\- the end\-all function when using
libuv.
.sp
The most common activity of systems programs is to deal with input and output,
rather than a lot of number\-crunching. The problem with using conventional
input/output functions (\fBread\fP, \fBfprintf\fP, etc.) is that they are
\fBblocking\fP\&. The actual write to a hard disk or reading from a network, takes
a disproportionately long time compared to the speed of the processor. The
functions don\(aqt return until the task is done, so that your program is doing
nothing. For programs which require high performance this is a major roadblock
as other activities and other I/O operations are kept waiting.
.sp
One of the standard solutions is to use threads. Each blocking I/O operation is
started in a separate thread (or in a thread pool). When the blocking function
gets invoked in the thread, the processor can schedule another thread to run,
which actually needs the CPU.
.sp
The approach followed by libuv uses another style, which is the \fBasynchronous,
non\-blocking\fP style. Most modern operating systems provide event notification
subsystems. For example, a normal \fBread\fP call on a socket would block until
the sender actually sent something. Instead, the application can request the
operating system to watch the socket and put an event notification in the
queue. The application can inspect the events at its convenience (perhaps doing
some number crunching before to use the processor to the maximum) and grab the
data. It is \fBasynchronous\fP because the application expressed interest at one
point, then used the data at another point (in time and space). It is
\fBnon\-blocking\fP because the application process was free to do other tasks.
This fits in well with libuv\(aqs event\-loop approach, since the operating system
events can be treated as just another libuv event. The non\-blocking ensures
that other events can continue to be handled as fast as they come in [1]\&.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
How the I/O is run in the background is not of our concern, but due to the
way our computer hardware works, with the thread as the basic unit of the
processor, libuv and OSes will usually run background/worker threads and/or
polling to perform tasks in a non\-blocking manner.
.UNINDENT
.UNINDENT
.sp
Bert Belder, one of the libuv core developers has a small video explaining the
architecture of libuv and its background. If you have no prior experience with
either libuv or libev, it is a quick, useful watch.
.sp
libuv\(aqs event loop is explained in more detail in the \fI\%documentation\fP\&.
.SS Hello World
.sp
With the basics out of the way, let\(aqs write our first libuv program. It does
nothing, except start a loop which will exit immediately.
.sp
helloworld/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <stdlib.h>
#include <uv.h>
int main() {
uv_loop_t *loop = malloc(sizeof(uv_loop_t));
uv_loop_init(loop);
printf("Now quitting.\en");
uv_run(loop, UV_RUN_DEFAULT);
uv_loop_close(loop);
free(loop);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
This program quits immediately because it has no events to process. A libuv
event loop has to be told to watch out for events using the various API
functions.
.sp
Starting with libuv v1.0, users should allocate the memory for the loops before
initializing it with \fBuv_loop_init(uv_loop_t *)\fP\&. This allows you to plug in
custom memory management. Remember to de\-initialize the loop using
\fBuv_loop_close(uv_loop_t *)\fP and then delete the storage. The examples never
close loops since the program quits after the loop ends and the system will
reclaim memory. Production grade projects, especially long running systems
programs, should take care to release correctly.
.SS Default loop
.sp
A default loop is provided by libuv and can be accessed using
\fBuv_default_loop()\fP\&. You should use this loop if you only want a single
loop.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
node.js uses the default loop as its main loop. If you are writing bindings
you should be aware of this.
.UNINDENT
.UNINDENT
.SS Error handling
.sp
Initialization functions or synchronous functions which may fail return a negative number on error. Async functions that may fail will pass a status parameter to their callbacks. The error messages are defined as \fBUV_E*\fP \fI\%constants\fP\&.
.sp
You can use the \fBuv_strerror(int)\fP and \fBuv_err_name(int)\fP functions
to get a \fBconst char *\fP describing the error or the error name respectively.
.sp
I/O read callbacks (such as for files and sockets) are passed a parameter \fBnread\fP\&. If \fBnread\fP is less than 0, there was an error (UV_EOF is the end of file error, which you may want to handle differently).
.SS Handles and Requests
.sp
libuv works by the user expressing interest in particular events. This is
usually done by creating a \fBhandle\fP to an I/O device, timer or process.
Handles are opaque structs named as \fBuv_TYPE_t\fP where type signifies what the
handle is used for.
.sp
libuv watchers
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
/* Handle types. */
typedef struct uv_loop_s uv_loop_t;
typedef struct uv_handle_s uv_handle_t;
typedef struct uv_dir_s uv_dir_t;
typedef struct uv_stream_s uv_stream_t;
typedef struct uv_tcp_s uv_tcp_t;
typedef struct uv_udp_s uv_udp_t;
typedef struct uv_pipe_s uv_pipe_t;
typedef struct uv_tty_s uv_tty_t;
typedef struct uv_poll_s uv_poll_t;
typedef struct uv_timer_s uv_timer_t;
typedef struct uv_prepare_s uv_prepare_t;
typedef struct uv_check_s uv_check_t;
typedef struct uv_idle_s uv_idle_t;
typedef struct uv_async_s uv_async_t;
typedef struct uv_process_s uv_process_t;
typedef struct uv_fs_event_s uv_fs_event_t;
typedef struct uv_fs_poll_s uv_fs_poll_t;
typedef struct uv_signal_s uv_signal_t;
/* Request types. */
typedef struct uv_req_s uv_req_t;
typedef struct uv_getaddrinfo_s uv_getaddrinfo_t;
typedef struct uv_getnameinfo_s uv_getnameinfo_t;
typedef struct uv_shutdown_s uv_shutdown_t;
typedef struct uv_write_s uv_write_t;
typedef struct uv_connect_s uv_connect_t;
typedef struct uv_udp_send_s uv_udp_send_t;
typedef struct uv_fs_s uv_fs_t;
typedef struct uv_work_s uv_work_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Handles represent long\-lived objects. Async operations on such handles are
identified using \fBrequests\fP\&. A request is short\-lived (usually used across
only one callback) and usually indicates one I/O operation on a handle.
Requests are used to preserve context between the initiation and the callback
of individual actions. For example, an UDP socket is represented by
a \fBuv_udp_t\fP, while individual writes to the socket use a \fBuv_udp_send_t\fP
structure that is passed to the callback after the write is done.
.sp
Handles are setup by a corresponding:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_TYPE_init(uv_loop_t *, uv_TYPE_t *)
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
function.
.sp
Callbacks are functions which are called by libuv whenever an event the watcher
is interested in has taken place. Application specific logic will usually be
implemented in the callback. For example, an IO watcher\(aqs callback will receive
the data read from a file, a timer callback will be triggered on timeout and so
on.
.SS Idling
.sp
Here is an example of using an idle handle. The callback is called once on
every turn of the event loop. A use case for idle handles is discussed in
utilities\&. Let us use an idle watcher to look at the watcher life cycle
and see how \fBuv_run()\fP will now block because a watcher is present. The idle
watcher is stopped when the count is reached and \fBuv_run()\fP exits since no
event watchers are active.
.sp
idle\-basic/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <uv.h>
int64_t counter = 0;
void wait_for_a_while(uv_idle_t* handle) {
counter++;
if (counter >= 10e6)
uv_idle_stop(handle);
}
int main() {
uv_idle_t idler;
uv_idle_init(uv_default_loop(), &idler);
uv_idle_start(&idler, wait_for_a_while);
printf("Idling...\en");
uv_run(uv_default_loop(), UV_RUN_DEFAULT);
uv_loop_close(uv_default_loop());
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Storing context
.sp
In callback based programming style you\(aqll often want to pass some \(aqcontext\(aq \-\-
application specific information \-\- between the call site and the callback. All
handles and requests have a \fBvoid* data\fP member which you can set to the
context and cast back in the callback. This is a common pattern used throughout
the C library ecosystem. In addition \fBuv_loop_t\fP also has a similar data
member.
.sp
.ce
----
.ce 0
.sp
.IP [1] 5
Depending on the capacity of the hardware of course.
.SS Filesystem
.sp
Simple filesystem read/write is achieved using the \fBuv_fs_*\fP functions and the
\fBuv_fs_t\fP struct.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
The libuv filesystem operations are different from socket operations\&. Socket operations use the non\-blocking operations provided
by the operating system. Filesystem operations use blocking functions
internally, but invoke these functions in a \fI\%thread pool\fP and notify
watchers registered with the event loop when application interaction is
required.
.UNINDENT
.UNINDENT
.sp
All filesystem functions have two forms \- \fIsynchronous\fP and \fIasynchronous\fP\&.
.sp
The \fIsynchronous\fP forms automatically get called (and \fBblock\fP) if the
callback is null. The return value of functions is a libuv error code\&. This is usually only useful for synchronous calls.
The \fIasynchronous\fP form is called when a callback is passed and the return
value is 0.
.SS Reading/Writing files
.sp
A file descriptor is obtained using
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int uv_fs_open(uv_loop_t* loop, uv_fs_t* req, const char* path, int flags, int mode, uv_fs_cb cb)
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBflags\fP and \fBmode\fP are standard
\fI\%Unix flags\fP\&.
libuv takes care of converting to the appropriate Windows flags.
.sp
File descriptors are closed using
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int uv_fs_close(uv_loop_t* loop, uv_fs_t* req, uv_file file, uv_fs_cb cb)
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Filesystem operation callbacks have the signature:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void callback(uv_fs_t* req);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Let\(aqs see a simple implementation of \fBcat\fP\&. We start with registering
a callback for when the file is opened:
.sp
uvcat/main.c \- opening a file
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_open(uv_fs_t *req) {
// The request passed to the callback is the same as the one the call setup
// function was passed.
assert(req == &open_req);
if (req\->result >= 0) {
iov = uv_buf_init(buffer, sizeof(buffer));
uv_fs_read(uv_default_loop(), &read_req, req\->result,
&iov, 1, \-1, on_read);
}
else {
fprintf(stderr, "error opening file: %s\en", uv_strerror((int)req\->result));
}
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The \fBresult\fP field of a \fBuv_fs_t\fP is the file descriptor in case of the
\fBuv_fs_open\fP callback. If the file is successfully opened, we start reading it.
.sp
uvcat/main.c \- read callback
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_read(uv_fs_t *req) {
if (req\->result < 0) {
fprintf(stderr, "Read error: %s\en", uv_strerror(req\->result));
}
else if (req\->result == 0) {
uv_fs_t close_req;
// synchronous
uv_fs_close(uv_default_loop(), &close_req, open_req.result, NULL);
}
else if (req\->result > 0) {
iov.len = req\->result;
uv_fs_write(uv_default_loop(), &write_req, 1, &iov, 1, \-1, on_write);
}
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
In the case of a read call, you should pass an \fIinitialized\fP buffer which will
be filled with data before the read callback is triggered. The \fBuv_fs_*\fP
operations map almost directly to certain POSIX functions, so EOF is indicated
in this case by \fBresult\fP being 0. In the case of streams or pipes, the
\fBUV_EOF\fP constant would have been passed as a status instead.
.sp
Here you see a common pattern when writing asynchronous programs. The
\fBuv_fs_close()\fP call is performed synchronously. \fIUsually tasks which are
one\-off, or are done as part of the startup or shutdown stage are performed
synchronously, since we are interested in fast I/O when the program is going
about its primary task and dealing with multiple I/O sources\fP\&. For solo tasks
the performance difference usually is negligible and may lead to simpler code.
.sp
Filesystem writing is similarly simple using \fBuv_fs_write()\fP\&. \fIYour callback
will be triggered after the write is complete\fP\&. In our case the callback
simply drives the next read. Thus read and write proceed in lockstep via
callbacks.
.sp
uvcat/main.c \- write callback
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_write(uv_fs_t *req) {
if (req\->result < 0) {
fprintf(stderr, "Write error: %s\en", uv_strerror((int)req\->result));
}
else {
uv_fs_read(uv_default_loop(), &read_req, open_req.result, &iov, 1, \-1, on_read);
}
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
Due to the way filesystems and disk drives are configured for performance,
a write that \(aqsucceeds\(aq may not be committed to disk yet.
.UNINDENT
.UNINDENT
.sp
We set the dominos rolling in \fBmain()\fP:
.sp
uvcat/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main(int argc, char **argv) {
uv_fs_open(uv_default_loop(), &open_req, argv[1], O_RDONLY, 0, on_open);
uv_run(uv_default_loop(), UV_RUN_DEFAULT);
uv_fs_req_cleanup(&open_req);
uv_fs_req_cleanup(&read_req);
uv_fs_req_cleanup(&write_req);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
The \fBuv_fs_req_cleanup()\fP function must always be called on filesystem
requests to free internal memory allocations in libuv.
