Граф коммитов

2548 Коммитов

Автор SHA1 Сообщение Дата
brendan%mozilla.org 37094600f5 Forgot to remove JSPROP_ENUMERATE in one case for bug 293782 (thanks to jwatt for catching this). 2005-06-14 16:20:27 +00:00
brendan%mozilla.org b7b9a35bd5 Better fix for 296397 (r=shaver, sr=jst). 2005-06-11 01:33:44 +00:00
brendan%mozilla.org ac081d05f9 Remove JSPROP_ENUMERATE from function args/vars hidden property attributes (followup for 293782). 2005-06-09 23:45:54 +00:00
brendan%mozilla.org 7cacacf414 Check access when accessing function and regexp __proto__, etc. (296397, r=shaver, a=me). 2005-06-09 01:02:36 +00:00
brendan%mozilla.org 23ed3dd111 Followup fix for 255555 (r=mrbkap, a=me). 2005-06-08 18:08:53 +00:00
mrbkap%gmail.com 87a6415a5e bug 255555: undefined and missing values are passed to Array.sort()'s comparison function. r+a=brendan 2005-06-08 16:38:38 +00:00
brendan%mozilla.org 5f76fd552e Fix bugs 296772 and 262948 (r=shaver, sr=jst, a=me). 2005-06-08 02:13:10 +00:00
mrbkap%gmail.com c2553b4dfa bug 294195: Remove an old dependence on null-terminated strings in favor of an end-pointer to fix a crash trying to interpret regexp dollar backrefs. r+a=brendan 2005-06-07 22:16:48 +00:00
brendan%mozilla.org dd66686ab6 Update comment to track tricky AVL tree reference. 2005-06-07 16:22:58 +00:00
jst%mozilla.jstenback.com c22744cdd3 Fixing tinderbox orangeness (Thunderbird) and bug 296467. Get the function object off of the stack frame through fp->argv[-2] and not through fp->fun->object as the latter may be null if the function on the frame is a clone and the real function object no longer exists. reviews and approval pending. 2005-06-03 04:16:14 +00:00
jst%mozilla.jstenback.com 6b9590cac8 Fixing bug 294795. Don't leave references from cloned member functions to the scope where xpconnect creates the functions (safe context). r=bzbarsky@mit.edu, sr=brendan@mozilla.org, a=brendan@mozilla.org 2005-06-02 21:33:18 +00:00
shaver%mozilla.org 3f1770c2b7 Bug 296006: correctly handle the re-locking of shallow GCThings. r+a=brendan 2005-06-02 12:17:51 +00:00
bzbarsky%mit.edu bbd38864c2 Log creation of XPCNativeWrapper and XPCWrappedNative objects if
DEBUG_XPCNativeWrapper is defined.  Bug 295435, r+sr=jst, a=shaver.
2005-06-01 19:26:24 +00:00
timeless%mozdev.org 36419b53d9 Bug 295636 OOM Crash [@ Init]
r=jst sr=jst a=shaver
2005-06-01 18:36:25 +00:00
bzbarsky%mit.edu de25a79074 Fix toString on the XPCNativeWrapper prototype to return something sane instead
of crashing.  Bug 295430, r+sr=jst, a=shaver
2005-06-01 15:20:29 +00:00
cls%seawood.org 90fcdb00a0 Use dashes instead of backslashes for program options to avoid msys shell expansion.
Thanks to Howard Chu <hyc@symas.com> for the patch.
Bug #294122 r=cls a=shaver
2005-06-01 14:28:35 +00:00
bzbarsky%mit.edu 50ff79e209 Separate out deep-ness and auto-unwrapping of native wrappers (base the latter
on whether the wrapper is explicit).  Bug 295782, r+sr+a=brendan
2005-06-01 01:46:01 +00:00
bzbarsky%mit.edu 461744f884 Don't return XPCNativeWrappers from an nsIXPConnect::WrapNative call. Bug
295152, r+sr=jst, a=brendan
2005-06-01 01:21:23 +00:00
brendan%mozilla.org ac89cb0198 Fix gcPoke management to cope with nesty destroy-context order (296119, r=shaver, a=me). 2005-06-01 01:16:06 +00:00
jst%mozilla.jstenback.com 70890dce44 Fixing crash introduced with my earlier changes for bug 294893. Make nw and wn finalizers order independent. r+sr=bzbarsky@mit.edu, a=drivers 2005-05-26 21:35:27 +00:00
jst%mozilla.jstenback.com a81b370e90 More fixes for bug 294893. When a XPCWrappedNative's scope is cleared, make sure we clear the scope for all of its XPCNativeWrappers as well. r+sr=bzbarsky@mit.edu, a=drivers 2005-05-26 19:30:36 +00:00
bzbarsky%mit.edu a68e2d534e Better version of last patch -- don't pin the atoms, and clean up nicely when
they're finalized so as not to crash.  Bug 294893, assumed future r=brendan,
a=drivers
2005-05-26 17:21:29 +00:00
bzbarsky%mit.edu 495c01f9c8 Backing out to fix orange till I can sort it out. 2005-05-26 14:41:34 +00:00
bzbarsky%mit.edu 732a318bfe Don't pin object atoms, since those can entrain all sorts of other stuff. Bug
294893 some more, r=brendan, sr=jst, a=brendan
2005-05-26 13:27:57 +00:00
