igor@mir2.org
|
3d870e0a9e
|
Bug 413744: JS_GCMETER requires to recompile just js/src, not the whole browser. r,a=brendan
|
2008-01-29 14:17:25 -08:00 |
|
brendan@mozilla.org
|
a4625d2c2b
|
Revert last checkin (414452).
|
2008-01-29 00:45:14 -08:00 |
|
brendan@mozilla.org
|
f4f2be6738
|
Move guts of js_SetProtoOrParent to jsgc.c, unifying serialization and deadlock avoidance machinery (414452, r=igor).
|
2008-01-28 23:18:21 -08:00 |
|
igor@mir2.org
|
d43d9f85f7
|
Bug 413565: GC now put all free cells to free lists. r,a=brendan
|
2008-01-23 05:05:48 -08:00 |
|
brendan@mozilla.org
|
d56e272e19
|
* Menu of -D flags for enabling instrumentation, as a commented-out CFLAGS += setting for convenient testing. * js_FindProperty and js_LookupPropertyWithFlags return indexes into the scope and prototype chains, respectively, to support internal instrumentation, and to pave the way for the return of the property cache (bug 365851).. * jsutil.[ch] JSBasicStats struct and functions for computing mean/sigma/max and auto-scaling histogram. * JS_SCOPE_DEPTH_METER instrumentation for compile- and run-time scope chain length instrumentation: + At compile time, rt->hostenvScopeDepthStats and rt->lexicalScopeDepthStats meter scope chains passed into the compile and evaluate APIs. + At runtime, rt->protoLookupDepthStats and rt->scopeSearchDepthStats track steps along the prototype and scope chains until the sought-after property is found. * JS_ARENAMETER uses JSBasicStats now. * Added rt->liveScopePropsPreSweep to fix the property tree stats code that rotted when property tree sweeping moved to after the finalization phase. * Un-bitrotted some DEBUG_brendan code, turned some off for myself via XXX. * Mac OS X toolchain requires initialized data shared across dynamic library member files, outlaws common data, so initialize extern metering vars. * Old HASHMETER code in jshash.[ch] is now JS_HASHMETER-controlled and based on JSBasicStats. * DEBUG_scopemeters macro renamed JS_DUMP_SCOPE_METERS; uses JSBasicStats now. * Disentangle DEBUG and DUMP_SCOPE_STATS (now JS_DUMP_PROPTREE_STATS) and fix inconsistent thread safety for liveScopeProps (sometimes atomic-incremented, sometimes runtime-locked). * Compiler-modeled maxScopeDepth will propagate via JSScript to runtime for capability-based, interpreter-inlined cache hit qualifier bits, to bypass scope and prototype chain lookup by optimizing for common monomorphic get, set, and call site referencing a prototype property in a well-named object (no shadowing or mutation in 99.9% of the cases).
|
2008-01-12 16:31:31 -08:00 |
|
brendan@mozilla.org
|
4bef2e90ac
|
64-bit fixes for problems reported by edwin@cheatah.nl (many thanks to Edwin; 410941, r=igor).
|
2008-01-07 00:41:06 -08:00 |
|
brendan@mozilla.org
|
7d8fcab4dc
|
Remove DUMP_CALL_TABLE (preliminary patch for 365851, r=igor).
|
2007-12-14 13:36:02 -08:00 |
|
igor@mir2.org
|
1831710dfa
|
Bug 403678: GC thing callback is removed. r,a=brendan
|
2007-11-19 02:57:04 -08:00 |
|
peterv@propagandism.org
|
9989fb489f
|
Fix for bug 375808 (Replacing GCX_PRIVATE by GCX_FUNCTION). r=igor, sr=jst, a=schrep.
