lukechilds
/
node
mirror of https://github.com/lukechilds/node.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
8827 Commits
131 Branches
446 Tags
325 MiB
Tree: 881ac26f27
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '881ac26f27'
${ noResults }
node/src/node.cc

3107 lines
76 KiB
Raw Normal View History

Update copyright headers
14 years ago
// Copyright Joyent, Inc. and other Node contributors.
//
// Permission is hereby granted, free of charge, to any person obtaining a
// copy of this software and associated documentation files (the
// "Software"), to deal in the Software without restriction, including
// without limitation the rights to use, copy, modify, merge, publish,
// distribute, sublicense, and/or sell copies of the Software, and to permit
// persons to whom the Software is furnished to do so, subject to the
// following conditions:
//
// The above copyright notice and this permission notice shall be included
// in all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
// OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN
// NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
// OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
// USE OR OTHER DEALINGS IN THE SOFTWARE.
Clean up the way windows headers are included Plus make inclusion order a little more consistent in general
14 years ago
core: use proper #include directives
13 years ago
#include "node.h"
process: add _getActiveHandles(), _getActiveRequests() * process._getActiveHandles() returns a list containing all active handles (timers, sockets, etc.) that have not been unref'd. * process._getActiveRequests() returns a list of active requests (in-flight actions like connecting to a remote host, writing data to a socket, etc.).
13 years ago
#include "req_wrap.h"
#include "handle_wrap.h"
src: use StringBytes for DecodeWrite/DecodeBytes/Encode Bonus: this makes node::Encode actually work properly with base64, ucs2, hex, etc.
12 years ago
#include "string_bytes.h"
everything is changed. i've waited much too long to commit. this is awful. i'm sorry for being so messy.
16 years ago
node.cc: include ares.h
13 years ago
#include "ares.h"
core: use proper #include directives
13 years ago
#include "uv.h"
Integrate libuv into build system
14 years ago
core: use proper #include directives
13 years ago
#include "v8-debug.h"
tracing: add systemtap support
12 years ago
#if defined HAVE_DTRACE || defined HAVE_ETW || defined HAVE_SYSTEMTAP
core: use proper #include directives
13 years ago
# include "node_dtrace.h"
Don't compile in node_dtrace.cc when not supported
13 years ago
#endif
windows: add tracing with performance counters Patch by Henry Rawas and Scott Blomquist.
12 years ago
#if defined HAVE_PERFCTR
# include "node_counters.h"
#endif
Add process.now Faster way to find out current time.
15 years ago
Clean up the way windows headers are included Plus make inclusion order a little more consistent in general
14 years ago
#include <locale.h>
Batch of ev -&gt; uv changes
14 years ago
#include <signal.h>
Major refactoring: program name now &#34;node&#34; Trying to make a more moduler design. Two libraries currently &#34;TCP&#34; and &#34;HTTP&#34; each have their own file. Other major feature added here is multiple web servers! excitement.
16 years ago
#include <stdio.h>
Fix ::exit() compile issues. Thanks Matthias!
16 years ago
#include <stdlib.h>
Just add an extra include for strings.h - don&#39;t try to be cute
15 years ago
#include <string.h>
Basic VC++ compatibility work.
14 years ago
#if !defined(_MSC_VER)
#include <strings.h>
#else
#define strcasecmp _stricmp
#endif
Add limits.h to src/node.cc for PATH_MAX
16 years ago
#include <limits.h> /* PATH_MAX */
Major refactoring: program name now &#34;node&#34; Trying to make a more moduler design. Two libraries currently &#34;TCP&#34; and &#34;HTTP&#34; each have their own file. Other major feature added here is multiple web servers! excitement.
16 years ago
#include <assert.h>
Basic VC++ compatibility work.
14 years ago
#if !defined(_MSC_VER)
#include <unistd.h> /* setuid, getuid */
#else
#include <direct.h>
#include <process.h>
#define getpid _getpid
Fix MSVS building.
14 years ago
#include <io.h>
#define umask _umask
typedef int mode_t;
Basic VC++ compatibility work.
14 years ago
#endif
Feature: add node.cwd() to access the current working directory.
16 years ago
#include <errno.h>
Add process.setuid(), getuid()
15 years ago
#include <sys/types.h>
process: expose the zlib version in use in process.versions
13 years ago
#include "zlib.h"
exit program on top level exceptions (also lots of reorg.)
16 years ago
Clean up the way windows headers are included Plus make inclusion order a little more consistent in general
14 years ago
#ifdef __POSIX__
Fix headers for node.cc
14 years ago
# include <pwd.h> /* getpwnam() */
# include <grp.h> /* getgrnam() */
#endif
core: use proper #include directives
13 years ago
#include "node_buffer.h"
#include "node_file.h"
#include "node_http_parser.h"
#include "node_constants.h"
#include "node_javascript.h"
#include "node_version.h"
#include "node_string.h"
Basic VC++ compatibility work.
14 years ago
#if HAVE_OPENSSL
core: use proper #include directives
13 years ago
# include "node_crypto.h"
Initial openssl support for net2
15 years ago
#endif
tracing: add systemtap support
12 years ago
#if HAVE_SYSTEMTAP
#include "node_systemtap.h"
#endif
core: use proper #include directives
13 years ago
#include "node_script.h"
#include "v8_typed_array.h"
node.assert() was broken.
16 years ago
Major refactoring: program name now &#34;node&#34; Trying to make a more moduler design. Two libraries currently &#34;TCP&#34; and &#34;HTTP&#34; each have their own file. Other major feature added here is multiple web servers! excitement.
16 years ago
using namespace v8;
environ symbol fix for Mac OS X
14 years ago
# ifdef __APPLE__
# include <crt_externs.h>
# define environ (*_NSGetEnviron())
Basic VC++ compatibility work.
14 years ago
# elif !defined(_MSC_VER)
Add &#39;extern char **environ&#39; so that the new ENV code compiles. Problem appeared in dc39e8202456686513c43a4e2c4e5e505eb3a901.
16 years ago
extern char **environ;
environ symbol fix for Mac OS X
14 years ago
# endif
Add &#39;extern char **environ&#39; so that the new ENV code compiles. Problem appeared in dc39e8202456686513c43a4e2c4e5e505eb3a901.
16 years ago
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
namespace node {
Statically define symbols ~7% improvement in hello world HTTP response bench.
15 years ago
process: add _getActiveHandles(), _getActiveRequests() * process._getActiveHandles() returns a list containing all active handles (timers, sockets, etc.) that have not been unref&#39;d. * process._getActiveRequests() returns a list of active requests (in-flight actions like connecting to a remote host, writing data to a socket, etc.).
13 years ago
ngx_queue_t handle_wrap_queue = { &handle_wrap_queue, &handle_wrap_queue };
ngx_queue_t req_wrap_queue = { &req_wrap_queue, &req_wrap_queue };
req_wrap: share process_symbol, domain_symbol Share persistent strings process_symbol and domain_symbol across compilation units. Avoids redefinition errors when src/node.cc includes src/req_wrap.h.
13 years ago
// declared in req_wrap.h
Persistent<String> process_symbol;
Persistent<String> domain_symbol;
Add &#39;--&#39; to seperate v8 args from program args
16 years ago
cares: Set process._errno, not global.errno This makes test-internet pass
12 years ago
// declared in node_internals.h
Persistent<Object> process;
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
static Persistent<Function> process_tickFromSpinner;
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
static Persistent<Function> process_tickCallback;
Add support for TLS SNI Fixes #1411
14 years ago
addon: Pass module object to NODE_MODULE init function mainly to allow native addons to export single functions on rather than being restricted to operating on an existing object. Init functions now receive exports as the first argument, like before, but also the module object as the second argument, if they support it. Related to #4634 cc: @rvagg
12 years ago
static Persistent<String> exports_symbol;
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
static Persistent<String> errno_symbol;
static Persistent<String> syscall_symbol;
static Persistent<String> errpath_symbol;
static Persistent<String> code_symbol;
isolates: add process.tid property, pass args to isolate
13 years ago
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
static Persistent<String> rss_symbol;
static Persistent<String> heap_total_symbol;
static Persistent<String> heap_used_symbol;
Add support for TLS SNI Fixes #1411
14 years ago
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
static Persistent<String> fatal_exception_symbol;
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
static Persistent<String> enter_symbol;
static Persistent<String> exit_symbol;
static Persistent<String> disposed_symbol;
Domain hooks in ReqWrap&lt;T&gt; and MakeCallback
13 years ago
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
static bool print_eval = false;
cmd: add a -i / --interactive flag to force the REPL This is the only way to spawn a node child process in REPL mode, and will also be needed to be able to use the REPL in MinTTY.
13 years ago
static bool force_repl = false;
Add --no-deprecation and --trace-deprecation flags
13 years ago
static bool trace_deprecation = false;
node: Add --throw-deprecation Extremely handy when tracking down a flood of recursive nextTick warnings.
12 years ago
static bool throw_deprecation = false;
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
static char *eval_string = NULL;
static int option_end_index = 0;
static bool use_debug_agent = false;
static bool debug_wait_connect = false;
static int debug_port=5858;
static int max_stack_size = 0;
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
static bool using_domains = false;
Allow process.setuid() and process.setgid() to accept string names in lieu of numeric IDs
15 years ago
Apply --no-deprecation to C++ warnings as well
13 years ago
// used by C++ modules as well
bool no_deprecation = false;
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
static uv_idle_t tick_spinner;
static bool need_tick_cb;
static Persistent<String> tick_callback_sym;
Batch of ev -&gt; uv changes
14 years ago
timer: move setImmediate from timer to uv_check uv_check is the robust place to invoke setImmediate callbacks after process.nextTick and before timers(setTimeout/setInterval)
12 years ago
static uv_check_t check_immediate_watcher;
static uv_idle_t idle_immediate_dummy;
static bool need_immediate_cb;
static Persistent<String> immediate_callback_sym;
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
// for quick ref to tickCallback values
node: code cleanup, make tick_infobox static It&#39;s not used outside of src/node.cc so make it static.
12 years ago
static struct {
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
uint32_t length;
uint32_t index;
uint32_t depth;
} tick_infobox;
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
#ifdef OPENSSL_NPN_NEGOTIATED
static bool use_npn = true;
Add node::Loop() and don&#39;t inc node_isolate.h in *.cc node::Loop() replaces the NODE_LOOP macro. This avoids hitting v8::Isolate::GetCurrent() for each loop lookup when HAVE_ISOLATE==0
13 years ago
#else
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
static bool use_npn = false;
Add node::Loop() and don&#39;t inc node_isolate.h in *.cc node::Loop() replaces the NODE_LOOP macro. This avoids hitting v8::Isolate::GetCurrent() for each loop lookup when HAVE_ISOLATE==0
13 years ago
#endif
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
#ifdef SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
static bool use_sni = true;
#else
static bool use_sni = false;
#endif
// process-relative uptime base, initialized at start-up
static double prog_start_time;
src: make node_isolate global V8 3.15 has new API functions that let you specify the Isolate. V8 and node.js generally spend 0.5-3.5% of the time in pthread_getspecific(), looking up the current Isolate. Avoid that overhead by making &#34;our&#34; isolate global so we can pass it around. The change to the new API is introduced in follow-up commits.
12 years ago
static volatile bool debugger_running = false;
static uv_async_t dispatch_debug_messages_async;
// Declared in node_internals.h
Isolate* node_isolate = NULL;
libeio bugfix part 3 Finally (hopefully) fix the issue that Felix reported. It&#39;s only appearing on macintosh (test/mjsunit/test-eio-race3.js) The trick/hack is to call eio_poll() again before reentering the event loop. Additionally this commit implements a more complex method of calling eio_poll(), occasionally dropping to an ev_idle watcher. See also: 3f3977283419fe81e09aa23b91e59de959a84abd http://lists.schmorp.de/pipermail/libev/2010q1/000855.html http://groups.google.com/group/nodejs/browse_thread/thread/9f8db11c792a68bb/a89705f68971f53c
15 years ago
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
static void Spin(uv_idle_t* handle, int status) {
assert((uv_idle_t*) handle == &tick_spinner);
assert(status == 0);
Change nextTick implementation for the better Use a prepare and idle watcher to execute the nextTick callback more quickly. Test provided by Matt Ranney.
15 years ago
// Avoid entering a V8 scope.
if (!need_tick_cb) return;
need_tick_cb = false;
deps: upgrade libuv to a478847 The event loop&#39;s reference counting scheme in this version of libuv has changed. Update the libuv bindings to reflect that fact.
13 years ago
uv_idle_stop(&tick_spinner);
Change nextTick implementation for the better Use a prepare and idle watcher to execute the nextTick callback more quickly. Test provided by Matt Ranney.
15 years ago
HandleScope scope;
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
if (process_tickFromSpinner.IsEmpty()) {
Local<Value> cb_v = process->Get(String::New("_tickFromSpinner"));
if (!cb_v->IsFunction()) {
fprintf(stderr, "process._tickFromSpinner assigned to non-function\n");
abort();
}
Local<Function> cb = cb_v.As<Function>();
process_tickFromSpinner = Persistent<Function>::New(cb);
Change nextTick implementation for the better Use a prepare and idle watcher to execute the nextTick callback more quickly. Test provided by Matt Ranney.
15 years ago
}
TryCatch try_catch;
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
process_tickFromSpinner->Call(process, 0, NULL);
Change nextTick implementation for the better Use a prepare and idle watcher to execute the nextTick callback more quickly. Test provided by Matt Ranney.
15 years ago
if (try_catch.HasCaught()) {
FatalException(try_catch);
}
}
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
static Handle<Value> NeedTickCallback(const Arguments& args) {
Avoid tick spinner spinning forever
14 years ago
need_tick_cb = true;
deps: upgrade libuv to a478847 The event loop&#39;s reference counting scheme in this version of libuv has changed. Update the libuv bindings to reflect that fact.
13 years ago
uv_idle_start(&tick_spinner, Spin);
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Fix process.nextTick throw call sites This patch now reports the proper throw call site for exceptions triggered within process.nextTick. So instead of this: node.js:201 throw e; // process.nextTick error, or &#39;error&#39; event on first tick ^ You will now see: mydir/myscript.js:15 throw new Error(&#39;My Error&#39;); ^ From my testing this patch causes no performance regressions, but does greatly simplify processing the nextTickQueue.
13 years ago
}
Avoid tick spinner spinning forever
14 years ago
Call Tick() after coming out of select() Previously we would only call it before going into select(). This is needed to fix test/simple/test-next-tick-ordering2.js.
15 years ago
timer: move setImmediate from timer to uv_check uv_check is the robust place to invoke setImmediate callbacks after process.nextTick and before timers(setTimeout/setInterval)
12 years ago
static void CheckImmediate(uv_check_t* handle, int status) {
assert(handle == &check_immediate_watcher);
assert(status == 0);
HandleScope scope;
if (immediate_callback_sym.IsEmpty()) {
immediate_callback_sym = NODE_PSYMBOL("_immediateCallback");
}
MakeCallback(process, immediate_callback_sym, 0, NULL);
}
static void IdleImmediateDummy(uv_idle_t* handle, int status) {
// Do nothing. Only for maintaining event loop
assert(handle == &idle_immediate_dummy);
assert(status == 0);
}
Unify ErrnoException creation
15 years ago
static inline const char *errno_string(int errorno) {
#define ERRNO_CASE(e) case e: return #e;
switch (errorno) {
#ifdef EACCES
ERRNO_CASE(EACCES);
#endif
#ifdef EADDRINUSE
ERRNO_CASE(EADDRINUSE);
#endif
#ifdef EADDRNOTAVAIL
ERRNO_CASE(EADDRNOTAVAIL);
#endif
#ifdef EAFNOSUPPORT
ERRNO_CASE(EAFNOSUPPORT);
#endif
#ifdef EAGAIN
ERRNO_CASE(EAGAIN);
Resolve EWOULDBLOCK to string if different thatn EAGAIN
15 years ago
#endif
#ifdef EWOULDBLOCK
# if EAGAIN != EWOULDBLOCK
Unify ErrnoException creation
15 years ago
ERRNO_CASE(EWOULDBLOCK);
# endif
#endif
#ifdef EALREADY
ERRNO_CASE(EALREADY);
#endif
#ifdef EBADF
ERRNO_CASE(EBADF);
#endif
#ifdef EBADMSG
ERRNO_CASE(EBADMSG);
#endif
#ifdef EBUSY
ERRNO_CASE(EBUSY);
#endif
#ifdef ECANCELED
ERRNO_CASE(ECANCELED);
#endif
#ifdef ECHILD
ERRNO_CASE(ECHILD);
#endif
#ifdef ECONNABORTED
ERRNO_CASE(ECONNABORTED);
#endif
#ifdef ECONNREFUSED
ERRNO_CASE(ECONNREFUSED);
#endif
#ifdef ECONNRESET
ERRNO_CASE(ECONNRESET);
#endif
#ifdef EDEADLK
ERRNO_CASE(EDEADLK);
#endif
#ifdef EDESTADDRREQ
ERRNO_CASE(EDESTADDRREQ);
#endif
#ifdef EDOM
ERRNO_CASE(EDOM);
#endif
#ifdef EDQUOT
ERRNO_CASE(EDQUOT);
#endif
#ifdef EEXIST
ERRNO_CASE(EEXIST);
#endif
#ifdef EFAULT
ERRNO_CASE(EFAULT);
#endif
#ifdef EFBIG
ERRNO_CASE(EFBIG);
#endif
#ifdef EHOSTUNREACH
ERRNO_CASE(EHOSTUNREACH);
#endif
#ifdef EIDRM
ERRNO_CASE(EIDRM);
#endif
#ifdef EILSEQ
ERRNO_CASE(EILSEQ);
#endif
#ifdef EINPROGRESS
ERRNO_CASE(EINPROGRESS);
#endif
#ifdef EINTR
ERRNO_CASE(EINTR);
#endif
#ifdef EINVAL
ERRNO_CASE(EINVAL);
#endif
#ifdef EIO
ERRNO_CASE(EIO);
#endif
#ifdef EISCONN
ERRNO_CASE(EISCONN);
#endif
#ifdef EISDIR
ERRNO_CASE(EISDIR);
#endif
#ifdef ELOOP
ERRNO_CASE(ELOOP);
#endif
#ifdef EMFILE
ERRNO_CASE(EMFILE);
#endif
#ifdef EMLINK
ERRNO_CASE(EMLINK);
#endif
#ifdef EMSGSIZE
ERRNO_CASE(EMSGSIZE);
#endif
#ifdef EMULTIHOP
ERRNO_CASE(EMULTIHOP);
#endif
#ifdef ENAMETOOLONG
ERRNO_CASE(ENAMETOOLONG);
#endif
#ifdef ENETDOWN
ERRNO_CASE(ENETDOWN);
#endif
#ifdef ENETRESET
ERRNO_CASE(ENETRESET);
#endif
#ifdef ENETUNREACH
ERRNO_CASE(ENETUNREACH);
#endif
#ifdef ENFILE
ERRNO_CASE(ENFILE);
#endif
#ifdef ENOBUFS
ERRNO_CASE(ENOBUFS);
#endif
#ifdef ENODATA
ERRNO_CASE(ENODATA);
#endif
#ifdef ENODEV
ERRNO_CASE(ENODEV);
#endif
#ifdef ENOENT
ERRNO_CASE(ENOENT);
#endif
#ifdef ENOEXEC
ERRNO_CASE(ENOEXEC);
#endif
#ifdef ENOLINK
ERRNO_CASE(ENOLINK);
#endif
Fix signo_string for linux and solaris
15 years ago
#ifdef ENOLCK
# if ENOLINK != ENOLCK
ERRNO_CASE(ENOLCK);
# endif
#endif
Unify ErrnoException creation
15 years ago
#ifdef ENOMEM
ERRNO_CASE(ENOMEM);
#endif
#ifdef ENOMSG
ERRNO_CASE(ENOMSG);
#endif
#ifdef ENOPROTOOPT
ERRNO_CASE(ENOPROTOOPT);
#endif
#ifdef ENOSPC
ERRNO_CASE(ENOSPC);
#endif
#ifdef ENOSR
ERRNO_CASE(ENOSR);
#endif
#ifdef ENOSTR
ERRNO_CASE(ENOSTR);
#endif
#ifdef ENOSYS
ERRNO_CASE(ENOSYS);
#endif
#ifdef ENOTCONN
ERRNO_CASE(ENOTCONN);
#endif
#ifdef ENOTDIR
ERRNO_CASE(ENOTDIR);
#endif
#ifdef ENOTEMPTY
ERRNO_CASE(ENOTEMPTY);
#endif
#ifdef ENOTSOCK
ERRNO_CASE(ENOTSOCK);
#endif
#ifdef ENOTSUP
ERRNO_CASE(ENOTSUP);
#else
# ifdef EOPNOTSUPP
ERRNO_CASE(EOPNOTSUPP);
# endif
#endif
#ifdef ENOTTY
ERRNO_CASE(ENOTTY);
#endif
#ifdef ENXIO
ERRNO_CASE(ENXIO);
#endif
#ifdef EOVERFLOW
ERRNO_CASE(EOVERFLOW);
#endif
#ifdef EPERM
ERRNO_CASE(EPERM);
#endif
#ifdef EPIPE
ERRNO_CASE(EPIPE);
#endif
#ifdef EPROTO
ERRNO_CASE(EPROTO);
#endif
#ifdef EPROTONOSUPPORT
ERRNO_CASE(EPROTONOSUPPORT);
#endif
#ifdef EPROTOTYPE
ERRNO_CASE(EPROTOTYPE);
#endif
#ifdef ERANGE
ERRNO_CASE(ERANGE);
#endif
#ifdef EROFS
ERRNO_CASE(EROFS);
#endif
#ifdef ESPIPE
ERRNO_CASE(ESPIPE);
#endif
#ifdef ESRCH
ERRNO_CASE(ESRCH);
