c8fbffcb9d
Bug 899367 - Have nsGlobalWindow go through nsJSContext to access the outer. r=mccr8
2013-09-04 14:06:55 -07:00
d4eab5f3af
Bug 899367 - Explicitly traverse outer windows, rather than doing it via JSContext iteration. r=mccr8
...
Note that this machinery will go away in further patches. But we want to be
deliberate about each of these steps.
2013-09-04 14:06:54 -07:00
6188191d87
Bug 899367 - Explicitly trace outer windows, rather than doing it via JSContext iteration. r=mccr8
2013-09-04 14:06:54 -07:00
7367a3a5b8
Bug 899367 - Only use JSOPTION_UNROOTED_GLOBAL for DOM JSContexts. r=mccr8
...
We don't cycle collect any other kind, so there's no difference between marking
them gray and letting the JS engine mark them black. This also gets rid of that
nasty code which reset the flag, which billm theorizes has to do with faulty
logic in ContextHolder when creating ephemeral sandbox cxes. The assertion in
this patch should catch us if anything goes wrong.
2013-09-04 14:06:54 -07:00
01af9297c7
Bug 883920 - use templates for {Hold,Drop}JSObjects. r=peterv
2013-08-16 13:10:17 -07:00
b8e47c2682
Bug 903419 part 2. Report unhandled rejections in promises. r=smaug,bholley,luke
2013-08-29 00:30:06 -04:00
08a8354b88
Bug 903419 part 1. Change js::GetErrorTypeName to take a JSRuntime, not a JSContext. r=luke
2013-08-29 00:30:06 -04:00
d34342055c
Bug 910291 - Hide TraceMalloc debug functions from content. r=bz
2013-08-29 07:48:43 +09:00
3c5b0c6b29
Bug 909597 (part 2) - Rename jsdbgapi.{h,cpp} as {js,vm}OldDebugAPI.{h,cpp}. r=luke.
...
--HG--
rename : js/src/jsdbgapi.h => js/public/OldDebugAPI.h
rename : js/src/jsdbgapi.cpp => js/src/vm/OldDebugAPI.cpp
extra : rebase_source : 9cd65a698e726c43d88d650f40823367bf98feae
2013-08-26 19:05:20 -07:00
6f0d3838c9
Backed out 12 changesets (bug 899367) for Windows and OSX mochitest failures on a CLOSED TREE.
...
Backed out changeset 46cf4c3eb447 (bug 899367)
Backed out changeset d8a876219fc7 (bug 899367)
Backed out changeset d930333f95a7 (bug 899367)
Backed out changeset efae8cc0fff8 (bug 899367)
Backed out changeset 1dd262d146a6 (bug 899367)
Backed out changeset 4c396b8a51d0 (bug 899367)
Backed out changeset c8c30176639a (bug 899367)
Backed out changeset aaa8fbcf9aaf (bug 899367)
Backed out changeset d1a782044a4b (bug 899367)
Backed out changeset b2672ab55046 (bug 899367)
Backed out changeset fc4deb0b06fa (bug 899367)
Backed out changeset b9f1018a609c (bug 899367)
2013-08-27 21:14:54 -04:00
702e2628d5
Bug 908903 - Add testing function to toggle jit compiler options. r=jandem
2013-08-27 16:32:04 -07:00
3ef78b339d
Bug 899367 - Make GetNativeGlobal consumers call GetWindowProxy and rm the former. r=mccr8
...
Some of the uses of this are a bit nonsensical now, but that's a problem for
another day.
2013-08-27 15:21:40 -07:00
0542144683
Bug 899367 - Remove xpc_UnmarkGrayContext. r=mccr8
...
Default compartment objects are never gray anymore.
2013-08-27 15:21:39 -07:00
0a90738ba7
Bug 899367 - Stop storing the window proxy on the JSContext. r=mccr8
...
We need to simultaneously report the new Heap Pointer to the cycle collector,
and remove the old machinery for cycle-collecting the default object via the
JSContext.