.UNINDENT
.UNINDENT
.SS Filesystem operations
.sp
All the standard filesystem operations like \fBunlink\fP, \fBrmdir\fP, \fBstat\fP are
supported asynchronously and have intuitive argument order. They follow the
same patterns as the read/write/open calls, returning the result in the
\fBuv_fs_t.result\fP field. The full list:
.sp
Filesystem operations
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int uv_fs_close(uv_loop_t* loop, uv_fs_t* req, uv_file file, uv_fs_cb cb);
int uv_fs_open(uv_loop_t* loop, uv_fs_t* req, const char* path, int flags, int mode, uv_fs_cb cb);
int uv_fs_read(uv_loop_t* loop, uv_fs_t* req, uv_file file, const uv_buf_t bufs[], unsigned int nbufs, int64_t offset, uv_fs_cb cb);
int uv_fs_unlink(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_fs_cb cb);
int uv_fs_write(uv_loop_t* loop, uv_fs_t* req, uv_file file, const uv_buf_t bufs[], unsigned int nbufs, int64_t offset, uv_fs_cb cb);
int uv_fs_copyfile(uv_loop_t* loop, uv_fs_t* req, const char* path, const char* new_path, int flags, uv_fs_cb cb);
int uv_fs_mkdir(uv_loop_t* loop, uv_fs_t* req, const char* path, int mode, uv_fs_cb cb);
int uv_fs_mkdtemp(uv_loop_t* loop, uv_fs_t* req, const char* tpl, uv_fs_cb cb);
int uv_fs_rmdir(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_fs_cb cb);
int uv_fs_scandir(uv_loop_t* loop, uv_fs_t* req, const char* path, int flags, uv_fs_cb cb);
int uv_fs_scandir_next(uv_fs_t* req, uv_dirent_t* ent);
int uv_fs_opendir(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_fs_cb cb);
int uv_fs_readdir(uv_loop_t* loop, uv_fs_t* req, uv_dir_t* dir, uv_fs_cb cb);
int uv_fs_closedir(uv_loop_t* loop, uv_fs_t* req, uv_dir_t* dir, uv_fs_cb cb);
int uv_fs_stat(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_fs_cb cb);
int uv_fs_fstat(uv_loop_t* loop, uv_fs_t* req, uv_file file, uv_fs_cb cb);
int uv_fs_rename(uv_loop_t* loop, uv_fs_t* req, const char* path, const char* new_path, uv_fs_cb cb);
int uv_fs_fsync(uv_loop_t* loop, uv_fs_t* req, uv_file file, uv_fs_cb cb);
int uv_fs_fdatasync(uv_loop_t* loop, uv_fs_t* req, uv_file file, uv_fs_cb cb);
int uv_fs_ftruncate(uv_loop_t* loop, uv_fs_t* req, uv_file file, int64_t offset, uv_fs_cb cb);
int uv_fs_sendfile(uv_loop_t* loop, uv_fs_t* req, uv_file out_fd, uv_file in_fd, int64_t in_offset, size_t length, uv_fs_cb cb);
int uv_fs_access(uv_loop_t* loop, uv_fs_t* req, const char* path, int mode, uv_fs_cb cb);
int uv_fs_chmod(uv_loop_t* loop, uv_fs_t* req, const char* path, int mode, uv_fs_cb cb);
int uv_fs_utime(uv_loop_t* loop, uv_fs_t* req, const char* path, double atime, double mtime, uv_fs_cb cb);
int uv_fs_futime(uv_loop_t* loop, uv_fs_t* req, uv_file file, double atime, double mtime, uv_fs_cb cb);
int uv_fs_lstat(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_fs_cb cb);
int uv_fs_link(uv_loop_t* loop, uv_fs_t* req, const char* path, const char* new_path, uv_fs_cb cb);
int uv_fs_symlink(uv_loop_t* loop, uv_fs_t* req, const char* path, const char* new_path, int flags, uv_fs_cb cb);
int uv_fs_readlink(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_fs_cb cb);
int uv_fs_realpath(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_fs_cb cb);
int uv_fs_fchmod(uv_loop_t* loop, uv_fs_t* req, uv_file file, int mode, uv_fs_cb cb);
int uv_fs_chown(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_uid_t uid, uv_gid_t gid, uv_fs_cb cb);
int uv_fs_fchown(uv_loop_t* loop, uv_fs_t* req, uv_file file, uv_uid_t uid, uv_gid_t gid, uv_fs_cb cb);
int uv_fs_lchown(uv_loop_t* loop, uv_fs_t* req, const char* path, uv_uid_t uid, uv_gid_t gid, uv_fs_cb cb);
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Buffers and Streams
.sp
The basic I/O handle in libuv is the stream (\fBuv_stream_t\fP). TCP sockets, UDP
sockets, and pipes for file I/O and IPC are all treated as stream subclasses.
.sp
Streams are initialized using custom functions for each subclass, then operated
upon using
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int uv_read_start(uv_stream_t*, uv_alloc_cb alloc_cb, uv_read_cb read_cb);
int uv_read_stop(uv_stream_t*);
int uv_write(uv_write_t* req, uv_stream_t* handle,
const uv_buf_t bufs[], unsigned int nbufs, uv_write_cb cb);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The stream based functions are simpler to use than the filesystem ones and
libuv will automatically keep reading from a stream when \fBuv_read_start()\fP is
called once, until \fBuv_read_stop()\fP is called.
.sp
The discrete unit of data is the buffer \-\- \fBuv_buf_t\fP\&. This is simply
a collection of a pointer to bytes (\fBuv_buf_t.base\fP) and the length
(\fBuv_buf_t.len\fP). The \fBuv_buf_t\fP is lightweight and passed around by value.
What does require management is the actual bytes, which have to be allocated
and freed by the application.
.sp
\fBERROR:\fP
.INDENT 0.0
.INDENT 3.5
THIS PROGRAM DOES NOT ALWAYS WORK, NEED SOMETHING BETTER**
.UNINDENT
.UNINDENT
.sp
To demonstrate streams we will need to use \fBuv_pipe_t\fP\&. This allows streaming
local files [2]\&. Here is a simple tee utility using libuv. Doing all operations
asynchronously shows the power of evented I/O. The two writes won\(aqt block each
other, but we have to be careful to copy over the buffer data to ensure we don\(aqt
free a buffer until it has been written.
.sp
The program is to be executed as:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
\&./uvtee <output_file>
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
We start off opening pipes on the files we require. libuv pipes to a file are
opened as bidirectional by default.
.sp
uvtee/main.c \- read on pipes
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main(int argc, char **argv) {
loop = uv_default_loop();
uv_pipe_init(loop, &stdin_pipe, 0);
uv_pipe_open(&stdin_pipe, 0);
uv_pipe_init(loop, &stdout_pipe, 0);
uv_pipe_open(&stdout_pipe, 1);
uv_fs_t file_req;
int fd = uv_fs_open(loop, &file_req, argv[1], O_CREAT | O_RDWR, 0644, NULL);
uv_pipe_init(loop, &file_pipe, 0);
uv_pipe_open(&file_pipe, fd);
uv_read_start((uv_stream_t*)&stdin_pipe, alloc_buffer, read_stdin);
uv_run(loop, UV_RUN_DEFAULT);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The third argument of \fBuv_pipe_init()\fP should be set to 1 for IPC using named
pipes. This is covered in processes\&. The \fBuv_pipe_open()\fP call
associates the pipe with the file descriptor, in this case \fB0\fP (standard
input).
.sp
We start monitoring \fBstdin\fP\&. The \fBalloc_buffer\fP callback is invoked as new
buffers are required to hold incoming data. \fBread_stdin\fP will be called with
these buffers.
.sp
uvtee/main.c \- reading buffers
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void alloc_buffer(uv_handle_t *handle, size_t suggested_size, uv_buf_t *buf) {
*buf = uv_buf_init((char*) malloc(suggested_size), suggested_size);
}
void read_stdin(uv_stream_t *stream, ssize_t nread, const uv_buf_t *buf) {
if (nread < 0){
if (nread == UV_EOF){
// end of file
uv_close((uv_handle_t *)&stdin_pipe, NULL);
uv_close((uv_handle_t *)&stdout_pipe, NULL);
uv_close((uv_handle_t *)&file_pipe, NULL);
}
} else if (nread > 0) {
write_data((uv_stream_t *)&stdout_pipe, nread, *buf, on_stdout_write);
write_data((uv_stream_t *)&file_pipe, nread, *buf, on_file_write);
}
// OK to free buffer as write_data copies it.
if (buf\->base)
free(buf\->base);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The standard \fBmalloc\fP is sufficient here, but you can use any memory allocation
scheme. For example, node.js uses its own slab allocator which associates
buffers with V8 objects.
.sp
The read callback \fBnread\fP parameter is less than 0 on any error. This error
might be EOF, in which case we close all the streams, using the generic close
function \fBuv_close()\fP which deals with the handle based on its internal type.
Otherwise \fBnread\fP is a non\-negative number and we can attempt to write that
many bytes to the output streams. Finally remember that buffer allocation and
deallocation is application responsibility, so we free the data.
.sp
The allocation callback may return a buffer with length zero if it fails to
allocate memory. In this case, the read callback is invoked with error
UV_ENOBUFS. libuv will continue to attempt to read the stream though, so you
must explicitly call \fBuv_close()\fP if you want to stop when allocation fails.
.sp
The read callback may be called with \fBnread = 0\fP, indicating that at this
point there is nothing to be read. Most applications will just ignore this.
.sp
uvtee/main.c \- Write to pipe
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct {
uv_write_t req;
uv_buf_t buf;
} write_req_t;
void free_write_req(uv_write_t *req) {
write_req_t *wr = (write_req_t*) req;
free(wr\->buf.base);
free(wr);
}
void on_stdout_write(uv_write_t *req, int status) {
free_write_req(req);
}
void on_file_write(uv_write_t *req, int status) {
free_write_req(req);
}
void write_data(uv_stream_t *dest, size_t size, uv_buf_t buf, uv_write_cb cb) {
write_req_t *req = (write_req_t*) malloc(sizeof(write_req_t));
req\->buf = uv_buf_init((char*) malloc(size), size);
memcpy(req\->buf.base, buf.base, size);
uv_write((uv_write_t*) req, (uv_stream_t*)dest, &req\->buf, 1, cb);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBwrite_data()\fP makes a copy of the buffer obtained from read. This buffer
does not get passed through to the write callback trigged on write completion. To
get around this we wrap a write request and a buffer in \fBwrite_req_t\fP and
unwrap it in the callbacks. We make a copy so we can free the two buffers from
the two calls to \fBwrite_data\fP independently of each other. While acceptable
for a demo program like this, you\(aqll probably want smarter memory management,
like reference counted buffers or a pool of buffers in any major application.
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
If your program is meant to be used with other programs it may knowingly or
unknowingly be writing to a pipe. This makes it susceptible to \fI\%aborting on
receiving a SIGPIPE\fP\&. It is a good idea to insert:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
signal(SIGPIPE, SIG_IGN)
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
in the initialization stages of your application.
.UNINDENT
.UNINDENT
.SS File change events
.sp
All modern operating systems provide APIs to put watches on individual files or
directories and be informed when the files are modified. libuv wraps common
file change notification libraries [1]\&. This is one of the more
inconsistent parts of libuv. File change notification systems are themselves
extremely varied across platforms so getting everything working everywhere is
difficult. To demonstrate, I\(aqm going to build a simple utility which runs
a command whenever any of the watched files change:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
\&./onchange <command> <file1> [file2] ...
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The file change notification is started using \fBuv_fs_event_init()\fP:
.sp
onchange/main.c \- The setup
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main(int argc, char **argv) {
if (argc <= 2) {
fprintf(stderr, "Usage: %s <command> <file1> [file2 ...]\en", argv[0]);
return 1;
}
loop = uv_default_loop();
command = argv[1];
while (argc\-\- > 2) {
fprintf(stderr, "Adding watch on %s\en", argv[argc]);
uv_fs_event_t *fs_event_req = malloc(sizeof(uv_fs_event_t));
uv_fs_event_init(loop, fs_event_req);
// The recursive flag watches subdirectories too.
uv_fs_event_start(fs_event_req, run_command, argv[argc], UV_FS_EVENT_RECURSIVE);
}
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The third argument is the actual file or directory to monitor. The last
argument, \fBflags\fP, can be:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
/*
* Flags to be passed to uv_fs_event_start().
*/
enum uv_fs_event_flags {
UV_FS_EVENT_WATCH_ENTRY = 1,
UV_FS_EVENT_STAT = 2,
UV_FS_EVENT_RECURSIVE = 4
};
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBUV_FS_EVENT_WATCH_ENTRY\fP and \fBUV_FS_EVENT_STAT\fP don\(aqt do anything (yet).
\fBUV_FS_EVENT_RECURSIVE\fP will start watching subdirectories as well on
supported platforms.
.sp
The callback will receive the following arguments:
.INDENT 0.0
.INDENT 3.5
.INDENT 0.0
.IP 1. 3
\fBuv_fs_event_t *handle\fP \- The handle. The \fBpath\fP field of the handle
is the file on which the watch was set.
.IP 2. 3
\fBconst char *filename\fP \- If a directory is being monitored, this is the
file which was changed. Only non\-\fBnull\fP on Linux and Windows. May be \fBnull\fP
even on those platforms.
.IP 3. 3
.INDENT 3.0
.TP
.B \fBint flags\fP \- one of \fBUV_RENAME\fP or \fBUV_CHANGE\fP, or a bitwise OR of
both.
.UNINDENT
.IP 4. 3
\fBint status\fP \- Currently 0.
.UNINDENT
.UNINDENT
.UNINDENT
.sp
In our example we simply print the arguments and run the command using
\fBsystem()\fP\&.
.sp
onchange/main.c \- file change notification callback
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void run_command(uv_fs_event_t *handle, const char *filename, int events, int status) {
char path[1024];
size_t size = 1023;
// Does not handle error if path is longer than 1023.
uv_fs_event_getpath(handle, path, &size);
path[size] = \(aq\e0\(aq;
fprintf(stderr, "Change detected in %s: ", path);
if (events & UV_RENAME)
fprintf(stderr, "renamed");
if (events & UV_CHANGE)
fprintf(stderr, "changed");
fprintf(stderr, " %s\en", filename ? filename : "");
system(command);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
.ce
----
.ce 0
.sp
.IP [1] 5
inotify on Linux, FSEvents on Darwin, kqueue on BSDs,
ReadDirectoryChangesW on Windows, event ports on Solaris, unsupported on Cygwin
.IP [2] 5
see pipes
.SS Networking
.sp
Networking in libuv is not much different from directly using the BSD socket
interface, some things are easier, all are non\-blocking, but the concepts stay
the same. In addition libuv offers utility functions to abstract the annoying,
repetitive and low\-level tasks like setting up sockets using the BSD socket
structures, DNS lookup, and tweaking various socket parameters.
.sp
The \fBuv_tcp_t\fP and \fBuv_udp_t\fP structures are used for network I/O.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
The code samples in this chapter exist to show certain libuv APIs. They are
not examples of good quality code. They leak memory and don\(aqt always close
connections properly.
.UNINDENT
.UNINDENT
.SS TCP
.sp
TCP is a connection oriented, stream protocol and is therefore based on the
libuv streams infrastructure.
.SS Server
.sp
Server sockets proceed by:
.INDENT 0.0
.IP 1. 3
\fBuv_tcp_init\fP the TCP handle.
.IP 2. 3
\fBuv_tcp_bind\fP it.
.IP 3. 3
Call \fBuv_listen\fP on the handle to have a callback invoked whenever a new
connection is established by a client.
.IP 4. 3
Use \fBuv_accept\fP to accept the connection.
.IP 5. 3
Use stream operations to communicate with the
client.
