Commit Graph

27900 Commits

Author SHA1 Message Date
Jon Coppeard
7345e3e7f2 Bug 891966 - 2 - Comment calls to Handle::fromMarkedLocation r=bz 2013-07-23 10:58:28 +01:00
Jon Coppeard
b216cf9603 Bug 891966 - 1 - Don't allow construction of Handle<T> from Heap<T> r=bz 2013-07-23 10:58:27 +01:00
Jon Coppeard
da1bfb1578 Bug 888338 - 5 - post barrier remaining heap-based GC things in XPConnect r=bholley 2013-07-23 10:58:27 +01:00
Jon Coppeard
cde9c3d3cd Bug 888338 - 4 - post barrier CTypes r=terrence 2013-07-23 10:58:27 +01:00
Jon Coppeard
3b22308eee Bug 888338 - 3 - don't call gray root tracer for minor GCs r=terrence 2013-07-23 10:58:26 +01:00
Jon Coppeard
74ca3cb51e Bug 888338 - 2 - give post barrier callback function an extra data argument r=terrence 2013-07-23 10:58:26 +01:00
Jon Coppeard
6c0077c7f6 Bug 888338 - 1 - Add TenuredHeap<T> class r=terrence r=bz 2013-07-23 10:58:26 +01:00
Shu-yu Guo
247953b3b2 Bug 860965 - Part 1: Copy 1D ParallelArray operations to Array. (r=luke,nmatsakis) 2013-05-11 22:39:46 -07:00
Shu-yu Guo
89c3be7c0d Bug 895782 - Fix callsite cloning interaction with inline dispatch in Ion redux. (r=jandem) 2013-07-22 16:12:06 -07:00
Eric Faust
efc59990b1 Bug 895985 - Remove CallInfo lambda data. (r=djvj) 2013-07-22 15:33:19 -07:00
Ryan VanderMeulen
8a4f4ebc64 Backed out changeset 190563720411 (bug 895465) for Linux32 jit-test failures. 2013-07-22 15:23:19 -04:00
Dan Gohman
30044709ad Bug 895465 - IonMonkey: Optimize codegen for test(and(x, y)). r=mrosenberg 2013-07-22 10:36:40 -07:00
Jon Coppeard
15a8352927 Bug 895883 - IonMonkey: (ARM) Use hardware divide instruction for unsigned division where possible r=mjrosenb 2013-07-22 10:14:16 +01:00
Nicholas Nethercote
884fcb4ecd Bug 887558 (part 4) - Introduce WrapperObject and CrossCrompartmentWrapperObject. r=jorendorff.
--HG--
extra : rebase_source : a4749bd143a2c668072bfeefa436f5ada9b8d374
2013-06-27 22:42:30 -07:00
Daniel Holbert
1d993233e4 Bug 896018 - Remove unused field 'context' from CompartmentChecker. r=bhackett 2013-07-19 21:20:29 -04:00
Matt Brubeck
5ee97af944 Bug 896028 - Add a standard Makefile variable for mochitest-metro files [r=gps] 2013-07-19 16:59:23 -07:00
Masatoshi Kimura
f851214c0b Bug 895231 - Make sure the Components shim is consistent with the real Components. r=bholley 2013-07-20 07:53:51 +09:00
Steve Fink
c885385f27 Bug 887420 - Do not read off the end of the transfer map, r=Waldo
--HG--
extra : rebase_source : aae0289af6d58d9ff785afa959accc44342762ea
2013-07-17 15:12:19 -07:00
Luke Wagner
dc8eda6706 Bug 893434 - OdinMonkey: make accessing non-data properties a link-time validation error (r=bbouvier)
--HG--
extra : rebase_source : beb308b3455d0fae18a6d6d5fee744f8b59dc2f2
2013-07-17 17:22:07 -05:00
Hannes Verschore
effbefd655 Bug 891808: IonMonkey: Improve bailing information for JSOP_FUNAPPLY and JSOP_FUNCALL, r=jandem 2013-07-19 13:48:48 -07:00
Jeff Walden
8958d51b99 Bug 895723 - Fix ion/AsmJS.h to not depend on bootlegged dependencies. r=luke
--HG--
extra : rebase_source : 3473fa9c42de65fdb579225c41e534d79c42d7fe
2013-07-19 13:05:11 -07:00
Jeff Walden
e80e3dd197 Bug 895723 - Don't open the mozilla namespace in AsmJS.cpp -- use particular things from it instead. This eases the pain of dealing with js/mozilla namespace conflicts, particularly for Vector. r=luke
--HG--
extra : rebase_source : 666d3e9d340f6c882a21230c63783a72ba25a9e5
2013-07-19 13:05:10 -07:00
Hannes Verschore
0d2f2c46c9 Bug 894582 - IonMonkey: Constant fold arithm operations with double output that fits in int32. r=jandem 2013-07-19 16:21:05 -04:00
Brian Hackett
d1baa91aaf Bug 875125 - Construct worker thread's PerThreadData before spawning actual thread on a CLOSED TREE 2013-07-19 12:57:12 -06:00
Roberta De Viti
2f60adc1ac Bug 894717 - Fix 'plase upgrade' typo in autoconf files. r=khuey 2013-07-19 10:47:15 -07:00
Hannes Verschore
10740fb602 Bug 895019 - Tracelogger: Add logger for minor gc, r=jandem 2013-07-19 10:41:32 -07:00
Brian Hackett
f869ad1378 Bug 875125 - Fix typo. 2013-07-19 11:01:46 -06:00
Hannes Verschore
e9b8bc2c9e Bug 893853 - IonMonkey: Don't set typeset of |this| when type is unknown, r=bhackett 2013-07-19 09:46:23 -07:00
Hannes Verschore
2da4132fa4 Bug 895019: Cleanup and make the tracelogger work again, r=jandem 2013-07-19 09:31:12 -07:00
Terrence Cole
abb6379863 Bug 895654 - Remove references to Unrooted from the static exact rooting analysis; r=bhackett,sfink,terrence
DONTBUILD