#endif
#ifdef ESTALE
ERRNO_CASE(ESTALE);
#endif
#ifdef ETIME
ERRNO_CASE(ETIME);
#endif
#ifdef ETIMEDOUT
ERRNO_CASE(ETIMEDOUT);
#endif
#ifdef ETXTBSY
ERRNO_CASE(ETXTBSY);
#endif
#ifdef EXDEV
ERRNO_CASE(EXDEV);
#endif
default: return "";
}
}
Properly handle child process exit codes The child process &#39;exit&#39; was returning the status of the process, rather than the exit code. This patch properly deconstructs the status into the exit code and the term signal a process may have received. See: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#Watcher_Specific_Functions_and_Data_-5 and waitpid(2)
15 years ago
const char *signo_string(int signo) {
#define SIGNO_CASE(e) case e: return #e;
switch (signo) {
#ifdef SIGHUP
SIGNO_CASE(SIGHUP);
#endif
#ifdef SIGINT
SIGNO_CASE(SIGINT);
#endif
#ifdef SIGQUIT
SIGNO_CASE(SIGQUIT);
#endif
#ifdef SIGILL
SIGNO_CASE(SIGILL);
#endif
#ifdef SIGTRAP
SIGNO_CASE(SIGTRAP);
#endif
#ifdef SIGABRT
SIGNO_CASE(SIGABRT);
#endif
#ifdef SIGIOT
# if SIGABRT != SIGIOT
SIGNO_CASE(SIGIOT);
# endif
#endif
#ifdef SIGBUS
SIGNO_CASE(SIGBUS);
#endif
#ifdef SIGFPE
SIGNO_CASE(SIGFPE);
#endif
#ifdef SIGKILL
SIGNO_CASE(SIGKILL);
#endif
#ifdef SIGUSR1
SIGNO_CASE(SIGUSR1);
#endif
#ifdef SIGSEGV
SIGNO_CASE(SIGSEGV);
#endif
#ifdef SIGUSR2
SIGNO_CASE(SIGUSR2);
#endif
#ifdef SIGPIPE
SIGNO_CASE(SIGPIPE);
#endif
#ifdef SIGALRM
SIGNO_CASE(SIGALRM);
#endif
SIGNO_CASE(SIGTERM);
There&#39;s no SIGCHLD on windows
14 years ago
#ifdef SIGCHLD
Properly handle child process exit codes The child process &#39;exit&#39; was returning the status of the process, rather than the exit code. This patch properly deconstructs the status into the exit code and the term signal a process may have received. See: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#Watcher_Specific_Functions_and_Data_-5 and waitpid(2)
15 years ago
SIGNO_CASE(SIGCHLD);
There&#39;s no SIGCHLD on windows
14 years ago
#endif
Properly handle child process exit codes The child process &#39;exit&#39; was returning the status of the process, rather than the exit code. This patch properly deconstructs the status into the exit code and the term signal a process may have received. See: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#Watcher_Specific_Functions_and_Data_-5 and waitpid(2)
15 years ago
#ifdef SIGSTKFLT
SIGNO_CASE(SIGSTKFLT);
#endif
#ifdef SIGCONT
SIGNO_CASE(SIGCONT);
#endif
#ifdef SIGSTOP
SIGNO_CASE(SIGSTOP);
#endif
#ifdef SIGTSTP
SIGNO_CASE(SIGTSTP);
#endif
process: use uv_signal instead of ev_signal
13 years ago
#ifdef SIGBREAK
SIGNO_CASE(SIGBREAK);
#endif
Properly handle child process exit codes The child process &#39;exit&#39; was returning the status of the process, rather than the exit code. This patch properly deconstructs the status into the exit code and the term signal a process may have received. See: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#Watcher_Specific_Functions_and_Data_-5 and waitpid(2)
15 years ago
#ifdef SIGTTIN
SIGNO_CASE(SIGTTIN);
#endif
#ifdef SIGTTOU
SIGNO_CASE(SIGTTOU);
#endif
#ifdef SIGURG
SIGNO_CASE(SIGURG);
#endif
#ifdef SIGXCPU
SIGNO_CASE(SIGXCPU);
#endif
#ifdef SIGXFSZ
SIGNO_CASE(SIGXFSZ);
#endif
#ifdef SIGVTALRM
SIGNO_CASE(SIGVTALRM);
#endif
#ifdef SIGPROF
SIGNO_CASE(SIGPROF);
#endif
#ifdef SIGWINCH
SIGNO_CASE(SIGWINCH);
#endif
#ifdef SIGIO
SIGNO_CASE(SIGIO);
#endif
#ifdef SIGPOLL
Fix signo_string for linux and solaris
15 years ago
# if SIGPOLL != SIGIO
Properly handle child process exit codes The child process &#39;exit&#39; was returning the status of the process, rather than the exit code. This patch properly deconstructs the status into the exit code and the term signal a process may have received. See: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#Watcher_Specific_Functions_and_Data_-5 and waitpid(2)
15 years ago
SIGNO_CASE(SIGPOLL);
Fix signo_string for linux and solaris
15 years ago
# endif
Properly handle child process exit codes The child process &#39;exit&#39; was returning the status of the process, rather than the exit code. This patch properly deconstructs the status into the exit code and the term signal a process may have received. See: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#Watcher_Specific_Functions_and_Data_-5 and waitpid(2)
15 years ago
#endif
#ifdef SIGLOST
SIGNO_CASE(SIGLOST);
#endif
#ifdef SIGPWR
Port to cygwin
15 years ago
# if SIGPWR != SIGLOST
Properly handle child process exit codes The child process &#39;exit&#39; was returning the status of the process, rather than the exit code. This patch properly deconstructs the status into the exit code and the term signal a process may have received. See: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#Watcher_Specific_Functions_and_Data_-5 and waitpid(2)
15 years ago
SIGNO_CASE(SIGPWR);
Port to cygwin
15 years ago
# endif
Properly handle child process exit codes The child process &#39;exit&#39; was returning the status of the process, rather than the exit code. This patch properly deconstructs the status into the exit code and the term signal a process may have received. See: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#Watcher_Specific_Functions_and_Data_-5 and waitpid(2)
15 years ago
#endif
#ifdef SIGSYS
SIGNO_CASE(SIGSYS);
#endif
default: return "";
}
}
Unify ErrnoException creation
15 years ago
Local<Value> ErrnoException(int errorno,
const char *syscall,
Added better error messages for async and sync fs calls with paths
15 years ago
const char *msg,
const char *path) {
Local<Value> e;
Unify ErrnoException creation
15 years ago
Local<String> estring = String::NewSymbol(errno_string(errorno));
Improve error reporting on windows
14 years ago
if (!msg[0]) {
msg = strerror(errorno);
}
Unify ErrnoException creation
15 years ago
Local<String> message = String::NewSymbol(msg);
Local<String> cons1 = String::Concat(estring, String::NewSymbol(", "));
Local<String> cons2 = String::Concat(cons1, message);
Display sys_errno when UV_UNKNOWN is returned
13 years ago
if (syscall_symbol.IsEmpty()) {
Unify ErrnoException creation
15 years ago
syscall_symbol = NODE_PSYMBOL("syscall");
errno_symbol = NODE_PSYMBOL("errno");
Added better error messages for async and sync fs calls with paths
15 years ago
errpath_symbol = NODE_PSYMBOL("path");
Expose errno with a string.
14 years ago
code_symbol = NODE_PSYMBOL("code");
Unify ErrnoException creation
15 years ago
}
Added better error messages for async and sync fs calls with paths
15 years ago
if (path) {
Local<String> cons3 = String::Concat(cons2, String::NewSymbol(" '"));
Local<String> cons4 = String::Concat(cons3, String::New(path));
Local<String> cons5 = String::Concat(cons4, String::NewSymbol("'"));
e = Exception::Error(cons5);
} else {
e = Exception::Error(cons2);
}
Local<Object> obj = e->ToObject();
src, test: downgrade to v8 3.14 api
12 years ago
obj->Set(errno_symbol, Integer::New(errorno));
Expose errno with a string.
14 years ago
obj->Set(code_symbol, estring);
Added better error messages for async and sync fs calls with paths
15 years ago
if (path) obj->Set(errpath_symbol, String::New(path));
Unify ErrnoException creation
15 years ago
if (syscall) obj->Set(syscall_symbol, String::NewSymbol(syscall));
return e;
}
Rename FSError to UVException and move to node.cc
13 years ago
static const char* get_uv_errno_string(int errorno) {
uv_err_t err;
memset(&err, 0, sizeof err);
err.code = (uv_err_code)errorno;
return uv_err_name(err);
}
static const char* get_uv_errno_message(int errorno) {
uv_err_t err;
memset(&err, 0, sizeof err);
err.code = (uv_err_code)errorno;
return uv_strerror(err);
}
// hack alert! copy of ErrnoException, tuned for uv errors
Local<Value> UVException(int errorno,
const char *syscall,
const char *msg,
const char *path) {
if (syscall_symbol.IsEmpty()) {
syscall_symbol = NODE_PSYMBOL("syscall");
errno_symbol = NODE_PSYMBOL("errno");
errpath_symbol = NODE_PSYMBOL("path");
code_symbol = NODE_PSYMBOL("code");
}
if (!msg || !msg[0])
msg = get_uv_errno_message(errorno);
Fix UVException errno bug
13 years ago
Local<String> estring = String::NewSymbol(get_uv_errno_string(errorno));
Rename FSError to UVException and move to node.cc
13 years ago
Local<String> message = String::NewSymbol(msg);
Local<String> cons1 = String::Concat(estring, String::NewSymbol(", "));
Local<String> cons2 = String::Concat(cons1, message);
Local<Value> e;
Local<String> path_str;
if (path) {
#ifdef _WIN32
if (strncmp(path, "\\\\?\\UNC\\", 8) == 0) {
path_str = String::Concat(String::New("\\\\"), String::New(path + 8));
} else if (strncmp(path, "\\\\?\\", 4) == 0) {
path_str = String::New(path + 4);
} else {
path_str = String::New(path);
}
#else
path_str = String::New(path);
#endif
Local<String> cons3 = String::Concat(cons2, String::NewSymbol(" '"));
Local<String> cons4 = String::Concat(cons3, path_str);
Local<String> cons5 = String::Concat(cons4, String::NewSymbol("'"));
e = Exception::Error(cons5);
} else {
e = Exception::Error(cons2);
}
Local<Object> obj = e->ToObject();
// TODO errno should probably go
src, test: downgrade to v8 3.14 api
12 years ago
obj->Set(errno_symbol, Integer::New(errorno));
Rename FSError to UVException and move to node.cc
13 years ago
obj->Set(code_symbol, estring);
if (path) obj->Set(errpath_symbol, path_str);
if (syscall) obj->Set(syscall_symbol, String::NewSymbol(syscall));
return e;
}
Windows: ErrnoException shouldn&#39;t mix up crt and win32 errors
13 years ago
#ifdef _WIN32
Remove platform files, and use uv platform api
13 years ago
// Does about the same as strerror(),
// but supports all windows error messages
static const char *winapi_strerror(const int errorno) {
char *errmsg = NULL;
FormatMessage(FORMAT_MESSAGE_ALLOCATE_BUFFER | FORMAT_MESSAGE_FROM_SYSTEM |
FORMAT_MESSAGE_IGNORE_INSERTS, NULL, errorno,
MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), (LPTSTR)&errmsg, 0, NULL);
if (errmsg) {
// Remove trailing newlines
for (int i = strlen(errmsg) - 1;
i >= 0 && (errmsg[i] == '\n' || errmsg[i] == '\r'); i--) {
errmsg[i] = '\0';
}
return errmsg;
} else {
// FormatMessage failed
return "Unknown error";
}
}
Windows: ErrnoException shouldn&#39;t mix up crt and win32 errors
13 years ago
Local<Value> WinapiErrnoException(int errorno,
Implement process._debugProcess
13 years ago
const char* syscall,
const char* msg,
const char* path) {
Windows: ErrnoException shouldn&#39;t mix up crt and win32 errors
13 years ago
Local<Value> e;
Implement process._debugProcess
13 years ago
if (!msg || !msg[0]) {
Windows: ErrnoException shouldn&#39;t mix up crt and win32 errors
13 years ago
msg = winapi_strerror(errorno);
}
Local<String> message = String::NewSymbol(msg);
if (syscall_symbol.IsEmpty()) {
syscall_symbol = NODE_PSYMBOL("syscall");
errno_symbol = NODE_PSYMBOL("errno");
errpath_symbol = NODE_PSYMBOL("path");
code_symbol = NODE_PSYMBOL("code");
}
if (path) {
Local<String> cons1 = String::Concat(message, String::NewSymbol(" '"));
Local<String> cons2 = String::Concat(cons1, String::New(path));
Local<String> cons3 = String::Concat(cons2, String::NewSymbol("'"));
e = Exception::Error(cons3);
} else {
e = Exception::Error(message);
}
Local<Object> obj = e->ToObject();
src, test: downgrade to v8 3.14 api
12 years ago
obj->Set(errno_symbol, Integer::New(errorno));
Windows: ErrnoException shouldn&#39;t mix up crt and win32 errors
13 years ago
if (path) obj->Set(errpath_symbol, String::New(path));
if (syscall) obj->Set(syscall_symbol, String::NewSymbol(syscall));
return e;
}
#endif
node: don&#39;t malloc in FromConstructorTemplate * allocate space for argv on the stack * move the declaration to node_internals.h
12 years ago
Handle<Value> FromConstructorTemplate(Persistent<FunctionTemplate> t,
Safe constructor for ObjectWrapped classes New() methods should be invoked as constructors, not regular functions. Corner cases like Script::New() may cause a SIGSEGV when the GC is run. More details: http://groups.google.com/group/nodejs/browse_thread/thread/a7e5db68d4cd6356
14 years ago
const Arguments& args) {
HandleScope scope;
node: don&#39;t malloc in FromConstructorTemplate * allocate space for argv on the stack * move the declaration to node_internals.h
12 years ago
Local<Value> argv[32];
unsigned argc = args.Length();
if (argc > ARRAY_SIZE(argv)) argc = ARRAY_SIZE(argv);
for (unsigned i = 0; i < argc; ++i) argv[i] = args[i];
return scope.Close(t->GetFunction()->NewInstance(argc, argv));
Safe constructor for ObjectWrapped classes New() methods should be invoked as constructors, not regular functions. Corner cases like Script::New() may cause a SIGSEGV when the GC is run. More details: http://groups.google.com/group/nodejs/browse_thread/thread/a7e5db68d4cd6356
14 years ago
}
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
Handle<Value> UsingDomains(const Arguments& args) {
HandleScope scope;
if (using_domains)
return scope.Close(Undefined());
using_domains = true;
Local<Value> tdc_v = process->Get(String::New("_tickDomainCallback"));
Local<Value> ndt_v = process->Get(String::New("_nextDomainTick"));
if (!tdc_v->IsFunction()) {
fprintf(stderr, "process._tickDomainCallback assigned to non-function\n");
abort();
}
if (!ndt_v->IsFunction()) {
fprintf(stderr, "process._nextDomainTick assigned to non-function\n");
abort();
}
Local<Function> tdc = tdc_v.As<Function>();
Local<Function> ndt = ndt_v.As<Function>();
process->Set(String::New("_tickCallback"), tdc);
src: domain should not replace nextTick function Previously if you cached process.nextTick and then require(&#39;domain&#39;) subsequent nextTick() calls would not be caught because enqueued functions were taking the wrong path. This keeps nextTick to a single function reference and changes the implementation details after domain has been required.
11 years ago
process->Set(String::New("_currentTickHandler"), ndt);
src: fix memory leak in UsingDomains() UsingDomains() assigned process_tickCallback when it had already been set by MakeCallback() a few frames down the call stack. Dispose the handle first or we&#39;ll retain whatever is in the lexical closure of the old process._tickCallback function. Fixes #5795.
12 years ago
process_tickCallback.Dispose(); // Possibly already set by MakeCallback().
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
process_tickCallback = Persistent<Function>::New(tdc);
return Undefined();
}
MakeCallback: Return the callback return value
13 years ago
Handle<Value>
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
MakeDomainCallback(const Handle<Object> object,
node: revert removal of MakeCallback In 0168109 an implementation of MakeCallback was accidently removed. It has been re-added.
12 years ago
const Handle<Function> callback,
int argc,
Handle<Value> argv[]) {
// TODO Hook for long stack traces to be made here.
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
// lazy load domain specific symbols
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
if (enter_symbol.IsEmpty()) {
enter_symbol = NODE_PSYMBOL("enter");
exit_symbol = NODE_PSYMBOL("exit");
disposed_symbol = NODE_PSYMBOL("_disposed");
}
Local<Value> domain_v = object->Get(domain_symbol);
Local<Object> domain;
Local<Function> enter;
Local<Function> exit;
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
TryCatch try_catch;
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
bool has_domain = domain_v->IsObject();
if (has_domain) {
domain = domain_v->ToObject();
assert(!domain.IsEmpty());
if (domain->Get(disposed_symbol)->IsTrue()) {
// domain has been disposed of.
return Undefined();
}
enter = Local<Function>::Cast(domain->Get(enter_symbol));
assert(!enter.IsEmpty());
enter->Call(domain, 0, NULL);
Domain hooks in ReqWrap&lt;T&gt; and MakeCallback
13 years ago
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
if (try_catch.HasCaught()) {
FatalException(try_catch);
return Undefined();
}
Domain hooks in ReqWrap&lt;T&gt; and MakeCallback
13 years ago
}
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
Local<Value> ret = callback->Call(object, argc, argv);
Move MakeCallback and SetErrno to node.cc
14 years ago
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
if (try_catch.HasCaught()) {
FatalException(try_catch);
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
}
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
if (has_domain) {
exit = Local<Function>::Cast(domain->Get(exit_symbol));
assert(!exit.IsEmpty());
exit->Call(domain, 0, NULL);
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
if (try_catch.HasCaught()) {
FatalException(try_catch);
return Undefined();
}
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
}
Domain hooks in ReqWrap&lt;T&gt; and MakeCallback
13 years ago
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
if (tick_infobox.length == 0) {
tick_infobox.index = 0;
tick_infobox.depth = 0;
return ret;
}
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
// process nextTicks after call
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
process_tickCallback->Call(process, 0, NULL);
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
if (try_catch.HasCaught()) {
FatalException(try_catch);
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
}
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
return ret;
}
Handle<Value>
MakeCallback(const Handle<Object> object,
node: revert removal of MakeCallback In 0168109 an implementation of MakeCallback was accidently removed. It has been re-added.
12 years ago
const Handle<Function> callback,
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
int argc,
Handle<Value> argv[]) {
// TODO Hook for long stack traces to be made here.
node: call MakeDomainCallback in all domain cases Previously there was no way to pass a Function callback directly to MakeCallback and support domains. The check has been added so that users never have to worry about supporting domains while using MakeCallback.
12 years ago
if (using_domains)
return MakeDomainCallback(object, callback, argc, argv);
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
// lazy load no domain next tick callbacks
Revert &#34;Move MakeCallback to JS&#34; This reverts commit 0109a9f90acdfdb287436676f2384f7b072fbb6a. Also included: Port all the changes to process._makeCallback into the C++ version. Immediate nextTick, etc. This yields a slight boost in several benchmarks. V8 is optimizing and deoptimizing process._makeCallback repeatedly.
12 years ago
if (process_tickCallback.IsEmpty()) {
Local<Value> cb_v = process->Get(String::New("_tickCallback"));
if (!cb_v->IsFunction()) {
fprintf(stderr, "process._tickCallback assigned to non-function\n");
abort();
}
Local<Function> cb = cb_v.As<Function>();
process_tickCallback = Persistent<Function>::New(cb);
}
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
TryCatch try_catch;
Local<Value> ret = callback->Call(object, argc, argv);
if (try_catch.HasCaught()) {
FatalException(try_catch);
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
}
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
if (tick_infobox.length == 0) {
tick_infobox.index = 0;
tick_infobox.depth = 0;
node: revert removal of MakeCallback In 0168109 an implementation of MakeCallback was accidently removed. It has been re-added.
12 years ago
return ret;
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
}
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
// process nextTicks after call
process_tickCallback->Call(process, 0, NULL);
Domain hooks in ReqWrap&lt;T&gt; and MakeCallback
13 years ago
if (try_catch.HasCaught()) {
FatalException(try_catch);
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Move MakeCallback and SetErrno to node.cc
14 years ago
}
MakeCallback: Return the callback return value
13 years ago
node: revert removal of MakeCallback In 0168109 an implementation of MakeCallback was accidently removed. It has been re-added.
12 years ago
return ret;
}
Handle<Value>
MakeCallback(const Handle<Object> object,
const Handle<String> symbol,
int argc,
Handle<Value> argv[]) {
HandleScope scope;
Local<Function> callback = object->Get(symbol).As<Function>();
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
if (using_domains)
node: revert removal of MakeCallback In 0168109 an implementation of MakeCallback was accidently removed. It has been re-added.