2013-08-27 15:21:39 -07:00
a322d019a4
Bug 899367 - Have nsGlobalWindow go through nsJSContext to access the outer. r=mccr8
2013-08-27 15:21:39 -07:00
d8ec53aa09
Bug 899367 - Explicitly traverse outer windows, rather than doing it via JSContext iteration. r=mccr8
...
Note that this machinery will go away in further patches. But we want to be
deliberate about each of these steps.
2013-08-27 15:21:38 -07:00
a9a3af4e97
Bug 899367 - Explicitly trace outer windows, rather than doing it via JSContext iteration. r=mccr8
2013-08-27 15:21:37 -07:00
ba0f5bdd30
Bug 899367 - Only use JSOPTION_UNROOTED_GLOBAL for DOM JSContexts. r=mccr8
...
We don't cycle collect any other kind, so there's no difference between marking
them gray and letting the JS engine mark them black. This also gets rid of that
nasty code which reset the flag, which billm theorizes has to do with faulty
logic in ContextHolder when creating ephemeral sandbox cxes. The assertion in
this patch should catch us if anything goes wrong.
2013-08-27 15:21:37 -07:00
8955ed2808
Bug 908576 part 3. Get rid of PrototypeIDMap structs. r=smaug
2013-08-23 22:42:41 -04:00
e6289f59db
Bug 907777 - Add preference for off thread parsing, coalesce with preference for off thread Ion compilation, r=billm.
2013-08-22 07:22:33 -06:00
913bab58b0
Merge PGO-green changeset from inbound to m-c.
2013-08-22 10:31:44 +02:00
05f294ce31
Bug 906503 - Remove nsIExceptionProvider; r=bholley
...
It is replaced by a direct special case in the caller.
2013-08-22 08:56:02 +02:00
e5e63a48e9
Bug 906269, don't start a GC cycle so likely while we're in middle of a CC cycle, r=mccr8
...
--HG--
extra : rebase_source : 90d3ad3d19159613f2cb83f0d28dd37a1a429b75
2013-08-22 00:02:12 +03:00
b6f4813baa
Bug 907201 - Enable activating TI and IonMonkey for chrome scripts via about:config. r=jandem
...
--HG--
extra : rebase_source : ee45b0108b4516cff12ecc91014cb566d432a4fa
2013-08-20 11:26:44 +02:00
9d466e12cc
Bug 901106 - Eliminate the nsJSRuntime namespace. r=mrbkap
2013-08-19 16:24:29 -07:00
b1f301a7a7
Bug 901106 - Make nsJSRuntime initialization infallible and do it implicitly in the nsJSContext constructor. r=mrbkap
...
We also move the runtime out of the namespace, which requires some updates.
2013-08-19 16:24:29 -07:00
ae0d0745b3
Bug 901106 - Get rid of vestigial nsJSRuntime instance and make it a namespace. r=mrbkap
...
We'll rename the namespace shortly.
2013-08-19 16:24:29 -07:00
c334ce29a3
Bug 901106 - Make sure sure the nsDOMScriptObjectFactory gets instantiated.
...
This used to happen implicitly when we created the first nsJSContext, but that
goes away in the next patch. Without this, we get a couple of odd exception-
related failures in the IPC indexedDB tests.
2013-08-19 16:24:28 -07:00
b4dde9a0ba
Bug 901106 - Remove nsIScriptRuntime. r=mrbkap
...
All this stuff is now for nsGlobalWindow's use only.
2013-08-19 16:24:28 -07:00
9ccecb62c8
Bug 901106 - Inline and remove nsIScriptContext::{Serialize,Deserialize}. r=smaug
2013-08-19 16:24:27 -07:00
8812f84312
Bug 884410 - Remove JS_ValueToECMAInt32 and replace it with JS::ToInt32. r=terrence
2013-08-15 17:41:51 -04:00
1ed269270a
Backed out changeset 3d9664b78ca7 (bug 884410) because of build bustage on a CLOSED TREE
2013-08-15 18:20:00 -04:00
63f6d15be9
Bug 884410 - Remove JS_ValueToECMAInt32 and replace it with JS::ToInt32. r=terrence
2013-08-15 17:41:51 -04:00
99058b577c
Bug 905353 - Minimize the #includes in dom/base; r=jst
...