.UNINDENT
.sp
Here is a simple echo server
.sp
tcp\-echo\-server/main.c \- The listen socket
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_close((uv_handle_t*) client, on_close);
}
}
int main() {
loop = uv_default_loop();
uv_tcp_t server;
uv_tcp_init(loop, &server);
uv_ip4_addr("0.0.0.0", DEFAULT_PORT, &addr);
uv_tcp_bind(&server, (const struct sockaddr*)&addr, 0);
int r = uv_listen((uv_stream_t*) &server, DEFAULT_BACKLOG, on_new_connection);
if (r) {
fprintf(stderr, "Listen error %s\en", uv_strerror(r));
return 1;
}
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
You can see the utility function \fBuv_ip4_addr\fP being used to convert from
a human readable IP address, port pair to the sockaddr_in structure required by
the BSD socket APIs. The reverse can be obtained using \fBuv_ip4_name\fP\&.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
There are \fBuv_ip6_*\fP analogues for the ip4 functions.
.UNINDENT
.UNINDENT
.sp
Most of the setup functions are synchronous since they are CPU\-bound.
\fBuv_listen\fP is where we return to libuv\(aqs callback style. The second
arguments is the backlog queue \-\- the maximum length of queued connections.
.sp
When a connection is initiated by clients, the callback is required to set up
a handle for the client socket and associate the handle using \fBuv_accept\fP\&.
In this case we also establish interest in reading from this stream.
.sp
tcp\-echo\-server/main.c \- Accepting the client
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
free(buf\->base);
}
void on_new_connection(uv_stream_t *server, int status) {
if (status < 0) {
fprintf(stderr, "New connection error %s\en", uv_strerror(status));
// error!
return;
}
uv_tcp_t *client = (uv_tcp_t*) malloc(sizeof(uv_tcp_t));
uv_tcp_init(loop, client);
if (uv_accept(server, (uv_stream_t*) client) == 0) {
uv_read_start((uv_stream_t*) client, alloc_buffer, echo_read);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The remaining set of functions is very similar to the streams example and can
be found in the code. Just remember to call \fBuv_close\fP when the socket isn\(aqt
required. This can be done even in the \fBuv_listen\fP callback if you are not
interested in accepting the connection.
.SS Client
.sp
Where you do bind/listen/accept on the server, on the client side it\(aqs simply
a matter of calling \fBuv_tcp_connect\fP\&. The same \fBuv_connect_cb\fP style
callback of \fBuv_listen\fP is used by \fBuv_tcp_connect\fP\&. Try:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_tcp_t* socket = (uv_tcp_t*)malloc(sizeof(uv_tcp_t));
uv_tcp_init(loop, socket);
uv_connect_t* connect = (uv_connect_t*)malloc(sizeof(uv_connect_t));
struct sockaddr_in dest;
uv_ip4_addr("127.0.0.1", 80, &dest);
uv_tcp_connect(connect, socket, (const struct sockaddr*)&dest, on_connect);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
where \fBon_connect\fP will be called after the connection is established. The
callback receives the \fBuv_connect_t\fP struct, which has a member \fB\&.handle\fP
pointing to the socket.
.SS UDP
.sp
The \fI\%User Datagram Protocol\fP offers connectionless, unreliable network
communication. Hence libuv doesn\(aqt offer a stream. Instead libuv provides
non\-blocking UDP support via the \fIuv_udp_t\fP handle (for receiving) and
\fIuv_udp_send_t\fP request (for sending) and related functions. That said, the
actual API for reading/writing is very similar to normal stream reads. To look
at how UDP can be used, the example shows the first stage of obtaining an IP
address from a \fI\%DHCP\fP server \-\- DHCP Discover.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
You will have to run \fIudp\-dhcp\fP as \fBroot\fP since it uses well known port
numbers below 1024.
.UNINDENT
.UNINDENT
.sp
udp\-dhcp/main.c \- Setup and send UDP packets
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_loop_t *loop;
uv_udp_t send_socket;
uv_udp_t recv_socket;
int main() {
loop = uv_default_loop();
uv_udp_init(loop, &recv_socket);
struct sockaddr_in recv_addr;
uv_ip4_addr("0.0.0.0", 68, &recv_addr);
uv_udp_bind(&recv_socket, (const struct sockaddr *)&recv_addr, UV_UDP_REUSEADDR);
uv_udp_recv_start(&recv_socket, alloc_buffer, on_read);
uv_udp_init(loop, &send_socket);
struct sockaddr_in broadcast_addr;
uv_ip4_addr("0.0.0.0", 0, &broadcast_addr);
uv_udp_bind(&send_socket, (const struct sockaddr *)&broadcast_addr, 0);
uv_udp_set_broadcast(&send_socket, 1);
uv_udp_send_t send_req;
uv_buf_t discover_msg = make_discover_msg();
struct sockaddr_in send_addr;
uv_ip4_addr("255.255.255.255", 67, &send_addr);
uv_udp_send(&send_req, &send_socket, &discover_msg, 1, (const struct sockaddr *)&send_addr, on_send);
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
The IP address \fB0.0.0.0\fP is used to bind to all interfaces. The IP
address \fB255.255.255.255\fP is a broadcast address meaning that packets
will be sent to all interfaces on the subnet. port \fB0\fP means that the OS
randomly assigns a port.
.UNINDENT
.UNINDENT
.sp
First we setup the receiving socket to bind on all interfaces on port 68 (DHCP
client) and start a read on it. This will read back responses from any DHCP
server that replies. We use the UV_UDP_REUSEADDR flag to play nice with any
other system DHCP clients that are running on this computer on the same port.
Then we setup a similar send socket and use \fBuv_udp_send\fP to send
a \fIbroadcast message\fP on port 67 (DHCP server).
.sp
It is \fBnecessary\fP to set the broadcast flag, otherwise you will get an
\fBEACCES\fP error [1]\&. The exact message being sent is not relevant to this
book and you can study the code if you are interested. As usual the read and
write callbacks will receive a status code of < 0 if something went wrong.
.sp
Since UDP sockets are not connected to a particular peer, the read callback
receives an extra parameter about the sender of the packet.
.sp
\fBnread\fP may be zero if there is no more data to be read. If \fBaddr\fP is NULL,
it indicates there is nothing to read (the callback shouldn\(aqt do anything), if
not NULL, it indicates that an empty datagram was received from the host at
\fBaddr\fP\&. The \fBflags\fP parameter may be \fBUV_UDP_PARTIAL\fP if the buffer
provided by your allocator was not large enough to hold the data. \fIIn this case
the OS will discard the data that could not fit\fP (That\(aqs UDP for you!).
.sp
udp\-dhcp/main.c \- Reading packets
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_read(uv_udp_t *req, ssize_t nread, const uv_buf_t *buf, const struct sockaddr *addr, unsigned flags) {
if (nread < 0) {
fprintf(stderr, "Read error %s\en", uv_err_name(nread));
uv_close((uv_handle_t*) req, NULL);
free(buf\->base);
return;
}
char sender[17] = { 0 };
uv_ip4_name((const struct sockaddr_in*) addr, sender, 16);
fprintf(stderr, "Recv from %s\en", sender);
// ... DHCP specific code
unsigned int *as_integer = (unsigned int*)buf\->base;
unsigned int ipbin = ntohl(as_integer[4]);
unsigned char ip[4] = {0};
int i;
for (i = 0; i < 4; i++)
ip[i] = (ipbin >> i*8) & 0xff;
fprintf(stderr, "Offered IP %d.%d.%d.%d\en", ip[3], ip[2], ip[1], ip[0]);
free(buf\->base);
uv_udp_recv_stop(req);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS UDP Options
.SS Time\-to\-live
.sp
The TTL of packets sent on the socket can be changed using \fBuv_udp_set_ttl\fP\&.
.SS IPv6 stack only
.sp
IPv6 sockets can be used for both IPv4 and IPv6 communication. If you want to
restrict the socket to IPv6 only, pass the \fBUV_UDP_IPV6ONLY\fP flag to
\fBuv_udp_bind\fP [2]\&.
.SS Multicast
.sp
A socket can (un)subscribe to a multicast group using:
.sp
where \fBmembership\fP is \fBUV_JOIN_GROUP\fP or \fBUV_LEAVE_GROUP\fP\&.
.sp
The concepts of multicasting are nicely explained in \fI\%this guide\fP\&.
.sp
Local loopback of multicast packets is enabled by default [3], use
\fBuv_udp_set_multicast_loop\fP to switch it off.
.sp
The packet time\-to\-live for multicast packets can be changed using
\fBuv_udp_set_multicast_ttl\fP\&.
.SS Querying DNS
.sp
libuv provides asynchronous DNS resolution. For this it provides its own
\fBgetaddrinfo\fP replacement [4]\&. In the callback you can
perform normal socket operations on the retrieved addresses. Let\(aqs connect to
Freenode to see an example of DNS resolution.
.sp
dns/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main() {
loop = uv_default_loop();
struct addrinfo hints;
hints.ai_family = PF_INET;
hints.ai_socktype = SOCK_STREAM;
hints.ai_protocol = IPPROTO_TCP;
hints.ai_flags = 0;
uv_getaddrinfo_t resolver;
fprintf(stderr, "irc.freenode.net is... ");
int r = uv_getaddrinfo(loop, &resolver, on_resolved, "irc.freenode.net", "6667", &hints);
if (r) {
fprintf(stderr, "getaddrinfo call error %s\en", uv_err_name(r));
return 1;
}
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
If \fBuv_getaddrinfo\fP returns non\-zero, something went wrong in the setup and
your callback won\(aqt be invoked at all. All arguments can be freed immediately
after \fBuv_getaddrinfo\fP returns. The \fIhostname\fP, \fIservname\fP and \fIhints\fP
structures are documented in \fI\%the getaddrinfo man page\fP\&. The
callback can be \fBNULL\fP in which case the function will run synchronously.
.sp
In the resolver callback, you can pick any IP from the linked list of \fBstruct
addrinfo(s)\fP\&. This also demonstrates \fBuv_tcp_connect\fP\&. It is necessary to
call \fBuv_freeaddrinfo\fP in the callback.
.sp
dns/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_resolved(uv_getaddrinfo_t *resolver, int status, struct addrinfo *res) {
if (status < 0) {
fprintf(stderr, "getaddrinfo callback error %s\en", uv_err_name(status));
return;
}
char addr[17] = {\(aq\e0\(aq};
uv_ip4_name((struct sockaddr_in*) res\->ai_addr, addr, 16);
fprintf(stderr, "%s\en", addr);
uv_connect_t *connect_req = (uv_connect_t*) malloc(sizeof(uv_connect_t));
uv_tcp_t *socket = (uv_tcp_t*) malloc(sizeof(uv_tcp_t));
uv_tcp_init(loop, socket);
uv_tcp_connect(connect_req, socket, (const struct sockaddr*) res\->ai_addr, on_connect);
uv_freeaddrinfo(res);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
libuv also provides the inverse \fI\%uv_getnameinfo\fP\&.
.SS Network interfaces
.sp
Information about the system\(aqs network interfaces can be obtained through libuv
using \fBuv_interface_addresses\fP\&. This simple program just prints out all the
interface details so you get an idea of the fields that are available. This is
useful to allow your service to bind to IP addresses when it starts.
.sp
interfaces/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <uv.h>
int main() {
char buf[512];
uv_interface_address_t *info;
int count, i;
uv_interface_addresses(&info, &count);
i = count;
printf("Number of interfaces: %d\en", count);
while (i\-\-) {
uv_interface_address_t interface = info[i];
printf("Name: %s\en", interface.name);
printf("Internal? %s\en", interface.is_internal ? "Yes" : "No");
if (interface.address.address4.sin_family == AF_INET) {
uv_ip4_name(&interface.address.address4, buf, sizeof(buf));
printf("IPv4 address: %s\en", buf);
}
else if (interface.address.address4.sin_family == AF_INET6) {
uv_ip6_name(&interface.address.address6, buf, sizeof(buf));
printf("IPv6 address: %s\en", buf);
}
printf("\en");
}
uv_free_interface_addresses(info, count);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBis_internal\fP is true for loopback interfaces. Note that if a physical
interface has multiple IPv4/IPv6 addresses, the name will be reported multiple
times, with each address being reported once.
.sp
.ce
----
.ce 0
.sp
.IP [1] 5
\fI\%https://beej.us/guide/bgnet/html/#broadcast\-packetshello\-world\fP
.IP [2] 5
on Windows only supported on Windows Vista and later.
.IP [3] 5
\fI\%https://www.tldp.org/HOWTO/Multicast\-HOWTO\-6.html#ss6.1\fP
.IP [4] 5
libuv use the system \fBgetaddrinfo\fP in the libuv threadpool. libuv
v0.8.0 and earlier also included \fI\%c\-ares\fP as an alternative, but this has been
removed in v0.9.0.
.SS Threads
.sp
Wait a minute? Why are we on threads? Aren\(aqt event loops supposed to be \fBthe
way\fP to do \fIweb\-scale programming\fP? Well... no. Threads are still the medium in
which processors do their jobs. Threads are therefore mighty useful sometimes, even
though you might have to wade through various synchronization primitives.
.sp
Threads are used internally to fake the asynchronous nature of all of the system
calls. libuv also uses threads to allow you, the application, to perform a task
asynchronously that is actually blocking, by spawning a thread and collecting
the result when it is done.
.sp
Today there are two predominant thread libraries: the Windows threads
implementation and POSIX\(aqs \fI\%pthreads(7)\fP\&. libuv\(aqs thread API is analogous to
the pthreads API and often has similar semantics.
.sp
A notable aspect of libuv\(aqs thread facilities is that it is a self contained
section within libuv. Whereas other features intimately depend on the event
loop and callback principles, threads are complete agnostic, they block as
required, signal errors directly via return values, and, as shown in the
\fI\%first example\fP, don\(aqt even require a running
event loop.
.sp
libuv\(aqs thread API is also very limited since the semantics and syntax of
threads are different on all platforms, with different levels of completeness.
.sp
This chapter makes the following assumption: \fBThere is only one event loop,
running in one thread (the main thread)\fP\&. No other thread interacts
with the event loop (except using \fBuv_async_send\fP).
.SS Core thread operations
.sp
There isn\(aqt much here, you just start a thread using \fBuv_thread_create()\fP and
wait for it to close using \fBuv_thread_join()\fP\&.