12 years ago
return scope.Close(MakeDomainCallback(object, callback, argc, argv));
return scope.Close(MakeCallback(object, callback, argc, argv));
Move MakeCallback and SetErrno to node.cc
14 years ago
}
process: streamlining tick callback logic * Callbacks from spinner now calls its own function, separate from the tickCallback logic * MakeCallback will call a domain specific function if a domain is detected * _tickCallback assumes no domains, until nextTick receives a callback with a domain. After that _tickCallback is overridden with the domain specific implementation. * _needTickCallback runs in startup() instead of nextTick (isaacs) * Fix bug in _fatalException where exit would be called twice (isaacs) * Process.domain has a default value of null * Manually track nextTickQueue.length (will be useful later) * Update tests to reflect internal api changes
12 years ago
Handle<Value>
MakeCallback(const Handle<Object> object,
const char* method,
int argc,
Handle<Value> argv[]) {
HandleScope scope;
Handle<Value> ret =
MakeCallback(object, String::NewSymbol(method), argc, argv);
return scope.Close(ret);
}
Display sys_errno when UV_UNKNOWN is returned
13 years ago
void SetErrno(uv_err_t err) {
HandleScope scope;
lib, src: remove errno global Remove the errno global. It&#39;s a property on the process object now. Fixes #3095.
12 years ago
static Persistent<String> errno_symbol;
Display sys_errno when UV_UNKNOWN is returned
13 years ago
if (errno_symbol.IsEmpty()) {
lib, src: remove errno global Remove the errno global. It&#39;s a property on the process object now. Fixes #3095.
12 years ago
errno_symbol = NODE_PSYMBOL("_errno");
Display sys_errno when UV_UNKNOWN is returned
13 years ago
}
if (err.code == UV_UNKNOWN) {
char errno_buf[100];
snprintf(errno_buf, 100, "Unknown system errno %d", err.sys_errno_);
lib, src: remove errno global Remove the errno global. It&#39;s a property on the process object now. Fixes #3095.
12 years ago
process->Set(errno_symbol, String::New(errno_buf));
Display sys_errno when UV_UNKNOWN is returned
13 years ago
} else {
lib, src: remove errno global Remove the errno global. It&#39;s a property on the process object now. Fixes #3095.
12 years ago
process->Set(errno_symbol, String::NewSymbol(uv_err_name(err)));
Display sys_errno when UV_UNKNOWN is returned
13 years ago
}
Move MakeCallback and SetErrno to node.cc
14 years ago
}
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
enum encoding ParseEncoding(Handle<Value> encoding_v, enum encoding _default) {
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
HandleScope scope;
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
if (!encoding_v->IsString()) return _default;
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
Avoiding unnecessary ToString() calls String::Utf8Value and String::AsciiValue constructors take Handle&lt;Value&gt; So no need to convert to Handle&lt;String&gt;
13 years ago
String::Utf8Value encoding(encoding_v);
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
if (strcasecmp(*encoding, "utf8") == 0) {
return UTF8;
Also accept &#39;utf-8&#39; to specify utf8 encodings.
15 years ago
} else if (strcasecmp(*encoding, "utf-8") == 0) {
return UTF8;
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
} else if (strcasecmp(*encoding, "ascii") == 0) {
return ASCII;
Buffer: adjust buffer size so the base64-decoded input fits snugly. Stops Valgrind from complaining about uninitialized memory access.
15 years ago
} else if (strcasecmp(*encoding, "base64") == 0) {
return BASE64;
UCS-2 support Closes GH-644.
14 years ago
} else if (strcasecmp(*encoding, "ucs2") == 0) {
return UCS2;
} else if (strcasecmp(*encoding, "ucs-2") == 0) {
return UCS2;
buffer: fix byteLength with UTF-16LE Fixes #4075.
12 years ago
} else if (strcasecmp(*encoding, "utf16le") == 0) {
return UCS2;
} else if (strcasecmp(*encoding, "utf-16le") == 0) {
return UCS2;
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
} else if (strcasecmp(*encoding, "binary") == 0) {
return BINARY;
crypto: bring module into modern age Introduce &#39;buffer&#39; encoding, allow returning and giving buffers as arguments of &#39;crypto&#39; routines. Fix #3278
13 years ago
} else if (strcasecmp(*encoding, "buffer") == 0) {
return BUFFER;
Closes GH-695 Add &#39;hex&#39; encoding to Buffer
14 years ago
} else if (strcasecmp(*encoding, "hex") == 0) {
return HEX;
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
} else if (strcasecmp(*encoding, "raw") == 0) {
Add --no-deprecation and --trace-deprecation flags
13 years ago
if (!no_deprecation) {
fprintf(stderr, "'raw' (array of integers) has been removed. "
"Use 'binary'.\n");
}
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
return BINARY;
} else if (strcasecmp(*encoding, "raws") == 0) {
Add --no-deprecation and --trace-deprecation flags
13 years ago
if (!no_deprecation) {
fprintf(stderr, "'raws' encoding has been renamed to 'binary'. "
"Please update your code.\n");
}
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
return BINARY;
} else {
return _default;
cpplint.py node.cc and node.h
16 years ago
}
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
}
Local<Value> Encode(const void *buf, size_t len, enum encoding encoding) {
src: use StringBytes for DecodeWrite/DecodeBytes/Encode Bonus: this makes node::Encode actually work properly with base64, ucs2, hex, etc.
12 years ago
return StringBytes::Encode(static_cast<const char*>(buf),
len,
encoding);
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
}
cpplint.py node.cc and node.h
16 years ago
// Returns -1 if the handle was not valid for decoding
ssize_t DecodeBytes(v8::Handle<v8::Value> val, enum encoding encoding) {
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
HandleScope scope;
if (val->IsArray()) {
Revert &#34;Fix #3242 Actually deprecate &#39;binary&#39; buffer encoding&#34; This reverts commit 5979f096d1b702ca2ba95664a0bbc8210109775b. Related: - #3279 - #3278
13 years ago
fprintf(stderr, "'raw' encoding (array of integers) has been removed. "
"Use 'binary'.\n");
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
assert(0);
return -1;
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
}
src: use StringBytes for DecodeWrite/DecodeBytes/Encode Bonus: this makes node::Encode actually work properly with base64, ucs2, hex, etc.
12 years ago
return StringBytes::Size(val, encoding);
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
}
#ifndef MIN
cpplint.py node.cc and node.h
16 years ago
# define MIN(a, b) ((a) < (b) ? (a) : (b))
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
#endif
// Returns number of bytes written.
Fix test-http-chunked. Need to check for \0 at end of utf8 strings
15 years ago
ssize_t DecodeWrite(char *buf,
size_t buflen,
cpplint.py node.cc and node.h
16 years ago
v8::Handle<v8::Value> val,
enum encoding encoding) {
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
HandleScope scope;
if (val->IsArray()) {
src: use StringBytes for DecodeWrite/DecodeBytes/Encode Bonus: this makes node::Encode actually work properly with base64, ucs2, hex, etc.
12 years ago
fprintf(stderr, "'raw' encoding (array of integers) has been removed.\n");
Remove &#34;raw&#34; encoding. Rename &#34;raws&#34; to &#34;binary&#34;. Deprecation warnings have been added to help the conversion to this new API.
16 years ago
assert(0);
return -1;
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
}
src: use StringBytes for DecodeWrite/DecodeBytes/Encode Bonus: this makes node::Encode actually work properly with base64, ucs2, hex, etc.
12 years ago
return StringBytes::Write(buf, buflen, val, encoding, NULL);
Byte stream to V8 decoding and encoding. This does not (should not) change behavior. Pulls those two functions (encode/decode) out into node.cc.
16 years ago
}
Saner, yet still ugly, fix to displaying syntax errors
14 years ago
void DisplayExceptionLine (TryCatch &try_catch) {
Report errors properly from --eval and stdin
13 years ago
// Prevent re-entry into this function. For example, if there is
// a throw from a program in vm.runInThisContext(code, filename, true),
// then we want to show the original failure, not the secondary one.
static bool displayed_error = false;
if (displayed_error) return;
displayed_error = true;
Saner, yet still ugly, fix to displaying syntax errors
14 years ago
HandleScope scope;
Fix whitespace and warnings in node.cc
15 years ago
Handle<Message> message = try_catch.Message();
Bugfix: ReportException shouldn&#39;t forget the top frame.
16 years ago
Remove tty_legacy
13 years ago
uv_tty_reset_mode();
Only one line break before stacktrace
15 years ago
fprintf(stderr, "\n");
Add binding to termios, implement readline for repl
15 years ago
Saner, yet still ugly, fix to displaying syntax errors
14 years ago
if (!message.IsEmpty()) {
Simplify exception reporting
15 years ago
// Print (filename):(line number): (message).
String::Utf8Value filename(message->GetScriptResourceName());
Saner, yet still ugly, fix to displaying syntax errors
14 years ago
const char* filename_string = *filename;
Simplify exception reporting
15 years ago
int linenum = message->GetLineNumber();
fprintf(stderr, "%s:%i\n", filename_string, linenum);
// Print line of source code.
String::Utf8Value sourceline(message->GetSourceLine());
Saner, yet still ugly, fix to displaying syntax errors
14 years ago
const char* sourceline_string = *sourceline;
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
Report errors properly from --eval and stdin
13 years ago
// Because of how node modules work, all scripts are wrapped with a
// "function (module, exports, __filename, ...) {"
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
// to provide script local variables.
//
// When reporting errors on the first line of a script, this wrapper
Report errors properly from --eval and stdin
13 years ago
// function is leaked to the user. There used to be a hack here to
// truncate off the first 62 characters, but it caused numerous other
// problems when vm.runIn*Context() methods were used for non-module
// code.
//
// If we ever decide to re-instate such a hack, the following steps
// must be taken:
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
//
Report errors properly from --eval and stdin
13 years ago
// 1. Pass a flag around to say "this code was wrapped"
// 2. Update the stack frame output so that it is also correct.
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
//
Report errors properly from --eval and stdin
13 years ago
// It would probably be simpler to add a line rather than add some
// number of characters to the first line, since V8 truncates the
// sourceline to 78 characters, and we end up not providing very much
// useful debugging info to the user if we remove 62 characters.
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
Simplify exception reporting
15 years ago
int start = message->GetStartColumn();
Report errors properly from --eval and stdin
13 years ago
int end = message->GetEndColumn();
// fprintf(stderr, "---\nsourceline:%s\noffset:%d\nstart:%d\nend:%d\n---\n", sourceline_string, start, end);
fprintf(stderr, "%s\n", sourceline_string);
// Print wavy underline (GetUnderline is deprecated).
for (int i = 0; i < start; i++) {
Check for tabs in source line in DisplayExceptionLine Fix for joyent/node#3280
13 years ago
fputc((sourceline_string[i] == '\t') ? '\t' : ' ', stderr);
Simplify exception reporting
15 years ago
}
for (int i = start; i < end; i++) {
Check for tabs in source line in DisplayExceptionLine Fix for joyent/node#3280
13 years ago
fputc('^', stderr);
Simplify exception reporting
15 years ago
}
Check for tabs in source line in DisplayExceptionLine Fix for joyent/node#3280
13 years ago
fputc('\n', stderr);
Hack to display more useful SyntaxError exceptions. For some reason v8 doesn&#39;t include the frame with the syntax error in the stack trace - so have to special case it.
15 years ago
}
Saner, yet still ugly, fix to displaying syntax errors
14 years ago
}
static void ReportException(TryCatch &try_catch, bool show_line) {
HandleScope scope;
if (show_line) DisplayExceptionLine(try_catch);
Simplify exception reporting
15 years ago
Simplify stack trace reporting using new V8 API
15 years ago
String::Utf8Value trace(try_catch.StackTrace());
Close #974 Properly report traceless errors. Also, tests for the same.
14 years ago
// range errors have a trace member set to undefined
if (trace.length() > 0 && !try_catch.StackTrace()->IsUndefined()) {
Stack traces for mjsunit errors, better error reporting function. The error reporting function tries to look at the &#34;stack&#34; element of the exception.
16 years ago
fprintf(stderr, "%s\n", *trace);
Report &#34;weird&#34; errors a little better. There are a few kinds of errors that are very confusing. 1. Errors raised in nextTick 2. Errors emitted on the &#34;error&#34; event 3. RangeErrors that crash the program (or anything without a stack trace) Long traces will make make these better, of course. In the meantime, this adds a few handy signposts (in the form of better error reporting and comments on the otherwise inscrutable code printed to the terminal) that can help new users find the cause, or at least, ask for help more effectively.
14 years ago
} else {
// this really only happens for RangeErrors, since they're the only
Close #974 Properly report traceless errors. Also, tests for the same.
14 years ago
// kind that won't have all this info in the trace, or when non-Error
// objects are thrown manually.
Report &#34;weird&#34; errors a little better. There are a few kinds of errors that are very confusing. 1. Errors raised in nextTick 2. Errors emitted on the &#34;error&#34; event 3. RangeErrors that crash the program (or anything without a stack trace) Long traces will make make these better, of course. In the meantime, this adds a few handy signposts (in the form of better error reporting and comments on the otherwise inscrutable code printed to the terminal) that can help new users find the cause, or at least, ask for help more effectively.
14 years ago
Local<Value> er = try_catch.Exception();
Close #974 Properly report traceless errors. Also, tests for the same.
14 years ago
bool isErrorObject = er->IsObject() &&
!(er->ToObject()->Get(String::New("message"))->IsUndefined()) &&
!(er->ToObject()->Get(String::New("name"))->IsUndefined());
if (isErrorObject) {
String::Utf8Value name(er->ToObject()->Get(String::New("name")));
fprintf(stderr, "%s: ", *name);
}
Avoiding unnecessary ToString() calls String::Utf8Value and String::AsciiValue constructors take Handle&lt;Value&gt; So no need to convert to Handle&lt;String&gt;
13 years ago
String::Utf8Value msg(!isErrorObject ? er
: er->ToObject()->Get(String::New("message")));
Report &#34;weird&#34; errors a little better. There are a few kinds of errors that are very confusing. 1. Errors raised in nextTick 2. Errors emitted on the &#34;error&#34; event 3. RangeErrors that crash the program (or anything without a stack trace) Long traces will make make these better, of course. In the meantime, this adds a few handy signposts (in the form of better error reporting and comments on the otherwise inscrutable code printed to the terminal) that can help new users find the cause, or at least, ask for help more effectively.
14 years ago
fprintf(stderr, "%s\n", *msg);
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
}
Report &#34;weird&#34; errors a little better. There are a few kinds of errors that are very confusing. 1. Errors raised in nextTick 2. Errors emitted on the &#34;error&#34; event 3. RangeErrors that crash the program (or anything without a stack trace) Long traces will make make these better, of course. In the meantime, this adds a few handy signposts (in the form of better error reporting and comments on the otherwise inscrutable code printed to the terminal) that can help new users find the cause, or at least, ask for help more effectively.
14 years ago
Add a few fflush(stderr) calls to track down missing stacktraces.
16 years ago
fflush(stderr);
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
}
cpplint.py node.cc and node.h
16 years ago
// Executes a str within the current v8 context.
Add string class that uses ExternalAsciiStringResource. Change the natives to use this class instead of creating completely new strings. Reduces memory usage by about 1 MB.
14 years ago
Local<Value> ExecuteString(Handle<String> source, Handle<Value> filename) {
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
HandleScope scope;
debugging/improving the module framework
16 years ago
TryCatch try_catch;
Script class with eval-function-family in binding(&#39;evals&#39;) plus tests.
15 years ago
Local<v8::Script> script = v8::Script::Compile(source, filename);
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
if (script.IsEmpty()) {
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
ReportException(try_catch, true);
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
exit(3);
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
}
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
Local<Value> result = script->Run();
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
if (result.IsEmpty()) {
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
ReportException(try_catch, true);
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
exit(4);
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
}
return scope.Close(result);
}
Temporary function to determine str byte length Will need a better place later on
15 years ago
process: add _getActiveHandles(), _getActiveRequests() * process._getActiveHandles() returns a list containing all active handles (timers, sockets, etc.) that have not been unref&#39;d. * process._getActiveRequests() returns a list of active requests (in-flight actions like connecting to a remote host, writing data to a socket, etc.).
13 years ago
static Handle<Value> GetActiveRequests(const Arguments& args) {
HandleScope scope;
Local<Array> ary = Array::New();
ngx_queue_t* q = NULL;
int i = 0;
ngx_queue_foreach(q, &req_wrap_queue) {
ReqWrap<uv_req_t>* w = container_of(q, ReqWrap<uv_req_t>, req_wrap_queue_);
if (w->object_.IsEmpty()) continue;
ary->Set(i++, w->object_);
}
return scope.Close(ary);
}
// Non-static, friend of HandleWrap. Could have been a HandleWrap method but
// implemented here for consistency with GetActiveRequests().
Handle<Value> GetActiveHandles(const Arguments& args) {
HandleScope scope;
Local<Array> ary = Array::New();
ngx_queue_t* q = NULL;
int i = 0;
node: make _getActiveHandles() return user objects Before this commit, process._getActiveHandles() returned a list of internal handles. Now, it returns the user objects that handles are attached to. For example, a tcp_wrap handle will now return its parent net.Socket object. It works for all handle types except timers because timer handles are shared across multiple user objects.
13 years ago
Local<String> owner_sym = String::New("owner");
process: add _getActiveHandles(), _getActiveRequests() * process._getActiveHandles() returns a list containing all active handles (timers, sockets, etc.) that have not been unref&#39;d. * process._getActiveRequests() returns a list of active requests (in-flight actions like connecting to a remote host, writing data to a socket, etc.).
13 years ago
ngx_queue_foreach(q, &handle_wrap_queue) {
HandleWrap* w = container_of(q, HandleWrap, handle_wrap_queue_);
handle_wrap: replace unref_ field with flags_ field Prep work for a follow-up commit that adds support for close callbacks.
12 years ago
if (w->object_.IsEmpty() || (w->flags_ & HandleWrap::kUnref)) continue;
node: make _getActiveHandles() return user objects Before this commit, process._getActiveHandles() returned a list of internal handles. Now, it returns the user objects that handles are attached to. For example, a tcp_wrap handle will now return its parent net.Socket object. It works for all handle types except timers because timer handles are shared across multiple user objects.
13 years ago
Local<Value> obj = w->object_->Get(owner_sym);
if (obj->IsUndefined()) obj = *w->object_;
ary->Set(i++, obj);
process: add _getActiveHandles(), _getActiveRequests() * process._getActiveHandles() returns a list containing all active handles (timers, sockets, etc.) that have not been unref&#39;d. * process._getActiveRequests() returns a list of active requests (in-flight actions like connecting to a remote host, writing data to a socket, etc.).
13 years ago
}
return scope.Close(ary);
}
Add process.abort
13 years ago
static Handle<Value> Abort(const Arguments& args) {
abort();
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Add process.abort
13 years ago
}
Added process.chdir()
15 years ago
static Handle<Value> Chdir(const Arguments& args) {
HandleScope scope;
Fix whitespace and warnings in node.cc
15 years ago
Added process.chdir()
15 years ago
if (args.Length() != 1 || !args[0]->IsString()) {
return ThrowException(Exception::Error(String::New("Bad argument.")));
}
Fix whitespace and warnings in node.cc
15 years ago
Avoiding unnecessary ToString() calls String::Utf8Value and String::AsciiValue constructors take Handle&lt;Value&gt; So no need to convert to Handle&lt;String&gt;
13 years ago
String::Utf8Value path(args[0]);
Fix whitespace and warnings in node.cc
15 years ago
Win: make process.cwd and chdir support non-ansi characters Closes GH-2215
13 years ago
uv_err_t r = uv_chdir(*path);
Fix whitespace and warnings in node.cc
15 years ago
Win: make process.cwd and chdir support non-ansi characters Closes GH-2215
13 years ago
if (r.code != UV_OK) {
return ThrowException(UVException(r.code, "uv_chdir"));
Added process.chdir()
15 years ago
}
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Added process.chdir()
15 years ago
}
Don&#39;t die when arguments are passed to process.cwd()
13 years ago
cpplint.py node.cc and node.h
16 years ago
static Handle<Value> Cwd(const Arguments& args) {
Feature: add node.cwd() to access the current working directory.
16 years ago
HandleScope scope;
Win: make process.cwd and chdir support non-ansi characters Closes GH-2215
13 years ago
#ifdef _WIN32
/* MAX_PATH is in characters, not bytes. Make sure we have enough headroom. */
char buf[MAX_PATH * 4 + 1];
#else
char buf[PATH_MAX + 1];
#endif
Feature: add node.cwd() to access the current working directory.
16 years ago
Win: make process.cwd and chdir support non-ansi characters Closes GH-2215
13 years ago
uv_err_t r = uv_cwd(buf, ARRAY_SIZE(buf) - 1);
if (r.code != UV_OK) {
return ThrowException(UVException(r.code, "uv_cwd"));
Feature: add node.cwd() to access the current working directory.
16 years ago
}
Coverity fixes: src/node.cc
15 years ago
Win: make process.cwd and chdir support non-ansi characters Closes GH-2215
13 years ago
buf[ARRAY_SIZE(buf) - 1] = '\0';
Local<String> cwd = String::New(buf);
Feature: add node.cwd() to access the current working directory.
16 years ago
return scope.Close(cwd);
}
Win: make process.cwd and chdir support non-ansi characters Closes GH-2215
13 years ago
windows: fix test-umask
14 years ago
static Handle<Value> Umask(const Arguments& args) {
add process.umask()
15 years ago
HandleScope scope;
Updated patch of node.cc for supporting reading of umask
15 years ago
unsigned int old;
Support octal strings for modes This allows the various fs utilities and process.umask to be used in ECMAScript 5 Strict Mode, where the octal literal format is verboten, without requiring users to litter their code with a bunch of parseInt calls.
14 years ago
windows: fix test-umask
14 years ago
if (args.Length() < 1 || args[0]->IsUndefined()) {
Updated patch of node.cc for supporting reading of umask
15 years ago
old = umask(0);
umask((mode_t)old);
Support octal strings for modes This allows the various fs utilities and process.umask to be used in ECMAScript 5 Strict Mode, where the octal literal format is verboten, without requiring users to litter their code with a bunch of parseInt calls.