--HG--
extra : rebase_source : 32046960a39679cf82916dc7d35ea3374de700f8
2013-07-18 15:12:46 -07:00
Steve Fink
0f2fac7086 Bug 895647 - new GC type: js::LazyScript, r=terrence
DONTBUILD

--HG--
extra : rebase_source : 78d019afe1f46c056c561c2faf4dac351ffb4454
2013-07-18 15:10:23 -07:00
Felix S. Klock II
404372d3dd Bug 894854: env var disabling censoring of self-hosted script frames (r=till)
It is only aenabled for debug builds.  The env var is: MOZ_SHOW_ALL_JS_FRAMES
2013-07-17 09:58:00 +02:00
Gervase Markham
83c0913922 Bug 715549 - remove last vestiges of tri-licence. DONTBUILD. 2013-07-19 16:08:33 +01:00
Douglas Crosher
324fb85dbe Bug 895830 - OdinMonkey: fix compilation of the perf spewer support. r=luke 2013-07-19 20:42:38 +10:00
Sankha Narayan Guria
a0acaa5dce Bug 866847 - Implement map#forEach and Set#forEach. r=evilpie 2013-07-19 08:19:44 +05:30
Brian Hackett
668a5b27fe Bug 875125 - Allow scripts to be parsed/emitted off the main thread, r=billm. 2013-07-19 08:06:02 -06:00
Sean Stangl
1512378066 Bug 868674 - Make supportsFloatingPoint() a static function. r=h4writer
---
 js/src/assembler/assembler/MacroAssemblerARM.h    |  6 +++---
 js/src/assembler/assembler/MacroAssemblerARMv7.h  |  6 +++---
 js/src/assembler/assembler/MacroAssemblerMIPS.h   |  6 +++---
 js/src/assembler/assembler/MacroAssemblerSparc.h  |  6 +++---
 js/src/assembler/assembler/MacroAssemblerX86.h    | 12 ++++--------
 js/src/assembler/assembler/MacroAssemblerX86_64.h |  6 +++---
 js/src/ion/AsmJS.cpp                              |  4 ++--
 js/src/jsapi.cpp                                  |  2 +-
 8 files changed, 22 insertions(+), 26 deletions(-)

---
 js/src/assembler/assembler/MacroAssemblerARM.h    |  6 +++---
 js/src/assembler/assembler/MacroAssemblerARMv7.h  |  6 +++---
 js/src/assembler/assembler/MacroAssemblerMIPS.h   |  6 +++---
 js/src/assembler/assembler/MacroAssemblerSparc.h  |  6 +++---
 js/src/assembler/assembler/MacroAssemblerX86.h    | 12 ++++--------
 js/src/assembler/assembler/MacroAssemblerX86_64.h |  6 +++---
 js/src/ion/AsmJS.cpp                              |  4 ++--
 js/src/jsapi.cpp                                  |  2 +-
 8 files changed, 22 insertions(+), 26 deletions(-)