.sp
thread\-create/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main() {
int tracklen = 10;
uv_thread_t hare_id;
uv_thread_t tortoise_id;
uv_thread_create(&hare_id, hare, &tracklen);
uv_thread_create(&tortoise_id, tortoise, &tracklen);
uv_thread_join(&hare_id);
uv_thread_join(&tortoise_id);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBTIP:\fP
.INDENT 0.0
.INDENT 3.5
\fBuv_thread_t\fP is just an alias for \fBpthread_t\fP on Unix, but this is an
implementation detail, avoid depending on it to always be true.
.UNINDENT
.UNINDENT
.sp
The second parameter is the function which will serve as the entry point for
the thread, the last parameter is a \fBvoid *\fP argument which can be used to pass
custom parameters to the thread. The function \fBhare\fP will now run in a separate
thread, scheduled pre\-emptively by the operating system:
.sp
thread\-create/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void hare(void *arg) {
int tracklen = *((int *) arg);
while (tracklen) {
tracklen\-\-;
sleep(1);
fprintf(stderr, "Hare ran another step\en");
}
fprintf(stderr, "Hare done running!\en");
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Unlike \fBpthread_join()\fP which allows the target thread to pass back a value to
the calling thread using a second parameter, \fBuv_thread_join()\fP does not. To
send values use \fI\%Inter\-thread communication\fP\&.
.SS Synchronization Primitives
.sp
This section is purposely spartan. This book is not about threads, so I only
catalogue any surprises in the libuv APIs here. For the rest you can look at
the \fI\%pthreads(7)\fP man pages.
.SS Mutexes
.sp
The mutex functions are a \fBdirect\fP map to the pthread equivalents.
.sp
libuv mutex functions
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int uv_mutex_init(uv_mutex_t* handle);
int uv_mutex_init_recursive(uv_mutex_t* handle);
void uv_mutex_destroy(uv_mutex_t* handle);
void uv_mutex_lock(uv_mutex_t* handle);
int uv_mutex_trylock(uv_mutex_t* handle);
void uv_mutex_unlock(uv_mutex_t* handle);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The \fBuv_mutex_init()\fP, \fBuv_mutex_init_recursive()\fP and \fBuv_mutex_trylock()\fP
functions will return 0 on success, and an error code otherwise.
.sp
If \fIlibuv\fP has been compiled with debugging enabled, \fBuv_mutex_destroy()\fP,
\fBuv_mutex_lock()\fP and \fBuv_mutex_unlock()\fP will \fBabort()\fP on error.
Similarly \fBuv_mutex_trylock()\fP will abort if the error is anything \fIother
than\fP \fBEAGAIN\fP or \fBEBUSY\fP\&.
.sp
Recursive mutexes are supported, but you should not rely on them. Also, they
should not be used with \fBuv_cond_t\fP variables.
.sp
The default BSD mutex implementation will raise an error if a thread which has
locked a mutex attempts to lock it again. For example, a construct like:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_mutex_init(a_mutex);
uv_mutex_lock(a_mutex);
uv_thread_create(thread_id, entry, (void *)a_mutex);
uv_mutex_lock(a_mutex);
// more things here
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
can be used to wait until another thread initializes some stuff and then
unlocks \fBa_mutex\fP but will lead to your program crashing if in debug mode, or
return an error in the second call to \fBuv_mutex_lock()\fP\&.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
Mutexes on Windows are always recursive.
.UNINDENT
.UNINDENT
.SS Locks
.sp
Read\-write locks are a more granular access mechanism. Two readers can access
shared memory at the same time. A writer may not acquire the lock when it is
held by a reader. A reader or writer may not acquire a lock when a writer is
holding it. Read\-write locks are frequently used in databases. Here is a toy
example.
.sp
locks/main.c \- simple rwlocks
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <uv.h>
uv_barrier_t blocker;
uv_rwlock_t numlock;
int shared_num;
void reader(void *n)
{
int num = *(int *)n;
int i;
for (i = 0; i < 20; i++) {
uv_rwlock_rdlock(&numlock);
printf("Reader %d: acquired lock\en", num);
printf("Reader %d: shared num = %d\en", num, shared_num);
uv_rwlock_rdunlock(&numlock);
printf("Reader %d: released lock\en", num);
}
uv_barrier_wait(&blocker);
}
void writer(void *n)
{
int num = *(int *)n;
int i;
for (i = 0; i < 20; i++) {
uv_rwlock_wrlock(&numlock);
printf("Writer %d: acquired lock\en", num);
shared_num++;
printf("Writer %d: incremented shared num = %d\en", num, shared_num);
uv_rwlock_wrunlock(&numlock);
printf("Writer %d: released lock\en", num);
}
uv_barrier_wait(&blocker);
}
int main()
{
uv_barrier_init(&blocker, 4);
shared_num = 0;
uv_rwlock_init(&numlock);
uv_thread_t threads[3];
int thread_nums[] = {1, 2, 1};
uv_thread_create(&threads[0], reader, &thread_nums[0]);
uv_thread_create(&threads[1], reader, &thread_nums[1]);
uv_thread_create(&threads[2], writer, &thread_nums[2]);
uv_barrier_wait(&blocker);
uv_barrier_destroy(&blocker);
uv_rwlock_destroy(&numlock);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Run this and observe how the readers will sometimes overlap. In case of
multiple writers, schedulers will usually give them higher priority, so if you
add two writers, you\(aqll see that both writers tend to finish first before the
readers get a chance again.
.sp
We also use barriers in the above example so that the main thread can wait for
all readers and writers to indicate they have ended.
.SS Others
.sp
libuv also supports \fI\%semaphores\fP, \fI\%condition variables\fP and \fI\%barriers\fP with APIs
very similar to their pthread counterparts.
.sp
In addition, libuv provides a convenience function \fBuv_once()\fP\&. Multiple
threads can attempt to call \fBuv_once()\fP with a given guard and a function
pointer, \fBonly the first one will win, the function will be called once and
only once\fP:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
/* Initialize guard */
static uv_once_t once_only = UV_ONCE_INIT;
int i = 0;
void increment() {
i++;
}
void thread1() {
/* ... work */
uv_once(once_only, increment);
}
void thread2() {
/* ... work */
uv_once(once_only, increment);
}
int main() {
/* ... spawn threads */
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
After all threads are done, \fBi == 1\fP\&.
.sp
libuv v0.11.11 onwards also added a \fBuv_key_t\fP struct and \fI\%api\fP for
thread\-local storage.
.SS libuv work queue
.sp
\fBuv_queue_work()\fP is a convenience function that allows an application to run
a task in a separate thread, and have a callback that is triggered when the
task is done. A seemingly simple function, what makes \fBuv_queue_work()\fP
tempting is that it allows potentially any third\-party libraries to be used
with the event\-loop paradigm. When you use event loops, it is \fIimperative to
make sure that no function which runs periodically in the loop thread blocks
when performing I/O or is a serious CPU hog\fP, because this means that the loop
slows down and events are not being handled at full capacity.
.sp
However, a lot of existing code out there features blocking functions (for example
a routine which performs I/O under the hood) to be used with threads if you
want responsiveness (the classic \(aqone thread per client\(aq server model), and
getting them to play with an event loop library generally involves rolling your
own system of running the task in a separate thread. libuv just provides
a convenient abstraction for this.
.sp
Here is a simple example inspired by \fI\%node.js is cancer\fP\&. We are going to
calculate fibonacci numbers, sleeping a bit along the way, but run it in
a separate thread so that the blocking and CPU bound task does not prevent the
event loop from performing other activities.
.sp
queue\-work/main.c \- lazy fibonacci
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void fib(uv_work_t *req) {
int n = *(int *) req\->data;
if (random() % 2)
sleep(1);
else
sleep(3);
long fib = fib_(n);
fprintf(stderr, "%dth fibonacci is %lu\en", n, fib);
}
void after_fib(uv_work_t *req, int status) {
fprintf(stderr, "Done calculating %dth fibonacci\en", *(int *) req\->data);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The actual task function is simple, nothing to show that it is going to be
run in a separate thread. The \fBuv_work_t\fP structure is the clue. You can pass
arbitrary data through it using the \fBvoid* data\fP field and use it to
communicate to and from the thread. But be sure you are using proper locks if
you are changing things while both threads may be running.
.sp
The trigger is \fBuv_queue_work\fP:
.sp
queue\-work/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main() {
loop = uv_default_loop();
int data[FIB_UNTIL];
uv_work_t req[FIB_UNTIL];
int i;
for (i = 0; i < FIB_UNTIL; i++) {
data[i] = i;
req[i].data = (void *) &data[i];
uv_queue_work(loop, &req[i], fib, after_fib);
}
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The thread function will be launched in a separate thread, passed the
\fBuv_work_t\fP structure and once the function returns, the \fIafter\fP function
will be called on the thread the event loop is running in. It will be passed
the same structure.
.sp
For writing wrappers to blocking libraries, a common pattern
is to use a baton to exchange data.
.sp
Since libuv version \fI0.9.4\fP an additional function, \fBuv_cancel()\fP, is
available. This allows you to cancel tasks on the libuv work queue. Only tasks
that \fIare yet to be started\fP can be cancelled. If a task has \fIalready started
executing, or it has finished executing\fP, \fBuv_cancel()\fP \fBwill fail\fP\&.
.sp
\fBuv_cancel()\fP is useful to cleanup pending tasks if the user requests
termination. For example, a music player may queue up multiple directories to
be scanned for audio files. If the user terminates the program, it should quit
quickly and not wait until all pending requests are run.
.sp
Let\(aqs modify the fibonacci example to demonstrate \fBuv_cancel()\fP\&. We first set
up a signal handler for termination.
.sp
queue\-cancel/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main() {
loop = uv_default_loop();
int data[FIB_UNTIL];
int i;
for (i = 0; i < FIB_UNTIL; i++) {
data[i] = i;
fib_reqs[i].data = (void *) &data[i];
uv_queue_work(loop, &fib_reqs[i], fib, after_fib);
}
uv_signal_t sig;
uv_signal_init(loop, &sig);
uv_signal_start(&sig, signal_handler, SIGINT);
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
When the user triggers the signal by pressing \fBCtrl+C\fP we send
\fBuv_cancel()\fP to all the workers. \fBuv_cancel()\fP will return \fB0\fP for those that are already executing or finished.
.sp
queue\-cancel/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void signal_handler(uv_signal_t *req, int signum)
{
printf("Signal received!\en");
int i;
for (i = 0; i < FIB_UNTIL; i++) {
uv_cancel((uv_req_t*) &fib_reqs[i]);
}
uv_signal_stop(req);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
For tasks that do get cancelled successfully, the \fIafter\fP function is called
with \fBstatus\fP set to \fBUV_ECANCELED\fP\&.
.sp
queue\-cancel/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void after_fib(uv_work_t *req, int status) {
if (status == UV_ECANCELED)
fprintf(stderr, "Calculation of %d cancelled.\en", *(int *) req\->data);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBuv_cancel()\fP can also be used with \fBuv_fs_t\fP and \fBuv_getaddrinfo_t\fP
requests. For the filesystem family of functions, \fBuv_fs_t.errorno\fP will be
set to \fBUV_ECANCELED\fP\&.
.sp
\fBTIP:\fP
.INDENT 0.0
.INDENT 3.5
A well designed program would have a way to terminate long running workers
that have already started executing. Such a worker could periodically check
for a variable that only the main process sets to signal termination.
.UNINDENT
.UNINDENT
.SS Inter\-thread communication
.sp
Sometimes you want various threads to actually send each other messages \fIwhile\fP
they are running. For example you might be running some long duration task in
a separate thread (perhaps using \fBuv_queue_work\fP) but want to notify progress
to the main thread. This is a simple example of having a download manager
informing the user of the status of running downloads.
.sp
progress/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_loop_t *loop;
uv_async_t async;
int main() {
loop = uv_default_loop();
uv_work_t req;
int size = 10240;
req.data = (void*) &size;
uv_async_init(loop, &async, print_progress);
uv_queue_work(loop, &req, fake_download, after);
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The async thread communication works \fIon loops\fP so although any thread can be
the message sender, only threads with libuv loops can be receivers (or rather
the loop is the receiver). libuv will invoke the callback (\fBprint_progress\fP)
with the async watcher whenever it receives a message.
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
It is important to realize that since the message send is \fIasync\fP, the callback
may be invoked immediately after \fBuv_async_send\fP is called in another
thread, or it may be invoked after some time. libuv may also combine
multiple calls to \fBuv_async_send\fP and invoke your callback only once. The
only guarantee that libuv makes is \-\- The callback function is called \fIat
least once\fP after the call to \fBuv_async_send\fP\&. If you have no pending
calls to \fBuv_async_send\fP, the callback won\(aqt be called. If you make two
or more calls, and libuv hasn\(aqt had a chance to run the callback yet, it
\fImay\fP invoke your callback \fIonly once\fP for the multiple invocations of
\fBuv_async_send\fP\&. Your callback will never be called twice for just one
event.
.UNINDENT
.UNINDENT
.sp
progress/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
double percentage;
void fake_download(uv_work_t *req) {
int size = *((int*) req\->data);
int downloaded = 0;
while (downloaded < size) {
percentage = downloaded*100.0/size;
async.data = (void*) &percentage;
uv_async_send(&async);
sleep(1);
downloaded += (200+random())%1000; // can only download max 1000bytes/sec,
// but at least a 200;
}
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
In the download function, we modify the progress indicator and queue the message
for delivery with \fBuv_async_send\fP\&. Remember: \fBuv_async_send\fP is also
non\-blocking and will return immediately.
.sp
progress/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void print_progress(uv_async_t *handle) {
double percentage = *((double*) handle\->data);
fprintf(stderr, "Downloaded %.2f%%\en", percentage);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The callback is a standard libuv pattern, extracting the data from the watcher.
.sp
Finally it is important to remember to clean up the watcher.
.sp
progress/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void after(uv_work_t *req, int status) {
fprintf(stderr, "Download complete\en");
uv_close((uv_handle_t*) &async, NULL);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
After this example, which showed the abuse of the \fBdata\fP field, \fI\%bnoordhuis\fP
pointed out that using the \fBdata\fP field is not thread safe, and
\fBuv_async_send()\fP is actually only meant to wake up the event loop. Use
a mutex or rwlock to ensure accesses are performed in the right order.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
mutexes and rwlocks \fBDO NOT\fP work inside a signal handler, whereas
\fBuv_async_send\fP does.