14 years ago
} else if(!args[0]->IsInt32() && !args[0]->IsString()) {
add process.umask()
15 years ago
return ThrowException(Exception::TypeError(
Support octal strings for modes This allows the various fs utilities and process.umask to be used in ECMAScript 5 Strict Mode, where the octal literal format is verboten, without requiring users to litter their code with a bunch of parseInt calls.
14 years ago
String::New("argument must be an integer or octal string.")));
} else {
int oct;
if(args[0]->IsInt32()) {
oct = args[0]->Uint32Value();
} else {
oct = 0;
String::Utf8Value str(args[0]);
// Parse the octal string.
for (int i = 0; i < str.length(); i++) {
char c = (*str)[i];
if (c > '7' || c < '0') {
return ThrowException(Exception::TypeError(
String::New("invalid octal string")));
}
oct *= 8;
oct += c - '0';
}
}
old = umask(static_cast<mode_t>(oct));
Updated patch of node.cc for supporting reading of umask
15 years ago
}
Support octal strings for modes This allows the various fs utilities and process.umask to be used in ECMAScript 5 Strict Mode, where the octal literal format is verboten, without requiring users to litter their code with a bunch of parseInt calls.
14 years ago
add process.umask()
15 years ago
return scope.Close(Uint32::New(old));
}
Add process.setuid(), getuid()
15 years ago
windows: fix test-umask
14 years ago
#ifdef __POSIX__
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
static const uid_t uid_not_found = static_cast<uid_t>(-1);
static const gid_t gid_not_found = static_cast<gid_t>(-1);
static uid_t uid_by_name(const char* name) {
struct passwd pwd;
struct passwd* pp;
char buf[8192];
errno = 0;
pp = NULL;
src: remove unused variables
12 years ago
if (getpwnam_r(name, &pwd, buf, sizeof(buf), &pp) == 0 && pp != NULL) {
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
return pp->pw_uid;
}
return uid_not_found;
}
static char* name_by_uid(uid_t uid) {
struct passwd pwd;
struct passwd* pp;
char buf[8192];
int rc;
errno = 0;
pp = NULL;
if ((rc = getpwuid_r(uid, &pwd, buf, sizeof(buf), &pp)) == 0 && pp != NULL) {
return strdup(pp->pw_name);
}
if (rc == 0) {
errno = ENOENT;
}
return NULL;
}
static gid_t gid_by_name(const char* name) {
struct group pwd;
struct group* pp;
char buf[8192];
errno = 0;
pp = NULL;
src: remove unused variables
12 years ago
if (getgrnam_r(name, &pwd, buf, sizeof(buf), &pp) == 0 && pp != NULL) {
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
return pp->gr_gid;
}
return gid_not_found;
}
#if 0 // For future use.
static const char* name_by_gid(gid_t gid) {
struct group pwd;
struct group* pp;
char buf[8192];
int rc;
errno = 0;
pp = NULL;
if ((rc = getgrgid_r(gid, &pwd, buf, sizeof(buf), &pp)) == 0 && pp != NULL) {
return strdup(pp->gr_name);
}
if (rc == 0) {
errno = ENOENT;
}
return NULL;
}
#endif
static uid_t uid_by_name(Handle<Value> value) {
if (value->IsUint32()) {
return static_cast<uid_t>(value->Uint32Value());
} else {
String::Utf8Value name(value);
return uid_by_name(*name);
}
}
static gid_t gid_by_name(Handle<Value> value) {
if (value->IsUint32()) {
return static_cast<gid_t>(value->Uint32Value());
} else {
String::Utf8Value name(value);
return gid_by_name(*name);
}
}
Add process.setuid(), getuid()
15 years ago
static Handle<Value> GetUid(const Arguments& args) {
HandleScope scope;
src: fix process.getuid() return value And process.getgid() too. Commit ed80638 changed fs.chown() and fs.fchown() to only accept unsigned integers. Make process.getuid() and process.getgid() follow suit. This commit should unbreak npm on OS X - it&#39;s hitting the new &#39;uid must be an unsigned int&#39; check when installing as e.g. user &#39;nobody&#39; (which has an UID of -2 in /etc/passwd or 4294967294 when cast to an uid_t.) Fixes #5904.
12 years ago
uid_t uid = getuid();
return scope.Close(Integer::NewFromUnsigned(uid));
Add process.setuid(), getuid()
15 years ago
}
windows: fix test-umask
14 years ago
Add setgid,getgid
15 years ago
static Handle<Value> GetGid(const Arguments& args) {
HandleScope scope;
src: fix process.getuid() return value And process.getgid() too. Commit ed80638 changed fs.chown() and fs.fchown() to only accept unsigned integers. Make process.getuid() and process.getgid() follow suit. This commit should unbreak npm on OS X - it&#39;s hitting the new &#39;uid must be an unsigned int&#39; check when installing as e.g. user &#39;nobody&#39; (which has an UID of -2 in /etc/passwd or 4294967294 when cast to an uid_t.) Fixes #5904.
12 years ago
gid_t gid = getgid();
return scope.Close(Integer::NewFromUnsigned(gid));
Add setgid,getgid
15 years ago
}
static Handle<Value> SetGid(const Arguments& args) {
HandleScope scope;
Fix gcc warnings in node.cc
15 years ago
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
if (!args[0]->IsUint32() && !args[0]->IsString()) {
return ThrowTypeError("setgid argument must be a number or a string");
Add setgid,getgid
15 years ago
}
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
gid_t gid = gid_by_name(args[0]);
Allow process.setuid() and process.setgid() to accept string names in lieu of numeric IDs
15 years ago
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
if (gid == gid_not_found) {
return ThrowError("setgid group id does not exist");
Allow process.setuid() and process.setgid() to accept string names in lieu of numeric IDs
15 years ago
}
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
if (setgid(gid)) {
Allow process.setuid() and process.setgid() to accept string names in lieu of numeric IDs
15 years ago
return ThrowException(ErrnoException(errno, "setgid"));
Add setgid,getgid
15 years ago
}
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Add setgid,getgid
15 years ago
}
Add process.setuid(), getuid()
15 years ago
windows: fix test-umask
14 years ago
Add process.setuid(), getuid()
15 years ago
static Handle<Value> SetUid(const Arguments& args) {
HandleScope scope;
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
if (!args[0]->IsUint32() && !args[0]->IsString()) {
return ThrowTypeError("setuid argument must be a number or a string");
Add process.setuid(), getuid()
15 years ago
}
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
uid_t uid = uid_by_name(args[0]);
Allow process.setuid() and process.setgid() to accept string names in lieu of numeric IDs
15 years ago
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
if (uid == uid_not_found) {
return ThrowError("setuid user id does not exist");
}
if (setuid(uid)) {
return ThrowException(ErrnoException(errno, "setuid"));
}
Allow process.setuid() and process.setgid() to accept string names in lieu of numeric IDs
15 years ago
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
}
static Handle<Value> GetGroups(const Arguments& args) {
HandleScope scope;
int ngroups = getgroups(0, NULL);
if (ngroups == -1) {
return ThrowException(ErrnoException(errno, "getgroups"));
}
gid_t* groups = new gid_t[ngroups];
ngroups = getgroups(ngroups, groups);
if (ngroups == -1) {
delete[] groups;
return ThrowException(ErrnoException(errno, "getgroups"));
}
Local<Array> groups_list = Array::New(ngroups);
bool seen_egid = false;
gid_t egid = getegid();
for (int i = 0; i < ngroups; i++) {
src, test: downgrade to v8 3.14 api
12 years ago
groups_list->Set(i, Integer::New(groups[i]));
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
if (groups[i] == egid) seen_egid = true;
}
delete[] groups;
if (seen_egid == false) {
src, test: downgrade to v8 3.14 api
12 years ago
groups_list->Set(ngroups, Integer::New(egid));
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
}
return scope.Close(groups_list);
}
static Handle<Value> SetGroups(const Arguments& args) {
HandleScope scope;
if (!args[0]->IsArray()) {
return ThrowTypeError("argument 1 must be an array");
}
Local<Array> groups_list = args[0].As<Array>();
size_t size = groups_list->Length();
gid_t* groups = new gid_t[size];
for (size_t i = 0; i < size; i++) {
gid_t gid = gid_by_name(groups_list->Get(i));
if (gid == gid_not_found) {
delete[] groups;
return ThrowError("group name not found");
Allow process.setuid() and process.setgid() to accept string names in lieu of numeric IDs
15 years ago
}
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
groups[i] = gid;
Allow process.setuid() and process.setgid() to accept string names in lieu of numeric IDs
15 years ago
}
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
int rc = setgroups(size, groups);
delete[] groups;
if (rc == -1) {
return ThrowException(ErrnoException(errno, "setgroups"));
Add process.setuid(), getuid()
15 years ago
}
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Add process.setuid(), getuid()
15 years ago
}
windows: fix test-umask
14 years ago
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
static Handle<Value> InitGroups(const Arguments& args) {
HandleScope scope;
if (!args[0]->IsUint32() && !args[0]->IsString()) {
return ThrowTypeError("argument 1 must be a number or a string");
}
if (!args[1]->IsUint32() && !args[1]->IsString()) {
return ThrowTypeError("argument 2 must be a number or a string");
}
String::Utf8Value arg0(args[0]);
gid_t extra_group;
bool must_free;
char* user;
if (args[0]->IsUint32()) {
user = name_by_uid(args[0]->Uint32Value());
must_free = true;
} else {
user = *arg0;
must_free = false;
}
if (user == NULL) {
return ThrowError("initgroups user not found");
}
extra_group = gid_by_name(args[1]);
if (extra_group == gid_not_found) {
if (must_free) free(user);
return ThrowError("initgroups extra group not found");
}
int rc = initgroups(user, extra_group);
if (must_free) {
free(user);
}
if (rc) {
return ThrowException(ErrnoException(errno, "initgroups"));
}
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
}
Disable node.cc functions that are not supported on windows
14 years ago
#endif // __POSIX__
Add process.setuid(), getuid()
15 years ago
cpplint.py node.cc and node.h
16 years ago
v8::Handle<v8::Value> Exit(const v8::Arguments& args) {
Add handlescope and flush to process.reallyExit()
15 years ago
HandleScope scope;
exit(args[0]->IntegerValue());
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
remove process.{cc,h} process.exit() now exit() the process object might return in the future but for now it is going away.
16 years ago
}
Change GC idle notify so that it runs along side setInterval Doesn&#39;t appear any slower. Also checks for high memory usage and tries to force more notifications.
15 years ago
Add process.uptime().
14 years ago
static Handle<Value> Uptime(const Arguments& args) {
HandleScope scope;
Remove platform files, and use uv platform api
13 years ago
double uptime;
Add process.uptime().
14 years ago
src: use monotonic time for process.uptime() `process.uptime()` interface will return the amount of time the current process has been running. To achieve this it was caching the `uv_uptime` value at program start, and then on the call to `process.uptime()` returning the delta between the two values. `uv_uptime` is defined as the number of seconds the operating system has been up since last boot. On sunos this interface uses `kstat`s which can be a significantly expensive operation as it requires exclusive access, but because of the design of `process.uptime()` node *had* to always call this on start. As a result if you had many node processes all starting at the same time you would suffer lock contention as they all tried to read kstats. Instead of using `uv_uptime` to achieve this, the libuv loop already has a concept of current loop time in the form of `uv_now()` which is in fact monotonically increasing, and already stored directly on the loop. By using this value at start every platform performs at least one fewer syscall during initialization. Since the interface to `uv_uptime` is defined as seconds, in the call to `process.uptime()` we now `uv_update_time` get our delta, divide by 1000 to get seconds, and then convert to an `Integer`. In 0.12 we can move back to `Number::New` instead and not lose precision. Caveat: For some platforms `uv_uptime` reports time monotonically increasing regardless of system hibernation, `uv_now` interface is also monotonically increasing but may not reflect time spent in hibernation.
11 years ago
uv_update_time(uv_default_loop());
uptime = uv_now(uv_default_loop()) - prog_start_time;
Add process.uptime().
14 years ago
src: use monotonic time for process.uptime() `process.uptime()` interface will return the amount of time the current process has been running. To achieve this it was caching the `uv_uptime` value at program start, and then on the call to `process.uptime()` returning the delta between the two values. `uv_uptime` is defined as the number of seconds the operating system has been up since last boot. On sunos this interface uses `kstat`s which can be a significantly expensive operation as it requires exclusive access, but because of the design of `process.uptime()` node *had* to always call this on start. As a result if you had many node processes all starting at the same time you would suffer lock contention as they all tried to read kstats. Instead of using `uv_uptime` to achieve this, the libuv loop already has a concept of current loop time in the form of `uv_now()` which is in fact monotonically increasing, and already stored directly on the loop. By using this value at start every platform performs at least one fewer syscall during initialization. Since the interface to `uv_uptime` is defined as seconds, in the call to `process.uptime()` we now `uv_update_time` get our delta, divide by 1000 to get seconds, and then convert to an `Integer`. In 0.12 we can move back to `Number::New` instead and not lose precision. Caveat: For some platforms `uv_uptime` reports time monotonically increasing regardless of system hibernation, `uv_now` interface is also monotonically increasing but may not reflect time spent in hibernation.
11 years ago
return scope.Close(Integer::New(uptime / 1000));
Add process.uptime().
14 years ago
}
Change GC idle notify so that it runs along side setInterval Doesn&#39;t appear any slower. Also checks for high memory usage and tries to force more notifications.
15 years ago
Add process.uvCounters() for debugging
13 years ago
Move memoryUsage() into C on Linux
15 years ago
v8::Handle<v8::Value> MemoryUsage(const v8::Arguments& args) {
HandleScope scope;
Remove process.memoryUsage().vsize Not meaningful cross platform; unused.
13 years ago
size_t rss;
Move memoryUsage() into C on Linux
15 years ago
Remove platform files, and use uv platform api
13 years ago
uv_err_t err = uv_resident_set_memory(&rss);
Move memoryUsage() into C on Linux
15 years ago
Remove platform files, and use uv platform api
13 years ago
if (err.code != UV_OK) {
return ThrowException(UVException(err.code, "uv_resident_set_memory"));
Add sys.memoryUsage()
15 years ago
}
Local<Object> info = Object::New();
Statically define symbols ~7% improvement in hello world HTTP response bench.
15 years ago
if (rss_symbol.IsEmpty()) {
rss_symbol = NODE_PSYMBOL("rss");
heap_total_symbol = NODE_PSYMBOL("heapTotal");
heap_used_symbol = NODE_PSYMBOL("heapUsed");
}
fixed problem with rss memory reporting incorrect number over 4 gigs
13 years ago
info->Set(rss_symbol, Number::New(rss));
Add sys.memoryUsage()
15 years ago
Add V8 heap info to process.memoryUsage()
15 years ago
// V8 memory usage
HeapStatistics v8_heap_stats;
V8::GetHeapStatistics(&v8_heap_stats);
Statically define symbols ~7% improvement in hello world HTTP response bench.
15 years ago
info->Set(heap_total_symbol,
src, test: downgrade to v8 3.14 api
12 years ago
Integer::NewFromUnsigned(v8_heap_stats.total_heap_size()));
Statically define symbols ~7% improvement in hello world HTTP response bench.
15 years ago
info->Set(heap_used_symbol,
src, test: downgrade to v8 3.14 api
12 years ago
Integer::NewFromUnsigned(v8_heap_stats.used_heap_size()));
Add V8 heap info to process.memoryUsage()
15 years ago
Add sys.memoryUsage()
15 years ago
return scope.Close(info);
}
Move memoryUsage() into C on Linux
15 years ago
Fix whitespace errors introduced by porting efforts
14 years ago
Move constants out of process object
15 years ago
Handle<Value> Kill(const Arguments& args) {
Added external interface for signal handlers. Also process.pid and node.kill().
15 years ago
HandleScope scope;
Fix whitespace and warnings in node.cc
15 years ago
process.kill coerces args to integers
14 years ago
if (args.Length() != 2) {
Added external interface for signal handlers. Also process.pid and node.kill().
15 years ago
return ThrowException(Exception::Error(String::New("Bad argument.")));
}
Fix whitespace and warnings in node.cc
15 years ago
process.kill to use uv_kill
13 years ago
int pid = args[0]->IntegerValue();
Move constants out of process object
15 years ago
int sig = args[1]->Int32Value();
process.kill to use uv_kill
13 years ago
uv_err_t err = uv_kill(pid, sig);
Added external interface for signal handlers. Also process.pid and node.kill().
15 years ago
process.kill to use uv_kill
13 years ago
if (err.code != UV_OK) {
SetErrno(err);
src, test: downgrade to v8 3.14 api
12 years ago
return scope.Close(Integer::New(-1));
process.kill to use uv_kill
13 years ago
}
Added external interface for signal handlers. Also process.pid and node.kill().
15 years ago
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Added external interface for signal handlers. Also process.pid and node.kill().
15 years ago
}
process: add `process.hrtime()` This commit adds a high-resolution timer function.
13 years ago
// used in Hrtime() below
#define NANOS_PER_SEC 1000000000
// Hrtime exposes libuv's uv_hrtime() high-resolution timer.
// The value returned by uv_hrtime() is a 64-bit int representing nanoseconds,
// so this function instead returns an Array with 2 entries representing seconds
// and nanoseconds, to avoid any integer overflow possibility.
// Pass in an Array from a previous hrtime() call to instead get a time diff.
Handle<Value> Hrtime(const v8::Arguments& args) {
HandleScope scope;
uint64_t t = uv_hrtime();
if (args.Length() > 0) {
// return a time diff tuple
process: throw a TypeError when anything but an Array is passed to hrtime() Fixes #3664.
13 years ago
if (!args[0]->IsArray()) {
Local<Value> exception = Exception::TypeError(
String::New("process.hrtime() only accepts an Array tuple."));
return ThrowException(exception);
}
process: add `process.hrtime()` This commit adds a high-resolution timer function.
13 years ago
Local<Array> inArray = Local<Array>::Cast(args[0]);
uint64_t seconds = inArray->Get(0)->Uint32Value();
uint64_t nanos = inArray->Get(1)->Uint32Value();
t -= (seconds * NANOS_PER_SEC) + nanos;
}
Local<Array> tuple = Array::New(2);
src, test: downgrade to v8 3.14 api
12 years ago
tuple->Set(0, Integer::NewFromUnsigned(t / NANOS_PER_SEC));
tuple->Set(1, Integer::NewFromUnsigned(t % NANOS_PER_SEC));
process: add `process.hrtime()` This commit adds a high-resolution timer function.
13 years ago
return scope.Close(tuple);
}
Move constants out of process object
15 years ago
Make node::DLOpen use uv_dlopen
13 years ago
typedef void (UV_DYNAMIC* extInit)(Handle<Object> exports);
First attempt at node.dlopen Compiled first working &#39;hello world&#39; module with this config
16 years ago
addon: Pass module object to NODE_MODULE init function mainly to allow native addons to export single functions on rather than being restricted to operating on an existing object. Init functions now receive exports as the first argument, like before, but also the module object as the second argument, if they support it. Related to #4634 cc: @rvagg
12 years ago
// DLOpen is process.dlopen(module, filename).
// Used to load 'module.node' dynamically shared objects.
cpplint.py node.cc and node.h
16 years ago
Handle<Value> DLOpen(const v8::Arguments& args) {
First attempt at node.dlopen Compiled first working &#39;hello world&#39; module with this config
16 years ago
HandleScope scope;
Make node::DLOpen use uv_dlopen
13 years ago
char symbol[1024], *base, *pos;
uv_lib_t lib;
int r;
First attempt at node.dlopen Compiled first working &#39;hello world&#39; module with this config
16 years ago
Make node::DLOpen use uv_dlopen
13 years ago
if (args.Length() < 2) {
Make error reporting from node::DLOpen more consistent
13 years ago
Local<Value> exception = Exception::Error(
String::New("process.dlopen takes exactly 2 arguments."));
return ThrowException(exception);
Make node::DLOpen use uv_dlopen
13 years ago
}
Build static executable. - Fix a few errors with node.dlopen() - Report errors to stderr (this should probably be a separate commit, but whatever)
16 years ago
addon: Pass module object to NODE_MODULE init function mainly to allow native addons to export single functions on rather than being restricted to operating on an existing object. Init functions now receive exports as the first argument, like before, but also the module object as the second argument, if they support it. Related to #4634 cc: @rvagg
12 years ago
Local<Object> module = args[0]->ToObject(); // Cast
String::Utf8Value filename(args[1]); // Cast
if (exports_symbol.IsEmpty()) {
exports_symbol = NODE_PSYMBOL("exports");
}
Local<Object> exports = module->Get(exports_symbol)->ToObject();
First attempt at node.dlopen Compiled first working &#39;hello world&#39; module with this config
16 years ago
deps: upgrade libuv to a478847 The event loop&#39;s reference counting scheme in this version of libuv has changed. Update the libuv bindings to reflect that fact.
13 years ago
if (uv_dlopen(*filename, &lib)) {
Local<String> errmsg = String::New(uv_dlerror(&lib));
#ifdef _WIN32
// Windows needs to add the filename into the error message
core: Append filename properly in dlopen on windows Fixes simple/test-module-loading on win32
12 years ago
errmsg = String::Concat(errmsg, args[1]->ToString());
process: show detailed error message in process.dlopen()
13 years ago
#endif
deps: upgrade libuv to a478847 The event loop&#39;s reference counting scheme in this version of libuv has changed. Update the libuv bindings to reflect that fact.
