Commit Graph

64539 Commits

Author SHA1 Message Date
Makoto Kato
7a5360fad1 Bug 636296 - Fix forceReturn trampoline on Win64/Solaris (r=dvander)
--HG--
extra : rebase_source : 85846343b38031931adafb7e44bc8983fdd3472c
2011-03-28 17:33:53 -07:00
Phil Ringnalda
8d4385b7fe Back out bd821ea0ad41 (638324) for mochitest-chrome crashes 2011-03-31 19:55:10 -07:00
Steve Fink
3b49dbe49e Backout bogus part of c46b4248ee36 (--rebase deleted a random file; thanks) (r=flames) 2011-03-31 15:48:25 -07:00
Andreas Gal
e69500d1bc Maintain a list of IdArrays as we hand them out via JS API (638324, r=brendan). 2011-03-31 14:42:33 -07:00
Jan de Mooij
759437fbc3 [INFER] Refactor RegExp.exec -> RegExp.test optimization, use in both TM and JM. 2011-03-28 17:06:49 -07:00
Jan de Mooij
465c36c7ff [INFER] Handle jsop_binary overflow in inc/dec ops, bug 644970. r=bhackett 2011-03-28 15:44:30 -07:00
Brian Hackett
977162f22a [INFER] More robust handling of pushed singletons in CALLPROP when the lhs could be null/undefined, bug 639099. 2011-03-27 19:21:59 -07:00
Brian Hackett
dc5232e83a [INFER] Inline callees at polymorphic call sites, bug 639099. 2011-03-27 18:25:05 -07:00
Brian Hackett
02df7e9a1d [INFER] Wait for code to get hot before inlining calls, bug 639099. 2011-03-27 07:48:03 -07:00
Brian Hackett
d31340f90f [INFER] Fix Linux build again. 2011-03-26 20:34:56 -07:00
Brian Hackett
1d7bcffc21 [INFER] Fix Linux build. 2011-03-26 20:23:15 -07:00
Brian Hackett
7074cd94d6 [INFER] Don't clone functions with singleton types. 2011-03-26 20:18:58 -07:00
Brian Hackett
94473000bc [INFER] Fix x64 build, revert accidental manifest.py changes, bug 639099. 2011-03-26 19:56:49 -07:00
Brian Hackett
74bba0be53 [INFER] Inline scripted calls, bug 639099. 2011-03-26 19:07:13 -07:00
Steve Fink
f24c9d54bb Bug 639943 - Convert IDL formatting to JS engine house style (r=timeless) 2011-03-26 12:05:40 -07:00
Igor Bukanov
d56f220bed bug 643548 - followup to fix js shell builld brekage on Linux 2011-03-31 15:07:48 +02:00
Paul Biggar
43b850291d Bug 643548: Remove mozalloc_undef_macro_wrappers hack from JS engine. (r=luke,rs=brendan)
mozalloc_undef_macro_wrappers are brittle and have side-effects that are hard
to debug and fix. The alternative is the just stick an underscore on the end of
malloc, free, etc, which is a comparatively small burden.
2011-03-31 01:14:12 -07:00
Paul Biggar
72c7701093 Bug 634155: Account for NewCompartment's memory, and change allocation APIs (r=nnethercote)
This changes the allocation API, in the following way:

  js_malloc -> {cx->,rt->,OffTheBooks::}malloc
  js_calloc -> {cx->,rt->,OffTheBooks::}calloc
  js_realloc -> {cx->,rt->,OffTheBooks::}realloc
  js_free -> {cx->,rt->,Foreground::,UnwantedForeground::}free
  js_new -> {cx->,rt->,OffTheBooks::}new_
  js_new_array -> {cx->,rt->,OffTheBooks::}new_array
  js_delete -> {cx->,rt->,Foreground::,UnwantedForeground::}delete_

This is to move as many allocations as possible through a JSContext (so that they may be aken into account by gcMallocBytes) and to move as many deallocations to the background as possible (except on error paths).
2011-03-31 01:13:49 -07:00
Nicholas Nethercote
e2c62437e0 Disable warnings-as-errors, which was introduced by bug 609532. See the bug for justification and follow-up. r=pbiggar. 2011-03-30 17:54:31 -07:00
Gary Kwong
79f0ef8539 Bug 643217 - Replace "ASSERTION FAILED" message with "Assertion failure" in assembler. r=njn
--HG--
extra : rebase_source : b9bd22ee029ebeea7b19327d38305ad1f3aba341
2011-03-31 08:44:57 +08:00
Brian Hackett
8e50870411 Transform RegExp.exec to RegExp.test in native call IC, bug 645889. r=dmandelin 2011-03-30 17:43:36 -07:00
David Mandelin
125dbc8ece Bug 645184: normalize id in addprop IC handler, r=dvander 2011-03-30 16:57:44 -07:00
Gregor Wagner
ee36f9fb20 Fix atomsCompartment assertion. No Bug, r=billm 2011-03-30 14:02:21 -07:00
Jim Blandy
f1e6365c99 Bug 645826: Include standard output and standard error in exception text when testing manifest conditions. r=dmandelin 2011-03-30 12:49:47 -07:00
Jim Blandy
d7f8d15385 Bug 643222: Don't assert that we never cached adds to call objects. r=jorendorff
This assertion should not have been added. Although most adds to call
objects are done by DEFVAR operations, which don't create property cache
entries, the test case shows a situation in which a SETNAME, which is
cached, does the add. Since the object whose property should receive the
value of an assignment's RHS is chosen before the RHS is evaluated, it is
correct to put the value of the 'delete' expression on the call.