.UNINDENT
.UNINDENT
.sp
One use case where \fBuv_async_send\fP is required is when interoperating with
libraries that require thread affinity for their functionality. For example in
node.js, a v8 engine instance, contexts and its objects are bound to the thread
that the v8 instance was started in. Interacting with v8 data structures from
another thread can lead to undefined results. Now consider some node.js module
which binds a third party library. It may go something like this:
.INDENT 0.0
.IP 1. 3
In node, the third party library is set up with a JavaScript callback to be
invoked for more information:
.INDENT 3.0
.INDENT 3.5
.sp
.nf
.ft C
var lib = require(\(aqlib\(aq);
lib.on_progress(function() {
console.log("Progress");
});
lib.do();
// do other stuff
.ft P
.fi
.UNINDENT
.UNINDENT
.IP 2. 3
\fBlib.do\fP is supposed to be non\-blocking but the third party lib is
blocking, so the binding uses \fBuv_queue_work\fP\&.
.IP 3. 3
The actual work being done in a separate thread wants to invoke the progress
callback, but cannot directly call into v8 to interact with JavaScript. So
it uses \fBuv_async_send\fP\&.
.IP 4. 3
The async callback, invoked in the main loop thread, which is the v8 thread,
then interacts with v8 to invoke the JavaScript callback.
.UNINDENT
.sp
.ce
----
.ce 0
.sp
.SS Processes
.sp
libuv offers considerable child process management, abstracting the platform
differences and allowing communication with the child process using streams or
named pipes.
.sp
A common idiom in Unix is for every process to do one thing and do it well. In
such a case, a process often uses multiple child processes to achieve tasks
(similar to using pipes in shells). A multi\-process model with messages
may also be easier to reason about compared to one with threads and shared
memory.
.sp
A common refrain against event\-based programs is that they cannot take
advantage of multiple cores in modern computers. In a multi\-threaded program
the kernel can perform scheduling and assign different threads to different
cores, improving performance. But an event loop has only one thread. The
workaround can be to launch multiple processes instead, with each process
running an event loop, and each process getting assigned to a separate CPU
core.
.SS Spawning child processes
.sp
The simplest case is when you simply want to launch a process and know when it
exits. This is achieved using \fBuv_spawn\fP\&.
.sp
spawn/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_loop_t *loop;
uv_process_t child_req;
uv_process_options_t options;
int main() {
loop = uv_default_loop();
char* args[3];
args[0] = "mkdir";
args[1] = "test\-dir";
args[2] = NULL;
options.exit_cb = on_exit;
options.file = "mkdir";
options.args = args;
int r;
if ((r = uv_spawn(loop, &child_req, &options))) {
fprintf(stderr, "%s\en", uv_strerror(r));
return 1;
} else {
fprintf(stderr, "Launched process with ID %d\en", child_req.pid);
}
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
\fBoptions\fP is implicitly initialized with zeros since it is a global
variable. If you change \fBoptions\fP to a local variable, remember to
initialize it to null out all unused fields:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_process_options_t options = {0};
.ft P
.fi
.UNINDENT
.UNINDENT
.UNINDENT
.UNINDENT
.sp
The \fBuv_process_t\fP struct only acts as the handle, all options are set via
\fBuv_process_options_t\fP\&. To simply launch a process, you need to set only the
\fBfile\fP and \fBargs\fP fields. \fBfile\fP is the program to execute. Since
\fBuv_spawn\fP uses \fI\%execvp(3)\fP internally, there is no need to supply the full
path. Finally as per underlying conventions, \fBthe arguments array has to be
one larger than the number of arguments, with the last element being NULL\fP\&.
.sp
After the call to \fBuv_spawn\fP, \fBuv_process_t.pid\fP will contain the process
ID of the child process.
.sp
The exit callback will be invoked with the \fIexit status\fP and the type of \fIsignal\fP
which caused the exit.
.sp
spawn/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_exit(uv_process_t *req, int64_t exit_status, int term_signal) {
fprintf(stderr, "Process exited with status %" PRId64 ", signal %d\en", exit_status, term_signal);
uv_close((uv_handle_t*) req, NULL);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
It is \fBrequired\fP to close the process watcher after the process exits.
.SS Changing process parameters
.sp
Before the child process is launched you can control the execution environment
using fields in \fBuv_process_options_t\fP\&.
.SS Change execution directory
.sp
Set \fBuv_process_options_t.cwd\fP to the corresponding directory.
.SS Set environment variables
.sp
\fBuv_process_options_t.env\fP is a null\-terminated array of strings, each of the
form \fBVAR=VALUE\fP used to set up the environment variables for the process. Set
this to \fBNULL\fP to inherit the environment from the parent (this) process.
.SS Option flags
.sp
Setting \fBuv_process_options_t.flags\fP to a bitwise OR of the following flags,
modifies the child process behaviour:
.INDENT 0.0
.IP \(bu 2
\fBUV_PROCESS_SETUID\fP \- sets the child\(aqs execution user ID to \fBuv_process_options_t.uid\fP\&.
.IP \(bu 2
\fBUV_PROCESS_SETGID\fP \- sets the child\(aqs execution group ID to \fBuv_process_options_t.gid\fP\&.
.UNINDENT
.sp
Changing the UID/GID is only supported on Unix, \fBuv_spawn\fP will fail on
Windows with \fBUV_ENOTSUP\fP\&.
.INDENT 0.0
.IP \(bu 2
\fBUV_PROCESS_WINDOWS_VERBATIM_ARGUMENTS\fP \- No quoting or escaping of
\fBuv_process_options_t.args\fP is done on Windows. Ignored on Unix.
.IP \(bu 2
\fBUV_PROCESS_DETACHED\fP \- Starts the child process in a new session, which
will keep running after the parent process exits. See example below.
.UNINDENT
.SS Detaching processes
.sp
Passing the flag \fBUV_PROCESS_DETACHED\fP can be used to launch daemons, or
child processes which are independent of the parent so that the parent exiting
does not affect it.
.sp
detach/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main() {
loop = uv_default_loop();
char* args[3];
args[0] = "sleep";
args[1] = "100";
args[2] = NULL;
options.exit_cb = NULL;
options.file = "sleep";
options.args = args;
options.flags = UV_PROCESS_DETACHED;
int r;
if ((r = uv_spawn(loop, &child_req, &options))) {
fprintf(stderr, "%s\en", uv_strerror(r));
return 1;
}
fprintf(stderr, "Launched sleep with PID %d\en", child_req.pid);
uv_unref((uv_handle_t*) &child_req);
return uv_run(loop, UV_RUN_DEFAULT);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Just remember that the handle is still monitoring the child, so your program
won\(aqt exit. Use \fBuv_unref()\fP if you want to be more \fIfire\-and\-forget\fP\&.
.SS Sending signals to processes
.sp
libuv wraps the standard \fBkill(2)\fP system call on Unix and implements one
with similar semantics on Windows, with \fIone caveat\fP: all of \fBSIGTERM\fP,
\fBSIGINT\fP and \fBSIGKILL\fP, lead to termination of the process. The signature
of \fBuv_kill\fP is:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_err_t uv_kill(int pid, int signum);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
For processes started using libuv, you may use \fBuv_process_kill\fP instead,
which accepts the \fBuv_process_t\fP watcher as the first argument, rather than
the pid. In this case, \fBremember to call\fP \fBuv_close\fP on the watcher.
.SS Signals
.sp
libuv provides wrappers around Unix signals with \fI\%some Windows support\fP as well.
.sp
Use \fBuv_signal_init()\fP to initialize
a handle and associate it with a loop. To listen for particular signals on
that handler, use \fBuv_signal_start()\fP with the handler function. Each handler
can only be associated with one signal number, with subsequent calls to
\fBuv_signal_start()\fP overwriting earlier associations. Use \fBuv_signal_stop()\fP to
stop watching. Here is a small example demonstrating the various possibilities:
.sp
signal/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <uv.h>
uv_loop_t* create_loop()
{
uv_loop_t *loop = malloc(sizeof(uv_loop_t));
if (loop) {
uv_loop_init(loop);
}
return loop;
}
void signal_handler(uv_signal_t *handle, int signum)
{
printf("Signal received: %d\en", signum);
uv_signal_stop(handle);
}
// two signal handlers in one loop
void thread1_worker(void *userp)
{
uv_loop_t *loop1 = create_loop();
uv_signal_t sig1a, sig1b;
uv_signal_init(loop1, &sig1a);
uv_signal_start(&sig1a, signal_handler, SIGUSR1);
uv_signal_init(loop1, &sig1b);
uv_signal_start(&sig1b, signal_handler, SIGUSR1);
uv_run(loop1, UV_RUN_DEFAULT);
}
// two signal handlers, each in its own loop
void thread2_worker(void *userp)
{
uv_loop_t *loop2 = create_loop();
uv_loop_t *loop3 = create_loop();
uv_signal_t sig2;
uv_signal_init(loop2, &sig2);
uv_signal_start(&sig2, signal_handler, SIGUSR1);
uv_signal_t sig3;
uv_signal_init(loop3, &sig3);
uv_signal_start(&sig3, signal_handler, SIGUSR1);
while (uv_run(loop2, UV_RUN_NOWAIT) || uv_run(loop3, UV_RUN_NOWAIT)) {
}
}
int main()
{
printf("PID %d\en", getpid());
uv_thread_t thread1, thread2;
uv_thread_create(&thread1, thread1_worker, 0);
uv_thread_create(&thread2, thread2_worker, 0);
uv_thread_join(&thread1);
uv_thread_join(&thread2);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
\fBuv_run(loop, UV_RUN_NOWAIT)\fP is similar to \fBuv_run(loop, UV_RUN_ONCE)\fP
in that it will process only one event. UV_RUN_ONCE blocks if there are no
pending events, while UV_RUN_NOWAIT will return immediately. We use NOWAIT
so that one of the loops isn\(aqt starved because the other one has no pending
activity.
.UNINDENT
.UNINDENT
.sp
Send \fBSIGUSR1\fP to the process, and you\(aqll find the handler being invoked
4 times, one for each \fBuv_signal_t\fP\&. The handler just stops each handle,
so that the program exits. This sort of dispatch to all handlers is very
useful. A server using multiple event loops could ensure that all data was
safely saved before termination, simply by every loop adding a watcher for
\fBSIGINT\fP\&.
.SS Child Process I/O
.sp
A normal, newly spawned process has its own set of file descriptors, with 0,
1 and 2 being \fBstdin\fP, \fBstdout\fP and \fBstderr\fP respectively. Sometimes you
may want to share file descriptors with the child. For example, perhaps your
applications launches a sub\-command and you want any errors to go in the log
file, but ignore \fBstdout\fP\&. For this you\(aqd like to have \fBstderr\fP of the
child be the same as the stderr of the parent. In this case, libuv supports
\fIinheriting\fP file descriptors. In this sample, we invoke the test program,
which is:
.sp
proc\-streams/test.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
int main()
{
fprintf(stderr, "This is stderr\en");
printf("This is stdout\en");
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The actual program \fBproc\-streams\fP runs this while sharing only \fBstderr\fP\&.
The file descriptors of the child process are set using the \fBstdio\fP field in
\fBuv_process_options_t\fP\&. First set the \fBstdio_count\fP field to the number of
file descriptors being set. \fBuv_process_options_t.stdio\fP is an array of
\fBuv_stdio_container_t\fP, which is:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
typedef struct uv_stdio_container_s {
uv_stdio_flags flags;
union {
uv_stream_t* stream;
int fd;
} data;
} uv_stdio_container_t;
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
where flags can have several values. Use \fBUV_IGNORE\fP if it isn\(aqt going to be
used. If the first three \fBstdio\fP fields are marked as \fBUV_IGNORE\fP they\(aqll
redirect to \fB/dev/null\fP\&.
.sp
Since we want to pass on an existing descriptor, we\(aqll use \fBUV_INHERIT_FD\fP\&.
Then we set the \fBfd\fP to \fBstderr\fP\&.
.sp
proc\-streams/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int main() {
loop = uv_default_loop();
/* ... */
options.stdio_count = 3;
uv_stdio_container_t child_stdio[3];
child_stdio[0].flags = UV_IGNORE;
child_stdio[1].flags = UV_IGNORE;
child_stdio[2].flags = UV_INHERIT_FD;
child_stdio[2].data.fd = 2;
options.stdio = child_stdio;
options.exit_cb = on_exit;
options.file = args[0];
options.args = args;
int r;
if ((r = uv_spawn(loop, &child_req, &options))) {
fprintf(stderr, "%s\en", uv_strerror(r));
return 1;
}
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
If you run \fBproc\-stream\fP you\(aqll see that only the line "This is stderr" will
be displayed. Try marking \fBstdout\fP as being inherited and see the output.
.sp
It is dead simple to apply this redirection to streams. By setting \fBflags\fP
to \fBUV_INHERIT_STREAM\fP and setting \fBdata.stream\fP to the stream in the
parent process, the child process can treat that stream as standard I/O. This
can be used to implement something like \fI\%CGI\fP\&.
.sp
A sample CGI script/executable is:
.sp
cgi/tick.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <unistd.h>
int main() {
int i;
for (i = 0; i < 10; i++) {
printf("tick\en");
fflush(stdout);
sleep(1);
}
printf("BOOM!\en");
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The CGI server combines the concepts from this chapter and networking so
that every client is sent ten ticks after which that connection is closed.
.sp
cgi/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_new_connection(uv_stream_t *server, int status) {
if (status == \-1) {
// error!
return;
}
uv_tcp_t *client = (uv_tcp_t*) malloc(sizeof(uv_tcp_t));
uv_tcp_init(loop, client);
if (uv_accept(server, (uv_stream_t*) client) == 0) {
invoke_cgi_script(client);
}
else {
uv_close((uv_handle_t*) client, NULL);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Here we simply accept the TCP connection and pass on the socket (\fIstream\fP) to
\fBinvoke_cgi_script\fP\&.
