{ "source": "doc/api/process.markdown", "globals": [ { "textRaw": "process", "name": "process", "type": "global", "desc": "
The process
object is a global object and can be accessed from anywhere.\nIt is an instance of [EventEmitter][].\n\n
Node will normally exit with a 0
status code when no more async\noperations are pending. The following status codes are used in other\ncases:\n\n
1
Uncaught Fatal Exception - There was an uncaught exception,\nand it was not handled by a domain or an uncaughtException
event\nhandler.2
- Unused (reserved by Bash for builtin misuse)3
Internal JavaScript Parse Error - The JavaScript source code\ninternal in Node's bootstrapping process caused a parse error. This\nis extremely rare, and generally can only happen during development\nof Node itself.4
Internal JavaScript Evaluation Failure - The JavaScript\nsource code internal in Node's bootstrapping process failed to\nreturn a function value when evaluated. This is extremely rare, and\ngenerally can only happen during development of Node itself.5
Fatal Error - There was a fatal unrecoverable error in V8.\nTypically a message will be printed to stderr with the prefix FATAL\nERROR
.6
Non-function Internal Exception Handler - There was an\nuncaught exception, but the internal fatal exception handler\nfunction was somehow set to a non-function, and could not be called.7
Internal Exception Handler Run-Time Failure - There was an\nuncaught exception, and the internal fatal exception handler\nfunction itself threw an error while attempting to handle it. This\ncan happen, for example, if a process.on('uncaughtException')
or\ndomain.on('error')
handler throws an error.8
- Unused. In previous versions of Node, exit code 8 sometimes\nindicated an uncaught exception.9
- Invalid Argument - Either an unknown option was specified,\nor an option requiring a value was provided without a value.10
Internal JavaScript Run-Time Failure - The JavaScript\nsource code internal in Node's bootstrapping process threw an error\nwhen the bootstrapping function was called. This is extremely rare,\nand generally can only happen during development of Node itself.12
Invalid Debug Argument - The --debug
and/or --debug-brk
\noptions were set, but an invalid port number was chosen.>128
Signal Exits - If Node receives a fatal signal such as\nSIGKILL
or SIGHUP
, then its exit code will be 128
plus the\nvalue of the signal code. This is a standard Unix practice, since\nexit codes are defined to be 7-bit integers, and signal exits set\nthe high-order bit, and then contain the value of the signal code.Emitted when the process is about to exit. There is no way to prevent the\nexiting of the event loop at this point, and once all exit
listeners have\nfinished running the process will exit. Therefore you must only perform\nsynchronous operations in this handler. This is a good hook to perform\nchecks on the module's state (like for unit tests). The callback takes one\nargument, the code the process is exiting with.\n\n
Example of listening for exit
:\n\n
process.on('exit', function(code) {\n // do *NOT* do this\n setTimeout(function() {\n console.log('This will not run');\n }, 0);\n console.log('About to exit with code:', code);\n});
\n",
"params": []
},
{
"textRaw": "Event: 'beforeExit'",
"type": "event",
"name": "beforeExit",
"desc": "This event is emitted when node empties it's event loop and has nothing else to\nschedule. Normally, node exits when there is no work scheduled, but a listener\nfor 'beforeExit' can make asynchronous calls, and cause node to continue.\n\n
\n'beforeExit' is not emitted for conditions causing explicit termination, such as\nprocess.exit()
or uncaught exceptions, and should not be used as an\nalternative to the 'exit' event unless the intention is to schedule more work.\n\n\n
Emitted when an exception bubbles all the way back to the event loop. If a\nlistener is added for this exception, the default action (which is to print\na stack trace and exit) will not occur.\n\n
\nExample of listening for uncaughtException
:\n\n
