Commit Graph

16031 Commits

Author SHA1 Message Date
Chris Leary
0371422967 Bug 712169: Make aligned/unaligned LifoAlloc consistent. (r=luke)
--HG--
extra : rebase_source : cff51d01e46f5b09d642368da532b9b6a8649053
2011-12-20 14:19:45 -08:00
Michael Wu
37d2a6827a Bug 701662 - Serialize filenames in JS_XDRFunctionObject, r=igor 2011-12-20 11:31:45 -08:00
Ed Morley
5676601bd2 Bug 709792 - Remove deprecated --enable-tracejit; r=khuey 2011-12-20 12:19:36 +00:00
Phil Ringnalda
a30061699c Back out 316d6a49a603 and 98617f7b667b (bug 624621) for jsreftest failures 2011-12-19 18:40:16 -08:00
Boris Zbarsky
9e927dd338 Bug 624621 part 3. Use the pre-redirect filename as the script filename and the channel principal as the origin principal, and base our cross-origin check on the origin principal. r=mrbkap 2011-12-19 12:48:12 -05:00
Igor Bukanov
d3409cbe70 bug 711623 - rt->gcNumFreeArenas is not updated properly. r=terrence
--HG--
extra : rebase_source : 5abd5e5c3ce789e912aa63ef21001d70893912b7
2011-12-19 23:07:24 +01:00
Andrew McCreight
70623bf5d3 Bug 710492 - add special cycle collector shape tracing path. r=bhackett 2011-12-19 10:24:56 -08:00
Tom Schuster
1f4f431168 Bug 708819 - Avoid recursion through Error.prototype.toString/toSource r=luke
--HG--
extra : rebase_source : 7e6437b497886bbd8109a6c8a392923fff537c84
2011-12-19 14:17:00 +01:00
Tom Schuster
46fa50e1b0 Bug 708695 - Fix issues found by clang's scan-build. r=Waldo
--HG--
extra : rebase_source : 5b1970539b2f5b211ebeb97765d8c549371cb888
2011-12-19 14:16:53 +01:00
Jeff Walden
95c80127e7 Bug 711775 - Implement MOZ_NORETURN to encapsulate this-function-doesn't-return semantics. r=cjones
--HG--
extra : rebase_source : d56f3ef47e11c42b5fe6a442e13fd6202208914f
2011-12-17 16:45:29 -05:00
Jeff Walden
6bb9efafe4 Bug 711672 - Break mfbt's dependency on jstypes.h by moving various API macros out of JS and into mfbt. r=luke, r=cjones
--HG--
extra : rebase_source : f18993e1eec1a5e5bcdb9cd97fa7d7c6ad261e2c
2011-12-13 14:26:58 -05:00
Nicholas Nethercote
41607c5826 Bug 709653 - Fix GC heap memory reporters. r=bhackett.
--HG--
extra : rebase_source : 6805671750bd6a9e9661e70832a3837d4064f6ea
2011-12-11 21:21:18 -08:00
Nicholas Nethercote
2fa63e0548 Bug 707865 - Convert nsTArray::SizeOf() to nsTArray::SizeOfExcludingThis(). r=jlebar.
--HG--
extra : rebase_source : d802d58bc7dedda2490878793923adc0ab55f779
2011-12-15 14:59:53 -08:00
Jeff Walden
24ef5dcc5b Bug 711672 - Remove DEFINE_LOCAL_CLASS_OF_STATIC_FUNCTION as it's unused. r=luke 2011-12-13 14:26:57 -05:00
Ralph Giles
1ea8cea5fd Bug 697881 - Fall back to gcc on darwin if there's no gcc-4.2. r=ted
Since Bug 513353, we've defaulted to gcc-4.2 on darwin.
Unfortunately the default was unconditional, and while
XCode 4.1 provided a gcc-4.2 alias, XCode 4.2 does not,
although 'gcc' is llvm-gcc 4.2.1. As such, the configure
script immediately fails.

The patch still prefers gcc-4.2, but falls back to plain
gcc if it's not available.