.sp
cgi/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
args[1] = NULL;
/* ... finding the executable path and setting up arguments ... */
options.stdio_count = 3;
uv_stdio_container_t child_stdio[3];
child_stdio[0].flags = UV_IGNORE;
child_stdio[1].flags = UV_INHERIT_STREAM;
child_stdio[1].data.stream = (uv_stream_t*) client;
child_stdio[2].flags = UV_IGNORE;
options.stdio = child_stdio;
options.exit_cb = cleanup_handles;
options.file = args[0];
options.args = args;
// Set this so we can close the socket after the child process exits.
child_req.data = (void*) client;
int r;
if ((r = uv_spawn(loop, &child_req, &options))) {
fprintf(stderr, "%s\en", uv_strerror(r));
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The \fBstdout\fP of the CGI script is set to the socket so that whatever our tick
script prints, gets sent to the client. By using processes, we can offload the
read/write buffering to the operating system, so in terms of convenience this
is great. Just be warned that creating processes is a costly task.
.SS Parent\-child IPC
.sp
A parent and child can have one or two way communication over a pipe created by
settings \fBuv_stdio_container_t.flags\fP to a bit\-wise combination of
\fBUV_CREATE_PIPE\fP and \fBUV_READABLE_PIPE\fP or \fBUV_WRITABLE_PIPE\fP\&. The
read/write flag is from the perspective of the child process. In this case,
the \fBuv_stream_t* stream\fP field must be set to point to an initialized,
unopened \fBuv_pipe_t\fP instance.
.SS New stdio Pipes
.sp
The \fBuv_pipe_t\fP structure represents more than just \fI\%pipe(7)\fP (or \fB|\fP),
but supports any streaming file\-like objects. On Windows, the only object of
that description is the \fI\%Named Pipe\fP\&. On Unix, this could be any of \fI\%Unix
Domain Socket\fP, or derived from \fI\%mkfifo(1)\fP, or it could actually be a
\fI\%pipe(7)\fP\&. When \fBuv_spawn\fP initializes a \fBuv_pipe_t\fP due to the
\fIUV_CREATE_PIPE\fP flag, it opts for creating a \fI\%socketpair(2)\fP\&.
.sp
This is intended for the purpose of allowing multiple libuv processes to
communicate with IPC. This is discussed below.
.SS Arbitrary process IPC
.sp
Since domain sockets [1] can have a well known name and a location in the
file\-system they can be used for IPC between unrelated processes. The \fI\%D\-BUS\fP
system used by open source desktop environments uses domain sockets for event
notification. Various applications can then react when a contact comes online
or new hardware is detected. The MySQL server also runs a domain socket on
which clients can interact with it.
.sp
When using domain sockets, a client\-server pattern is usually followed with the
creator/owner of the socket acting as the server. After the initial setup,
messaging is no different from TCP, so we\(aqll re\-use the echo server example.
.sp
pipe\-echo\-server/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void remove_sock(int sig) {
uv_fs_t req;
uv_fs_unlink(loop, &req, PIPENAME, NULL);
exit(0);
}
int main() {
loop = uv_default_loop();
uv_pipe_t server;
uv_pipe_init(loop, &server, 0);
signal(SIGINT, remove_sock);
int r;
if ((r = uv_pipe_bind(&server, PIPENAME))) {
fprintf(stderr, "Bind error %s\en", uv_err_name(r));
return 1;
}
if ((r = uv_listen((uv_stream_t*) &server, 128, on_new_connection))) {
fprintf(stderr, "Listen error %s\en", uv_err_name(r));
return 2;
}
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
We name the socket \fBecho.sock\fP which means it will be created in the local
directory. This socket now behaves no different from TCP sockets as far as
the stream API is concerned. You can test this server using \fI\%socat\fP:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
$ socat \- /path/to/socket
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
A client which wants to connect to a domain socket will use:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void uv_pipe_connect(uv_connect_t *req, uv_pipe_t *handle, const char *name, uv_connect_cb cb);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
where \fBname\fP will be \fBecho.sock\fP or similar. On Unix systems, \fBname\fP must
point to a valid file (e.g. \fB/tmp/echo.sock\fP). On Windows, \fBname\fP follows a
\fB\e\e?\epipe\eecho.sock\fP format.
.SS Sending file descriptors over pipes
.sp
The cool thing about domain sockets is that file descriptors can be exchanged
between processes by sending them over a domain socket. This allows processes
to hand off their I/O to other processes. Applications include load\-balancing
servers, worker processes and other ways to make optimum use of CPU. libuv only
supports sending \fBTCP sockets or other pipes\fP over pipes for now.
.sp
To demonstrate, we will look at a echo server implementation that hands of
clients to worker processes in a round\-robin fashion. This program is a bit
involved, and while only snippets are included in the book, it is recommended
to read the full code to really understand it.
.sp
The worker process is quite simple, since the file\-descriptor is handed over to
it by the master.
.sp
multi\-echo\-server/worker.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_loop_t *loop;
uv_pipe_t queue;
int main() {
loop = uv_default_loop();
uv_pipe_init(loop, &queue, 1 /* ipc */);
uv_pipe_open(&queue, 0);
uv_read_start((uv_stream_t*)&queue, alloc_buffer, on_new_connection);
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBqueue\fP is the pipe connected to the master process on the other end, along
which new file descriptors get sent. It is important to set the \fBipc\fP
argument of \fBuv_pipe_init\fP to 1 to indicate this pipe will be used for
inter\-process communication! Since the master will write the file handle to the
standard input of the worker, we connect the pipe to \fBstdin\fP using
\fBuv_pipe_open\fP\&.
.sp
multi\-echo\-server/worker.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_new_connection(uv_stream_t *q, ssize_t nread, const uv_buf_t *buf) {
if (nread < 0) {
if (nread != UV_EOF)
fprintf(stderr, "Read error %s\en", uv_err_name(nread));
uv_close((uv_handle_t*) q, NULL);
return;
}
uv_pipe_t *pipe = (uv_pipe_t*) q;
if (!uv_pipe_pending_count(pipe)) {
fprintf(stderr, "No pending count\en");
return;
}
uv_handle_type pending = uv_pipe_pending_type(pipe);
assert(pending == UV_TCP);
uv_tcp_t *client = (uv_tcp_t*) malloc(sizeof(uv_tcp_t));
uv_tcp_init(loop, client);
if (uv_accept(q, (uv_stream_t*) client) == 0) {
uv_os_fd_t fd;
uv_fileno((const uv_handle_t*) client, &fd);
fprintf(stderr, "Worker %d: Accepted fd %d\en", getpid(), fd);
uv_read_start((uv_stream_t*) client, alloc_buffer, echo_read);
}
else {
uv_close((uv_handle_t*) client, NULL);
}
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
First we call \fBuv_pipe_pending_count()\fP to ensure that a handle is available
to read out. If your program could deal with different types of handles,
\fBuv_pipe_pending_type()\fP can be used to determine the type.
Although \fBaccept\fP seems odd in this code, it actually makes sense. What
\fBaccept\fP traditionally does is get a file descriptor (the client) from
another file descriptor (The listening socket). Which is exactly what we do
here. Fetch the file descriptor (\fBclient\fP) from \fBqueue\fP\&. From this point
the worker does standard echo server stuff.
.sp
Turning now to the master, let\(aqs take a look at how the workers are launched to
allow load balancing.
.sp
multi\-echo\-server/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
struct child_worker {
uv_process_t req;
uv_process_options_t options;
uv_pipe_t pipe;
} *workers;
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The \fBchild_worker\fP structure wraps the process, and the pipe between the
master and the individual process.
.sp
multi\-echo\-server/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void setup_workers() {
round_robin_counter = 0;
// ...
// launch same number of workers as number of CPUs
uv_cpu_info_t *info;
int cpu_count;
uv_cpu_info(&info, &cpu_count);
uv_free_cpu_info(info, cpu_count);
child_worker_count = cpu_count;
workers = calloc(cpu_count, sizeof(struct child_worker));
while (cpu_count\-\-) {
struct child_worker *worker = &workers[cpu_count];
uv_pipe_init(loop, &worker\->pipe, 1);
uv_stdio_container_t child_stdio[3];
child_stdio[0].flags = UV_CREATE_PIPE | UV_READABLE_PIPE;
child_stdio[0].data.stream = (uv_stream_t*) &worker\->pipe;
child_stdio[1].flags = UV_IGNORE;
child_stdio[2].flags = UV_INHERIT_FD;
child_stdio[2].data.fd = 2;
worker\->options.stdio = child_stdio;
worker\->options.stdio_count = 3;
worker\->options.exit_cb = close_process_handle;
worker\->options.file = args[0];
worker\->options.args = args;
uv_spawn(loop, &worker\->req, &worker\->options);
fprintf(stderr, "Started worker %d\en", worker\->req.pid);
}
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
In setting up the workers, we use the nifty libuv function \fBuv_cpu_info\fP to
get the number of CPUs so we can launch an equal number of workers. Again it is
important to initialize the pipe acting as the IPC channel with the third
argument as 1. We then indicate that the child process\(aq \fBstdin\fP is to be
a readable pipe (from the point of view of the child). Everything is
straightforward till here. The workers are launched and waiting for file
descriptors to be written to their standard input.
.sp
It is in \fBon_new_connection\fP (the TCP infrastructure is initialized in
\fBmain()\fP), that we accept the client socket and pass it along to the next
worker in the round\-robin.
.sp
multi\-echo\-server/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_new_connection(uv_stream_t *server, int status) {
if (status == \-1) {
// error!
return;
}
uv_tcp_t *client = (uv_tcp_t*) malloc(sizeof(uv_tcp_t));
uv_tcp_init(loop, client);
if (uv_accept(server, (uv_stream_t*) client) == 0) {
uv_write_t *write_req = (uv_write_t*) malloc(sizeof(uv_write_t));
dummy_buf = uv_buf_init("a", 1);
struct child_worker *worker = &workers[round_robin_counter];
uv_write2(write_req, (uv_stream_t*) &worker\->pipe, &dummy_buf, 1, (uv_stream_t*) client, NULL);
round_robin_counter = (round_robin_counter + 1) % child_worker_count;
}
else {
uv_close((uv_handle_t*) client, NULL);
}
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The \fBuv_write2\fP call handles all the abstraction and it is simply a matter of
passing in the handle (\fBclient\fP) as the right argument. With this our
multi\-process echo server is operational.
.sp
Thanks to Kyle for \fI\%pointing out\fP that \fBuv_write2()\fP requires a non\-empty
buffer even when sending handles.
.sp
.ce
----
.ce 0
.sp
.IP [1] 5
In this section domain sockets stands in for named pipes on Windows as
well.
.SS Advanced event loops
.sp
libuv provides considerable user control over event loops, and you can achieve
interesting results by juggling multiple loops. You can also embed libuv\(aqs
event loop into another event loop based library \-\- imagine a Qt based UI, and
Qt\(aqs event loop driving a libuv backend which does intensive system level
tasks.
.SS Stopping an event loop
.sp
\fBuv_stop()\fP can be used to stop an event loop. The earliest the loop will
stop running is \fIon the next iteration\fP, possibly later. This means that events
that are ready to be processed in this iteration of the loop will still be
processed, so \fBuv_stop()\fP can\(aqt be used as a kill switch. When \fBuv_stop()\fP
is called, the loop \fBwon\(aqt\fP block for i/o on this iteration. The semantics of
these things can be a bit difficult to understand, so let\(aqs look at
\fBuv_run()\fP where all the control flow occurs.
.sp
src/unix/core.c \- uv_run
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
static void uv__run_closing_handles(uv_loop_t* loop) {
uv_handle_t* p;
uv_handle_t* q;
p = loop\->closing_handles;
loop\->closing_handles = NULL;
while (p) {
q = p\->next_closing;
uv__finish_close(p);
p = q;
}
}
int uv_is_closing(const uv_handle_t* handle) {
return uv__is_closing(handle);
}
int uv_backend_fd(const uv_loop_t* loop) {
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBstop_flag\fP is set by \fBuv_stop()\fP\&. Now all libuv callbacks are invoked
within the event loop, which is why invoking \fBuv_stop()\fP in them will still
lead to this iteration of the loop occurring. First libuv updates timers, then
runs pending timer, idle and prepare callbacks, and invokes any pending I/O
callbacks. If you were to call \fBuv_stop()\fP in any of them, \fBstop_flag\fP
would be set. This causes \fBuv_backend_timeout()\fP to return \fB0\fP, which is
why the loop does not block on I/O. If on the other hand, you called
\fBuv_stop()\fP in one of the check handlers, I/O has already finished and is not
affected.
.sp
\fBuv_stop()\fP is useful to shutdown a loop when a result has been computed or
there is an error, without having to ensure that all handlers are stopped one
by one.
.sp
Here is a simple example that stops the loop and demonstrates how the current
iteration of the loop still takes places.
.sp
uvstop/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <uv.h>
int64_t counter = 0;
void idle_cb(uv_idle_t *handle) {
printf("Idle callback\en");
counter++;
if (counter >= 5) {
uv_stop(uv_default_loop());
printf("uv_stop() called\en");
}
}
void prep_cb(uv_prepare_t *handle) {
printf("Prep callback\en");
}
int main() {
uv_idle_t idler;
uv_prepare_t prep;
uv_idle_init(uv_default_loop(), &idler);
uv_idle_start(&idler, idle_cb);
uv_prepare_init(uv_default_loop(), &prep);
uv_prepare_start(&prep, prep_cb);
uv_run(uv_default_loop(), UV_RUN_DEFAULT);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Utilities
.sp
This chapter catalogues tools and techniques which are useful for common tasks.
The \fI\%libev man page\fP already covers some patterns which can be adopted to
libuv through simple API changes. It also covers parts of the libuv API that
don\(aqt require entire chapters dedicated to them.
.SS Timers
.sp
Timers invoke the callback after a certain time has elapsed since the timer was
started. libuv timers can also be set to invoke at regular intervals instead of
just once.
.sp
Simple use is to init a watcher and start it with a \fBtimeout\fP, and optional \fBrepeat\fP\&.
Timers can be stopped at any time.
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_timer_t timer_req;
uv_timer_init(loop, &timer_req);
uv_timer_start(&timer_req, callback, 5000, 2000);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
will start a repeating timer, which first starts 5 seconds (the \fBtimeout\fP) after the execution
of \fBuv_timer_start\fP, then repeats every 2 seconds (the \fBrepeat\fP). Use:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_timer_stop(&timer_req);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
to stop the timer. This can be used safely from within the callback as well.
.sp
The repeat interval can be modified at any time with:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_timer_set_repeat(uv_timer_t *timer, int64_t repeat);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
which will take effect \fBwhen possible\fP\&. If this function is called from
a timer callback, it means:
.INDENT 0.0
.IP \(bu 2
If the timer was non\-repeating, the timer has already been stopped. Use
\fBuv_timer_start\fP again.