process.on('uncaughtException', function(err) {\n console.log('Caught exception: ' + err);\n});\n\nsetTimeout(function() {\n console.log('This will still run.');\n}, 500);\n\n// Intentionally cause an exception, but don't catch it.\nnonexistentFunc();\nconsole.log('This will not run.');
\nNote that uncaughtException
is a very crude mechanism for exception\nhandling.\n\n
Don't use it, use domains instead. If you do use it, restart\nyour application after every unhandled exception!\n\n
\nDo not use it as the node.js equivalent of On Error Resume Next
. An\nunhandled exception means your application - and by extension node.js itself -\nis in an undefined state. Blindly resuming means anything could happen.\n\n
Think of resuming as pulling the power cord when you are upgrading your system.\nNine out of ten times nothing happens - but the 10th time, your system is bust.\n\n
\nYou have been warned.\n\n
\n", "params": [] }, { "textRaw": "Signal Events", "name": "SIGINT, SIGHUP, etc.", "type": "event", "desc": "Emitted when the processes receives a signal. See sigaction(2) for a list of\nstandard POSIX signal names such as SIGINT, SIGHUP, etc.\n\n
\nExample of listening for SIGINT
:\n\n
// Start reading from stdin so we don't exit.\nprocess.stdin.resume();\n\nprocess.on('SIGINT', function() {\n console.log('Got SIGINT. Press Control-D to exit.');\n});
\nAn easy way to send the SIGINT
signal is with Control-C
in most terminal\nprograms.\n\n
Note:\n\n
\nSIGUSR1
is reserved by node.js to start the debugger. It's possible to\ninstall a listener but that won't stop the debugger from starting.SIGTERM
and SIGINT
have default handlers on non-Windows platforms that resets\nthe terminal mode before exiting with code 128 + signal number
. If one of\nthese signals has a listener installed, its default behaviour will be removed\n(node will no longer exit).SIGPIPE
is ignored by default, it can have a listener installed.SIGHUP
is generated on Windows when the console window is closed, and on other\nplatforms under various similar conditions, see signal(7). It can have a\nlistener installed, however node will be unconditionally terminated by Windows\nabout 10 seconds later. On non-Windows platforms, the default behaviour of\nSIGHUP
is to terminate node, but once a listener has been installed its\ndefault behaviour will be removed.SIGTERM
is not supported on Windows, it can be listened on.SIGINT
from the terminal is supported on all platforms, and can usually be\ngenerated with CTRL+C
(though this may be configurable). It is not generated\nwhen terminal raw mode is enabled.SIGBREAK
is delivered on Windows when CTRL+BREAK
is pressed, on non-Windows\nplatforms it can be listened on, but there is no way to send or generate it.SIGWINCH
is delivered when the console has been resized. On Windows, this will\nonly happen on write to the console when the cursor is being moved, or when a\nreadable tty is used in raw mode.SIGKILL
cannot have a listener installed, it will unconditionally terminate\nnode on all platforms.SIGSTOP
cannot have a listener installed.Note that Windows does not support sending Signals, but node offers some\nemulation with process.kill()
, and child_process.kill()
:\n- Sending signal 0
can be used to search for the existence of a process\n- Sending SIGINT
, SIGTERM
, and SIGKILL
cause the unconditional exit of the\n target process.\n\n
A Writable Stream
to stdout
(on fd 1
).\n\n
Example: the definition of console.log
\n\n
console.log = function(d) {\n process.stdout.write(d + '\\n');\n};
\nprocess.stderr
and process.stdout
are unlike other streams in Node in\nthat they cannot be closed (end()
will throw), they never emit the finish
\nevent and that writes are usually blocking.\n\n
To check if Node is being run in a TTY context, read the isTTY
property\non process.stderr
, process.stdout
, or process.stdin
:\n\n
$ node -p "Boolean(process.stdin.isTTY)"\ntrue\n$ echo "foo" | node -p "Boolean(process.stdin.isTTY)"\nfalse\n\n$ node -p "Boolean(process.stdout.isTTY)"\ntrue\n$ node -p "Boolean(process.stdout.isTTY)" | cat\nfalse