Note that we try $CC and $CXX from the environment first
so it is still possible to override this default.
2011-12-18 20:20:28 +01:00
Ms2ger
6464cec1ae No bug - Remove unused JS::Value::setObjectOrUndefined; r=luke 2011-12-18 11:06:43 +01:00
John Schoenick
34b5116fdd Bug 687679 - Rename JS_CompileFile to JS_CompileUTF8File, fix users. r=jorendorff,jst 2011-12-16 11:08:59 -08:00
Jeff Walden
52e1a4c28f Bug 711611 - NSPR's obsolete int32, int64, etc. types get defined without an opt-out #define -- add that #define to SpiderMonkey so int32/etc. can't be (mis-)used. r=bhackett
--HG--
extra : rebase_source : 18c73b832cb9f4587c4a1114f237edcaba657199
2011-12-16 17:03:07 -05:00
Bobby Holley
c5d0be548b Bug 711288 - Unconditionally use the new value of JSPROP_READONLY in accessor -> data transformations. r=Waldo 2011-12-16 14:32:39 -08:00
Bill McCloskey
45769c3f0c Bug 710947 - Allow hash table elements to be removed during GC (r=luke) 2011-12-16 14:01:44 -08:00
Brian Hackett
18192c5284 Fix lingering uses of int32 types, no bug. 2011-12-16 13:21:02 -08:00
Brian Hackett
3983e0f31e Add interface for accessing PC counter information from chrome code, bug 687134. r=sfink,waldo 2011-12-16 13:11:08 -08:00
Ms2ger
b6f42e130c Backed out changeset ba447ace2594 (bug 687679) for bustage. 2011-12-16 21:17:16 +01:00
John Schoenick
d2b02fa02e Bug 687679 - Rename JS_CompileFile to JS_CompileUTF8File, fix users. r=jorendorff,jst 2011-12-16 11:08:59 -08:00
Bill McCloskey
d39ab52638 Bug 711158 - Avoid triggering write barrier during verifier (r=luke) 2011-12-16 10:31:56 -08:00
Jeff Walden
a7d6a6d6e8 Bug 708735 - Really remove the uint32/JSUint32 typedefs now that all users have been fully corrected. r=luke, r=dmandelin from the original change, landing in a CLOSED TREE (which can open now!) 2011-12-16 10:00:08 -05:00
Jeff Walden
369c8705b9 Backed out changeset f879c54dffae: 5efcb9c3b375 introduced more JSAPI users that need fixing for the <stdint.h> types, and it's not worth trying to do fixups for that commit now while holding the tree closed. uint32/JSUint32 type removals will occur later today, after I reopen up a CLOSED TREE. 2011-12-16 08:20:21 -05:00
Jeff Walden
c1e1e8864c Bug 708735 - Remove the non-<stdint.h> {u,}int{8,16,32,64} and JS{Uint,Int}{8,16,32,64} integer types from the JSAPI. The public header added in the previous commit with the old typedefs should ease migration for embedders. r=luke, r=dmandelin, pushing into a CLOSED TREE so the landing is smooth 2011-12-15 21:51:08 -05:00
Jeff Walden
476c7d19c5 Bug 708735 - Use <stdint.h> types in JSAPI and throughout SpiderMonkey. Continue to provide the {u,}int{8,16,32,64} and JS{Uint,Int}{8,16,32,64} integer types through a single header, however, for a simpler backout strategy -- and also to ease the transition for embedders. r=timeless on switching the jsd API to use the <stdint.h> types, r=luke, r=dmandelin 2011-12-08 22:54:10 -05:00
Ed Morley
3691e9b488 Bug 710486 - Remove unnecessary mkdir objdir/js/src/nanojit; r=ted 2011-12-16 09:13:29 +00:00
Jim Blandy
0f12451749 Bug 708156: Ensure that JM compilation doesn't used out-of-date ScriptAnalysis structures. r=bhackett 2011-12-15 18:08:00 -08:00
Terrence Cole
8eb45701fc Bug 710946 - Add write barriers to getterObj and setterObj; r=billm
These fields are immutable, so they do not necessarily need a barrier for
incremental collection -- they do, however, need a barrier for generational
collection, and for reliable verification of the barriers.  It would be nice to
make these HeapPtr's, rather than manually calling the barrier functions, but
the fields are unioned with a function pointer, so this is not trivial.