--HG--
extra : rebase_source : f4ec6b6c9c9f94959311685150b77a63bb5fd9be
2013-07-18 14:45:00 +02:00
Terrence Cole
f6f7d9d82a Bug 894441 - Include WeakMap keys in the DumpHeap report; r=billm
--HG--
extra : rebase_source : 8091813cc3ab2a930a239558a1d9decfc039d9f0
2013-07-16 14:43:30 -07:00
Terrence Cole
190e92587f Bug 895144 - Treat asm.js Functions as long-lived by default; r=luke
--HG--
extra : rebase_source : 6e08bf3d0575d26de3ac6ba41d703168a1f4a429
2013-07-17 16:25:42 -07:00
Terrence Cole
ecd92774b9 Bug 888313 - Update generational GC tuning parameters for better EarlyBoyer performance; r=jandem
--HG--
extra : rebase_source : 0da9dfb64a6e63ff3eade490b0f20cc00528a161
2013-07-18 12:00:10 -07:00
Terrence Cole
61dde7e67c Bug 895177 - Inline object tracing for minor GC; r=jandem
--HG--
extra : rebase_source : 00086cacce4575d04595ca646777c8ce7403a06e
2013-07-17 17:17:43 -07:00
Eric Faust
15139fee40 Bug 895111 - Allow Ion inline checking of functions with lazy scripts. (r=djvj) 2013-07-18 14:03:05 -07:00
Terrence Cole
6b10689c60 Bug 895220 - Do not OOM in AtomizeAndCopyChars if using NoGC; r=bhackett
--HG--
extra : rebase_source : 208b4cb5e652ed7f854586d676e0d055c065b330
2013-07-18 13:55:39 -07:00
Luke Wagner
e55b0aa8d9 Bug 895015 - OdinMonkey: make asm.js entries show up builtin FF profiler (r=bbouvier)
--HG--
extra : rebase_source : 0f682fde81ef8e9dc89f0b9ad76e928d24358394
2013-07-17 17:22:39 -05:00
Brian Hackett
992fab173b Bug 893732 - Ensure indexes are integers when loading/storing to static typed arrays, r=jandem. 2013-07-18 13:33:44 -06:00
Till Schneidereit
17a39c5cdb Backed out changeset 65fd27c97676 (bug 894948) for regressions in Dromaeo DOM-traversal
--HG--
extra : rebase_source : e2bfa1f59ff8ab0a277883eabd67b7aff228907c
2013-07-18 19:58:14 +02:00
Brian Hackett
315234fcfe Bug 893752 - Clean up compartment compiler outputs after releasing all JIT code, r=jandem. 2013-07-18 13:14:20 -06:00
Ryan VanderMeulen
c86305f66e Backed out changeset c1025190b208 (bug 866847) for jit-test failures.
CLOSED TREE
2013-07-18 13:06:48 -04:00
Ryan VanderMeulen
0e1b1ebe69 Backed out 2 changesets (bug 895015, bug 893434) for Android test_asmjs timeouts on a CLOSED TREE.
Backed out changeset cadcb731a8fb (bug 895015)
Backed out changeset 17404ee140f3 (bug 893434)
2013-07-18 13:03:05 -04:00
Andrew McCreight
2fed751bf1 Bug 892405 - Print Max Pause before Total Time in GC logging. r=terrence 2013-07-18 09:07:28 -07:00
Andrew McCreight
6dbb4f19fa Bug 893375 - Make weak map GC tracing names more useful. r=billm 2013-07-18 09:07:27 -07:00
Ehsan Akhgari
c069003a4b Bug 895141 - Remove useless prlong.h inclusions from the tree; r=jcranmer 2013-07-18 12:06:38 -04:00
Ivaylo Dimitrov
9402c54541 Bug 893575 - Don't fail to build if HWCAP_IDIVA is not defined. r=jonco 2013-07-14 19:34:43 +01:00
Sankha Narayan Guria
69084ace53 Bug 866847 - Implement map#forEach and Set#forEach. r=evilpie 2013-07-18 20:37:29 +05:30
Brandon Benvie
b7a0876052 Bug 889627 - Make Debugger eval functions accept options for url and line number. r=jimb 2013-07-17 15:12:24 -07:00
Luke Wagner
4c8b555601 Bug 895015 - OdinMonkey: make asm.js entries show up builtin FF profiler (r=bbouvier)
--HG--
extra : rebase_source : b6b5391a1b39b3abeb0f194bfa4ca4a9f54ce90f
2013-07-17 17:22:39 -05:00
Luke Wagner
3115c9cae9 Bug 893434 - OdinMonkey: make accessing non-data properties a link-time validation error (r=bbouvier)
--HG--
extra : rebase_source : 5b827256536c16d9e2da15e9877efcf393e1a9ef
2013-07-17 17:22:07 -05:00
Tom Schuster
7f21dbc657 Bug 893858 - More testing for CPOWs and fix some small bugs. r=dvander 2013-07-17 09:27:49 -04:00
Jon Coppeard
86aadbaae4 Bug 888237 - IonMonkey: (ARM) Use hardware divide instruction to calculate modulus r=mjrosenb 2013-07-18 14:25:06 +01:00
Jon Coppeard
1c760fa01e Bug 890271 - Assertion failure: operand.isFloatReg(), at js/src/ion/arm/MoveEmitter-arm.cpp:76 r=dvander 2013-07-18 14:25:06 +01:00
Jan de Mooij
4af3e726ee Bug 894852 - Fix LSetPropertyPolymorphicT and LGetPropertyPolymorphicV bailout kind. r=bhackett 2013-07-18 14:22:53 +02:00
Till Schneidereit
74885aba61 Bug 894948 - Remove effectively-dead analysis-related code from interpreter. r=jandem
--HG--
extra : rebase_source : b1cd7b3232a2541d2282609cb698d6c9f825e3be
extra : amend_source : 97483ece12c0750e9fcff906ffdf6321617c106d
2013-07-17 17:51:06 +02:00
Shu-yu Guo
55fa92a274 Bug 892971 - Part 2: Unify names specific to parallel execution to use the 'Par' suffix. (r=nmatsakis) 2013-07-24 03:01:32 -07:00
Shu-yu Guo
cf04eee95a Bug 892971 - Part 1: Refactor calling multimodal VM functions. (r=nbp) 2013-07-24 03:01:32 -07:00
Jan de Mooij
3aa2907b36 Bug 894843 - Fix ConvertLinearSum to not add an unnecessary MAdd(x, 0) instruction. r=bhackett 2013-07-18 10:09:08 +02:00
Mike Hommey
92bf1ab8c2 Bug 894242 - Make C++0x/C++11 mode of gcc/clang mandatory. r=ted 2013-07-17 16:19:25 +09:00
Steve Fink
e9fcd10430 Bug 895027 - s/arguments/scriptArgs/. r=NPOTB
DONTBUILD