--HG--
extra : rebase_source : e7a1d2262bf6d8c9b78057b8eb6cdbfd98262f68
2013-08-15 14:17:48 -04:00
dc64810021
Bug 905300 - Remove dead code in SpiderMonkey. r=bhackett
2013-08-15 13:45:30 +02:00
27a832e507
Bug 903212 - Remove explicit compartment GC machinery. r=smaug
2013-08-14 14:00:03 -07:00
03c1b0ddf4
Bug 903212 - Remove nsIXPCScriptNotify machinery. r=mrbkap
2013-08-14 14:00:03 -07:00
0846601377
Bug 903212 - Remove ScriptEvaluated, and simplify callers. r=mrbkap
2013-08-14 14:00:03 -07:00
4c9559f8e1
Bug 903212 - Remove activity tracking from nsJSContext. r=smaug
...
This only has an impact when we enable explict compartmental GCs, which have
been preffed off for a while.
2013-08-14 14:00:02 -07:00
3ed72e3e52
Bug 901364 - Remove a bunch of now-unused modal state tracking machinery. r=mrbkap
2013-08-12 12:54:51 -07:00
80de28966e
Bug 901364 - Make operation callbacks runtime-wide. r=mrbkap
2013-08-12 12:54:50 -07:00
c73211b3fb
Bug 901364 - Switch to the XPConnect operation callback for DOM windows. r=mrbkap
2013-08-12 12:54:49 -07:00
3b168e2acb
Bug 901364 - Hoist slow script prompt machinery into nsGlobalWindow. r=mrbkap
2013-08-12 12:54:49 -07:00
5ec2086f99
Bug 898263 (part 4) - Move |JSTrapStatus| and related function typedefs out of jsprvtd.h. r=jorendorff.
...
--HG--
extra : rebase_source : 68c1d3aaaa40763e073ee791c1d0078c5d5cf87e
2013-08-08 21:30:15 -07:00
a4f15f4503
Bug 901362 - Refactor nsJSUtils::EvaluateString to take an EvaluateOptions structure. r=gabor
2013-08-08 16:51:35 -07:00
d7fb701374
Bug 901162 - Inline ExecuteScript in the one caller. r=bz
2013-08-08 16:51:34 -07:00
ad01da34e4
Bug 901162 - Inline CompileScript into the one caller. r=bz
2013-08-08 16:51:34 -07:00
1315953921
Bug 901162 - Hoist EvaluateString into nsJSUtils. r=bz
2013-08-08 16:51:34 -07:00
f2b057f9f0
Bug 901162 - Remove unused mExecuteDepth machinery. r=bz
2013-08-08 16:51:34 -07:00
d8f824e34e
Bug 898914 (part 1) - Remove JSBool.
...
--HG--
extra : rebase_source : 2d202e0e5005a7f54724b1540481c15cde3ad52e
2013-08-08 15:53:04 -07:00
8ce53f9ffd
Bug 902332 - Replace JS_{FALSE,TRUE} with {false,true} almost everywhere. r=luke,bz.
...
--HG--
extra : rebase_source : 25f4de1bfae830b0af6407d260a70b787ab1dc9b
2013-08-06 23:59:54 -07:00
9624be5713
Bug 901750 - Change return type of |JSNative| from |JSBool| to |bool|. code=nnethercote,jandem. r=luke,bz.
...
--HG--
extra : rebase_source : 5b3d7cc339af6d93bde078322e25c6e740d1b617
2013-08-02 00:41:57 -07:00
50634577d5
Merge m-c to inbound.
2013-08-02 11:07:57 +02:00
0f9d038985
Bug 897945 - Use an XPCOM string for nsContentUtils::ReportToConsole's category argument; r=bholley
2013-08-02 09:09:11 +02:00
2e729aa83d
Bug 881323 - Re-implement CycleCollectorParticipant with actual vtables, with constexpr to avoid static initializers. r=mccr8
2013-08-02 10:29:05 +09:00
b4ad5a68ec
Bug 899812 - Error reporters should consistently report errors to stderr (r=bholley)
2013-07-31 17:04:49 -07:00
98a0425f0f
Bug 899245 - Rename JS_GetGlobalForScopeChain to JS::CurrentGlobalOrNull. r=luke
...