13 years ago
return ThrowException(Exception::Error(errmsg));
First attempt at node.dlopen Compiled first working &#39;hello world&#39; module with this config
16 years ago
}
addon: Pass module object to NODE_MODULE init function mainly to allow native addons to export single functions on rather than being restricted to operating on an existing object. Init functions now receive exports as the first argument, like before, but also the module object as the second argument, if they support it. Related to #4634 cc: @rvagg
12 years ago
String::Utf8Value path(args[1]);
Make node::DLOpen use uv_dlopen
13 years ago
base = *path;
Add support for the module structure to process.dlopen.
15 years ago
Make node::DLOpen use uv_dlopen
13 years ago
/* Find the shared library filename within the full path. */
#ifdef __POSIX__
pos = strrchr(base, '/');
if (pos != NULL) {
core: fix off-by-one error in node::DLOpen() Fixes loading of modules that use the NODE_MODULE macro on POSIX platforms.
13 years ago
base = pos + 1;
Make node::DLOpen use uv_dlopen
13 years ago
}
#else // Windows
for (;;) {
pos = strpbrk(base, "\\/:");
if (pos == NULL) {
break;
Add support for the module structure to process.dlopen.
15 years ago
}
Make node::DLOpen use uv_dlopen
13 years ago
base = pos + 1;
}
#endif
/* Strip the .node extension. */
pos = strrchr(base, '.');
if (pos != NULL) {
*pos = '\0';
}
Add support for the module structure to process.dlopen.
15 years ago
Make node::DLOpen use uv_dlopen
13 years ago
/* Add the `_module` suffix to the extension name. */
r = snprintf(symbol, sizeof symbol, "%s_module", base);
node: fix signedness compiler warnings
13 years ago
if (r <= 0 || static_cast<size_t>(r) >= sizeof symbol) {
Make error reporting from node::DLOpen more consistent
13 years ago
Local<Value> exception =
Exception::Error(String::New("Out of memory."));
return ThrowException(exception);
Add support for the module structure to process.dlopen.
15 years ago
}
node: s/-/_/ in add-on symbol name Replace dashes with underscores. When loading foo-bar.node, look for foo_bar_module, not foo-bar_module. The latter is not a legal symbol name.
12 years ago
/* Replace dashes with underscores. When loading foo-bar.node,
* look for foo_bar_module, not foo-bar_module.
*/
for (pos = symbol; *pos != '\0'; ++pos) {
if (*pos == '-') *pos = '_';
}
Make node::DLOpen use uv_dlopen
13 years ago
node_module_struct *mod;
deps: upgrade libuv to a478847 The event loop&#39;s reference counting scheme in this version of libuv has changed. Update the libuv bindings to reflect that fact.
13 years ago
if (uv_dlsym(&lib, symbol, reinterpret_cast<void**>(&mod))) {
node: don&#39;t scan add-on for &#34;init&#34; symbol From this commit onwards, use of the NODE_MODULE macro is mandatory. This lets node guard against modules that are ABI incompatible.
13 years ago
char errmsg[1024];
snprintf(errmsg, sizeof(errmsg), "Symbol %s not found.", symbol);
return ThrowError(errmsg);
Add support for the module structure to process.dlopen.
15 years ago
}
if (mod->version != NODE_MODULE_VERSION) {
node: don&#39;t scan add-on for &#34;init&#34; symbol From this commit onwards, use of the NODE_MODULE macro is mandatory. This lets node guard against modules that are ABI incompatible.
13 years ago
char errmsg[1024];
snprintf(errmsg,
sizeof(errmsg),
"Module version mismatch. Expected %d, got %d.",
NODE_MODULE_VERSION, mod->version);
return ThrowError(errmsg);
First attempt at node.dlopen Compiled first working &#39;hello world&#39; module with this config
16 years ago
}
Add comments to node.cc
15 years ago
// Execute the C++ module
addon: Pass module object to NODE_MODULE init function mainly to allow native addons to export single functions on rather than being restricted to operating on an existing object. Init functions now receive exports as the first argument, like before, but also the module object as the second argument, if they support it. Related to #4634 cc: @rvagg
12 years ago
mod->register_func(exports, module);
First attempt at node.dlopen Compiled first working &#39;hello world&#39; module with this config
16 years ago
Coverity fixes: src/node.cc
15 years ago
// Tell coverity that 'handle' should not be freed when we return.
// coverity[leaked_storage]
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
First attempt at node.dlopen Compiled first working &#39;hello world&#39; module with this config
16 years ago
}
Initial implementation of process.evalcx
15 years ago
cpplint.py node.cc and node.h
16 years ago
static void OnFatalError(const char* location, const char* message) {
Simplify exception reporting
15 years ago
if (location) {
fprintf(stderr, "FATAL ERROR: %s %s\n", location, message);
} else {
fprintf(stderr, "FATAL ERROR: %s\n", message);
}
src: OnFatalError handler must abort() We are in an unrecoverable state if v8 throws a FatalError, actually ask the operating system to dump core in this case. Fixes #6836
11 years ago
abort();
exit program on top level exceptions (also lots of reorg.)
16 years ago
}
cpplint.py node.cc and node.h
16 years ago
void FatalException(TryCatch &try_catch) {
Implement process &#34;uncaughtException&#34; event This event can be used to overwrite the default exception mechanism which reports the exception and kills the node process. See google group post: http://groups.google.com/group/nodejs/browse_thread/thread/9721dc3a2638446f
15 years ago
HandleScope scope;
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
if (fatal_exception_symbol.IsEmpty())
fatal_exception_symbol = NODE_PSYMBOL("_fatalException");
Implement process &#34;uncaughtException&#34; event This event can be used to overwrite the default exception mechanism which reports the exception and kills the node process. See google group post: http://groups.google.com/group/nodejs/browse_thread/thread/9721dc3a2638446f
15 years ago
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
Local<Value> fatal_v = process->Get(fatal_exception_symbol);
Implement process &#34;uncaughtException&#34; event This event can be used to overwrite the default exception mechanism which reports the exception and kills the node process. See google group post: http://groups.google.com/group/nodejs/browse_thread/thread/9721dc3a2638446f
15 years ago
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
if (!fatal_v->IsFunction()) {
// failed before the process._fatalException function was added!
// this is probably pretty bad. Nothing to do but report and exit.
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
ReportException(try_catch, true);
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
exit(6);
Implement process &#34;uncaughtException&#34; event This event can be used to overwrite the default exception mechanism which reports the exception and kills the node process. See google group post: http://groups.google.com/group/nodejs/browse_thread/thread/9721dc3a2638446f
15 years ago
}
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
Local<Function> fatal_f = Local<Function>::Cast(fatal_v);
Implement process &#34;uncaughtException&#34; event This event can be used to overwrite the default exception mechanism which reports the exception and kills the node process. See google group post: http://groups.google.com/group/nodejs/browse_thread/thread/9721dc3a2638446f
15 years ago
Local<Value> error = try_catch.Exception();
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
Local<Value> argv[] = { error };
Implement process &#34;uncaughtException&#34; event This event can be used to overwrite the default exception mechanism which reports the exception and kills the node process. See google group post: http://groups.google.com/group/nodejs/browse_thread/thread/9721dc3a2638446f
15 years ago
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
TryCatch fatal_try_catch;
Merge remote-tracking branch &#39;ry/v0.6&#39; into v0.6-merge Conflicts: Makefile lib/zlib.js src/node.cc src/node.js
13 years ago
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
// this will return true if the JS layer handled it, false otherwise
Local<Value> caught = fatal_f->Call(process, ARRAY_SIZE(argv), argv);
if (fatal_try_catch.HasCaught()) {
// the fatal exception function threw, so we must exit
ReportException(fatal_try_catch, true);
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
exit(7);
domain: Do not use uncaughtException handler This adds a process._fatalException method which is called into from C++ in order to either emit the &#39;uncaughtException&#39; method, or emit &#39;error&#39; on the active domain. The &#39;uncaughtException&#39; event is an implementation detail that it would be nice to deprecate one day, so exposing it as part of the domain machinery is not ideal. Fix #4375
12 years ago
}
if (false == caught->BooleanValue()) {
ReportException(try_catch, true);
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
exit(8);
Report errors thrown from uncaughtException handers
13 years ago
}
better error messages
16 years ago
}
enable support for the thread pool
16 years ago
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
Persistent<Object> binding_cache;
Persistent<Array> module_load_list;
Load c++ modules on demand
15 years ago
static Handle<Value> Binding(const Arguments& args) {
HandleScope scope;
Local<String> module = args[0]->ToString();
String::Utf8Value module_v(module);
Register builtin extensions via a macro, rather than a manual strcmp Set the stage for making the builtin modules more dynamic. Note: this only converts crypto and net, I will add more extensions in a later commit. * node.h: Add utility macro for converting macro values to strings. * node.h: Include the actual module name inside the module structure, not just the file it was built from. * node.h: New Macro, NODE_MODULE_DECL, for declaring an external reference to a module structure. * node_extensions.cc: New File, implements get_builtin_module, which iterates over the module structures that are compiled into node. * node.cc(node::Binding): Use the new module lookup function to find modules. * node_{net,crypto}.c: Add NODE_MODULEs to generate the module structure.
15 years ago
node_module_struct* modp;
Load c++ modules on demand
15 years ago
if (binding_cache.IsEmpty()) {
binding_cache = Persistent<Object>::New(Object::New());
}
Local<Object> exports;
Register builtin extensions via a macro, rather than a manual strcmp Set the stage for making the builtin modules more dynamic. Note: this only converts crypto and net, I will add more extensions in a later commit. * node.h: Add utility macro for converting macro values to strings. * node.h: Include the actual module name inside the module structure, not just the file it was built from. * node.h: New Macro, NODE_MODULE_DECL, for declaring an external reference to a module structure. * node_extensions.cc: New File, implements get_builtin_module, which iterates over the module structures that are compiled into node. * node.cc(node::Binding): Use the new module lookup function to find modules. * node_{net,crypto}.c: Add NODE_MODULEs to generate the module structure.
15 years ago
if (binding_cache->Has(module)) {
exports = binding_cache->Get(module)->ToObject();
Add process.moduleLoadList for better startup transparency
14 years ago
return scope.Close(exports);
}
// Append a string to process.moduleLoadList
char buf[1024];
snprintf(buf, 1024, "Binding %s", *module_v);
uint32_t l = module_load_list->Length();
module_load_list->Set(l, String::New(buf));
Move constants out of process object
15 years ago
Add process.moduleLoadList for better startup transparency
14 years ago
if ((modp = get_builtin_module(*module_v)) != NULL) {
Register builtin extensions via a macro, rather than a manual strcmp Set the stage for making the builtin modules more dynamic. Note: this only converts crypto and net, I will add more extensions in a later commit. * node.h: Add utility macro for converting macro values to strings. * node.h: Include the actual module name inside the module structure, not just the file it was built from. * node.h: New Macro, NODE_MODULE_DECL, for declaring an external reference to a module structure. * node_extensions.cc: New File, implements get_builtin_module, which iterates over the module structures that are compiled into node. * node.cc(node::Binding): Use the new module lookup function to find modules. * node_{net,crypto}.c: Add NODE_MODULEs to generate the module structure.
15 years ago
exports = Object::New();
addon: Pass module object to NODE_MODULE init function mainly to allow native addons to export single functions on rather than being restricted to operating on an existing object. Init functions now receive exports as the first argument, like before, but also the module object as the second argument, if they support it. Related to #4634 cc: @rvagg
12 years ago
// Internal bindings don't have a "module" object,
// only exports.
modp->register_func(exports, Undefined());
Register builtin extensions via a macro, rather than a manual strcmp Set the stage for making the builtin modules more dynamic. Note: this only converts crypto and net, I will add more extensions in a later commit. * node.h: Add utility macro for converting macro values to strings. * node.h: Include the actual module name inside the module structure, not just the file it was built from. * node.h: New Macro, NODE_MODULE_DECL, for declaring an external reference to a module structure. * node_extensions.cc: New File, implements get_builtin_module, which iterates over the module structures that are compiled into node. * node.cc(node::Binding): Use the new module lookup function to find modules. * node_{net,crypto}.c: Add NODE_MODULEs to generate the module structure.
15 years ago
binding_cache->Set(module, exports);
Move constants out of process object
15 years ago
} else if (!strcmp(*module_v, "constants")) {
exports = Object::New();
DefineConstants(exports);
binding_cache->Set(module, exports);
Move native js files into binding object
15 years ago
} else if (!strcmp(*module_v, "natives")) {
Use the top level check for the bindings cache (rather than each potential module having its own if check)
15 years ago
exports = Object::New();
Dynamically load native scripts
14 years ago
DefineJavaScript(exports);
Use the top level check for the bindings cache (rather than each potential module having its own if check)
15 years ago
binding_cache->Set(module, exports);
Dynamically load native scripts
14 years ago
Load c++ modules on demand
15 years ago
} else {
Move constants out of process object
15 years ago
Load c++ modules on demand
15 years ago
return ThrowException(Exception::Error(String::New("No such module")));
}
return scope.Close(exports);
}
Enable debugging. Use the --debug command line flag to enable. It appears that d8 sucks. Luckily it can be rewritten rather easily with the repl and tcp client libraries. Node&#39;s CL option parsing is getting rather unwieldy - needs refactor.
15 years ago
Stub out process.title
15 years ago
static Handle<Value> ProcessTitleGetter(Local<String> property,
const AccessorInfo& info) {
HandleScope scope;
Remove platform files, and use uv platform api
13 years ago
char buffer[512];
uv_get_process_title(buffer, sizeof(buffer));
return scope.Close(String::New(buffer));
Stub out process.title
15 years ago
}
static void ProcessTitleSetter(Local<String> property,
Local<Value> value,
const AccessorInfo& info) {
HandleScope scope;
Avoiding unnecessary ToString() calls String::Utf8Value and String::AsciiValue constructors take Handle&lt;Value&gt; So no need to convert to Handle&lt;String&gt;
13 years ago
String::Utf8Value title(value);
Remove platform files, and use uv platform api
13 years ago
// TODO: protect with a lock
uv_set_process_title(*title);
Stub out process.title
15 years ago
}
Make writes to process.env update the real environment. Tests included.
14 years ago
static Handle<Value> EnvGetter(Local<String> property,
const AccessorInfo& info) {
Windows: use unicode environment
13 years ago
HandleScope scope;
#ifdef __POSIX__
Make writes to process.env update the real environment. Tests included.
14 years ago
String::Utf8Value key(property);
const char* val = getenv(*key);
if (val) {
return scope.Close(String::New(val));
}
Windows: use unicode environment
13 years ago
#else // _WIN32
String::Value key(property);
WCHAR buffer[32767]; // The maximum size allowed for environment variables.
DWORD result = GetEnvironmentVariableW(reinterpret_cast<WCHAR*>(*key),
buffer,
ARRAY_SIZE(buffer));
// If result >= sizeof buffer the buffer was too small. That should never
// happen. If result == 0 and result != ERROR_SUCCESS the variable was not
// not found.
if ((result > 0 || GetLastError() == ERROR_SUCCESS) &&
result < ARRAY_SIZE(buffer)) {
return scope.Close(String::New(reinterpret_cast<uint16_t*>(buffer), result));
}
#endif
Re-apply commit e307468. The V8 assert got triggered by a missing HandleScope::Close().
13 years ago
// Not found. Fetch from prototype.
return scope.Close(info.Data().As<Object>()->Get(property));
Make writes to process.env update the real environment. Tests included.
14 years ago
}
static Handle<Value> EnvSetter(Local<String> property,
Local<Value> value,
const AccessorInfo& info) {
win: fix test-process-env Remove support for setting process.env.TZ as it doesn&#39;t seem we can do it x-platform without fixing V8.
14 years ago
HandleScope scope;
Windows: use unicode environment
13 years ago
#ifdef __POSIX__
Make writes to process.env update the real environment. Tests included.
14 years ago
String::Utf8Value key(property);
String::Utf8Value val(value);
setenv(*key, *val, 1);
Windows: use unicode environment
13 years ago
#else // _WIN32
String::Value key(property);
String::Value val(value);
WCHAR* key_ptr = reinterpret_cast<WCHAR*>(*key);
// Environment variables that start with '=' are read-only.
if (key_ptr[0] != L'=') {
SetEnvironmentVariableW(key_ptr, reinterpret_cast<WCHAR*>(*val));
win: fix test-process-env Remove support for setting process.env.TZ as it doesn&#39;t seem we can do it x-platform without fixing V8.
14 years ago
}
Disable node.cc functions that are not supported on windows
14 years ago
#endif
Windows: use unicode environment
13 years ago
// Whether it worked or not, always return rval.
return scope.Close(value);
Make writes to process.env update the real environment. Tests included.
14 years ago
}
static Handle<Integer> EnvQuery(Local<String> property,
const AccessorInfo& info) {
Windows: use unicode environment
13 years ago
HandleScope scope;
#ifdef __POSIX__
Make writes to process.env update the real environment. Tests included.
14 years ago
String::Utf8Value key(property);
if (getenv(*key)) {
src, test: downgrade to v8 3.14 api
12 years ago
return scope.Close(Integer::New(0));
Make writes to process.env update the real environment. Tests included.
14 years ago
}
Windows: use unicode environment
13 years ago
#else // _WIN32
String::Value key(property);
WCHAR* key_ptr = reinterpret_cast<WCHAR*>(*key);
if (GetEnvironmentVariableW(key_ptr, NULL, 0) > 0 ||
GetLastError() == ERROR_SUCCESS) {
if (key_ptr[0] == L'=') {
// Environment variables that start with '=' are hidden and read-only.
return scope.Close(Integer::New(v8::ReadOnly ||
v8::DontDelete ||
src, test: downgrade to v8 3.14 api
12 years ago
v8::DontEnum));
Windows: use unicode environment
13 years ago
} else {
src, test: downgrade to v8 3.14 api
12 years ago
return scope.Close(Integer::New(0));
Windows: use unicode environment
13 years ago
}
}
#endif
// Not found
return scope.Close(Handle<Integer>());
Make writes to process.env update the real environment. Tests included.
14 years ago
}
static Handle<Boolean> EnvDeleter(Local<String> property,
const AccessorInfo& info) {
win: fix test-process-env Remove support for setting process.env.TZ as it doesn&#39;t seem we can do it x-platform without fixing V8.
14 years ago
HandleScope scope;
Disable node.cc functions that are not supported on windows
14 years ago
#ifdef __POSIX__
Windows: use unicode environment
13 years ago
String::Utf8Value key(property);
src, test: downgrade to v8 3.14 api
12 years ago
if (!getenv(*key)) return False();
node: don&#39;t check return value of unsetenv() It returns void on some platforms, notably FreeBSD.
13 years ago
unsetenv(*key); // can't check return value, it's void on some platforms
src, test: downgrade to v8 3.14 api
12 years ago
return True();
Disable node.cc functions that are not supported on windows
14 years ago
#else
Windows: use unicode environment
13 years ago
String::Value key(property);
WCHAR* key_ptr = reinterpret_cast<WCHAR*>(*key);
if (key_ptr[0] == L'=' || !SetEnvironmentVariableW(key_ptr, NULL)) {
// Deletion failed. Return true if the key wasn't there in the first place,
// false if it is still there.
bool rv = GetEnvironmentVariableW(key_ptr, NULL, NULL) == 0 &&
GetLastError() != ERROR_SUCCESS;
return scope.Close(Boolean::New(rv));
Make writes to process.env update the real environment. Tests included.
14 years ago
}
src, test: downgrade to v8 3.14 api
12 years ago
return True();
Windows: use unicode environment
13 years ago
#endif
Make writes to process.env update the real environment. Tests included.
14 years ago
}
static Handle<Array> EnvEnumerator(const AccessorInfo& info) {
HandleScope scope;
Windows: use unicode environment
13 years ago
#ifdef __POSIX__
Make writes to process.env update the real environment. Tests included.
14 years ago
int size = 0;
while (environ[size]) size++;
Local<Array> env = Array::New(size);
for (int i = 0; i < size; ++i) {
const char* var = environ[i];
const char* s = strchr(var, '=');
const int length = s ? s - var : strlen(var);
env->Set(i, String::New(var, length));
}
Windows: use unicode environment
13 years ago
#else // _WIN32
WCHAR* environment = GetEnvironmentStringsW();
if (environment == NULL) {
// This should not happen.
return scope.Close(Handle<Array>());
}
Local<Array> env = Array::New();
WCHAR* p = environment;
int i = 0;
while (*p != NULL) {
WCHAR *s;
if (*p == L'=') {
// If the key starts with '=' it is a hidden environment variable.
p += wcslen(p) + 1;
continue;
} else {
s = wcschr(p, L'=');
}
if (!s) {
s = p + wcslen(p);
}
env->Set(i++, String::New(reinterpret_cast<uint16_t*>(p), s - p));
p = s + wcslen(s) + 1;
}
FreeEnvironmentStringsW(environment);
#endif
Make writes to process.env update the real environment. Tests included.
14 years ago
return scope.Close(env);
}
process: add process.features, remove process.useUV Partially fixes #1385.
14 years ago
static Handle<Object> GetFeatures() {
HandleScope scope;
Local<Object> obj = Object::New();
Add process.features.debug; fixes simple/test-executable-path.js
14 years ago
obj->Set(String::NewSymbol("debug"),
#if defined(DEBUG) && DEBUG
src, test: downgrade to v8 3.14 api
12 years ago
True()
Add process.features.debug; fixes simple/test-executable-path.js
14 years ago
#else
src, test: downgrade to v8 3.14 api
12 years ago
False()
Add process.features.debug; fixes simple/test-executable-path.js
14 years ago
#endif
);
src, test: downgrade to v8 3.14 api
12 years ago
obj->Set(String::NewSymbol("uv"), True());
obj->Set(String::NewSymbol("ipv6"), True()); // TODO ping libuv
Add support for TLS SNI Fixes #1411
14 years ago
obj->Set(String::NewSymbol("tls_npn"), Boolean::New(use_npn));
obj->Set(String::NewSymbol("tls_sni"), Boolean::New(use_sni));
process: add process.features, remove process.useUV Partially fixes #1385.