--HG--
extra : rebase_source : 44431c33d65a4d5493063968fd9a4bef53db3ebd
2013-07-17 11:08:59 -07:00
Steve Fink
9f34d29b4c Bug 895027 - Add a --source argument to analyze.py to make it easier to run from arbitrary directories. r=NPOTB
DONTBUILD

--HG--
extra : rebase_source : 09bef9ba415f1292646cf3534c1d59cf132667d9
2013-07-17 11:08:35 -07:00
Steve Fink
7a4b796461 Bug 895027 - Use loadRelativeToScript to load helper .js files in static rooting analysis. r=NPOTB
DONTBUILD

--HG--
extra : rebase_source : 49f286d1dd99860be2d163e51063ae8c54c5630d
2013-07-17 11:08:20 -07:00
Steve Fink
75f7eb08db Bug 895027 - Print out commands executed to make it easier to debug individual scripts. r=NPOTB
DONTBUILD

--HG--
extra : rebase_source : 07f5869243cce0a5c684a9e6d2ac0e46dbcd1db6
2013-07-17 11:08:05 -07:00
Steve Fink
b65a7c6294 Bug 895027 - Switch to more accurate set of functions that indicate a GC call for static rooting analysis, r=terrence
DONTBUILD

--HG--
rename : mfbt/Vector.h => js/public/Vector.h
extra : rebase_source : bbadeb34bf976c0a47e35afdc2fba0dffb63da12
2013-06-06 10:37:53 -07:00
Jeff Walden
cb1fe089f2 Bug 895242 - Cast a bool-valued expression to uintptr_t, so that the bitwise-or expression containing it doesn't trigger compiler warnings on Windows. r=trivial 2013-07-17 22:55:55 -07:00
Jeff Walden
70f2069abc Back out 44347cbdf6a0 for triggering jsapi-test orange. r=bustage 2013-07-17 22:55:49 -07:00
Norbert Lindenberg
0be930ad22 Bug 891974 - Test ecma_3/Date/15.9.5.5-02.js fails; partially roll back the buggy commit for bug 890076 to fix. r=jwalden
--HG--
extra : rebase_source : 60a3a1241a98c32e7ffd7b332193bc985b72ac57
2013-07-17 12:21:26 -07:00
Jeff Walden
cfac041780 Bug 894653 - Self-host Error.prototype.toString, and fix an edge-case bug in the current implementation. r=till
--HG--
extra : rebase_source : 0d89b9c899accedd3a23841be850c83e33ec59b6
2013-07-17 12:14:39 -07:00
Jeff Walden
c35ce34f5e Bug 894653 - Make the IsObject self-hosted utility function actually correct. r=till
--HG--
extra : rebase_source : 3cdd7454efa3c5029b3bc908162c1f54547f4c7d
2013-07-16 15:57:57 -07:00
Jeff Walden
10bde77bda Bug 891177 - Implement mozilla/Vector.h, and make js/Vector.h implement js::Vector using mozilla::Vector's implementation of the functionality. r=terrence
--HG--
rename : js/public/Vector.h => mfbt/Vector.h
extra : rebase_source : d5f87a48485e3f2241228a4b003e80974c86fd5f
2013-07-09 16:33:29 -07:00
Jeff Walden
5e2dbd029b Bug 891177 - Remove Vector.h's js/TemplateLib.h dependency by introducing mfbt/TemplateLib.h with the necessary bits. r=terrence
--HG--
extra : rebase_source : e84231171d6bd6c1e2de8201b8c9563375723d01
2013-07-08 12:42:13 -07:00
Jeff Walden
4196fa2e0f Bug 891177 - Remove Vector.h's js/Utility.h dependency. r=terrence
--HG--
extra : rebase_source : d9be649a097b3b9b92cd62685d0fadfa8c0e2bc4
2013-07-03 16:07:43 -07:00
Jeff Walden
f52f37cddc Bug 891177 - Move leading/trailing-zero-bit counting functions, ceiling/floor log2 functions, and round-up-pow2 functions into MathAlgorithms.h. r=terrence
--HG--
extra : rebase_source : 8cfbd68b8cd4a0e21185dd864c7e827ccfa6b751
2013-07-03 15:46:51 -07:00
Benjamin Bouvier
2cf4360796 Bug 894781 - Set implicit truncate for MAbs only for int32 in asm.js mode; r=luke
--HG--
extra : rebase_source : e3bb8353d3e65181e35577f732482a961e88a061
2013-07-17 14:12:43 -07:00
Wes Kocher
cc90ec9c73 Backout 95a4b87a0583 (bug 894948) for causing an octane-crypto regression on a CLOSED TREE 2013-07-17 19:01:03 -07:00
Bobby Holley
09e84333e6 Bug 887334 - Enter a compartment in b2g-only unloading code. r=me CLOSED TREE 2013-07-17 14:28:54 -07:00
Ryan VanderMeulen
7ecd1a4b17 Backed out changeset 76e7ef98a9d1 (bug 867856) for jit-test failures.
CLOSED TREE
2013-07-17 15:54:44 -04:00
Bobby Holley
e96cffb55d Bug 887334 - Reorder some bookkeeping and assert in setCompartment that both the old and new compartments are marked as entered. r=luke 2013-07-17 11:53:55 -07:00