This was done with perl/grep.
2013-07-29 16:45:27 -07:00
43974de990
Bug 899245 - Rename JS_SetGlobalObject and move it into jsfriendapi. r=luke
2013-07-29 16:45:26 -07:00
acaaa3b1a2
Bug 899245 - Rename js::GetDefaultGlobalForContext to js::DefaultObjectForContextOrNull. r=luke
...
This is more correct, since the object very often is not a global.
2013-07-29 16:45:25 -07:00
ae46376c2e
Bug 897433 - Telemetry for SnowWhite and more async SnowWhite freeing (patch v4), r=mccr8
2013-07-27 13:48:45 +03:00
496513c7cd
Backed out changeset 300f5b7d72e1 (bug 897433) for intermittent Linux dromaeo crashes.
2013-07-25 21:18:20 -04:00
e05c81b4a4
Bug 897433, Telemetry for SnowWhite and more async SnowWhite freeing, r=mccr8
...
--HG--
extra : rebase_source : 487139a95f50e69d6273c2870ed6623f175544e7
2013-07-25 22:56:20 +03:00
050814f5b7
Bug 895340 - Fix error reporting of JS exceptions from C++ (r=bholley)
2013-07-23 08:32:36 -07:00
a0d710b2dd
Bug 895578 - Make SnowWhiteKiller to use fallible TArray, r=mccr8
2013-07-19 15:53:16 +03:00
055a4a328f
Bug 887334 - Miscellaneous JSAutoCompartments. r=luke
2013-07-17 11:53:53 -07:00
b002b30b2a
bug 789919, (snow-white) make addref/release of CCable objects faster by removing indirect refcnt increase/decrease, r=mccr8, test changes r=ehsan
...
--HG--
extra : rebase_source : 2a3b22425c14d6daedc91d62a652c34431acd2fb
2013-07-09 13:30:58 -04:00
e95fdd5ba8
Bug 890636 - Remove JSOPTION_PCCOUNT, r=jandem.
2013-07-08 09:17:35 -06:00
919587b427
Bug 860085 - Stop using XPConnect::ReleaseJSContext in nsJSEnvironment::DestroyJSContext. r=gabor,mccr8
...
We now have the invariant that any in-use cx must be pushed onto the JSContext
stack with one of our stack-scoped automatic nsCxPusher classes. These classes
hold a strong ref to the nsIScriptContext associated with the JSContext they
push (if any). This means that, if this cx is in use, we will always have at
least one strong reference to the nsJSContext coming from the stack, meaning
that neither the destructor nor the Unlink() implementation will be called.
So we don't need to do any deferred destruction of the cx anymore.
2013-07-03 11:05:19 -06:00
5fcbe40dc3
Bug 887125 - Allow 0 for Memory GC Preferences. r=bent
2013-06-27 14:40:17 +08:00
9d5fdcde72
Bug 829841 - Add an option to enable eager compilation in the browser. r=dvander
2013-06-24 05:13:05 -07:00
6bab803585
Bug 875790 - Add cx push back into nsJSContext::CompileScript. r=me
2013-06-21 09:10:27 -07:00
2e3db4ed09
Bug 871005 - Add a pref for GC decommit threshold (r=gregor)
2013-06-20 18:06:53 -07:00
3b1eda0cea
Bug 883450 - Miscellaneous requests. r=gabor
2013-06-20 11:05:34 -07:00
5b896f6665
Bug 877762 - GC: Post-barrier cycle collector participants - 5 Convert JS::Value to use Heap<T> r=smaug
2013-06-18 11:00:37 +01:00
defbf4358c
Bug 823978 - Rename JSOPTION_STRICT to JSOPTION_EXTRA_WARNINGS. r=Waldo
...