|
2007-11-18 06:10:28 -08:00 |
|
reed@reedloden.com
|
4c5746dc93
|
relanding bug 400687 since it doesn't look to be the source of the Ts regression
|
2007-11-14 20:16:35 -08:00 |
|
gavin@gavinsharp.com
|
939757e2f4
|
Back out the patch for bug 400687 to see whether it caused the Ts regression, a=sicking
|
2007-11-14 16:17:28 -08:00 |
|
igor@mir2.org
|
ca808e133e
|
Bug 400687: removal of dependency of xpconnect on internal JS GC thing types. r,a=brendan
|
2007-11-13 13:23:07 -08:00 |
|
igor@mir2.org
|
9bd984356d
|
Bug 402087: fixing iteraction between gczeal mode and scripts. r,a=brendan, aM9=beltzner
|
2007-11-06 03:05:00 -08:00 |
|
benjamin@smedbergs.us
|
b4a8d79ecb
|
Bug 396936 - FLAGP_TO_THING bustage from C++, r+a=brendan
|
2007-09-20 12:49:59 -07:00 |
|
igor@mir2.org
|
1dd9e7fca8
|
Bug 333236: merge sweep and free phases in GC. r=brendan
|
2007-09-19 15:04:47 -07:00 |
|
igor@mir2.org
|
cc6784b030
|
Bug 394551: no JS frames for fast native calls. r=brendan
|
2007-09-18 00:34:54 -07:00 |
|
igor@mir2.org
|
cd16ec9fa2
|
Bug 396049: the last ditch GC gcPoke checks are moved to js_NewGCThing. r=brendan
|
2007-09-18 00:22:07 -07:00 |
|
brendan@mozilla.org
|
19b6fcc0f0
|
Back-out due to test failure.
|
2007-09-17 18:22:20 -07:00 |
|
brendan@mozilla.org
|
2260fcfe2e
|
Make GCF_SYSTEM immutable per object (396487, r=igor).
|
2007-09-17 17:47:42 -07:00 |
|
igor@mir2.org
|
aacdc13f0f
|
Bug 394551: taking out the last patch as broke Windows build.
|
2007-09-17 12:28:14 -07:00 |
|
igor@mir2.org
|
065a93dd1c
|
Bug 394551: no JS frames for fast native calls. r=brendan
|
2007-09-17 12:08:46 -07:00 |
|
igor@mir2.org
|
f058abee03
|
Bug 392263: using mmap/VirtualAlloc for GC arenas. r=brendan
|
2007-09-16 06:03:17 -07:00 |
|
igor@mir2.org
|
f3d30850c6
|
Bug 392263: taking out the patch due to bug 396299.
|
2007-09-16 05:02:56 -07:00 |
|
igor@mir2.org
|
5855f137e7
|
Bug 392263: using mmap/VirualAlloc for GC arenas. r=brendan
|
2007-09-15 10:19:32 -07:00 |
|
igor@mir2.org
|
9343ce4a2c
|
Bug 392263: taking out the patch due to Mac build problems.
|
2007-09-15 08:26:30 -07:00 |
|
igor@mir2.org
|
43e3bb8cc3
|
Bug 392263: using mmap/VirualAlloc for GC arenas. r=brendan
|
2007-09-15 08:11:41 -07:00 |
|
igor@mir2.org
|
4606e8ce2b
|
Bug 394709: fixing memory leak with watch handlers. r+a=brendan
|
2007-09-07 14:02:20 -07:00 |
|
igor@mir2.org
|
153df36a83
|
Bug 393184: js_NewGCThing no longer zeros the allocated thing. r=brendan
|
2007-08-22 13:56:08 -07:00 |
|
igor@mir2.org
|
4773ca1cae
|
Bug 391290: mutability flag for strings is stored inside strings. r=brendan
|
2007-08-15 23:23:06 -07:00 |
|
igor@mir2.org
|
14e3457a58
|
Bug 386265: using double kashing for atoms. r=brendan
|
2007-08-11 13:25:16 -07:00 |
|
igor@mir2.org
|
dec81ea13e
|
Bug 391033: no frame manipulation when reporting errors. r=brendan
|
2007-08-10 23:12:00 -07:00 |
|
igor@mir2.org
|
4d68066c7e
|
Bug 387481, bug 389063: multithreading atom fixes and cleanups. r=brendan
|
2007-08-07 00:29:32 -07:00 |
|
dbaron@dbaron.org
|
b8ec995348
|
Fix leak of JSScript when a JSFunction is collected in a later GC than its function object. This changes GCX_PRIVATE to GCX_FUNCTION, and is essentially the same as the finalization part of the changes from bug 375808 (by igor). b=389757 r=igor a1.9=brendan
|
2007-08-04 00:00:43 -07:00 |
|
igor@mir2.org
|
ec402ee7a9
|
Bug 390078: Avoiding unnecessary GC-tracing of JSStackFrame members. r=brendan
|
2007-08-03 12:56:35 -07:00 |
|
brendan@mozilla.org
|
ba3ad2ce7a
|
Fast (frame-less) native call optimizations (385393, r=igor).