Bobby Holley
1571ff2ce4 Bug 887334 - Require cx->compartment() to be null when destroying a context. r=luke
If it's non-null, that means we've got a JSAutoCompartment on the stack, which
is going to run into trouble when it tries to restore the old compartment on
the now-dead cx.
2013-07-17 11:53:54 -07:00
Bobby Holley
8e0afa92a2 Bug 887334 - Stop setting the compartment to defaultCompartmentObject_->compartment(). r=luke
With this change, defaultCompartmentObject_ is just an opaque (though traced)
piece of embedder state on the cx.
2013-07-17 11:53:54 -07:00
Bobby Holley
9883f9fcf9 Bug 887334 - Fix JSAPI test harness to not fail when the initial compartment is null. r=luke
I don't know why it's done this way, but it sure needs to change.
2013-07-17 11:53:53 -07:00
Bobby Holley
5420505253 Bug 887334 - Miscellaneous JSAutoCompartments. r=luke 2013-07-17 11:53:53 -07:00
Bobby Holley
a93493b71d Bug 887334 - Add a JSAutoCompartment to AutoCxPusher. r=luke,gabor
This should hopefully take care of any cases where consumers expect to be in
the default compartment.
2013-07-17 11:53:52 -07:00
Bobby Holley
e90537d5c6 Bug 887334 - Use the new AutoCompartment overload for the atoms compartment and remove AutoEnterAtomsCompartment. r=bhackett
We have to do some temporary hackiness to deal with some of the new PJS work.
This patch stays as true to the old world as possible, so that we can more
easily backport it.
2013-07-17 11:53:52 -07:00
Bobby Holley
0c5572cc76 Bug 887334 - Don't rethrow in quickstubs if there's already an exception pending. r=luke 2013-07-17 11:53:52 -07:00
Bobby Holley
c8c93ab150 Bug 887334 - Add a JSCompartment* to AutoCompartment and use it in JS_NewGlobalObject. r=luke
This gets rid of the wonky setCompartment usage in that function.
2013-07-17 11:53:52 -07:00
Bobby Holley
aee282ccae Bug 887334 - Move faulty JSAutoRequest in initSelfHosting. r=luke
The call to JS_SetGlobalObject causes cx->compartment_ to be set to the self-
hosting global, which means that the JSAutoCompartment picks up that compartment
as the 'previous' compartment. So despite the attempt to restore things with
JS_SetGlobalObject at the end of the function, the JSAutoCompartment destructor
actually ends up leaving cx in the self-hosting global's compartment at the end
of this function. Moving the JSAutoCompartment construction above the call to
JS_SetGlobalObject fixes the problem.
2013-07-17 11:53:51 -07:00
Bobby Holley
36f005e12f Bug 887334 - Null-check compartment() in JS_GetGlobalForScopeChain(). r=luke
cx->global() assumes a non-null compartment(). When we fix up various bugs
related to being in a compartment when we shouldn't be, we start to crash here.
Fix it.
2013-07-17 11:53:51 -07:00
Bobby Holley
a5bbbb322d Bug 887334 - Rejigger the string manipulation in OnJSContextNew to avoid depending on being in a compartment. r=luke
The current code makes calls that assume (implicitly, via assertions) that |cx|
is in a compartment, which isn't a valid assumption going forward.
2013-07-17 11:53:51 -07:00
Bobby Holley
089d93ca1d Bug 887334 - Remove AutoSwitchCompartment. r=luke
This thing is...yikes.
2013-07-17 11:53:50 -07:00
Nicolas B. Pierron
54c6ec750c Bug 867856 - Fix intermittent timeout of bug743094.js test case. r=h4writer 2013-07-17 11:51:19 -07:00
Nicolas B. Pierron
fbded8c6a4 Bug 892426 - Add an additional temp-Register to SetElementIC. r=jandem 2013-07-17 11:51:19 -07:00
Till Schneidereit
f564136588 Bug 894948 - Remove effectively-dead analysis-related code from interpreter. r=jandem
--HG--
extra : rebase_source : 037764b4c11746ee52d5347c76e2a5242b35afab
2013-07-17 19:21:10 +02:00
Luke Wagner
e4d0a480fa Bug 882008 - OdinMonkey: fix subtraction to take doublish, rm Use (r=bbouvier)
--HG--
extra : rebase_source : 819570ab142518328aa41a1c6b81ffa0a5d7b96f
2013-07-02 11:38:43 -07:00
Luke Wagner
d03c8ecb55 Bug 880538 - LifoAlloc::release after parsing every function (r=bbouvier)