--HG--
extra : rebase_source : 34f95fc6017d3f3f4f091083debacbd466b51bf6
2013-06-11 12:49:11 -07:00
48564710fc
Backed out changeset 1df122edcf0d (bug 823978) for bustage (again).
...
CLOSED TREE (again)
2013-06-11 15:27:46 -04:00
859ae9d3b2
Bug 823978 - Rename JSOPTION_STRICT to JSOPTION_EXTRA_WARNINGS. r=Waldo
...
--HG--
extra : rebase_source : 1396c8ec5bd99ab35343cfff2b87d22955671243
2013-06-11 11:30:18 -07:00
a580723853
Backed out changeset 7e69e102a432 (bug 823978) for bustage.
...
CLOSED TREE
2013-06-11 14:28:17 -04:00
6a8ee25998
Bug 823978 - Rename JSOPTION_STRICT to JSOPTION_EXTRA_WARNINGS. r=Waldo
...
--HG--
extra : rebase_source : f5c0e9c743dadb0431eacef6b46b6e274264d3aa
2013-06-11 10:55:21 -07:00
3eabb6ea4b
Bug 876805. Fix unsafe reference gc hazards in dom/ code. r=smaug
2013-05-29 16:16:04 -04:00
cd38f4feb9
Merge the last PGO-green cset on m-i to m-c
2013-05-27 09:44:40 -07:00
f7a96e327a
Bug 875790 - Revert AutoPushJSContext change in nsJSContext::CompileScript due to crashes. r=bholley
2013-05-27 06:40:37 -07:00
a07d2309af
Bug 874691 - Make CC participant tracing indirect r=mccr8 r=smaug r=bz
2013-05-27 12:50:49 +01:00
8f823bfe9a
Bug 868110 - Introduce a more private API for the consumers that really do need the default compartment object. r=luke
...
The primary consumer of this is the whole inner/outer DOM window setup, which
uses the default global to track the current inner. But there are few other
random ones as well.
We use this as an opportunity to convert a bunch of consumers from the two-step
GetNativeContext() -> JS_GetGlobalObject() into just |GetNativeGlobal()|. This
will make things much easier to convert when we start tracking the current inner
explicitly.
2013-05-22 15:42:44 -06:00
f2943090bd
Bug 868130 - Remove the lion's share of JSAutoRequests in gecko. r=gabor
...
There are still a handful that either are used with other runtimes, or that
happen very early/late in cx the lifetime of various things where it doesn't
necessarily make sense to have a cx on the stack. This should definitely ensure
that we're not doing double-duty with the nsCxPusher change, though.
2013-05-22 10:05:28 -06:00
cdde50556c
Bug 868130 - Include nsCxPusher.h everywhere we need it, and stop including it from nsContentUtils.h. r=gabor
2013-05-22 10:05:26 -06:00
750d6364fd
Bug 868130 - Make sure mContext is stack-top in the various places where nsJSEnvironment uses it. r=gabor
...
I also replaced some uses of mJSContext with AutoJSContext upon auditing that
they're only ever called by other nsJSContext functions that just pushed. This
makes the code much easier to audit because we know that we're just inheriting
the caller's cx (which is stack-top) rather than injecting a potentially-
unrelated cx.
2013-05-22 10:05:24 -06:00
e7d2092dbd
Bug 868130 - Make sure mContext is stack-top in nsJSContext::CompileScript. r=gabor
2013-05-22 10:05:23 -06:00
162bdc21e4
Bug 857845 part 3 - Remove JM JSAPI flags, memory reporters and browser prefs. r=djvj
...
--HG--
extra : rebase_source : 8b2ed85801fcd8c8fd0fc16a89b1a275add1a310
2013-05-22 14:50:45 +02:00
c1533a7e3b
Bug 841312 - Remove the termination function API. r=bz
...
\o/
2013-05-21 11:45:56 -06:00
cce2e96e58
Bug 860941 - Separate the handling of |dialogArguments| and |arguments|, and use IDL for the |dialogArguments| getter. r=jst
...