|
2007-08-01 21:33:52 -07:00 |
|
brendan@mozilla.org
|
18e7692117
|
Backing out, see bug 389122 and orange or red tinderboxes.
|
2007-07-21 16:21:20 -07:00 |
|
brendan@mozilla.org
|
5f038f69ae
|
Fast natives and related optimizations (385393, r=igor).
|
2007-07-21 14:39:42 -07:00 |
|
igor@mir2.org
|
79f3a28c2d
|
Bug 380469: Calling close on unreachable generators from GC is no longer supported. r=brendan
|
2007-07-13 00:28:47 -07:00 |
|
mrbkap@gmail.com
|
f90fd387a3
|
Clear dormantNext when we're done using it. bug 387725, r=igor
|
2007-07-11 14:57:05 -07:00 |
|
igor@mir2.org
|
ab560f0cb7
|
Bug 385729: objects and regexps are stored in separated tables in JSScript. r=brendan,mrbkap
|
2007-07-08 02:03:34 -07:00 |
|
crowder@fiverocks.com
|
daa4556abc
|
Bug 357016: spidermonkey should be buildable by a C++ compiler, patch by
crowder/Jason Orendorff, r=brendan
|
2007-07-05 13:37:47 -07:00 |
|
igor@mir2.org
|
9137bfaf4e
|
Bug 349326: for-in loop now always closes iterator objects. r=brendan
|
2007-07-02 05:13:23 -07:00 |
|
igor@mir2.org
|
64f1c8053d
|
Bug 384478: Patch from Mike Moening to fix warning when compiling with VC. r=me
|
2007-06-18 12:33:21 -07:00 |
|
peterv@propagandism.org
|
7e189d8549
|
Fix for bug 377884 (Switch nsXPConnect::Traverse to use tracing). r=jst/igor, sr=dbaron.
|
2007-05-25 01:42:08 -07:00 |
|
igor@mir2.org
|
e8eeb4f10c
|
Bug 379758: SETSP is removed
|
2007-05-24 00:51:46 -07:00 |
|
crowder@fiverocks.com
|
c66619c8af
|
Bug 308429: make tooMuchGC dynamic (runtime gczeal option), r=brendan
|
2007-05-15 16:27:20 -07:00 |
|
igor@mir2.org
|
6dca224ba8
|
Bug 379455: tracing API now let the tracer to know about the GC lock count.
|
2007-05-04 08:43:19 -07:00 |
|
igor@mir2.org
|
7f6e92e7d2
|
Bug 379146: New API to register application-specific GC roots. r=brendan
|
2007-05-01 03:09:46 -07:00 |
|
igor@mir2.org
|
68d7b74a96
|
Bug 378261: Replacing GC_MARK_DEBUG by DumpHeap. r=brendan
|
2007-04-25 06:43:18 -07:00 |
|
igor@mir2.org
|
d194d1ad7f
|
Bug 378492: proper checks for null and jsval type when tracing, r=brendan
|
2007-04-24 16:11:02 -07:00 |
|