\nSee the tty docs for more information.\n\n
\n" }, { "textRaw": "process.stderr", "name": "stderr", "desc": "A writable stream to stderr (on fd 2
).\n\n
process.stderr
and process.stdout
are unlike other streams in Node in\nthat they cannot be closed (end()
will throw), they never emit the finish
\nevent and that writes are usually blocking.\n\n
A Readable Stream
for stdin (on fd 0
).\n\n
Example of opening standard input and listening for both events:\n\n
\nprocess.stdin.setEncoding('utf8');\n\nprocess.stdin.on('readable', function() {\n var chunk = process.stdin.read();\n if (chunk !== null) {\n process.stdout.write('data: ' + chunk);\n }\n});\n\nprocess.stdin.on('end', function() {\n process.stdout.write('end');\n});
\nAs a Stream, process.stdin
can also be used in "old" mode that is compatible\nwith scripts written for node prior v0.10.\nFor more information see\nStream compatibility.\n\n
In "old" Streams mode the stdin stream is paused by default, so one\nmust call process.stdin.resume()
to read from it. Note also that calling\nprocess.stdin.resume()
itself would switch stream to "old" mode.\n\n
If you are starting a new project you should prefer a more recent "new" Streams\nmode over "old" one.\n\n
\n" }, { "textRaw": "process.argv", "name": "argv", "desc": "An array containing the command line arguments. The first element will be\n'node', the second element will be the name of the JavaScript file. The\nnext elements will be any additional command line arguments.\n\n
\n// print process.argv\nprocess.argv.forEach(function(val, index, array) {\n console.log(index + ': ' + val);\n});
\nThis will generate:\n\n
\n$ node process-2.js one two=three four\n0: node\n1: /Users/mjr/work/node/process-2.js\n2: one\n3: two=three\n4: four
\n"
},
{
"textRaw": "process.execPath",
"name": "execPath",
"desc": "This is the absolute pathname of the executable that started the process.\n\n
\nExample:\n\n
\n/usr/local/bin/node
\n"
},
{
"textRaw": "process.execArgv",
"name": "execArgv",
"desc": "This is the set of node-specific command line options from the\nexecutable that started the process. These options do not show up in\nprocess.argv
, and do not include the node executable, the name of\nthe script, or any options following the script name. These options\nare useful in order to spawn child processes with the same execution\nenvironment as the parent.\n\n
Example:\n\n
\n$ node --harmony script.js --version
\nresults in process.execArgv:\n\n
\n['--harmony']
\nand process.argv:\n\n
\n['/usr/local/bin/node', 'script.js', '--version']
\n"
},
{
"textRaw": "process.env",
"name": "env",
"desc": "An object containing the user environment. See environ(7).\n\n
\nAn example of this object looks like:\n\n
\n{ TERM: 'xterm-256color',\n SHELL: '/usr/local/bin/bash',\n USER: 'maciej',\n PATH: '~/.bin/:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin',\n PWD: '/Users/maciej',\n EDITOR: 'vim',\n SHLVL: '1',\n HOME: '/Users/maciej',\n LOGNAME: 'maciej',\n _: '/usr/local/bin/node' }
\nYou can write to this object, but changes won't be reflected outside of your\nprocess. That means that the following won't work:\n\n
\nnode -e 'process.env.foo = "bar"' && echo $foo
\nBut this will:\n\n
\nprocess.env.foo = 'bar';\nconsole.log(process.env.foo);
\n"
},
{
"textRaw": "process.exitCode",
"name": "exitCode",
"desc": "A number which will be the process exit code, when the process either\nexits gracefully, or is exited via process.exit()
without specifying\na code.\n\n
Specifying a code to process.exit(code)
will override any previous\nsetting of process.exitCode
.\n\n\n
A compiled-in property that exposes NODE_VERSION
.\n\n
console.log('Version: ' + process.version);
\n"
},
{
"textRaw": "process.versions",
"name": "versions",
"desc": "A property exposing version strings of node and its dependencies.\n\n
\nconsole.log(process.versions);
\nWill print something like:\n\n
\n{ http_parser: '1.0',\n node: '0.10.4',\n v8: '3.14.5.8',\n ares: '1.9.0-DEV',\n uv: '0.10.3',\n zlib: '1.2.3',\n modules: '11',\n openssl: '1.0.1e' }