14 years ago
obj->Set(String::NewSymbol("tls"),
Boolean::New(get_builtin_module("crypto") != NULL));
return scope.Close(obj);
}
debugger: export `debug_port` to `process` `process.debug_port` is useful for changing debugger port in runtime, before starting it (via SIGUSR1). Using `--port=` argument for debugger repl, tests will run debugger server on a `common.PORT` (as it usually does for any other servers). `process._debugEnd()` stops debugger and its server. * debugger: implemented process._debugEnd(), `node debug --port=5858 app.js` * test: start debugger repl on common.PORT * fixes #2613 * fixes #2614
13 years ago
static Handle<Value> DebugPortGetter(Local<String> property,
const AccessorInfo& info) {
HandleScope scope;
src, test: downgrade to v8 3.14 api
12 years ago
return scope.Close(Integer::NewFromUnsigned(debug_port));
debugger: export `debug_port` to `process` `process.debug_port` is useful for changing debugger port in runtime, before starting it (via SIGUSR1). Using `--port=` argument for debugger repl, tests will run debugger server on a `common.PORT` (as it usually does for any other servers). `process._debugEnd()` stops debugger and its server. * debugger: implemented process._debugEnd(), `node debug --port=5858 app.js` * test: start debugger repl on common.PORT * fixes #2613 * fixes #2614
13 years ago
}
static void DebugPortSetter(Local<String> property,
Local<Value> value,
const AccessorInfo& info) {
HandleScope scope;
debug_port = value->NumberValue();
}
Implement process._debugProcess
13 years ago
static Handle<Value> DebugProcess(const Arguments& args);
pause command
13 years ago
static Handle<Value> DebugPause(const Arguments& args);
debugger: export `debug_port` to `process` `process.debug_port` is useful for changing debugger port in runtime, before starting it (via SIGUSR1). Using `--port=` argument for debugger repl, tests will run debugger server on a `common.PORT` (as it usually does for any other servers). `process._debugEnd()` stops debugger and its server. * debugger: implemented process._debugEnd(), `node debug --port=5858 app.js` * test: start debugger repl on common.PORT * fixes #2613 * fixes #2614
13 years ago
static Handle<Value> DebugEnd(const Arguments& args);
Implement process._debugProcess
13 years ago
timer: move setImmediate from timer to uv_check uv_check is the robust place to invoke setImmediate callbacks after process.nextTick and before timers(setTimeout/setInterval)
12 years ago
Handle<Value> NeedImmediateCallbackGetter(Local<String> property,
const AccessorInfo& info) {
return Boolean::New(need_immediate_cb);
}
static void NeedImmediateCallbackSetter(Local<String> property,
Local<Value> value,
const AccessorInfo& info) {
HandleScope scope;
bool bool_value = value->BooleanValue();
if (need_immediate_cb == bool_value) return;
need_immediate_cb = bool_value;
if (need_immediate_cb) {
uv_check_start(&check_immediate_watcher, node::CheckImmediate);
// idle handle is needed only to maintain event loop
uv_idle_start(&idle_immediate_dummy, node::IdleImmediateDummy);
} else {
uv_check_stop(&check_immediate_watcher);
uv_idle_stop(&idle_immediate_dummy);
}
}
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
Handle<Object> SetupProcessObject(int argc, char *argv[]) {
Clean ups
16 years ago
HandleScope scope;
Add profile window around socket on_read.
16 years ago
Add openssl version to process.versions
14 years ago
int i, j;
Namespace EVERYTHING under process; introduce GLOBAL http://groups.google.com/group/nodejs/browse_thread/thread/1034fd2ad2cd93e8
15 years ago
Local<FunctionTemplate> process_template = FunctionTemplate::New();
node: change the constructor name of process from EventEmitter to process
13 years ago
process_template->SetClassName(String::NewSymbol("process"));
better error messages
16 years ago
node: change the constructor name of process from EventEmitter to process
13 years ago
process = Persistent<Object>::New(process_template->GetFunction()->NewInstance());
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
Stub out process.title
15 years ago
process->SetAccessor(String::New("title"),
ProcessTitleGetter,
ProcessTitleSetter);
Add process.platform
15 years ago
// process.version
Namespace EVERYTHING under process; introduce GLOBAL http://groups.google.com/group/nodejs/browse_thread/thread/1034fd2ad2cd93e8
15 years ago
process->Set(String::NewSymbol("version"), String::New(NODE_VERSION));
Add process.versions
15 years ago
Add process.moduleLoadList for better startup transparency
14 years ago
// process.moduleLoadList
module_load_list = Persistent<Array>::New(Array::New());
process->Set(String::NewSymbol("moduleLoadList"), module_load_list);
process: use NODE_STRINGIFY macro to set process.versions.uv
13 years ago
// process.versions
Add process.versions
15 years ago
Local<Object> versions = Object::New();
process->Set(String::NewSymbol("versions"), versions);
process: expose the http_parser version in process.versions
13 years ago
versions->Set(String::NewSymbol("http_parser"), String::New(
NODE_STRINGIFY(HTTP_PARSER_VERSION_MAJOR) "."
NODE_STRINGIFY(HTTP_PARSER_VERSION_MINOR)));
Add process.versions
15 years ago
// +1 to get rid of the leading 'v'
versions->Set(String::NewSymbol("node"), String::New(NODE_VERSION+1));
versions->Set(String::NewSymbol("v8"), String::New(V8::GetVersion()));
versions->Set(String::NewSymbol("ares"), String::New(ARES_VERSION_STR));
src: tie process.versions.uv to uv_version_string()
12 years ago
versions->Set(String::NewSymbol("uv"), String::New(uv_version_string()));
process: expose the zlib version in use in process.versions
13 years ago
versions->Set(String::NewSymbol("zlib"), String::New(ZLIB_VERSION));
process: expose NODE_MODULE_VERSION in process.versions
12 years ago
versions->Set(String::NewSymbol("modules"),
String::New(NODE_STRINGIFY(NODE_MODULE_VERSION)));
Basic VC++ compatibility work.
14 years ago
#if HAVE_OPENSSL
Add openssl version to process.versions
14 years ago
// Stupid code to slice out the version string.
int c, l = strlen(OPENSSL_VERSION_TEXT);
node: fix use of potentially uninitialized variable
13 years ago
for (i = j = 0; i < l; i++) {
Add openssl version to process.versions
14 years ago
c = OPENSSL_VERSION_TEXT[i];
if ('0' <= c && c <= '9') {
for (j = i + 1; j < l; j++) {
c = OPENSSL_VERSION_TEXT[j];
if (c == ' ') break;
}
break;
}
}
versions->Set(String::NewSymbol("openssl"),
String::New(OPENSSL_VERSION_TEXT + i, j - i));
#endif
Add process.versions
15 years ago
Implement `process.arch` to get a String of the current processor architecture, with docs.
14 years ago
// process.arch
process->Set(String::NewSymbol("arch"), String::New(ARCH));
Add process.platform
15 years ago
// process.platform
Fix process.platform
14 years ago
process->Set(String::NewSymbol("platform"), String::New(PLATFORM));
Add process.platform
15 years ago
Downcase process.ARGV/ENV to process.argv/env
15 years ago
// process.argv
Remove &#39;--&#39; from command line parsing
15 years ago
Local<Array> arguments = Array::New(argc - option_end_index + 1);
src, test: downgrade to v8 3.14 api
12 years ago
arguments->Set(Integer::New(0), String::New(argv[0]));
Fix options parsing The current node has a bug where it will fail to pass the option argument immediately preceding the first non-option argument to V8. That is the --perf flag will be ignored by V8 when running &#39;node --perf script.js&#39;.
15 years ago
for (j = 1, i = option_end_index; i < argc; j++, i++) {
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
Local<String> arg = String::New(argv[i]);
src, test: downgrade to v8 3.14 api
12 years ago
arguments->Set(Integer::New(j), arg);
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
}
Add comments to node.cc
15 years ago
// assign it
Downcase process.ARGV/ENV to process.argv/env
15 years ago
process->Set(String::NewSymbol("argv"), arguments);
everything is changed. i&#39;ve waited much too long to commit. this is awful. i&#39;m sorry for being so messy.
16 years ago
Expose original argv as process.execArgv for cluster and child_process.fork()
13 years ago
// process.execArgv
Local<Array> execArgv = Array::New(option_end_index - 1);
for (j = 1, i = 0; j < option_end_index; j++, i++) {
src, test: downgrade to v8 3.14 api
12 years ago
execArgv->Set(Integer::New(i), String::New(argv[j]));
Expose original argv as process.execArgv for cluster and child_process.fork()
13 years ago
}
// assign it
process->Set(String::NewSymbol("execArgv"), execArgv);
Downcase process.ARGV/ENV to process.argv/env
15 years ago
// create process.env
Make writes to process.env update the real environment. Tests included.
14 years ago
Local<ObjectTemplate> envTemplate = ObjectTemplate::New();
print warning on process.ENV
14 years ago
envTemplate->SetNamedPropertyHandler(EnvGetter,
EnvSetter,
EnvQuery,
EnvDeleter,
EnvEnumerator,
Re-apply commit e307468. The V8 assert got triggered by a missing HandleScope::Close().
13 years ago
Object::New());
Make writes to process.env update the real environment. Tests included.
14 years ago
Local<Object> env = envTemplate->NewInstance();
Downcase process.ARGV/ENV to process.argv/env
15 years ago
process->Set(String::NewSymbol("env"), env);
src, test: downgrade to v8 3.14 api
12 years ago
process->Set(String::NewSymbol("pid"), Integer::New(getpid()));
process: add process.features, remove process.useUV Partially fixes #1385.
14 years ago
process->Set(String::NewSymbol("features"), GetFeatures());
timer: move setImmediate from timer to uv_check uv_check is the robust place to invoke setImmediate callbacks after process.nextTick and before timers(setTimeout/setInterval)
12 years ago
process->SetAccessor(String::New("_needImmediateCallback"),
NeedImmediateCallbackGetter,
NeedImmediateCallbackSetter);
Add access to user environment via ENV
16 years ago
Added -e, --eval
14 years ago
// -e, --eval
if (eval_string) {
process->Set(String::NewSymbol("_eval"), String::New(eval_string));
process: set _print_eval even when --eval is not passed This is for scripts being fed from stdin: $ echo &#34;{ foo: &#39;bar&#39; }&#34; | node -p
13 years ago
}
// -p, --print
if (print_eval) {
src, test: downgrade to v8 3.14 api
12 years ago
process->Set(String::NewSymbol("_print_eval"), True());
Added -e, --eval
14 years ago
}
process: comment for consistency
13 years ago
// -i, --interactive
cmd: add a -i / --interactive flag to force the REPL This is the only way to spawn a node child process in REPL mode, and will also be needed to be able to use the REPL in MinTTY.
13 years ago
if (force_repl) {
src, test: downgrade to v8 3.14 api
12 years ago
process->Set(String::NewSymbol("_forceRepl"), True());
cmd: add a -i / --interactive flag to force the REPL This is the only way to spawn a node child process in REPL mode, and will also be needed to be able to use the REPL in MinTTY.
13 years ago
}
Add --no-deprecation and --trace-deprecation flags
13 years ago
// --no-deprecation
if (no_deprecation) {
src, test: downgrade to v8 3.14 api
12 years ago
process->Set(String::NewSymbol("noDeprecation"), True());
Add --no-deprecation and --trace-deprecation flags
13 years ago
}
node: Add --throw-deprecation Extremely handy when tracking down a flood of recursive nextTick warnings.
12 years ago
// --throw-deprecation
if (throw_deprecation) {
process->Set(String::NewSymbol("throwDeprecation"), True());
}
Add --no-deprecation and --trace-deprecation flags
13 years ago
// --trace-deprecation
if (trace_deprecation) {
src, test: downgrade to v8 3.14 api
12 years ago
process->Set(String::NewSymbol("traceDeprecation"), True());
Add --no-deprecation and --trace-deprecation flags
13 years ago
}
Initial implementation of process.execPath Darwin, Linux, Solaris. FreeBSD still needs testing. TODO: Amend the tests where we use spawn with argv[0].
15 years ago
size_t size = 2*PATH_MAX;
Basic VC++ compatibility work.
14 years ago
char* execPath = new char[size];
Use uv_exepath
14 years ago
if (uv_exepath(execPath, &size) != 0) {
Initial implementation of process.execPath Darwin, Linux, Solaris. FreeBSD still needs testing. TODO: Amend the tests where we use spawn with argv[0].
15 years ago
// as a last ditch effort, fallback on argv[0] ?
process->Set(String::NewSymbol("execPath"), String::New(argv[0]));
} else {
Fix OS::GetExecutablePath for platform_none
14 years ago
process->Set(String::NewSymbol("execPath"), String::New(execPath, size));
Initial implementation of process.execPath Darwin, Linux, Solaris. FreeBSD still needs testing. TODO: Amend the tests where we use spawn with argv[0].
15 years ago
}
Basic VC++ compatibility work.
14 years ago
delete [] execPath;
Initial implementation of process.execPath Darwin, Linux, Solaris. FreeBSD still needs testing. TODO: Amend the tests where we use spawn with argv[0].
15 years ago
debugger: rename `process.debug_port` to `process.debugPort` We should comply to to camelCase naming convention.
13 years ago
process->SetAccessor(String::New("debugPort"),
debugger: export `debug_port` to `process` `process.debug_port` is useful for changing debugger port in runtime, before starting it (via SIGUSR1). Using `--port=` argument for debugger repl, tests will run debugger server on a `common.PORT` (as it usually does for any other servers). `process._debugEnd()` stops debugger and its server. * debugger: implemented process._debugEnd(), `node debug --port=5858 app.js` * test: start debugger repl on common.PORT * fixes #2613 * fixes #2614
13 years ago
DebugPortGetter,
DebugPortSetter);
Initial implementation of process.execPath Darwin, Linux, Solaris. FreeBSD still needs testing. TODO: Amend the tests where we use spawn with argv[0].
15 years ago
Add comments to node.cc
15 years ago
// define various internal methods
process: add _getActiveHandles(), _getActiveRequests() * process._getActiveHandles() returns a list containing all active handles (timers, sockets, etc.) that have not been unref&#39;d. * process._getActiveRequests() returns a list of active requests (in-flight actions like connecting to a remote host, writing data to a socket, etc.).
13 years ago
NODE_SET_METHOD(process, "_getActiveRequests", GetActiveRequests);
NODE_SET_METHOD(process, "_getActiveHandles", GetActiveHandles);
Change nextTick implementation for the better Use a prepare and idle watcher to execute the nextTick callback more quickly. Test provided by Matt Ranney.
15 years ago
NODE_SET_METHOD(process, "_needTickCallback", NeedTickCallback);
Namespace EVERYTHING under process; introduce GLOBAL http://groups.google.com/group/nodejs/browse_thread/thread/1034fd2ad2cd93e8
15 years ago
NODE_SET_METHOD(process, "reallyExit", Exit);
Add process.abort
13 years ago
NODE_SET_METHOD(process, "abort", Abort);
Added process.chdir()
15 years ago
NODE_SET_METHOD(process, "chdir", Chdir);
Namespace EVERYTHING under process; introduce GLOBAL http://groups.google.com/group/nodejs/browse_thread/thread/1034fd2ad2cd93e8
15 years ago
NODE_SET_METHOD(process, "cwd", Cwd);
Disable node.cc functions that are not supported on windows
14 years ago
windows: fix test-umask
14 years ago
NODE_SET_METHOD(process, "umask", Umask);
Disable node.cc functions that are not supported on windows
14 years ago
#ifdef __POSIX__
Add process.setuid(), getuid()
15 years ago
NODE_SET_METHOD(process, "getuid", GetUid);
NODE_SET_METHOD(process, "setuid", SetUid);
Add setgid,getgid
15 years ago
NODE_SET_METHOD(process, "setgid", SetGid);
NODE_SET_METHOD(process, "getgid", GetGid);
process: add getgroups(), setgroups(), initgroups() DRY the getuid(), getgid(), etc. functions while we&#39;re at it.
12 years ago
NODE_SET_METHOD(process, "getgroups", GetGroups);
NODE_SET_METHOD(process, "setgroups", SetGroups);
NODE_SET_METHOD(process, "initgroups", InitGroups);
process.kill to use uv_kill
13 years ago
#endif // __POSIX__
Add setgid,getgid
15 years ago
Move constants out of process object
15 years ago
NODE_SET_METHOD(process, "_kill", Kill);
Disable node.cc functions that are not supported on windows
14 years ago
Implement process._debugProcess
13 years ago
NODE_SET_METHOD(process, "_debugProcess", DebugProcess);
pause command
13 years ago
NODE_SET_METHOD(process, "_debugPause", DebugPause);
debugger: export `debug_port` to `process` `process.debug_port` is useful for changing debugger port in runtime, before starting it (via SIGUSR1). Using `--port=` argument for debugger repl, tests will run debugger server on a `common.PORT` (as it usually does for any other servers). `process._debugEnd()` stops debugger and its server. * debugger: implemented process._debugEnd(), `node debug --port=5858 app.js` * test: start debugger repl on common.PORT * fixes #2613 * fixes #2614
13 years ago
NODE_SET_METHOD(process, "_debugEnd", DebugEnd);
Implement process._debugProcess
13 years ago
process: add `process.hrtime()` This commit adds a high-resolution timer function.
13 years ago
NODE_SET_METHOD(process, "hrtime", Hrtime);
Make node::DLOpen use uv_dlopen
13 years ago
NODE_SET_METHOD(process, "dlopen", DLOpen);
Add process.uptime().
14 years ago
NODE_SET_METHOD(process, "uptime", Uptime);
Move memoryUsage() into C on Linux
15 years ago
NODE_SET_METHOD(process, "memoryUsage", MemoryUsage);
Clean ups
16 years ago
Load c++ modules on demand
15 years ago
NODE_SET_METHOD(process, "binding", Binding);
domain: fix domain callback from MakeCallback Since _tickCallback and _tickDomainCallback were both called from MakeCallback, it was possible for a callback to be called that required a domain directly to _tickCallback. The fix was to implement process.usingDomains(). This will set all applicable functions to their domain counterparts, and set a flag in cc to let MakeCallback know domain callbacks always need to be checked. Added test in own file. It&#39;s important that the test remains isolated.
12 years ago
NODE_SET_METHOD(process, "_usingDomains", UsingDomains);
process: allow ticker to cross communicate better Using external memory values allows for quick communication between js and cc land, so we can check if the js land callback needs to be run. (this is where I meant that manually tracking nextTickQueue.length would be helpful) Also did some minor cleanup of removing the old Tick and StartTickSpinner functions, and a few unneeded comments. Conflicts: src/node.cc
12 years ago
// values use to cross communicate with processNextTick
Local<Object> info_box = Object::New();
info_box->SetIndexedPropertiesToExternalArrayData(&tick_infobox,
kExternalUnsignedIntArray,
3);
process->Set(String::NewSymbol("_tickInfoBox"), info_box);
events: _events to object and undefined not null By making sure the _events is always an object there is one less check that needs to be performed by emit. Use undefined instead of null. typeof checks are a lot faster than isArray. There are a few places where the this._events check cannot be removed because it is possible for the user to call those methods after using utils.extend to create their own EventEmitter, but before it has actually been instantiated.
12 years ago
// pre-set _events object for faster emit checks
process->Set(String::NewSymbol("_events"), Object::New());
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
return process;
}
static void AtExit() {
Fixes #1805. Reset TTY mode on exit.
13 years ago
uv_tty_reset_mode();
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
}
static void SignalExit(int signal) {
Fixes #1805. Reset TTY mode on exit.
13 years ago
uv_tty_reset_mode();
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
_exit(128 + signal);
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
}
Deshadow, denamespace variables in node.cc Prep for isolates.
13 years ago
void Load(Handle<Object> process_l) {
node: move symbol caching to Load()
12 years ago
process_symbol = NODE_PSYMBOL("process");
domain_symbol = NODE_PSYMBOL("domain");
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
// Compile, execute the src/node.js file. (Which was included as static C
// string in node_natives.h. 'natve_node' is the string containing that
// source code.)
Fix whitespace and warnings in node.cc
15 years ago
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
// The node.js file returns a function 'f'
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
atexit(AtExit);
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
TryCatch try_catch;
Add string class that uses ExternalAsciiStringResource. Change the natives to use this class instead of creating completely new strings. Reduces memory usage by about 1 MB.
14 years ago
Local<Value> f_value = ExecuteString(MainSource(),
IMMUTABLE_STRING("node.js"));
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
if (try_catch.HasCaught()) {
Improve error reporting - No more single line &#34;node.js:176:9&#34; errors - No more strange output when error happens on first line due to module wrapper function. - A few tests to check these things
15 years ago
ReportException(try_catch, true);
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
exit(10);
}
assert(f_value->IsFunction());
Local<Function> f = Local<Function>::Cast(f_value);
// Now we call 'f' with the 'process' variable that we've built up with
// all our bindings. Inside node.js we'll take care of assigning things to
// their places.
Fix whitespace and warnings in node.cc
15 years ago
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
// We start the process this way in order to be more modular. Developers
// who do not like how 'src/node.js' setups the module system but do like
// Node's I/O bindings may want to replace 'f' with their own function.