--HG--
extra : rebase_source : 9ead29e3dc5110f2b1f30c5218c41cbbc26555e5
2013-06-21 16:20:47 -07:00
Luke Wagner
de23e334eb Bug 880538 - move 'hasDestructuringArgs' flag into the FunctionBox (r=bhackett)
--HG--
extra : rebase_source : d65025cd591c7faf9e64f4768da0b2b8352ee393
2013-06-28 10:29:58 -07:00
Luke Wagner
be562454f2 Bug 880538 - store a pointer from ParseContext to it's PNK_FUNCTION ParseNode (r=bhackett)
--HG--
extra : rebase_source : acb4b0e5d94d0c1777173a4b0f378c46591805b3
2013-06-28 10:29:58 -07:00
Luke Wagner
e5cfdc10f1 Bug 880538 - set pn->pn_funbox immediately (r=bhackett)
--HG--
extra : rebase_source : 11b265e64f8355989632bd2556ed119a397cb10a
2013-06-21 16:07:25 -07:00
Luke Wagner
1974b8aa53 Bug 880538 - OdinMonkey: don't create an IonContext in MacroAssembler (r=sstangl)
--HG--
extra : rebase_source : c949b8337141681eb983ca4792849a726a059e18
2013-06-28 10:29:57 -07:00
Luke Wagner
f50432075d Bug 880538 - OdinMonkey: make a single pass over the parse tree (r=bbouvier)
--HG--
extra : rebase_source : c5aafde69f44906750d1d62d6878bfbab9e048ee
2013-06-13 11:48:47 -07:00
Ryan VanderMeulen
97083e557a Backed out 4 changesets (bug 893858, bug 868859, bug 886903) for Linux debug mochitest-other crashes on a CLOSED TREE.
Backed out changeset 0ded4854067b (bug 886903)
Backed out changeset ca8afb897e13 (bug 893858)
Backed out changeset c15f503a96cc (bug 868859)
Backed out changeset 6a788f399a7f (bug 868859)
2013-07-17 12:43:59 -04:00
Ed Morley
8887f367e6 Backed out changeset ad74eb485a87 (bug 891695) for startup crashes on a CLOSED TREE 2013-07-17 17:18:11 +01:00
Dan Gohman
75373a534a Bug 891739 - IonMonkey: Make or_ and xor_ range analysis even more precise, and avoid calling js_bitscan_clz32 with a zero operand. r=nbp,waldo 2013-07-17 08:16:05 -07:00
Dan Gohman
4a52fe74a1 Bug 891695 - IonMonkey: Avoid passing pointers by const reference. r=rpearl 2013-07-17 07:32:48 -07:00
Jan de Mooij
7b0eb9b451 Bug 894447 - Avoid unnecessary monitor instructions and bailouts with JSOP_SETELEM accessing named properties and backout bug 894463. r=bhackett 2013-07-17 16:21:22 +02:00
Tom Schuster
78d54d5af2 Bug 893858 - More testing for CPOWs and two small bug fixes. r=dvander 2013-07-17 09:27:49 -04:00
Shu-yu Guo
0af2e2f204 Bug 894782 - Fix callsite cloning interaction with inline dispatch in Ion. (r=jandem) 2013-07-17 06:11:18 -07:00
Brian Hackett
86d8fe99e8 Bug 894669 - Add analysis for finding variables unnecessarily entrained by inner functions, r=luke. 2013-07-16 18:54:47 -06:00
Ryan VanderMeulen
5739ad1b00 Backed out changeset 56423bbe9a13 (bug 891695) because it didn't un-bust itself automagically. 2013-07-17 08:54:01 -04:00
Carsten "Tomcat" Book
d98e2cd471 Relanding after backout Bug 891695 - IonMonkey: Avoid passing pointers by const reference. r=rpearl 2013-07-17 14:16:17 +02:00
Carsten "Tomcat" Book
4fa20cfa2f Backed out changeset 5939320824b1 (bug 894242) "CLOSED TREE" 2013-07-17 10:36:12 +02:00
Mike Hommey
31df024bde Bug 894242 - Make C++0x/C++11 mode of gcc/clang mandatory. r=ted 2013-07-17 16:19:25 +09:00
Mike Hommey
4df35f912c Backout changeset cd67ffb5ca47 (bug 883381), effectively restoring bug 877937 2013-07-17 16:18:46 +09:00
Mike Hommey
521e88720d Bug 894240 - Fix pointer dereference in js::cli::OptionParser::extractValue. r=Waldo 2013-07-17 16:16:55 +09:00
Wes Kocher
85273bb749 Backout 1372b813d76f (bug 891695) for suspicion of causing bustage on a CLOSED TREE
--HG--
extra : amend_source : 9e9d1d7792df7060a2d61d5e8721b67602e14451
2013-07-16 20:47:17 -07:00
Bobby Holley
26926302d8 Bug 889911 - Switch xpcshell to SystemErrorReporter with a little bit of special magic. r=mrbkap
XPCShell currently overrides all the JSContexts whose creation it observes with
its own custom error reporter. This reporter does all sorts of funny things which
we try to clean up for the most part. But there are a few very intricate
considerations at play.