brendan%mozilla.org 6fdd65d11d Fix null deref crash on s.match.apply (295052, r=shaver). 2005-05-25 23:28:00 +00:00
jst%mozilla.jstenback.com a12a67e400 More fixes for leak bug 294893. Make sure to clear the scope of the native wrapper when a wrapped native's scope is cleared. r+sr=bzbarsky@mit.edu, a=drivers 2005-05-25 03:40:24 +00:00
jst%mozilla.jstenback.com 679306af53 More attempts at fixing leak bug 294893 by fixing up parenting of XPCNativeWrapper objects, and fixing problem with accessing a deleted wrapper in some cases. r+sr=bzbarsky@mit.edu, a=asa@mozilla.org 2005-05-24 20:05:43 +00:00
jst%mozilla.jstenback.com 762423d761 Fixing (hopefully) leak bug 294893. Don't mark mNativeWrapper when an XPCWrappedNative is marked, let XPCNativeWrapper's die when all script references to them are gone even if their XPCWrappedNative is still referenced. r/sr/a=asa@mozilla.org (will get real reviews once people are reachable). 2005-05-24 06:42:30 +00:00
bzbarsky%mit.edu 3456bb0fa9 Fix deep-wrapping to reuse existing wrappers if it can. Bug 295301, r+sr=jst,
a=asa
2005-05-24 05:11:20 +00:00
timeless%mozdev.org 7e3c34656b Bug 293992 printf passed user controlled buffer
r=shaver sr=bz a=asa
2005-05-23 21:00:06 +00:00
jst%mozilla.jstenback.com b732bd9de4 More fixes for bug 295101. Don't leave *pob2 dangling (caught by bz). r+sr=bzbarsky@mit.edu, a=brendan@mozilla.org (on irc) 2005-05-23 16:59:29 +00:00
bzbarsky%mit.edu 239b86a89a Check in jst's patch to auto-mark our new objects before we do things that
might trigger gc.  Bug 295101 partial patch, r+sr=bzbarsky/brendan, a=brendan
2005-05-23 14:14:45 +00:00
brendan%mozilla.org 8d11e72c0e Bypass XPCNativeWrapper's delProperty hook (295122, r+sr=jst). 2005-05-23 05:58:05 +00:00
bzbarsky%mit.edu 55bdaeb6a4 Stop forwarding addProperty to the wrapped native. That's not the right thing
to do in general (since we use OBJ_DEFINE_PROPERTY in this case in resolve
anyway), and it breaks some things.  Bug 295122, r/sr=jst/brendan, a=brendan
2005-05-22 20:48:00 +00:00
brendan%mozilla.org fe8d068db1 Fix UMR regression introduced in 3.110 (295116, tachyonal r=shaver). 2005-05-22 18:25:08 +00:00
brendan%mozilla.org 1b3f9fa8ee Fix 294957 for good. 2005-05-21 17:35:16 +00:00
brendan%mozilla.org ac6c6c55fd Fix from bz for glitch in last checkin. 2005-05-21 14:55:54 +00:00
brendan%mozilla.org dc516e2d5f Fix stale comment added in last rev. 2005-05-21 06:32:12 +00:00
brendan%mozilla.org 931fdb098f Fix adblock and many other non-system chrome users of sytem-chrome xbl (294960, r+sr=bz). 2005-05-21 06:24:08 +00:00
brendan%mozilla.org 2d8d326848 Argh, I made JS_IsSystemObject return a bogo-boolean -- thanks to bz for catching this. 2005-05-21 05:52:20 +00:00
brendan%mozilla.org 3c1d4852bc Fix frames[0] case where DOM doesn't use an item method (294960, r+sr=bz). 2005-05-20 21:46:24 +00:00
brendan%mozilla.org 01554c0578 Fix 294957, fallout from the 281988 landing (r+sr=bz). 2005-05-20 20:06:22 +00:00
bzbarsky%mit.edu fc7844f174 Flag our private as ISUPPORTS so object principals, etc, become happy. Bug
281988 ongoing, r+sr+a=brendan.
2005-05-20 19:00:14 +00:00
brendan%mozilla.org 39161400fd Followup for 281988, r+sr=bz, to make deep XPCNativeWrapper __parent__ mirror wrapped-native parent linkage. 2005-05-20 18:51:57 +00:00
bzbarsky%mit.edu 50a41f5d6b Make sure to not use an XPCNativeWrapper as the parent for an
XPCWrappedNative.  Get its underlying XPCWrappedNative and use that instead.
Bug 281988 fun continues, r+sr+a=brendan
2005-05-20 17:35:11 +00:00
bzbarsky%mit.edu 194ef290a9 Make GetScopeOfObject() know about XPCNativeWrapper and get the scope of the
underlying XPCWrappedNative.  Bug 281988, r+sr+a=brendan
2005-05-20 05:09:56 +00:00
bzbarsky%mit.edu 195fa43291 Fixing build bustage. Thanks to swalker for the heads-up 2005-05-20 04:47:28 +00:00
brendan%mozilla.org fa453c7bb5 Big patch from jst and me to beef up and automate, configurably, XPCNativeWrapper (281988, r+sr=bz/shaver, a=drivers). 2005-05-20 03:12:22 +00:00
brendan%mozilla.org e1232f3586 Unregress __noSuchMethod__ for 1.8b2. 2005-05-17 20:01:28 +00:00
brendan%mozilla.org 4c3b59e700 Another 293782 followup, to restore lost perf in call_enumerate (called from js_PutCallObject, imp. for closure perf). 2005-05-17 17:44:57 +00:00