Remove process.global
15 years ago
// Add a reference to the global object
Local<Object> global = v8::Context::GetCurrent()->Global();
src, test: downgrade to v8 3.14 api
12 years ago
Local<Value> args[1] = { Local<Value>::New(process_l) };
Configurable debug ports, and --debug-brk
15 years ago
tracing: add systemtap support
12 years ago
#if defined HAVE_DTRACE || defined HAVE_ETW || defined HAVE_SYSTEMTAP
Land Cantrill&#39;s DTrace patch only works on solaris
14 years ago
InitDTrace(global);
New DTrace probes from CA team
14 years ago
#endif
Land Cantrill&#39;s DTrace patch only works on solaris
14 years ago
windows: add tracing with performance counters Patch by Henry Rawas and Scott Blomquist.
12 years ago
#if defined HAVE_PERFCTR
InitPerfCounters(global);
#endif
Fix whitespace and warnings in node.cc
15 years ago
f->Call(global, 1, args);
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
Revert &#34;Fix &#39;uncaughtException&#39; for top level exceptions&#34; This reverts commit 8f8dcf8ed63b19a6c20915e12af83a3ad792f1d2.
15 years ago
if (try_catch.HasCaught()) {
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
FatalException(try_catch);
Reorganize the start-up process - assign &#39;GLOBAL&#39; (and now &#39;global&#39;) inside src/node.js - position for eventually allowing replacements to src/node.js for people like Kris Kowal who want the nice libev and libeio bindings but not necessarily node&#39;s choices of modules or promises.
15 years ago
}
Clean ups
16 years ago
}
Major refactoring: program name now &#34;node&#34; Trying to make a more moduler design. Two libraries currently &#34;TCP&#34; and &#34;HTTP&#34; each have their own file. Other major feature added here is multiple web servers! excitement.
16 years ago
Configurable debug ports, and --debug-brk
15 years ago
static void PrintHelp();
static void ParseDebugOpt(const char* arg) {
const char *p = 0;
use_debug_agent = true;
if (!strcmp (arg, "--debug-brk")) {
debug_wait_connect = true;
return;
} else if (!strcmp(arg, "--debug")) {
return;
} else if (strstr(arg, "--debug-brk=") == arg) {
debug_wait_connect = true;
p = 1 + strchr(arg, '=');
debug_port = atoi(p);
} else if (strstr(arg, "--debug=") == arg) {
p = 1 + strchr(arg, '=');
debug_port = atoi(p);
}
if (p && debug_port > 1024 && debug_port < 65536)
return;
fprintf(stderr, "Bad debug option.\n");
if (p) fprintf(stderr, "Debug port must be in range 1025 to 65535.\n");
PrintHelp();
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
exit(12);
Configurable debug ports, and --debug-brk
15 years ago
}
cpplint.py node.cc and node.h
16 years ago
static void PrintHelp() {
Improve --help message and Fix -e/--eval switch --help should at least give some hint that node takes the -e switch Update help message to include -e and swap eval block with module load block so argv works correctly [steve@sane node](master)$ ./node -e &#39;process.argv&#39; foo bar [ &#39;/home/steve/code/node/node&#39;, &#39;foo&#39;, &#39;bar&#39; ] :) Add simple test for -e option. Fixes #1311.
14 years ago
printf("Usage: node [options] [ -e script | script.js ] [arguments] \n"
introduce node cluster
13 years ago
" node debug script.js [arguments] \n"
debugger: error out without script TODO: debug&gt; run script.js
14 years ago
"\n"
Add environmental varibles to help text
15 years ago
"Options:\n"
Add --max-stack-size flag. v8 doesn&#39;t expose a command-line flag to set the stack size, so this adds a new flag that node understands how to handle and uses v8&#39;s ResourceConstraints API.
14 years ago
" -v, --version print node's version\n"
Improve --help message and Fix -e/--eval switch --help should at least give some hint that node takes the -e switch Update help message to include -e and swap eval block with module load block so argv works correctly [steve@sane node](master)$ ./node -e &#39;process.argv&#39; foo bar [ &#39;/home/steve/code/node/node&#39;, &#39;foo&#39;, &#39;bar&#39; ] :) Add simple test for -e option. Fixes #1311.
14 years ago
" -e, --eval script evaluate script\n"
node: update description of --print
12 years ago
" -p, --print evaluate script and print result\n"
cmd: add a -i / --interactive flag to force the REPL This is the only way to spawn a node child process in REPL mode, and will also be needed to be able to use the REPL in MinTTY.
13 years ago
" -i, --interactive always enter the REPL even if stdin\n"
" does not appear to be a terminal\n"
Add --no-deprecation and --trace-deprecation flags
13 years ago
" --no-deprecation silence deprecation warnings\n"
" --trace-deprecation show stack traces on deprecations\n"
Add --max-stack-size flag. v8 doesn&#39;t expose a command-line flag to set the stack size, so this adds a new flag that node understands how to handle and uses v8&#39;s ResourceConstraints API.
14 years ago
" --v8-options print v8 command line options\n"
" --max-stack-size=val set max v8 stack size (bytes)\n"
Add environmental varibles to help text
15 years ago
"\n"
cli: fix output of --help
13 years ago
"Environment variables:\n"
#ifdef _WIN32
"NODE_PATH ';'-separated list of directories\n"
#else
Add --max-stack-size flag. v8 doesn&#39;t expose a command-line flag to set the stack size, so this adds a new flag that node understands how to handle and uses v8&#39;s ResourceConstraints API.
14 years ago
"NODE_PATH ':'-separated list of directories\n"
cli: fix output of --help
13 years ago
#endif
Finish removing require.paths Fix require() completion bug in repl, and correct man output
14 years ago
" prefixed to the module search path.\n"
Add --max-stack-size flag. v8 doesn&#39;t expose a command-line flag to set the stack size, so this adds a new flag that node understands how to handle and uses v8&#39;s ResourceConstraints API.
14 years ago
"NODE_MODULE_CONTEXTS Set to 1 to load modules in their own\n"
" global contexts.\n"
debugger: error out without script TODO: debug&gt; run script.js
14 years ago
"NODE_DISABLE_COLORS Set to 1 to disable colors in the REPL\n"
Add environmental varibles to help text
15 years ago
"\n"
debugger: error out without script TODO: debug&gt; run script.js
14 years ago
"Documentation can be found at http://nodejs.org/\n");
Add simple command line switches
16 years ago
}
Add comments to node.cc
15 years ago
// Parse node command line arguments.
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
static void ParseArgs(int argc, char **argv) {
Fix options parsing The current node has a bug where it will fail to pass the option argument immediately preceding the first non-option argument to V8. That is the --perf flag will be ignored by V8 when running &#39;node --perf script.js&#39;.
15 years ago
int i;
Enable debugging. Use the --debug command line flag to enable. It appears that d8 sucks. Luckily it can be rewritten rather easily with the repl and tcp client libraries. Node&#39;s CL option parsing is getting rather unwieldy - needs refactor.
15 years ago
// TODO use parse opts
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
for (i = 1; i < argc; i++) {
Add simple command line switches
16 years ago
const char *arg = argv[i];
Configurable debug ports, and --debug-brk
15 years ago
if (strstr(arg, "--debug") == arg) {
ParseDebugOpt(arg);
Removed deprecation errors in src/node.cc Removed constness on string litterals. This should cause no problem since we&#39;re not modifying them.
15 years ago
argv[i] = const_cast<char*>("");
Add &#39;--&#39; to seperate v8 args from program args
16 years ago
} else if (strcmp(arg, "--version") == 0 || strcmp(arg, "-v") == 0) {
Add simple command line switches
16 years ago
printf("%s\n", NODE_VERSION);
exit(0);
Add --max-stack-size flag. v8 doesn&#39;t expose a command-line flag to set the stack size, so this adds a new flag that node understands how to handle and uses v8&#39;s ResourceConstraints API.
14 years ago
} else if (strstr(arg, "--max-stack-size=") == arg) {
const char *p = 0;
p = 1 + strchr(arg, '=');
max_stack_size = atoi(p);
argv[i] = const_cast<char*>("");
Add simple command line switches
16 years ago
} else if (strcmp(arg, "--help") == 0 || strcmp(arg, "-h") == 0) {
PrintHelp();
exit(0);
cli: make -p equivalent to -pe Fixes #3938.
13 years ago
} else if (strcmp(arg, "--eval") == 0 ||
strcmp(arg, "-e") == 0 ||
strcmp(arg, "--print") == 0 ||
strcmp(arg, "-pe") == 0 ||
strcmp(arg, "-p") == 0) {
cli: make argument to -p optional
13 years ago
bool is_eval = strchr(arg, 'e') != NULL;
bool is_print = strchr(arg, 'p') != NULL;
// argument to -p and --print is optional
if (is_eval == true && i + 1 >= argc) {
cli: make -p equivalent to -pe Fixes #3938.
13 years ago
fprintf(stderr, "Error: %s requires an argument\n", arg);
core: Mark exit() calls with status codes Also, exit with 128+n for signal exit n, as is The Unix Way.
12 years ago
exit(13);
Added -e, --eval
14 years ago
}
cli: make argument to -p optional
13 years ago
print_eval = print_eval || is_print;
Added -e, --eval
14 years ago
argv[i] = const_cast<char*>("");
cli: make argument to -p optional
13 years ago
// --eval, -e and -pe always require an argument
if (is_eval == true) {
eval_string = argv[++i];
continue;
}
// next arg is the expression to evaluate unless it starts with:
// - a dash, then it's another switch
// - "\\-", then it's an escaped expression, drop the backslash
if (argv[i + 1] == NULL) continue;
if (argv[i + 1][0] == '-') continue;
Added -e, --eval
14 years ago
eval_string = argv[++i];
cli: make argument to -p optional
13 years ago
if (strncmp(eval_string, "\\-", 2) == 0) ++eval_string;
cmd: add a -i / --interactive flag to force the REPL This is the only way to spawn a node child process in REPL mode, and will also be needed to be able to use the REPL in MinTTY.
13 years ago
} else if (strcmp(arg, "--interactive") == 0 || strcmp(arg, "-i") == 0) {
force_repl = true;
argv[i] = const_cast<char*>("");
Add simple command line switches
16 years ago
} else if (strcmp(arg, "--v8-options") == 0) {
Removed deprecation errors in src/node.cc Removed constness on string litterals. This should cause no problem since we&#39;re not modifying them.
15 years ago
argv[i] = const_cast<char*>("--help");
Add --no-deprecation and --trace-deprecation flags
13 years ago
} else if (strcmp(arg, "--no-deprecation") == 0) {
argv[i] = const_cast<char*>("");
no_deprecation = true;
} else if (strcmp(arg, "--trace-deprecation") == 0) {
argv[i] = const_cast<char*>("");
trace_deprecation = true;
node: Add --throw-deprecation Extremely handy when tracking down a flood of recursive nextTick warnings.
12 years ago
} else if (strcmp(arg, "--throw-deprecation") == 0) {
argv[i] = const_cast<char*>("");
throw_deprecation = true;
Remove &#39;--&#39; from command line parsing
15 years ago
} else if (argv[i][0] != '-') {
break;
Add simple command line switches
16 years ago
}
}
Fix options parsing The current node has a bug where it will fail to pass the option argument immediately preceding the first non-option argument to V8. That is the --perf flag will be ignored by V8 when running &#39;node --perf script.js&#39;.
15 years ago
option_end_index = i;
Add simple command line switches
16 years ago
}
Implement process._debugProcess
13 years ago
debugger: wake up the event loop when a debugger command is dispatched When the event loop was blocked in epoll / kqueue or similar, debugger commands wouldn&#39;t be processed. This patch fixes that by adding an uv_async handle which is triggered when a debugger command is dispatched. The async handle&#39;s callback makes sure that V8 is entered. Closes GH-3626 Closes GH-3718
13 years ago
// Called from the main thread.
static void DispatchDebugMessagesAsyncCallback(uv_async_t* handle, int status) {
v8::Debug::ProcessDebugMessages();
}
// Called from V8 Debug Agent TCP thread.
static void DispatchMessagesDebugAgentCallback() {
uv_async_send(&dispatch_debug_messages_async);
}
start debug agent on SIGUSR1
14 years ago
static void EnableDebug(bool wait_connect) {
Implement process._debugProcess
13 years ago
// If we're called from another thread, make sure to enter the right
// v8 isolate.
node_isolate->Enter();
debugger: wake up the event loop when a debugger command is dispatched When the event loop was blocked in epoll / kqueue or similar, debugger commands wouldn&#39;t be processed. This patch fixes that by adding an uv_async handle which is triggered when a debugger command is dispatched. The async handle&#39;s callback makes sure that V8 is entered. Closes GH-3626 Closes GH-3718
13 years ago
v8::Debug::SetDebugMessageDispatchHandler(DispatchMessagesDebugAgentCallback,
false);
uv_async_init(uv_default_loop(),
&dispatch_debug_messages_async,
DispatchDebugMessagesAsyncCallback);
uv_unref((uv_handle_t*) &dispatch_debug_messages_async);
start debug agent on SIGUSR1
14 years ago
// Start the debug thread and it's associated TCP server on port 5858.
debugger: use v8 api to wait for a connection Use v8::Debug::EnableAgent(_, _, true) to wait for incoming debugger-client connection before emitting any break (or other) events. This commit should fix test/simple/test-debugger-repl faults.
13 years ago
bool r = v8::Debug::EnableAgent("node " NODE_VERSION,
debug_port,
wait_connect);
start debug agent on SIGUSR1
14 years ago
// Crappy check that everything went well. FIXME
assert(r);
// Print out some information.
EnableDebug immediatly on SIGUSR1 Don&#39;t wait for script to break somewhere, because that may not happen if execution is inside event-loop, not in v8. Add &#39;\n&#39; to the end of &#39;debugger listening...&#39; message
13 years ago
fprintf(stderr, "debugger listening on port %d\n", debug_port);
Implement process._debugProcess
13 years ago
fflush(stderr);
more cli options * node debug localhost:5858 - connects to remote debugger * node debug -p `pgrep node` - connects to running process * Fixed double-run of debugger on SIGUSR1
13 years ago
debugger_running = true;
Implement process._debugProcess
13 years ago
node_isolate->Exit();
start debug agent on SIGUSR1
14 years ago
}
Ctrl+Break on windows starts debugger server
13 years ago
#ifdef __POSIX__
core: Use a uv_signal for debug listener Starting the debugger directly in the SIGUSR1 signal handler results in a malloc lock contention ~1% of the time. It hangs the test, which is annoying on a daily basis to all of us, but it also is pretty terrible if you actually want to debug a node process that has gone sideways. Credit to @bnoordhuis for most of this. I just added the unref which keeps it from messing up the event loop for other stuff.
12 years ago
static void EnableDebugSignalHandler(uv_signal_t* handle, int) {
EnableDebug immediatly on SIGUSR1 Don&#39;t wait for script to break somewhere, because that may not happen if execution is inside event-loop, not in v8. Add &#39;\n&#39; to the end of &#39;debugger listening...&#39; message
13 years ago
// Break once process will return execution to v8
Implement process._debugProcess
13 years ago
v8::Debug::DebugBreak(node_isolate);
SIGUSR1 should break the VM without delay
14 years ago
EnableDebug immediatly on SIGUSR1 Don&#39;t wait for script to break somewhere, because that may not happen if execution is inside event-loop, not in v8. Add &#39;\n&#39; to the end of &#39;debugger listening...&#39; message
13 years ago
if (!debugger_running) {
SIGUSR1 should break the VM without delay
14 years ago
fprintf(stderr, "Hit SIGUSR1 - starting debugger agent.\n");
EnableDebug(false);
}
start debug agent on SIGUSR1
14 years ago
}
Implement process._debugProcess
13 years ago
static void RegisterSignalHandler(int signal, void (*handler)(int)) {
struct sigaction sa;
memset(&sa, 0, sizeof(sa));
sa.sa_handler = handler;
sigfillset(&sa.sa_mask);
sigaction(signal, &sa, NULL);
}
Handle<Value> DebugProcess(const Arguments& args) {
HandleScope scope;
if (args.Length() != 1) {
return ThrowException(Exception::Error(
String::New("Invalid number of arguments.")));
}
pid_t pid;
int r;
pid = args[0]->IntegerValue();
r = kill(pid, SIGUSR1);
if (r != 0) {
return ThrowException(ErrnoException(errno, "kill"));
}
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Implement process._debugProcess
13 years ago
}
Ctrl+Break on windows starts debugger server
13 years ago
#endif // __POSIX__
Implement process._debugProcess
13 years ago
#ifdef _WIN32
DWORD WINAPI EnableDebugThreadProc(void* arg) {
Ctrl+Break on windows starts debugger server
13 years ago
// Break once process will return execution to v8
Implement process._debugProcess
13 years ago
if (!debugger_running) {
for (int i = 0; i < 1; i++) {
fprintf(stderr, "Starting debugger agent.\r\n");
fflush(stderr);
EnableDebug(false);
}
}
Ctrl+Break on windows starts debugger server
13 years ago
v8::Debug::DebugBreak();
Implement process._debugProcess
13 years ago
return 0;
}
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
static int GetDebugSignalHandlerMappingName(DWORD pid, wchar_t* buf,
size_t buf_len) {
return _snwprintf(buf, buf_len, L"node-debug-handler-%u", pid);
Implement process._debugProcess
13 years ago
}
static int RegisterDebugSignalHandler() {
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
wchar_t mapping_name[32];
Implement process._debugProcess
13 years ago
HANDLE mapping_handle;
DWORD pid;
LPTHREAD_START_ROUTINE* handler;
pid = GetCurrentProcessId();
if (GetDebugSignalHandlerMappingName(pid,
mapping_name,
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
ARRAY_SIZE(mapping_name)) < 0) {
Implement process._debugProcess
13 years ago
return -1;
}
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
mapping_handle = CreateFileMappingW(INVALID_HANDLE_VALUE,
Implement process._debugProcess
13 years ago
NULL,
PAGE_READWRITE,
0,
sizeof *handler,
mapping_name);
if (mapping_handle == NULL) {
return -1;
Ctrl+Break on windows starts debugger server
13 years ago
}
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
handler = reinterpret_cast<LPTHREAD_START_ROUTINE*>(
MapViewOfFile(mapping_handle,
FILE_MAP_ALL_ACCESS,
0,
0,
sizeof *handler));
Implement process._debugProcess
13 years ago
if (handler == NULL) {
CloseHandle(mapping_handle);
return -1;
}
*handler = EnableDebugThreadProc;
UnmapViewOfFile((void*) handler);
return 0;
Ctrl+Break on windows starts debugger server
13 years ago
}
start debug agent on SIGUSR1
14 years ago
Implement process._debugProcess
13 years ago
static Handle<Value> DebugProcess(const Arguments& args) {
HandleScope scope;
src, test: downgrade to v8 3.14 api
12 years ago
Handle<Value> rv = Undefined();
Implement process._debugProcess
13 years ago
DWORD pid;
Rename some occurrences of `process_1` back to `process`
13 years ago
HANDLE process = NULL;
Implement process._debugProcess
13 years ago
HANDLE thread = NULL;
HANDLE mapping = NULL;
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
wchar_t mapping_name[32];
Implement process._debugProcess
13 years ago
LPTHREAD_START_ROUTINE* handler = NULL;
start debug agent on SIGUSR1
14 years ago
Implement process._debugProcess
13 years ago
if (args.Length() != 1) {
rv = ThrowException(Exception::Error(String::New("Invalid number of arguments.")));
goto out;
}
Add signal handlers so we clean up before exiting. Add SIGTERM and SIGINT signal handlers so that we run the exit handlers before exiting when getting these signals. Fixes an issue where we couldn&#39;t run vi after CTRL+C&#39;ing node because the stdin fd was left non-blocking.
14 years ago
Implement process._debugProcess
13 years ago
pid = (DWORD) args[0]->IntegerValue();
Rename some occurrences of `process_1` back to `process`
13 years ago
process = OpenProcess(PROCESS_CREATE_THREAD | PROCESS_QUERY_INFORMATION |
Implement process._debugProcess
13 years ago
PROCESS_VM_OPERATION | PROCESS_VM_WRITE |
PROCESS_VM_READ,
FALSE,
pid);
Rename some occurrences of `process_1` back to `process`
13 years ago
if (process == NULL) {
Implement process._debugProcess
13 years ago
rv = ThrowException(WinapiErrnoException(GetLastError(), "OpenProcess"));
goto out;
}
if (GetDebugSignalHandlerMappingName(pid,
mapping_name,
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
ARRAY_SIZE(mapping_name)) < 0) {
Implement process._debugProcess
13 years ago
rv = ThrowException(ErrnoException(errno, "sprintf"));
goto out;
}
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
mapping = OpenFileMappingW(FILE_MAP_READ, FALSE, mapping_name);
Implement process._debugProcess
13 years ago
if (mapping == NULL) {
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
rv = ThrowException(WinapiErrnoException(GetLastError(),
"OpenFileMappingW"));
Implement process._debugProcess
13 years ago
goto out;
}
Make win32 ansi api usage explicit Use widechar versions in a couple of places. Don&#39;t use C-style cast in C++ code.