.IP \(bu 2
If the timer is repeating, the next timeout has already been scheduled, so
the old repeat interval will be used once more before the timer switches to
the new interval.
.UNINDENT
.sp
The utility function:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int uv_timer_again(uv_timer_t *)
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
applies \fBonly to repeating timers\fP and is equivalent to stopping the timer
and then starting it with both initial \fBtimeout\fP and \fBrepeat\fP set to the
old \fBrepeat\fP value. If the timer hasn\(aqt been started it fails (error code
\fBUV_EINVAL\fP) and returns \-1.
.sp
An actual timer example is in the \fI\%reference count section\fP\&.
.SS Event loop reference count
.sp
The event loop only runs as long as there are active handles. This system
works by having every handle increase the reference count of the event loop
when it is started and decreasing the reference count when stopped. It is also
possible to manually change the reference count of handles using:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void uv_ref(uv_handle_t*);
void uv_unref(uv_handle_t*);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
These functions can be used to allow a loop to exit even when a watcher is
active or to use custom objects to keep the loop alive.
.sp
The latter can be used with interval timers. You might have a garbage collector
which runs every X seconds, or your network service might send a heartbeat to
others periodically, but you don\(aqt want to have to stop them along all clean
exit paths or error scenarios. Or you want the program to exit when all your
other watchers are done. In that case just unref the timer immediately after
creation so that if it is the only watcher running then \fBuv_run\fP will still
exit.
.sp
This is also used in node.js where some libuv methods are being bubbled up to
the JS API. A \fBuv_handle_t\fP (the superclass of all watchers) is created per
JS object and can be ref/unrefed.
.sp
ref\-timer/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_loop_t *loop;
uv_timer_t gc_req;
uv_timer_t fake_job_req;
int main() {
loop = uv_default_loop();
uv_timer_init(loop, &gc_req);
uv_unref((uv_handle_t*) &gc_req);
uv_timer_start(&gc_req, gc, 0, 2000);
// could actually be a TCP download or something
uv_timer_init(loop, &fake_job_req);
uv_timer_start(&fake_job_req, fake_job, 9000, 0);
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
We initialize the garbage collector timer, then immediately \fBunref\fP it.
Observe how after 9 seconds, when the fake job is done, the program
automatically exits, even though the garbage collector is still running.
.SS Idler pattern
.sp
The callbacks of idle handles are invoked once per event loop. The idle
callback can be used to perform some very low priority activity. For example,
you could dispatch a summary of the daily application performance to the
developers for analysis during periods of idleness, or use the application\(aqs
CPU time to perform SETI calculations :) An idle watcher is also useful in
a GUI application. Say you are using an event loop for a file download. If the
TCP socket is still being established and no other events are present your
event loop will pause (\fBblock\fP), which means your progress bar will freeze
and the user will face an unresponsive application. In such a case queue up and
idle watcher to keep the UI operational.
.sp
idle\-compute/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_loop_t *loop;
uv_fs_t stdin_watcher;
uv_idle_t idler;
char buffer[1024];
int main() {
loop = uv_default_loop();
uv_idle_init(loop, &idler);
uv_buf_t buf = uv_buf_init(buffer, 1024);
uv_fs_read(loop, &stdin_watcher, 0, &buf, 1, \-1, on_type);
uv_idle_start(&idler, crunch_away);
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Here we initialize the idle watcher and queue it up along with the actual
events we are interested in. \fBcrunch_away\fP will now be called repeatedly
until the user types something and presses Return. Then it will be interrupted
for a brief amount as the loop deals with the input data, after which it will
keep calling the idle callback again.
.sp
idle\-compute/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void crunch_away(uv_idle_t* handle) {
// Compute extra\-terrestrial life
// fold proteins
// computer another digit of PI
// or similar
fprintf(stderr, "Computing PI...\en");
// just to avoid overwhelming your terminal emulator
uv_idle_stop(handle);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Passing data to worker thread
.sp
When using \fBuv_queue_work\fP you\(aqll usually need to pass complex data through
to the worker thread. The solution is to use a \fBstruct\fP and set
\fBuv_work_t.data\fP to point to it. A slight variation is to have the
\fBuv_work_t\fP itself as the first member of this struct (called a baton [1]).
This allows cleaning up the work request and all the data in one free call.
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
struct ftp_baton {
uv_work_t req;
char *host;
int port;
char *username;
char *password;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
ftp_baton *baton = (ftp_baton*) malloc(sizeof(ftp_baton));
baton\->req.data = (void*) baton;
baton\->host = strdup("my.webhost.com");
baton\->port = 21;
// ...
uv_queue_work(loop, &baton\->req, ftp_session, ftp_cleanup);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Here we create the baton and queue the task.
.sp
Now the task function can extract the data it needs:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void ftp_session(uv_work_t *req) {
ftp_baton *baton = (ftp_baton*) req\->data;
fprintf(stderr, "Connecting to %s\en", baton\->host);
}
void ftp_cleanup(uv_work_t *req) {
ftp_baton *baton = (ftp_baton*) req\->data;
free(baton\->host);
// ...
free(baton);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
We then free the baton which also frees the watcher.
.SS External I/O with polling
.sp
Usually third\-party libraries will handle their own I/O, and keep track of
their sockets and other files internally. In this case it isn\(aqt possible to use
the standard stream I/O operations, but the library can still be integrated
into the libuv event loop. All that is required is that the library allow you
to access the underlying file descriptors and provide functions that process
tasks in small increments as decided by your application. Some libraries though
will not allow such access, providing only a standard blocking function which
will perform the entire I/O transaction and only then return. It is unwise to
use these in the event loop thread, use the threadpool instead. Of
course, this will also mean losing granular control on the library.
.sp
The \fBuv_poll\fP section of libuv simply watches file descriptors using the
operating system notification mechanism. In some sense, all the I/O operations
that libuv implements itself are also backed by \fBuv_poll\fP like code. Whenever
the OS notices a change of state in file descriptors being polled, libuv will
invoke the associated callback.
.sp
Here we will walk through a simple download manager that will use \fI\%libcurl\fP to
download files. Rather than give all control to libcurl, we\(aqll instead be
using the libuv event loop, and use the non\-blocking, async \fI\%multi\fP interface to
progress with the download whenever libuv notifies of I/O readiness.
.sp
uvwget/main.c \- The setup
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <assert.h>
#include <stdio.h>
#include <stdlib.h>
#include <uv.h>
#include <curl/curl.h>
uv_loop_t *loop;
CURLM *curl_handle;
uv_timer_t timeout;
}
int main(int argc, char **argv) {
loop = uv_default_loop();
if (argc <= 1)
return 0;
if (curl_global_init(CURL_GLOBAL_ALL)) {
fprintf(stderr, "Could not init cURL\en");
return 1;
}
uv_timer_init(loop, &timeout);
curl_handle = curl_multi_init();
curl_multi_setopt(curl_handle, CURLMOPT_SOCKETFUNCTION, handle_socket);
curl_multi_setopt(curl_handle, CURLMOPT_TIMERFUNCTION, start_timeout);
while (argc\-\- > 1) {
add_download(argv[argc], argc);
}
uv_run(loop, UV_RUN_DEFAULT);
curl_multi_cleanup(curl_handle);
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The way each library is integrated with libuv will vary. In the case of
libcurl, we can register two callbacks. The socket callback \fBhandle_socket\fP
is invoked whenever the state of a socket changes and we have to start polling
it. \fBstart_timeout\fP is called by libcurl to notify us of the next timeout
interval, after which we should drive libcurl forward regardless of I/O status.
This is so that libcurl can handle errors or do whatever else is required to
get the download moving.
.sp
Our downloader is to be invoked as:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
$ ./uvwget [url1] [url2] ...
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
So we add each argument as an URL
.sp
uvwget/main.c \- Adding urls
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void add_download(const char *url, int num) {
char filename[50];
sprintf(filename, "%d.download", num);
FILE *file;
file = fopen(filename, "w");
if (file == NULL) {
fprintf(stderr, "Error opening %s\en", filename);
return;
}
CURL *handle = curl_easy_init();
curl_easy_setopt(handle, CURLOPT_WRITEDATA, file);
curl_easy_setopt(handle, CURLOPT_URL, url);
curl_multi_add_handle(curl_handle, handle);
fprintf(stderr, "Added download %s \-> %s\en", url, filename);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
We let libcurl directly write the data to a file, but much more is possible if
you so desire.
.sp
\fBstart_timeout\fP will be called immediately the first time by libcurl, so
things are set in motion. This simply starts a libuv \fI\%timer\fP which
drives \fBcurl_multi_socket_action\fP with \fBCURL_SOCKET_TIMEOUT\fP whenever it
times out. \fBcurl_multi_socket_action\fP is what drives libcurl, and what we
call whenever sockets change state. But before we go into that, we need to poll
on sockets whenever \fBhandle_socket\fP is called.
.sp
uvwget/main.c \- Setting up polling
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void start_timeout(CURLM *multi, long timeout_ms, void *userp) {
if (timeout_ms <= 0)
timeout_ms = 1; /* 0 means directly call socket_action, but we\(aqll do it in a bit */
uv_timer_start(&timeout, on_timeout, timeout_ms, 0);
}
int handle_socket(CURL *easy, curl_socket_t s, int action, void *userp, void *socketp) {
curl_context_t *curl_context;
if (action == CURL_POLL_IN || action == CURL_POLL_OUT) {
if (socketp) {
curl_context = (curl_context_t*) socketp;
}
else {
curl_context = create_curl_context(s);
curl_multi_assign(curl_handle, s, (void *) curl_context);
}
}
switch (action) {
case CURL_POLL_IN:
uv_poll_start(&curl_context\->poll_handle, UV_READABLE, curl_perform);
break;
case CURL_POLL_OUT:
uv_poll_start(&curl_context\->poll_handle, UV_WRITABLE, curl_perform);
break;
case CURL_POLL_REMOVE:
if (socketp) {
uv_poll_stop(&((curl_context_t*)socketp)\->poll_handle);
destroy_curl_context((curl_context_t*) socketp);
curl_multi_assign(curl_handle, s, NULL);
}
break;
default:
abort();
}
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
We are interested in the socket fd \fBs\fP, and the \fBaction\fP\&. For every socket
we create a \fBuv_poll_t\fP handle if it doesn\(aqt exist, and associate it with the
socket using \fBcurl_multi_assign\fP\&. This way \fBsocketp\fP points to it whenever
the callback is invoked.
.sp
In the case that the download is done or fails, libcurl requests removal of the
poll. So we stop and free the poll handle.
.sp
Depending on what events libcurl wishes to watch for, we start polling with
\fBUV_READABLE\fP or \fBUV_WRITABLE\fP\&. Now libuv will invoke the poll callback
whenever the socket is ready for reading or writing. Calling \fBuv_poll_start\fP
multiple times on the same handle is acceptable, it will just update the events
mask with the new value. \fBcurl_perform\fP is the crux of this program.
.sp
uvwget/main.c \- Driving libcurl.
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void curl_perform(uv_poll_t *req, int status, int events) {
uv_timer_stop(&timeout);
int running_handles;
int flags = 0;
if (status < 0) flags = CURL_CSELECT_ERR;
if (!status && events & UV_READABLE) flags |= CURL_CSELECT_IN;
if (!status && events & UV_WRITABLE) flags |= CURL_CSELECT_OUT;
curl_context_t *context;
context = (curl_context_t*)req;
curl_multi_socket_action(curl_handle, context\->sockfd, flags, &running_handles);
check_multi_info();
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The first thing we do is to stop the timer, since there has been some progress
in the interval. Then depending on what event triggered the callback, we set
the correct flags. Then we call \fBcurl_multi_socket_action\fP with the socket
that progressed and the flags informing about what events happened. At this
point libcurl does all of its internal tasks in small increments, and will
attempt to return as fast as possible, which is exactly what an evented program
wants in its main thread. libcurl keeps queueing messages into its own queue
about transfer progress. In our case we are only interested in transfers that
are completed. So we extract these messages, and clean up handles whose
transfers are done.
.sp
uvwget/main.c \- Reading transfer status.
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void check_multi_info(void) {
char *done_url;
CURLMsg *message;
int pending;
while ((message = curl_multi_info_read(curl_handle, &pending))) {
switch (message\->msg) {
case CURLMSG_DONE:
curl_easy_getinfo(message\->easy_handle, CURLINFO_EFFECTIVE_URL,
&done_url);
printf("%s DONE\en", done_url);
curl_multi_remove_handle(curl_handle, message\->easy_handle);
curl_easy_cleanup(message\->easy_handle);
break;
default:
fprintf(stderr, "CURLMSG default\en");
abort();
}
}
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Check & Prepare watchers
.sp
TODO
.SS Loading libraries
.sp
libuv provides a cross platform API to dynamically load \fI\%shared libraries\fP\&.
This can be used to implement your own plugin/extension/module system and is
used by node.js to implement \fBrequire()\fP support for bindings. The usage is
quite simple as long as your library exports the right symbols. Be careful with
sanity and security checks when loading third party code, otherwise your
program will behave unpredictably. This example implements a very simple
plugin system which does nothing except print the name of the plugin.
.sp
Let us first look at the interface provided to plugin authors.
.sp
plugin/plugin.h
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#ifndef UVBOOK_PLUGIN_SYSTEM
#define UVBOOK_PLUGIN_SYSTEM
// Plugin authors should use this to register their plugins with mfp.
void mfp_register(const char *name);
#endif
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
You can similarly add more functions that plugin authors can use to do useful
things in your application [2]\&. A sample plugin using this API is:
.sp
plugin/hello.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include "plugin.h"
void initialize() {
mfp_register("Hello World!");
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Our interface defines that all plugins should have an \fBinitialize\fP function
which will be called by the application. This plugin is compiled as a shared
library and can be loaded by running our application:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
$ ./plugin libhello.dylib
Loading libhello.dylib
Registered plugin "Hello World!"
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
The shared library filename will be different depending on platforms. On
Linux it is \fBlibhello.so\fP\&.
.UNINDENT
.UNINDENT
.sp
This is done by using \fBuv_dlopen\fP to first load the shared library
\fBlibhello.dylib\fP\&. Then we get access to the \fBinitialize\fP function using
\fBuv_dlsym\fP and invoke it.