\n"
},
{
"textRaw": "process.config",
"name": "config",
"desc": "An Object containing the JavaScript representation of the configure options\nthat were used to compile the current node executable. This is the same as\nthe "config.gypi" file that was produced when running the ./configure
script.\n\n
An example of the possible output looks like:\n\n
\n{ target_defaults:\n { cflags: [],\n default_configuration: 'Release',\n defines: [],\n include_dirs: [],\n libraries: [] },\n variables:\n { host_arch: 'x64',\n node_install_npm: 'true',\n node_prefix: '',\n node_shared_cares: 'false',\n node_shared_http_parser: 'false',\n node_shared_libuv: 'false',\n node_shared_v8: 'false',\n node_shared_zlib: 'false',\n node_use_dtrace: 'false',\n node_use_openssl: 'true',\n node_shared_openssl: 'false',\n strict_aliasing: 'true',\n target_arch: 'x64',\n v8_use_snapshot: 'true' } }
\n"
},
{
"textRaw": "process.pid",
"name": "pid",
"desc": "The PID of the process.\n\n
\nconsole.log('This process is pid ' + process.pid);
\n"
},
{
"textRaw": "process.title",
"name": "title",
"desc": "Getter/setter to set what is displayed in 'ps'.\n\n
\nWhen used as a setter, the maximum length is platform-specific and probably\nshort.\n\n
\nOn Linux and OS X, it's limited to the size of the binary name plus the\nlength of the command line arguments because it overwrites the argv memory.\n\n
\nv0.8 allowed for longer process title strings by also overwriting the environ\nmemory but that was potentially insecure/confusing in some (rather obscure)\ncases.\n\n\n
\n" }, { "textRaw": "process.arch", "name": "arch", "desc": "What processor architecture you're running on: 'arm'
, 'ia32'
, or 'x64'
.\n\n
console.log('This processor architecture is ' + process.arch);
\n"
},
{
"textRaw": "process.platform",
"name": "platform",
"desc": "What platform you're running on:\n'darwin'
, 'freebsd'
, 'linux'
, 'sunos'
or 'win32'
\n\n
console.log('This platform is ' + process.platform);
\n"
},
{
"textRaw": "process.mainModule",
"name": "mainModule",
"desc": "Alternate way to retrieve\nrequire.main
.\nThe difference is that if the main module changes at runtime, require.main
\nmight still refer to the original main module in modules that were required\nbefore the change occurred. Generally it's safe to assume that the two refer\nto the same module.\n\n
As with require.main
, it will be undefined
if there was no entry script.\n\n
This causes node to emit an abort. This will cause node to exit and\ngenerate a core file.\n\n
\n", "signatures": [ { "params": [] } ] }, { "textRaw": "process.chdir(directory)", "type": "method", "name": "chdir", "desc": "Changes the current working directory of the process or throws an exception if that fails.\n\n
\nconsole.log('Starting directory: ' + process.cwd());\ntry {\n process.chdir('/tmp');\n console.log('New directory: ' + process.cwd());\n}\ncatch (err) {\n console.log('chdir: ' + err);\n}
\n",
"signatures": [
{
"params": [
{
"name": "directory"
}
]
}
]
},
{
"textRaw": "process.cwd()",
"type": "method",
"name": "cwd",
"desc": "Returns the current working directory of the process.\n\n
\nconsole.log('Current directory: ' + process.cwd());
\n",
"signatures": [
{
"params": []
}
]
},
{
"textRaw": "process.exit([code])",
"type": "method",
"name": "exit",
"desc": "Ends the process with the specified code
. If omitted, exit uses the\n'success' code 0
.\n\n
To exit with a 'failure' code:\n\n
\nprocess.exit(1);
\nThe shell that executed node should see the exit code as 1.\n\n\n
\n", "signatures": [ { "params": [ { "name": "code", "optional": true } ] } ] }, { "textRaw": "process.getgid()", "type": "method", "name": "getgid", "desc": "Note: this function is only available on POSIX platforms (i.e. not Windows,\nAndroid)\n\n
\nGets the group identity of the process. (See getgid(2).)\nThis is the numerical group id, not the group name.\n\n
\nif (process.getgid) {\n console.log('Current gid: ' + process.getgid());\n}
\n",
"signatures": [
{
"params": []
}
]
},
{
"textRaw": "process.setgid(id)",