This patch is bigger than I'd like it to be, but there are a lot of interlocked
dependencies and I eventually decided it was easier to just lump it together.
The semantics of |showModalDialog|/|window.dialogArguments| (an web-exposed
HTML5 feature) and |openDialog|/|window.arguments| (a XUL-proprietary feature)
are quite different. The former is essentially a security-checked JSVal, while
the latter gets converted into an array. We handled them together in the old
world, which led to a lot of confusion and muddled semantics. This patch
separates them.
This patch also eschews the roundabout resolve hook for dialogArguments in favor
of returning them directly from the XPIDL getter. This better matches the
behavior in the spec, especially because it allows dialogArguments to live on
the outer as they're supposed to, rather than the first inner that happens to
end up in the docshell. All in all, this should make this all very
straightforward to convert WebIDL when the time comes.
The current spec on the origin checks here is pretty fictional, so I've filed
https://www.w3.org/Bugs/Public/show_bug.cgi?id=21932 to fix it. This patch
should more or less preserve the current security behavior.
2013-05-17 10:43:19 -07:00
8765722e58
Bug 865320 - Move the logic for deciding when to doing a merging CC into the cycle collector. r=smaug
2013-04-29 16:41:41 -07:00
4521aa4875
Backed out changeset 0ad1a92ca568 (bug 860941)
2013-05-17 09:08:44 +01:00
7b7f24f6d5
Bug 860941 - Separate the handling of |dialogArguments| and |arguments|, and use IDL for the |dialogArguments| getter. r=jst
...
This patch is bigger than I'd like it to be, but there are a lot of interlocked
dependencies and I eventually decided it was easier to just lump it together.
The semantics of |showModalDialog|/|window.dialogArguments| (an web-exposed
HTML5 feature) and |openDialog|/|window.arguments| (a XUL-proprietary feature)
are quite different. The former is essentially a security-checked JSVal, while
the latter gets converted into an array. We handled them together in the old
world, which led to a lot of confusion and muddled semantics. This patch
separates them.
This patch also eschews the roundabout resolve hook for dialogArguments in favor
of returning them directly from the XPIDL getter. This better matches the
behavior in the spec, especially because it allows dialogArguments to live on
the outer as they're supposed to, rather than the first inner that happens to
end up in the docshell. All in all, this should make this all very
straightforward to convert WebIDL when the time comes.
The current spec on the origin checks here is pretty fictional, so I've filed
https://www.w3.org/Bugs/Public/show_bug.cgi?id=21932 to fix it. This patch
should more or less preserve the current security behavior.
2013-05-16 23:33:14 -07:00
bd907d5829
Bug 868312 - Rooting fixes for dom r=bz
2013-05-12 01:17:42 -04:00
480bf64980
Bug 868312 - Rooting fixes for dom r=bz
2013-05-10 14:22:21 -04:00
b287cc2cb5
Bug 868312: Root dom/workers r=bz
2013-05-09 03:27:40 -04:00
768211fa86
Bug 868635 - Make merging CC heuristics go through scx. r=mccr8
2013-05-07 14:18:03 -07:00
262b462c83
Bug 868184 - OdinMonkey: enable by default on beta/release (r=vlad)
...
--HG--
extra : rebase_source : c97bf37723eded74f95aae21e2baa692cf059c2e
2013-05-07 13:34:17 -07:00
31412343f1
Bug 868122 - Remove nsIScriptContextPrincipal. r=gabor
...
\o/
2013-05-06 16:53:10 -07:00
Bobby Holley
Andrew McCreight
Boris Zbarsky
Masatoshi Kimura
Nicholas Nethercote
Ryan VanderMeulen
Nicolas Pierron
Brian Hackett
Ms2ger
Olli Pettay
Till Schneidereit
Tom Schuster
Ehsan Akhgari
Jan de Mooij
Mike Hommey
Bill McCloskey
Gregor Wagner
Nicolas B. Pierron
Jon Coppeard
Benjamin Peterson
Phil Ringnalda
octoploid@yandex.com
Ed Morley
David Zbarsky
Luke Wagner