.sp
plugin/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include "plugin.h"
typedef void (*init_plugin_function)();
void mfp_register(const char *name) {
fprintf(stderr, "Registered plugin \e"%s\e"\en", name);
}
int main(int argc, char **argv) {
if (argc == 1) {
fprintf(stderr, "Usage: %s [plugin1] [plugin2] ...\en", argv[0]);
return 0;
}
uv_lib_t *lib = (uv_lib_t*) malloc(sizeof(uv_lib_t));
while (\-\-argc) {
fprintf(stderr, "Loading %s\en", argv[argc]);
if (uv_dlopen(argv[argc], lib)) {
fprintf(stderr, "Error: %s\en", uv_dlerror(lib));
continue;
}
init_plugin_function init_plugin;
if (uv_dlsym(lib, "initialize", (void **) &init_plugin)) {
fprintf(stderr, "dlsym error: %s\en", uv_dlerror(lib));
continue;
}
init_plugin();
}
return 0;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBuv_dlopen\fP expects a path to the shared library and sets the opaque
\fBuv_lib_t\fP pointer. It returns 0 on success, \-1 on error. Use \fBuv_dlerror\fP
to get the error message.
.sp
\fBuv_dlsym\fP stores a pointer to the symbol in the second argument in the third
argument. \fBinit_plugin_function\fP is a function pointer to the sort of
function we are looking for in the application\(aqs plugins.
.SS TTY
.sp
Text terminals have supported basic formatting for a long time, with a \fI\%pretty
standardised\fP command set. This formatting is often used by programs to
improve the readability of terminal output. For example \fBgrep \-\-colour\fP\&.
libuv provides the \fBuv_tty_t\fP abstraction (a stream) and related functions to
implement the ANSI escape codes across all platforms. By this I mean that libuv
converts ANSI codes to the Windows equivalent, and provides functions to get
terminal information.
.sp
The first thing to do is to initialize a \fBuv_tty_t\fP with the file descriptor
it reads/writes from. This is achieved with:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
int uv_tty_init(uv_loop_t*, uv_tty_t*, uv_file fd, int unused)
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The \fBunused\fP parameter is now auto\-detected and ignored. It previously needed
to be set to use \fBuv_read_start()\fP on the stream.
.sp
It is then best to use \fBuv_tty_set_mode\fP to set the mode to \fInormal\fP
which enables most TTY formatting, flow\-control and other settings. \fI\%Other\fP modes
are also available.
.sp
Remember to call \fBuv_tty_reset_mode\fP when your program exits to restore the
state of the terminal. Just good manners. Another set of good manners is to be
aware of redirection. If the user redirects the output of your command to
a file, control sequences should not be written as they impede readability and
\fBgrep\fP\&. To check if the file descriptor is indeed a TTY, call
\fBuv_guess_handle\fP with the file descriptor and compare the return value with
\fBUV_TTY\fP\&.
.sp
Here is a simple example which prints white text on a red background:
.sp
tty/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include <uv.h>
uv_loop_t *loop;
uv_tty_t tty;
int main() {
loop = uv_default_loop();
uv_tty_init(loop, &tty, STDOUT_FILENO, 0);
uv_tty_set_mode(&tty, UV_TTY_MODE_NORMAL);
if (uv_guess_handle(1) == UV_TTY) {
uv_write_t req;
uv_buf_t buf;
buf.base = "\e033[41;37m";
buf.len = strlen(buf.base);
uv_write(&req, (uv_stream_t*) &tty, &buf, 1, NULL);
}
uv_write_t req;
uv_buf_t buf;
buf.base = "Hello TTY\en";
buf.len = strlen(buf.base);
uv_write(&req, (uv_stream_t*) &tty, &buf, 1, NULL);
uv_tty_reset_mode();
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The final TTY helper is \fBuv_tty_get_winsize()\fP which is used to get the
width and height of the terminal and returns \fB0\fP on success. Here is a small
program which does some animation using the function and character position
escape codes.
.sp
tty\-gravity/main.c
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include <uv.h>
uv_loop_t *loop;
uv_tty_t tty;
uv_timer_t tick;
uv_write_t write_req;
int width, height;
int pos = 0;
char *message = " Hello TTY ";
void update(uv_timer_t *req) {
char data[500];
uv_buf_t buf;
buf.base = data;
buf.len = sprintf(data, "\e033[2J\e033[H\e033[%dB\e033[%luC\e033[42;37m%s",
pos,
(unsigned long) (width\-strlen(message))/2,
message);
uv_write(&write_req, (uv_stream_t*) &tty, &buf, 1, NULL);
pos++;
if (pos > height) {
uv_tty_reset_mode();
uv_timer_stop(&tick);
}
}
int main() {
loop = uv_default_loop();
uv_tty_init(loop, &tty, STDOUT_FILENO, 0);
uv_tty_set_mode(&tty, 0);
if (uv_tty_get_winsize(&tty, &width, &height)) {
fprintf(stderr, "Could not get TTY information\en");
uv_tty_reset_mode();
return 1;
}
fprintf(stderr, "Width %d, height %d\en", width, height);
uv_timer_init(loop, &tick);
uv_timer_start(&tick, update, 200, 200);
return uv_run(loop, UV_RUN_DEFAULT);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The escape codes are:
.TS
center;
|l|l|.
_
T{
Code
T} T{
Meaning
T}
_
T{
\fI2\fP J
T} T{
Clear part of the screen, 2 is entire screen
T}
_
T{
H
T} T{
Moves cursor to certain position, default top\-left
T}
_
T{
\fIn\fP B
T} T{
Moves cursor down by n lines
T}
_
T{
\fIn\fP C
T} T{
Moves cursor right by n columns
T}
_
T{
m
T} T{
Obeys string of display settings, in this case green background (40+2), white text (30+7)
T}
_
.TE
.sp
As you can see this is very useful to produce nicely formatted output, or even
console based arcade games if that tickles your fancy. For fancier control you
can try \fI\%ncurses\fP\&.
.sp
Changed in version 1.23.1:: the \fIreadable\fP parameter is now unused and ignored.
The appropriate value will now be auto\-detected from the kernel.
.sp
.ce
----
.ce 0
.sp
.IP [1] 5
I was first introduced to the term baton in this context, in Konstantin
Käfer\(aqs excellent slides on writing node.js bindings \-\-
\fI\%https://kkaefer.com/node\-cpp\-modules/#baton\fP
.IP [2] 5
mfp is My Fancy Plugin
.SS About
.sp
\fI\%Nikhil Marathe\fP started writing this book one
afternoon (June 16, 2012) when he didn\(aqt feel like programming. He had recently
been stung by the lack of good documentation on libuv while working on
\fI\%node\-taglib\fP\&. Although reference
documentation was present, there were no comprehensive tutorials. This book is
the output of that need and tries to be accurate. That said, the book may have
mistakes. Pull requests are encouraged.
.sp
Nikhil is indebted to Marc Lehmann\(aqs comprehensive \fI\%man page\fP about libev which
describes much of the semantics of the two libraries.
.sp
This book was made using \fI\%Sphinx\fP and \fI\%vim\fP\&.
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
In 2017 the libuv project incorporated the Nikhil\(aqs work into the official
documentation and it\(aqs maintained there henceforth.
.UNINDENT
.UNINDENT
.SS Upgrading
.sp
Migration guides for different libuv versions, starting with 1.0.
.SS libuv 0.10 \-> 1.0.0 migration guide
.sp
Some APIs changed quite a bit throughout the 1.0.0 development process. Here
is a migration guide for the most significant changes that happened after 0.10
was released.
.SS Loop initialization and closing
.sp
In libuv 0.10 (and previous versions), loops were created with \fIuv_loop_new\fP, which
allocated memory for a new loop and initialized it; and destroyed with \fIuv_loop_delete\fP,
which destroyed the loop and freed the memory. Starting with 1.0, those are deprecated
and the user is responsible for allocating the memory and then initializing the loop.
.sp
libuv 0.10
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_loop_t* loop = uv_loop_new();
\&...
uv_loop_delete(loop);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
libuv 1.0
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_loop_t* loop = malloc(sizeof *loop);
uv_loop_init(loop);
\&...
uv_loop_close(loop);
free(loop);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
Error handling was omitted for brevity. Check the documentation for \fBuv_loop_init()\fP
and \fBuv_loop_close()\fP\&.
.UNINDENT
.UNINDENT
.SS Error handling
.sp
Error handling had a major overhaul in libuv 1.0. In general, functions and status parameters
would get 0 for success and \-1 for failure on libuv 0.10, and the user had to use \fIuv_last_error\fP
to fetch the error code, which was a positive number.
.sp
In 1.0, functions and status parameters contain the actual error code, which is 0 for success, or
a negative number in case of error.
.sp
libuv 0.10
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
\&... assume \(aqserver\(aq is a TCP server which is already listening
r = uv_listen((uv_stream_t*) server, 511, NULL);
if (r == \-1) {
uv_err_t err = uv_last_error(uv_default_loop());
/* err.code contains UV_EADDRINUSE */
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
libuv 1.0
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
\&... assume \(aqserver\(aq is a TCP server which is already listening
r = uv_listen((uv_stream_t*) server, 511, NULL);
if (r < 0) {
/* r contains UV_EADDRINUSE */
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Threadpool changes
.sp
In libuv 0.10 Unix used a threadpool which defaulted to 4 threads, while Windows used the
\fIQueueUserWorkItem\fP API, which uses a Windows internal threadpool, which defaults to 512
threads per process.
.sp
In 1.0, we unified both implementations, so Windows now uses the same implementation Unix
does. The threadpool size can be set by exporting the \fBUV_THREADPOOL_SIZE\fP environment
variable. See threadpool\&.
.SS Allocation callback API change
.sp
In libuv 0.10 the callback had to return a filled \fBuv_buf_t\fP by value:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
uv_buf_t alloc_cb(uv_handle_t* handle, size_t size) {
return uv_buf_init(malloc(size), size);
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
In libuv 1.0 a pointer to a buffer is passed to the callback, which the user
needs to fill:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void alloc_cb(uv_handle_t* handle, size_t size, uv_buf_t* buf) {
buf\->base = malloc(size);
buf\->len = size;
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Unification of IPv4 / IPv6 APIs
.sp
libuv 1.0 unified the IPv4 and IPv6 APIS. There is no longer a \fIuv_tcp_bind\fP and \fIuv_tcp_bind6\fP
duality, there is only \fBuv_tcp_bind()\fP now.
.sp
IPv4 functions took \fBstruct sockaddr_in\fP structures by value, and IPv6 functions took
\fBstruct sockaddr_in6\fP\&. Now functions take a \fBstruct sockaddr*\fP (note it\(aqs a pointer).
It can be stack allocated.
.sp
libuv 0.10
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
struct sockaddr_in addr = uv_ip4_addr("0.0.0.0", 1234);
\&...
uv_tcp_bind(&server, addr)
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
libuv 1.0
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
struct sockaddr_in addr;
uv_ip4_addr("0.0.0.0", 1234, &addr)
\&...
uv_tcp_bind(&server, (const struct sockaddr*) &addr, 0);
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The IPv4 and IPv6 struct creating functions (\fBuv_ip4_addr()\fP and \fBuv_ip6_addr()\fP)
have also changed, make sure you check the documentation.
.INDENT 0.0
.TP
.B \&..note::
This change applies to all functions that made a distinction between IPv4 and IPv6
addresses.
.UNINDENT
.SS Streams / UDP data receive callback API change
.sp
The streams and UDP data receive callbacks now get a pointer to a \fBuv_buf_t\fP buffer,
not a structure by value.
.sp
libuv 0.10
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_read(uv_stream_t* handle,
ssize_t nread,
uv_buf_t buf) {
...
}
void recv_cb(uv_udp_t* handle,
ssize_t nread,
uv_buf_t buf,
struct sockaddr* addr,
unsigned flags) {
...
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
libuv 1.0
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_read(uv_stream_t* handle,
ssize_t nread,
const uv_buf_t* buf) {
...
}
void recv_cb(uv_udp_t* handle,
ssize_t nread,
const uv_buf_t* buf,
const struct sockaddr* addr,
unsigned flags) {
...
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Receiving handles over pipes API change
.sp
In libuv 0.10 (and earlier versions) the \fIuv_read2_start\fP function was used to start reading
data on a pipe, which could also result in the reception of handles over it. The callback
for such function looked like this:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_read(uv_pipe_t* pipe,
ssize_t nread,
uv_buf_t buf,
uv_handle_type pending) {
...
}
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
In libuv 1.0, \fIuv_read2_start\fP was removed, and the user needs to check if there are pending
handles using \fBuv_pipe_pending_count()\fP and \fBuv_pipe_pending_type()\fP while in
the read callback:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
void on_read(uv_stream_t* handle,
ssize_t nread,
const uv_buf_t* buf) {
...
while (uv_pipe_pending_count((uv_pipe_t*) handle) != 0) {
pending = uv_pipe_pending_type((uv_pipe_t*) handle);
...
}
...
}
.ft P
.fi
.UNINDENT
.UNINDENT
.SS Extracting the file descriptor out of a handle
.sp
While it wasn\(aqt supported by the API, users often accessed the libuv internals in
order to get access to the file descriptor of a TCP handle, for example.
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
fd = handle\->io_watcher.fd;
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
This is now properly exposed through the \fBuv_fileno()\fP function.
.SS uv_fs_readdir rename and API change
.sp
\fIuv_fs_readdir\fP returned a list of strings in the \fIreq\->ptr\fP field upon completion in
libuv 0.10. In 1.0, this function got renamed to \fBuv_fs_scandir()\fP, since it\(aqs
actually implemented using \fBscandir(3)\fP\&.
.sp
In addition, instead of allocating a full list strings, the user is able to get one
result at a time by using the \fBuv_fs_scandir_next()\fP function. This function
does not need to make a roundtrip to the threadpool, because libuv will keep the
list of \fIdents\fP returned by \fBscandir(3)\fP around.
.SH DOWNLOADS
.sp
libuv can be downloaded from \fI\%here\fP\&.
.SH INSTALLATION
.sp
Installation instructions can be found in \fI\%the README\fP\&.
.SH AUTHOR
libuv contributors
.SH COPYRIGHT
2014-present, libuv contributors
.\" Generated by docutils manpage writer.
.