"type": "method",
"name": "setgid",
"desc": "Note: this function is only available on POSIX platforms (i.e. not Windows,\nAndroid)\n\n
\nSets the group identity of the process. (See setgid(2).) This accepts either\na numerical ID or a groupname string. If a groupname is specified, this method\nblocks while resolving it to a numerical ID.\n\n
\nif (process.getgid && process.setgid) {\n console.log('Current gid: ' + process.getgid());\n try {\n process.setgid(501);\n console.log('New gid: ' + process.getgid());\n }\n catch (err) {\n console.log('Failed to set gid: ' + err);\n }\n}
\n",
"signatures": [
{
"params": [
{
"name": "id"
}
]
}
]
},
{
"textRaw": "process.getuid()",
"type": "method",
"name": "getuid",
"desc": "Note: this function is only available on POSIX platforms (i.e. not Windows,\nAndroid)\n\n
\nGets the user identity of the process. (See getuid(2).)\nThis is the numerical userid, not the username.\n\n
\nif (process.getuid) {\n console.log('Current uid: ' + process.getuid());\n}
\n",
"signatures": [
{
"params": []
}
]
},
{
"textRaw": "process.setuid(id)",
"type": "method",
"name": "setuid",
"desc": "Note: this function is only available on POSIX platforms (i.e. not Windows,\nAndroid)\n\n
\nSets the user identity of the process. (See setuid(2).) This accepts either\na numerical ID or a username string. If a username is specified, this method\nblocks while resolving it to a numerical ID.\n\n
\nif (process.getuid && process.setuid) {\n console.log('Current uid: ' + process.getuid());\n try {\n process.setuid(501);\n console.log('New uid: ' + process.getuid());\n }\n catch (err) {\n console.log('Failed to set uid: ' + err);\n }\n}
\n",
"signatures": [
{
"params": [
{
"name": "id"
}
]
}
]
},
{
"textRaw": "process.getgroups()",
"type": "method",
"name": "getgroups",
"desc": "Note: this function is only available on POSIX platforms (i.e. not Windows,\nAndroid)\n\n
\nReturns an array with the supplementary group IDs. POSIX leaves it unspecified\nif the effective group ID is included but node.js ensures it always is.\n\n\n
\n", "signatures": [ { "params": [] } ] }, { "textRaw": "process.setgroups(groups)", "type": "method", "name": "setgroups", "desc": "Note: this function is only available on POSIX platforms (i.e. not Windows,\nAndroid)\n\n
\nSets the supplementary group IDs. This is a privileged operation, meaning you\nneed to be root or have the CAP_SETGID capability.\n\n
\nThe list can contain group IDs, group names or both.\n\n\n
\n", "signatures": [ { "params": [ { "name": "groups" } ] } ] }, { "textRaw": "process.initgroups(user, extra_group)", "type": "method", "name": "initgroups", "desc": "Note: this function is only available on POSIX platforms (i.e. not Windows,\nAndroid)\n\n
\nReads /etc/group and initializes the group access list, using all groups of\nwhich the user is a member. This is a privileged operation, meaning you need\nto be root or have the CAP_SETGID capability.\n\n
\nuser
is a user name or user ID. extra_group
is a group name or group ID.\n\n
Some care needs to be taken when dropping privileges. Example:\n\n
\nconsole.log(process.getgroups()); // [ 0 ]\nprocess.initgroups('bnoordhuis', 1000); // switch user\nconsole.log(process.getgroups()); // [ 27, 30, 46, 1000, 0 ]\nprocess.setgid(1000); // drop root gid\nconsole.log(process.getgroups()); // [ 27, 30, 46, 1000 ]
\n",
"signatures": [
{
"params": [
{
"name": "user"
},
{
"name": "extra_group"
}
]
}
]
},
{
"textRaw": "process.kill(pid[, signal])",
"type": "method",
"name": "kill",
"desc": "Send a signal to a process. pid
is the process id and signal
is the\nstring describing the signal to send. Signal names are strings like\n'SIGINT' or 'SIGHUP'. If omitted, the signal will be 'SIGTERM'.\nSee Signal Events and kill(2) for more information.\n\n
Will throw an error if target does not exist, and as a special case, a signal of\n0
can be used to test for the existence of a process.\n\n
Note that just because the name of this function is process.kill
, it is\nreally just a signal sender, like the kill