It would be nice to ensure that this call object, which we know we are
adding bindings to, had been recognized as extensible at compile time by
the analysis added in bug 554955. However, we only set "extensible parents"
flag on the bindings of calls whose *parents* are extensible, not the
extensible calls themselves. So there isn't anything convenient to check
here.
2011-03-30 12:48:24 -07:00
Jan de Mooij
72e18941b6 [INFER] Don't evict lhs register in jsop_relational_int and jsop_equality, bug 643829. r=bhackett 2011-03-25 13:40:20 -07:00
Jan de Mooij
302719e425 [INFER] Merge JSOP_RSH with jsop_bitop, bug 643109. r=bhackett 2011-03-24 12:37:57 -07:00
Brian Hackett
9849da7614 [INFER] Don't evict existing type register when allocating data register in storeTop, bug 643653. 2011-03-24 12:12:09 -07:00
Paul Biggar
a8948c1eb5 [INFER] Fix typos that were part of the build failure on ARM. (r=redness) 2011-03-23 17:55:51 -07:00
Paul Biggar
2b96612a07 Fix build redness due to warnings-as-errors from parentheses. 2011-03-23 15:09:08 -07:00
Paul Biggar
9caae16818 Fix warnings-as-errors fallout from TM merge (r=redness) 2011-03-23 13:43:16 -07:00
Brian Hackett
d737bb3ddb Merge TM -> JM 2011-03-22 18:07:31 -07:00
Luke Wagner
6a818f48e3 Silence windows warning from 4290338c3956 (r=me) 2011-03-22 17:21:48 -07:00
Brian Hackett
732fe004fb [INFER] Don't immediately evict data register when storing top frame entry, bug 643653. 2011-03-22 16:54:15 -07:00
Luke Wagner
1b7a97d56f Put back nsISSLStatusProvider.idl. Seems like http://mercurial.selenic.com/bts/issue2471 strikes again (r=burning) 2011-03-22 16:52:59 -07:00
Brian Hackett
7933449143 [INFER] Don't store a integer result in a double slot in JSOP_NEG, bug 643913. 2011-03-22 16:06:09 -07:00
Brian Hackett
c834eae239 [INFER] Don't follow scripted caller across compartments, bug 643670. 2011-03-22 15:29:41 -07:00
Brian Hackett
989848101e [INFER] Rejoin during CALLPROP if either compilation did a GETPROP for string accesses, bug 643669. 2011-03-22 14:40:07 -07:00
Brian Hackett
bd9e58285b [INFER] Don't use pic->atom after possibly triggering recompilation in CallProp, bug 643552. 2011-03-22 14:21:56 -07:00
Brian Hackett
0c9d96a072 [INFER] Use type sets instead of known pushed types when looking for static overflows, bug 643376. 2011-03-22 13:41:56 -07:00
Brian Hackett
bad0d7d4a8 [INFER] Fix NULL deref in testSingletonPropertyTypes. 2011-03-22 12:37:01 -07:00
Jan de Mooij
49c9ac30f6 [INFER] Don't push a constant value in JSOP_TYPEOF, bug 643733. r=bhackett 2011-03-22 12:35:05 -07:00
Brian Hackett
bf929fa84d [INFER] Handle recompilations triggered in ICs by obj->lookupProperty, bug 643272. 2011-03-22 12:23:37 -07:00
Brian Hackett
356a520709 [INFER] Disable typed array ICs, bug 643266. 2011-03-22 11:57:15 -07:00
Brian Hackett
c4cfa981d7 [INFER] More robust handling of constant object frame entries. 2011-03-22 05:27:03 -07:00
Jan de Mooij
c633bccba3 [INFER] Inline some native functions, bug 621961. r=bhackett 2011-03-21 22:55:27 -07:00
Luke Wagner
c2d405ba7a Bug 613457 - use *ById JSAPIs in storage (r=sdwilsh)
--HG--
extra : rebase_source : bc2548573f500f664b37b24e95bdc3b9502ba28f
2011-03-21 11:37:47 -07:00
Chris Leary
2ef1aa6439 Bug 642274: Add loose equality to JSAPI. (r=Waldo)
--HG--
extra : rebase_source : 3001604e6e6bfc9c8e73aafc1498a71ae7ade9cd
2011-03-21 10:04:43 -07:00
Andreas Gal
707f928b25 Expose public APIs for converting doubles to signed/unsigned int per ECMA spec (bug 631132, r=bzbarsky). 2011-03-29 19:19:15 -07:00
David Mandelin
30d90dd975 Bug 646304: Fix MSVC warning, r=bent 2011-03-29 19:05:12 -07:00