--HG--
extra : rebase_source : f1f33198317e28fef3cc777f3e1e84725ba9cdc5
2011-12-15 18:03:20 -08:00
Andrew McCreight
541dcff9f2 Bug 680937, part 2 - Add native wrapper preservation hook, call it in WeakMap_set. r=billm 2011-12-15 17:45:21 -08:00
Andrew McCreight
bdf4ad1f26 Bug 680937, part 1 - Don't always mark wrapped native keys. r=billm 2011-12-15 17:44:34 -08:00
Terrence Cole
6191221385 Bug 711266 - JSRope::flatten needs unconditional post barriers; r=billm
For incremental barriers we check if compartment()->needsBarrier to defer the
cost when we are not in a GC cycle.  We cannot do this for cross-generation
barriers.  We need to remove the checks on the post barriers and rename the enum
to something  more specific.
2011-12-15 17:34:59 -08:00
Terrence Cole
8052b64a6e Bug 710950 - JSScript::function_ should be barriered; r=billm
The field is immutable, so we can return an unwrapped pointer and don't have to
worry about it for incremental barriers.  We do need post barriers on it during
init for detecting cross-generation pointers, however.  Just make it a normal
HeapPtr.

--HG--
extra : rebase_source : a52d3d74ab5bf3ae926b9da5212825ce1c0ae93d
2011-12-15 15:27:58 -08:00
Felix Fung
14c8124b0a Bug 710996 - Double-assignment in JS_GetValidFrameCalleeObject(). r=jorendorff 2011-12-15 00:28:11 -08:00
Igor Bukanov
908fd20e0d bug 711095] - make GC scanning compatible with the incremental GC. r=billm
--HG--
extra : rebase_source : 5328784491f322824977ba80dd302cc13e84520a
2011-12-15 21:07:45 +01:00
Igor Bukanov
eee1ff8508 bug 710581 - fixing GCC 4.6 warning. r=bhackett
--HG--
extra : rebase_source : ae76549e73945844a75c1d518682b5db4af4cd1a
2011-12-14 20:51:56 +01:00
Terrence Cole
073e795eed Bug 711165 - DenseArray are missing some barrier calls; r=billm
The copy path was caught, the init path was not.  This does not matter for
incremental barriers, but will for cross-generation barriers.

--HG--
extra : rebase_source : 4fa7c4cc860cd9265ccd598340941840ee4e23e9
2011-12-15 15:27:32 -08:00
Brian Hackett
3df3c83b7c Correctly compute fun.caller for cloned functions, bug 710780. r=luke 2011-12-15 14:09:49 -08:00
David Mandelin
c4568252b9 No bug, remove file needed only for nanojit landings 2011-12-15 14:07:59 -08:00
Chris Leary
0afa159f1a Bug 709622: Conditional with empty consequent causes strict strict warning. (r=Waldo) 2011-12-13 13:22:37 -08:00
Luke Wagner
80d107697c Bug 703645 - rm SRC_LOOPHEAD (r=waldo)
--HG--
extra : rebase_source : 9301aa0af4e81a8725f05c702145c89a4ece4db6
2011-11-22 11:37:57 -08:00
Luke Wagner
83117ca3ba Bug 704478 - Fix single-threaded runtime assertions in shell workers (r=jorendorff)
--HG--
extra : rebase_source : 13deb300ee2aa3d5518549d4f296b276a0da4e42
2011-11-22 11:37:57 -08:00
Luke Wagner
17636ec3af Bug 704351 - Fix ValueToScript in js shell (r=waldo)
--HG--
extra : rebase_source : 0c072081ed9699017ddae80404c24562c6de3536
2011-11-22 11:37:57 -08:00
Luke Wagner
e77e503d98 Bug 705879 - Avoid touching stale generator frame in js_CloneBlockObject (r=bhackett)
--HG--
extra : rebase_source : be217f5a36fb69c37fe09e41f34b7f031a3e8224
2011-12-07 17:14:35 -08:00
Luke Wagner
a5fc0b3da5 Bug 624621 - Allow embeddings to attach an extra 'origin' principal to scripts (r=mrbkap)
--HG--
extra : rebase_source : 85ed15c9f230bc342e7810e62abace18a7177c4b
2011-06-27 17:42:53 -07:00
Bobby Holley
ad05f2e304 Bug 702491 - Don't set JSPROP_READONLY for accessor properties. r=Waldo 2011-12-15 11:40:57 -08:00
Bill McCloskey
bc4234efe8 Bug 708303 - Back out 1deb23332fb5 for orange 2011-12-15 10:30:09 -08:00