system call. The signal sent\nmay do something other than kill the target process.\n\n
Example of sending a signal to yourself:\n\n
\nprocess.on('SIGHUP', function() {\n console.log('Got SIGHUP signal.');\n});\n\nsetTimeout(function() {\n console.log('Exiting.');\n process.exit(0);\n}, 100);\n\nprocess.kill(process.pid, 'SIGHUP');
\nNote: When SIGUSR1 is received by Node.js it starts the debugger, see\nSignal Events.\n\n
\n", "signatures": [ { "params": [ { "name": "pid" }, { "name": "signal", "optional": true } ] } ] }, { "textRaw": "process.memoryUsage()", "type": "method", "name": "memoryUsage", "desc": "Returns an object describing the memory usage of the Node process\nmeasured in bytes.\n\n
\nvar util = require('util');\n\nconsole.log(util.inspect(process.memoryUsage()));
\nThis will generate:\n\n
\n{ rss: 4935680,\n heapTotal: 1826816,\n heapUsed: 650472 }
\nheapTotal
and heapUsed
refer to V8's memory usage.\n\n\n
Once the current event loop turn runs to completion, call the callback\nfunction.\n\n
\nThis is not a simple alias to setTimeout(fn, 0)
, it's much more\nefficient. It runs before any additional I/O events (including\ntimers) fire in subsequent ticks of the event loop.\n\n
console.log('start');\nprocess.nextTick(function() {\n console.log('nextTick callback');\n});\nconsole.log('scheduled');\n// Output:\n// start\n// scheduled\n// nextTick callback
\nThis is important in developing APIs where you want to give the user the\nchance to assign event handlers after an object has been constructed,\nbut before any I/O has occurred.\n\n
\nfunction MyThing(options) {\n this.setupOptions(options);\n\n process.nextTick(function() {\n this.startDoingStuff();\n }.bind(this));\n}\n\nvar thing = new MyThing();\nthing.getReadyForStuff();\n\n// thing.startDoingStuff() gets called now, not before.
\nIt is very important for APIs to be either 100% synchronous or 100%\nasynchronous. Consider this example:\n\n
\n// WARNING! DO NOT USE! BAD UNSAFE HAZARD!\nfunction maybeSync(arg, cb) {\n if (arg) {\n cb();\n return;\n }\n\n fs.stat('file', cb);\n}
\nThis API is hazardous. If you do this:\n\n
\nmaybeSync(true, function() {\n foo();\n});\nbar();
\nthen it's not clear whether foo()
or bar()
will be called first.\n\n
This approach is much better:\n\n
\nfunction definitelyAsync(arg, cb) {\n if (arg) {\n process.nextTick(cb);\n return;\n }\n\n fs.stat('file', cb);\n}
\nNote: the nextTick queue is completely drained on each pass of the\nevent loop before additional I/O is processed. As a result,\nrecursively setting nextTick callbacks will block any I/O from\nhappening, just like a while(true);
loop.\n\n
Sets or reads the process's file mode creation mask. Child processes inherit\nthe mask from the parent process. Returns the old mask if mask
argument is\ngiven, otherwise returns the current mask.\n\n
var oldmask, newmask = 0022;\n\noldmask = process.umask(newmask);\nconsole.log('Changed umask from: ' + oldmask.toString(8) +\n ' to ' + newmask.toString(8));
\n",
"signatures": [
{
"params": [
{
"name": "mask",
"optional": true
}
]
}
]
},
{
"textRaw": "process.uptime()",
"type": "method",
"name": "uptime",
"desc": "Number of seconds Node has been running.\n\n\n
\n", "signatures": [ { "params": [] } ] }, { "textRaw": "process.hrtime()", "type": "method", "name": "hrtime", "desc": "Returns the current high-resolution real time in a [seconds, nanoseconds]
\ntuple Array. It is relative to an arbitrary time in the past. It is not\nrelated to the time of day and therefore not subject to clock drift. The\nprimary use is for measuring performance between intervals.\n\n
You may pass in the result of a previous call to process.hrtime()
to get\na diff reading, useful for benchmarks and measuring intervals:\n\n
var time = process.hrtime();\n// [ 1800216, 25 ]\n\nsetTimeout(function() {\n var diff = process.hrtime(time);\n // [ 1, 552 ]\n\n console.log('benchmark took %d nanoseconds', diff[0] * 1e9 + diff[1]);\n // benchmark took 1000000527 nanoseconds\n}, 1000);
\n",
"signatures": [
{
"params": []
}
]
}
]
}
]
}