First, the old xpcshell error reporter does some mumbo jumbo with the
XPCCallContext stack to try to guess whether some other code might catch the
exception. This is total garbage on a number of fronts, particularly because
the XPCCallContext stack has no concept of saved frame chains, nested event
loops, sandbox boundaries, origin boundaries, or any of the myriad of
complicating factors that determine whether or not an exception will propagate.

So we get rid of it. But this causes some crazy debugger tests to fail, because
they rely on an exception from uriloader/exthandler/nsHandlerService.js getting
squelched, and can't handle anybody reporting errors to the console service at
the particular moment of contortionism when the exception is raised. So we need
to introduce an explicit mechanism to disable the error reporter here to keep
things running.

Second, we have to be very careful about tracking the return status of the
xpcshell binary. The old code would simply flag an error code if the error
handler was invoked, and we can mostly continue to do that. But there are some
complications. See the comments.

Finally, we don't anything analogous in XPCShellEnvironment, because I have
patches in bug 889714 to remove its state-dependence on the error reporter.
I'll switch it to SystemErrorReporter in that bug.
2013-07-16 20:38:44 -07:00
Bobby Holley
b755828946 Bug 889911 - Replace mozJSLoaderErrorReporter with SystemErrorReporter and remove the former. r=mrbkap 2013-07-16 20:38:44 -07:00
Bobby Holley
a2deff3242 Bug 889911 - Introduce xpc::SystemErrorReporter, roughly based on mozJSComponentLoader's error reporter. r=mrbkap 2013-07-16 20:38:44 -07:00
Benoit Girard
82435a9c18 Bug 838803 - Remove HAVE_STDCALL to work around clang bug. r=glandium
--HG--
extra : rebase_source : e1cdb524a4f7f0742f2051e5915e5e21db4c0423
2013-07-16 22:01:50 -04:00
Wes Kocher
f1fea88196 Backout 399362c3e135 (bug 894669) for bustage 2013-07-16 18:41:52 -07:00
Ryan VanderMeulen
5333a265b8 Backed out 10 changesets (bug 889911, bug 889714) due to merge conflicts on a CLOSED TREE.
Backed out changeset 1a1a536121da (bug 889714)
Backed out changeset 2cd88ef9eea5 (bug 889714)
Backed out changeset 489723887eca (bug 889714)
Backed out changeset 2b38ce22cf97 (bug 889714)
Backed out changeset 87b0a59a5d51 (bug 889714)
Backed out changeset 13229bab2ba4 (bug 889714)
Backed out changeset 234bd6d1fbed (bug 889714)
Backed out changeset 4f5f62284917 (bug 889714)
Backed out changeset 18537c4436c7 (bug 889911)
Backed out changeset ca7060ab1588 (bug 889911)
2013-07-16 21:16:31 -04:00
Bobby Holley
b6f705b27a Bug 889911 - Switch xpcshell to SystemErrorReporter with a little bit of special magic. r=mrbkap
XPCShell currently overrides all the JSContexts whose creation it observes with
its own custom error reporter. This reporter does all sorts of funny things which
we try to clean up for the most part. But there are a few very intricate
considerations at play.

First, the old xpcshell error reporter does some mumbo jumbo with the
XPCCallContext stack to try to guess whether some other code might catch the
exception. This is total garbage on a number of fronts, particularly because
the XPCCallContext stack has no concept of saved frame chains, nested event
loops, sandbox boundaries, origin boundaries, or any of the myriad of
complicating factors that determine whether or not an exception will propagate.

So we get rid of it. But this causes some crazy debugger tests to fail, because
they rely on an exception from uriloader/exthandler/nsHandlerService.js getting
squelched, and can't handle anybody reporting errors to the console service at
the particular moment of contortionism when the exception is raised. So we need
to introduce an explicit mechanism to disable the error reporter here to keep
things running.

Second, we have to be very careful about tracking the return status of the
xpcshell binary. The old code would simply flag an error code if the error
handler was invoked, and we can mostly continue to do that. But there are some
complications. See the comments.