13 years ago
handler = reinterpret_cast<LPTHREAD_START_ROUTINE*>(
MapViewOfFile(mapping,
FILE_MAP_READ,
0,
0,
sizeof *handler));
Implement process._debugProcess
13 years ago
if (handler == NULL || *handler == NULL) {
rv = ThrowException(WinapiErrnoException(GetLastError(), "MapViewOfFile"));
goto out;
}
Rename some occurrences of `process_1` back to `process`
13 years ago
thread = CreateRemoteThread(process,
Implement process._debugProcess
13 years ago
NULL,
0,
*handler,
NULL,
0,
NULL);
if (thread == NULL) {
rv = ThrowException(WinapiErrnoException(GetLastError(),
"CreateRemoteThread"));
goto out;
}
// Wait for the thread to terminate
if (WaitForSingleObject(thread, INFINITE) != WAIT_OBJECT_0) {
rv = ThrowException(WinapiErrnoException(GetLastError(),
"WaitForSingleObject"));
goto out;
}
out:
Rename some occurrences of `process_1` back to `process`
13 years ago
if (process != NULL) {
CloseHandle(process);
Implement process._debugProcess
13 years ago
}
if (thread != NULL) {
CloseHandle(thread);
}
if (handler != NULL) {
UnmapViewOfFile(handler);
}
if (mapping != NULL) {
CloseHandle(mapping);
}
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
Add signal handlers so we clean up before exiting. Add SIGTERM and SIGINT signal handlers so that we run the exit handlers before exiting when getting these signals. Fixes an issue where we couldn&#39;t run vi after CTRL+C&#39;ing node because the stdin fd was left non-blocking.
14 years ago
}
Implement process._debugProcess
13 years ago
#endif // _WIN32
Add signal handlers so we clean up before exiting. Add SIGTERM and SIGINT signal handlers so that we run the exit handlers before exiting when getting these signals. Fixes an issue where we couldn&#39;t run vi after CTRL+C&#39;ing node because the stdin fd was left non-blocking.
14 years ago
pause command
13 years ago
static Handle<Value> DebugPause(const Arguments& args) {
v8::Debug::DebugBreak(node_isolate);
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
pause command
13 years ago
}
debugger: export `debug_port` to `process` `process.debug_port` is useful for changing debugger port in runtime, before starting it (via SIGUSR1). Using `--port=` argument for debugger repl, tests will run debugger server on a `common.PORT` (as it usually does for any other servers). `process._debugEnd()` stops debugger and its server. * debugger: implemented process._debugEnd(), `node debug --port=5858 app.js` * test: start debugger repl on common.PORT * fixes #2613 * fixes #2614
13 years ago
static Handle<Value> DebugEnd(const Arguments& args) {
if (debugger_running) {
v8::Debug::DisableAgent();
debugger_running = false;
}
src, test: downgrade to v8 3.14 api
12 years ago
return Undefined();
debugger: export `debug_port` to `process` `process.debug_port` is useful for changing debugger port in runtime, before starting it (via SIGUSR1). Using `--port=` argument for debugger repl, tests will run debugger server on a `common.PORT` (as it usually does for any other servers). `process._debugEnd()` stops debugger and its server. * debugger: implemented process._debugEnd(), `node debug --port=5858 app.js` * test: start debugger repl on common.PORT * fixes #2613 * fixes #2614
13 years ago
}
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
char** Init(int argc, char *argv[]) {
// Initialize prog_start_time to get relative uptime.
src: use monotonic time for process.uptime() `process.uptime()` interface will return the amount of time the current process has been running. To achieve this it was caching the `uv_uptime` value at program start, and then on the call to `process.uptime()` returning the delta between the two values. `uv_uptime` is defined as the number of seconds the operating system has been up since last boot. On sunos this interface uses `kstat`s which can be a significantly expensive operation as it requires exclusive access, but because of the design of `process.uptime()` node *had* to always call this on start. As a result if you had many node processes all starting at the same time you would suffer lock contention as they all tried to read kstats. Instead of using `uv_uptime` to achieve this, the libuv loop already has a concept of current loop time in the form of `uv_now()` which is in fact monotonically increasing, and already stored directly on the loop. By using this value at start every platform performs at least one fewer syscall during initialization. Since the interface to `uv_uptime` is defined as seconds, in the call to `process.uptime()` we now `uv_update_time` get our delta, divide by 1000 to get seconds, and then convert to an `Integer`. In 0.12 we can move back to `Number::New` instead and not lose precision. Caveat: For some platforms `uv_uptime` reports time monotonically increasing regardless of system hibernation, `uv_now` interface is also monotonically increasing but may not reflect time spent in hibernation.
11 years ago
prog_start_time = uv_now(uv_default_loop());
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
Inherited FDs should not accidentally be inherited by child processes This makes test-child-process-detached pass on Windows.
13 years ago
// Make inherited handles noninheritable.
uv_disable_stdio_inheritance();
Add comments to node.cc
15 years ago
// Parse a few arguments which are specific to Node.
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
node::ParseArgs(argc, argv);
Remove &#39;--&#39; from command line parsing
15 years ago
// Parse the rest of the args (up to the 'option_end_index' (where '--' was
Add comments to node.cc
15 years ago
// in the command line))
Deshadow, denamespace variables in node.cc Prep for isolates.
13 years ago
int v8argc = option_end_index;
remove node::CheckBreak in favor of using the v8 debugger js object
15 years ago
char **v8argv = argv;
Deshadow, denamespace variables in node.cc Prep for isolates.
13 years ago
if (debug_wait_connect) {
Revert &#34;Expose the V8 debug object process.debug&#34; This reverts commit d9fbb8a58071e339711e8c0ceeb31de5a645fd42.
15 years ago
// v8argv is a copy of argv up to the script file argument +2 if --debug-brk
Remove module.js - put code into src/node.js
15 years ago
// to expose the v8 debugger js object so that node.js can set
Revert &#34;Expose the V8 debug object process.debug&#34; This reverts commit d9fbb8a58071e339711e8c0ceeb31de5a645fd42.
15 years ago
// a breakpoint on the first line of the startup script
v8argc += 2;
v8argv = new char*[v8argc];
Fix -Wsizeof-array-argument compiler warning.
12 years ago
memcpy(v8argv, argv, sizeof(*argv) * option_end_index);
Deshadow, denamespace variables in node.cc Prep for isolates.
13 years ago
v8argv[option_end_index] = const_cast<char*>("--expose_debug_as");
v8argv[option_end_index + 1] = const_cast<char*>("v8debug");
Revert &#34;Expose the V8 debug object process.debug&#34; This reverts commit d9fbb8a58071e339711e8c0ceeb31de5a645fd42.
15 years ago
}
Add --max-stack-size flag. v8 doesn&#39;t expose a command-line flag to set the stack size, so this adds a new flag that node understands how to handle and uses v8&#39;s ResourceConstraints API.
14 years ago
// For the normal stack which moves from high to low addresses when frames
// are pushed, we can compute the limit as stack_size bytes below the
// the address of a stack variable (e.g. &stack_var) as an approximation
// of the start of the stack (we're assuming that we haven't pushed a lot
// of frames yet).
Deshadow, denamespace variables in node.cc Prep for isolates.
13 years ago
if (max_stack_size != 0) {
Add --max-stack-size flag. v8 doesn&#39;t expose a command-line flag to set the stack size, so this adds a new flag that node understands how to handle and uses v8&#39;s ResourceConstraints API.
14 years ago
uint32_t stack_var;
ResourceConstraints constraints;
Deshadow, denamespace variables in node.cc Prep for isolates.
13 years ago
uint32_t *stack_limit = &stack_var - (max_stack_size / sizeof(uint32_t));
Add --max-stack-size flag. v8 doesn&#39;t expose a command-line flag to set the stack size, so this adds a new flag that node understands how to handle and uses v8&#39;s ResourceConstraints API.
14 years ago
constraints.set_stack_limit(stack_limit);
SetResourceConstraints(&constraints); // Must be done before V8::Initialize
}
remove node::CheckBreak in favor of using the v8 debugger js object
15 years ago
V8::SetFlagsFromCommandLine(&v8argc, v8argv, false);
Move arg parsing to beginning of main() For faster &#34;./node -v&#34;
16 years ago
Disable node.cc functions that are not supported on windows
14 years ago
#ifdef __POSIX__
Ignore SIGPIPE Was inadvertently removed with evcom, http://github.com/ry/node/commit/c72967d3351c4a28c3f2434748598017aa0126d5#L10L1900
15 years ago
// Ignore SIGPIPE
Add signal handlers so we clean up before exiting. Add SIGTERM and SIGINT signal handlers so that we run the exit handlers before exiting when getting these signals. Fixes an issue where we couldn&#39;t run vi after CTRL+C&#39;ing node because the stdin fd was left non-blocking.
14 years ago
RegisterSignalHandler(SIGPIPE, SIG_IGN);
RegisterSignalHandler(SIGINT, SignalExit);
RegisterSignalHandler(SIGTERM, SignalExit);
Disable node.cc functions that are not supported on windows
14 years ago
#endif // __POSIX__
Ignore SIGPIPE Was inadvertently removed with evcom, http://github.com/ry/node/commit/c72967d3351c4a28c3f2434748598017aa0126d5#L10L1900
15 years ago
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
uv_idle_init(uv_default_loop(), &tick_spinner);
Revert &#34;Simplify next tick logic by looping around ev_loop&#34; This reverts commit 241ea7e5954e4a049af85cd8d51b84202cf0ea3a. Fixes GH-511: https://github.com/ry/node/issues/issue/511
14 years ago
timer: move setImmediate from timer to uv_check uv_check is the robust place to invoke setImmediate callbacks after process.nextTick and before timers(setTimeout/setInterval)
12 years ago
uv_check_init(uv_default_loop(), &check_immediate_watcher);
uv_unref((uv_handle_t*) &check_immediate_watcher);
uv_idle_init(uv_default_loop(), &idle_immediate_dummy);
Add comments to node.cc
15 years ago
V8::SetFatalErrorHandler(node::OnFatalError);
Implement process._debugProcess
13 years ago
// Fetch a reference to the main isolate, so we have a reference to it
// even when we need it to access it from another (debugger) thread.
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
node_isolate = Isolate::GetCurrent();
start debug agent on SIGUSR1
14 years ago
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
// If the --debug flag was specified then initialize the debug thread.
if (use_debug_agent) {
EnableDebug(debug_wait_connect);
} else {
Implement process._debugProcess
13 years ago
#ifdef _WIN32
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
RegisterDebugSignalHandler();
Implement process._debugProcess
13 years ago
#else // Posix
core: Use a uv_signal for debug listener Starting the debugger directly in the SIGUSR1 signal handler results in a malloc lock contention ~1% of the time. It hangs the test, which is annoying on a daily basis to all of us, but it also is pretty terrible if you actually want to debug a node process that has gone sideways. Credit to @bnoordhuis for most of this. I just added the unref which keeps it from messing up the event loop for other stuff.
12 years ago
static uv_signal_t signal_watcher;
uv_signal_init(uv_default_loop(), &signal_watcher);
uv_signal_start(&signal_watcher, EnableDebugSignalHandler, SIGUSR1);
uv_unref((uv_handle_t*)&signal_watcher);
Don&#39;t attempt to capture SIGUSR1
14 years ago
#endif // __POSIX__
Enable debugging. Use the --debug command line flag to enable. It appears that d8 sucks. Luckily it can be rewritten rather easily with the repl and tcp client libraries. Node&#39;s CL option parsing is getting rather unwieldy - needs refactor.
15 years ago
}
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
return argv;
}
isolates: add process.tid property, pass args to isolate
13 years ago
ref isolate loop
13 years ago
addon: add AtExit() function Lets native addons register exit hooks that run after the event loop has quit but before the VM is killed. Fixes #3147.
13 years ago
struct AtExitCallback {
AtExitCallback* next_;
void (*cb_)(void* arg);
void* arg_;
};
static AtExitCallback* at_exit_functions_;
void RunAtExit() {
AtExitCallback* p = at_exit_functions_;
at_exit_functions_ = NULL;
while (p) {
AtExitCallback* q = p->next_;
p->cb_(p->arg_);
delete p;
p = q;
}
}
void AtExit(void (*cb)(void* arg), void* arg) {
AtExitCallback* p = new AtExitCallback;
p->cb_ = cb;
p->arg_ = arg;
p->next_ = at_exit_functions_;
at_exit_functions_ = p;
}
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
void EmitExit(v8::Handle<v8::Object> process_l) {
// process.emit('exit')
src, test: downgrade to v8 3.14 api
12 years ago
process_l->Set(String::NewSymbol("_exiting"), True());
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
Local<Value> emit_v = process_l->Get(String::New("emit"));
assert(emit_v->IsFunction());
Local<Function> emit = Local<Function>::Cast(emit_v);
src, test: downgrade to v8 3.14 api
12 years ago
Local<Value> args[] = { String::New("exit"), Integer::New(0) };
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
TryCatch try_catch;
process: ensure that the &#34;exit&#34; event always has &#34;code&#34; given Upon &#34;normal&#34; exiting of Node (i.e. the event loop completes naturally), the &#34;code&#34; parameter was not being passed to the &#34;exit&#34; event listeners. Be consistent. Tests included.
13 years ago
emit->Call(process_l, 2, args);
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
if (try_catch.HasCaught()) {
FatalException(try_catch);
isolates: implement message passing Parent and child isolates can now pass arbitrary binary messages between each other. The messages are sent and received through a thread-safe queue that wakes up the event loop of the receiving thread.
13 years ago
}
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
}
isolates: implement message passing Parent and child isolates can now pass arbitrary binary messages between each other. The messages are sent and received through a thread-safe queue that wakes up the event loop of the receiving thread.
13 years ago
Expose original argv as process.execArgv for cluster and child_process.fork()
13 years ago
static char **copy_argv(int argc, char **argv) {
size_t strlen_sum;
char **argv_copy;
char *argv_data;
size_t len;
int i;
strlen_sum = 0;
for(i = 0; i < argc; i++) {
strlen_sum += strlen(argv[i]) + 1;
}
argv_copy = (char **) malloc(sizeof(char *) * (argc + 1) + strlen_sum);
if (!argv_copy) {
return NULL;
}
argv_data = (char *) argv_copy + sizeof(char *) * (argc + 1);
for(i = 0; i < argc; i++) {
argv_copy[i] = argv_data;
len = strlen(argv[i]) + 1;
memcpy(argv_data, argv[i], len);
argv_data += len;
}
argv_copy[argc] = NULL;
return argv_copy;
}
Export more functions for initializing and starting node. For greater flexibility in controlling node&#39;s initialization and startup, the following new functions are exported. - node::Init() - node::SetupProcessObject() - node::Load() - node::EmitExit() These are some of the major steps involved in node::Setup(). Exporting these functions allows an embedding program to write a replacement for node::Start(), and to have access to the node process object after it&#39;s created.
14 years ago
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
int Start(int argc, char *argv[]) {
process: fix process.title setter Commit 19fd530 broke the argv initialization logic that&#39;s used on linux and freebsd to update the process name (as displayed in tools like `top`). Fixes test/simple/test-setproctitle.js.
13 years ago
// Hack aroung with the argv pointer. Used for process.title = "blah".
argv = uv_setup_args(argc, argv);
Expose original argv as process.execArgv for cluster and child_process.fork()
13 years ago
// Logic to duplicate argv as Init() modifies arguments
// that are passed into it.
char **argv_copy = copy_argv(argc, argv);
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
// This needs to run *before* V8::Initialize()
Expose original argv as process.execArgv for cluster and child_process.fork()
13 years ago
// Use copy here as to not modify the original argv:
Init(argc, argv_copy);
Revert support for isolates. It was decided that the performance benefits that isolates offer (faster spin-up times for worker processes, faster inter-worker communication, possibly a lower memory footprint) are not actual bottlenecks for most people and do not outweigh the potential stability issues and intrusive changes to the code base that first-class support for isolates requires. Hence, this commit backs out all isolates-related changes. Good bye, isolates. We hardly knew ye.
13 years ago
node: un-revert 9a6012e The crashes in debug mode after adding Locker are *not* caused by Locker. Locker is merely exposing issues that already existed. Some of these issues have since been fixed in 70635753. This reverts commit 407181538b5193f6810bf06a2056a200265c0e93. This reapplies commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
V8::Initialize();
src: seed V8&#39;s random number generator at startup The default entropy source is /dev/urandom on UNIX platforms, which is okay but we can do better by seeding it from OpenSSL&#39;s entropy pool. On Windows we can certainly do better; on that platform, V8 seeds the random number generator using only the current system time. Fixes #6250. NB: This is a back-port of commit 7ac2391 from the master branch that for some reason never got back-ported to the v0.10 branch. The default on UNIX platforms in v0.10 is different and arguably worse than it is with master: if no entropy source is provided, V8 3.14 calls srandom() with a xor of the PID and the current time in microseconds. That means that on systems with a coarse system clock, the initial state of the PRNG may be easily guessable. The situation on Windows is even more dire because there the PRNG is seeded with only the current time... in milliseconds.
11 years ago
#if HAVE_OPENSSL
// V8 on Windows doesn't have a good source of entropy. Seed it from
// OpenSSL's pool.
V8::SetEntropySource(crypto::EntropySource);
#endif
node: un-revert 9a6012e The crashes in debug mode after adding Locker are *not* caused by Locker. Locker is merely exposing issues that already existed. Some of these issues have since been fixed in 70635753. This reverts commit 407181538b5193f6810bf06a2056a200265c0e93. This reapplies commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
{
Locker locker;
HandleScope handle_scope;
Revert &#34;Re-add top-level v8::Locker&#34; The locker makes node crash in debug mode sometimes. For example, test/simple/test-repl.js triggers it. This reverts commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
node: un-revert 9a6012e The crashes in debug mode after adding Locker are *not* caused by Locker. Locker is merely exposing issues that already existed. Some of these issues have since been fixed in 70635753. This reverts commit 407181538b5193f6810bf06a2056a200265c0e93. This reapplies commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
// Create the one and only Context.
Persistent<Context> context = Context::New();
Context::Scope context_scope(context);
Revert &#34;Re-add top-level v8::Locker&#34; The locker makes node crash in debug mode sometimes. For example, test/simple/test-repl.js triggers it. This reverts commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
node: un-revert 9a6012e The crashes in debug mode after adding Locker are *not* caused by Locker. Locker is merely exposing issues that already existed. Some of these issues have since been fixed in 70635753. This reverts commit 407181538b5193f6810bf06a2056a200265c0e93. This reapplies commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
// Use original argv, as we're just copying values out of it.
Handle<Object> process_l = SetupProcessObject(argc, argv);
v8_typed_array::AttachBindings(context->Global());
Revert &#34;Re-add top-level v8::Locker&#34; The locker makes node crash in debug mode sometimes. For example, test/simple/test-repl.js triggers it. This reverts commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
node: un-revert 9a6012e The crashes in debug mode after adding Locker are *not* caused by Locker. Locker is merely exposing issues that already existed. Some of these issues have since been fixed in 70635753. This reverts commit 407181538b5193f6810bf06a2056a200265c0e93. This reapplies commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
// Create all the objects, load modules, do everything.
// so your next reading stop should be node::Load()!
Load(process_l);
Revert &#34;Re-add top-level v8::Locker&#34; The locker makes node crash in debug mode sometimes. For example, test/simple/test-repl.js triggers it. This reverts commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
node: un-revert 9a6012e The crashes in debug mode after adding Locker are *not* caused by Locker. Locker is merely exposing issues that already existed. Some of these issues have since been fixed in 70635753. This reverts commit 407181538b5193f6810bf06a2056a200265c0e93. This reapplies commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
// All our arguments are loaded. We've evaluated all of the scripts. We
// might even have created TCP servers. Now we enter the main eventloop. If
// there are no watchers on the loop (except for the ones that were
// uv_unref'd) then this function exits. As long as there are active
// watchers, it blocks.
deps: upgrade libuv to 4ba03dd
12 years ago
uv_run(uv_default_loop(), UV_RUN_DEFAULT);
Revert &#34;Re-add top-level v8::Locker&#34; The locker makes node crash in debug mode sometimes. For example, test/simple/test-repl.js triggers it. This reverts commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
node: un-revert 9a6012e The crashes in debug mode after adding Locker are *not* caused by Locker. Locker is merely exposing issues that already existed. Some of these issues have since been fixed in 70635753. This reverts commit 407181538b5193f6810bf06a2056a200265c0e93. This reapplies commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
EmitExit(process_l);
RunAtExit();
isolates: add atexit() functionality for isolates
13 years ago
Add comments to node.cc
15 years ago
#ifndef NDEBUG
node: un-revert 9a6012e The crashes in debug mode after adding Locker are *not* caused by Locker. Locker is merely exposing issues that already existed. Some of these issues have since been fixed in 70635753. This reverts commit 407181538b5193f6810bf06a2056a200265c0e93. This reapplies commit 9a6012edd9330296b7476bc6b7fbda2cd5c8165d. Conflicts: src/node.cc
13 years ago
context.Dispose();
#endif
}
#ifndef NDEBUG
// Clean up. Not strictly necessary.
Clean ups
16 years ago
V8::Dispose();
Add comments to node.cc
15 years ago
#endif // NDEBUG
Statically link in dependencies for node.exe
14 years ago
Expose original argv as process.execArgv for cluster and child_process.fork()
13 years ago
// Clean up the copy:
free(argv_copy);
Implement onExit() hook for modules. onExit() is similar to the onLoad() callback. onExit() is called on each module just before the process exits. This can be used to check state in unit tests, but not to perform I/O. The process will forcibly exit as soon as all of the onExit callbacks are made.
16 years ago
return 0;
Major refactoring: program name now &#34;node&#34; Trying to make a more moduler design. Two libraries currently &#34;TCP&#34; and &#34;HTTP&#34; each have their own file. Other major feature added here is multiple web servers! excitement.
16 years ago
}
Add node::Start() first step towards building a libnode.a
15 years ago
} // namespace node