Finally, we don't anything analogous in XPCShellEnvironment, because I have
patches in bug 889714 to remove its state-dependence on the error reporter.
I'll switch it to SystemErrorReporter in that bug.
2013-07-16 18:04:49 -07:00
Bobby Holley
269151d7fe Bug 889911 - Introduce xpc::SystemErrorReporter, roughly based on mozJSComponentLoader's error reporter. r=mrbkap 2013-07-16 18:04:48 -07:00
Dan Gohman
1d9d94c40e Bug 891695 - IonMonkey: Avoid passing pointers by const reference. r=rpearl 2013-07-16 16:00:03 -07:00
Brian Hackett
0bb18aa410 Bug 894669 - Add analysis for finding variables unnecessarily entrained by inner functions, r=luke. 2013-07-16 18:54:47 -06:00
Jan de Mooij
0f3720ceef Bug 852421 - Remove MarkTypeObjectUnknownProperties call from Object.create. r=bhackett
--HG--
extra : rebase_source : efbc9855ea3b548838b7c2b74feac3db987814e5
2013-07-16 21:34:02 +02:00
Till Schneidereit
72188a392a Bug 827396 - rm TypeObject::CONTRIBUTION_LIMIT and TypeObject::contribution. r=bhackett
--HG--
extra : rebase_source : c564216435bafeea58c0bfb32bdae07a3587fa72
2013-07-16 20:42:44 +02:00
Norbert Lindenberg
7cedad89fd Bug 853706 - Fixed a test case that depended on incorrect number of significant digits for 0. r=jwalden
--HG--
extra : rebase_source : a0789030f25b4541d7c3262200f840bda8f10e52
2013-07-16 11:22:57 -07:00
Jeff Walden
6d1aa84530 Bug 853706 - Fix nits noted in review of a new test. r=me 2013-07-16 11:09:56 -07:00
Norbert Lindenberg
6743e38dc0 Bug 853706 - Added new test case for formatting 0 with significant digits. r=jwalden
--HG--
extra : rebase_source : 359a1afdbe2132ebf135809e3d1c121457c5b595
2013-07-16 10:40:33 -07:00
Norbert Lindenberg
fcfda47ad9 Bug 853706 - Enable previously failing conformance test. r=jwalden
--HG--
extra : rebase_source : f855b031368b37a10d819a3913003f8520228a12
2013-07-16 10:40:21 -07:00
Jeff Walden
54c2c4cbfe Bug 894181 - Convert a bunch of SHAPE_* macros to inline functions to eliminate warnings, enhance debuggability. r=terrence
--HG--
extra : rebase_source : 2b8a5ba43a3303470b99788b7637488c38050764
2013-07-16 08:14:57 -07:00
Jeff Walden
b5d72d1f65 Bug 894172 - Eliminate DO_NEXT_OP(len) in favor of an unadorned goto the label in question. This eliminates tautological |len == len| compares that clang+ccache warns about, and it eliminates the previous apparent possibility that any value could be passed to DO_NEXT_OP, when in fact only |len| (or a value equal to it) could be passed. r=terrence
--HG--
extra : rebase_source : 0e1561a7bcabc6811fea6b167251046c2871224d
2013-07-16 08:14:57 -07:00
Boris Zbarsky
6491b478e3 Bug 894463. Go back to deoptimizing vanilla objects and arrays more eagerly to unregress ss-tagcloud. r=bhackett 2013-07-16 13:42:03 -04:00
Jan de Mooij
4dec24de90 Bug 870356 - Fix IonMonkey Math.round bug on x86/x64 hardware without SSE 4.1. r=sstangl 2013-07-16 17:55:29 +02:00
Jeff Walden
2e37cdb786 Convert SN_MAKE_TERMINATOR and SN_IS_TERMINATOR to inline functions to avoid clang+ccache warnings about over-parenthesized equality expressions. No bug, rs=terrence
--HG--
extra : rebase_source : 50b8f70501a481a4dda715284ccd9a21bbfa0eb6
2013-07-15 15:26:04 -07:00
Jeff Walden
f424972f65 Bug 891437 - Implement mozilla/Array.h, a class suitable for use where a C array would be used, with additional debug bounds-checking. r=Ms2ger
--HG--
extra : rebase_source : 314232d817e67d9a26dfa215b5c897e0dd0be353
2013-02-07 15:32:20 -08:00
Ryan VanderMeulen
f01bee06e9 Backed out changeset 302fb81f0729 (bug 858940) for frequent Android test crashes. 2013-07-16 11:05:11 -04:00
Brian O'Keefe
f98dd45a72 Bug 883502 - Part 1: Move 'chromium_config.mk' includes after rules.mk. r=gps 2013-07-04 08:28:43 -04:00
Ryan VanderMeulen
7eb28cd59d Backed out 2 changesets (bug 893362) for check-sync-dir bustage.
Backed out changeset cb79166bb8fa (bug 893362)
Backed out changeset 9db461b687fd (bug 893362)
2013-07-16 09:57:12 -04:00
Ryan VanderMeulen
280d212e0c Bug 893362 - Fix check-sync-dirs bustage. 2013-07-16 09:47:16 -04:00
ISHIKAWA, Chiaki
7420bbd44f Bug 893362 - Align the target name for the START and FINISH in the build log. r=gps 2013-07-16 09:33:22 -04:00
Gabor Krizsanits
c9c9a69587 Bug 874158 - Crash in GetNativeForGlobal. r=bholley 2013-07-16 15:04:28 +02:00