Граф коммитов

100 Коммитов

Автор SHA1 Сообщение Дата
Benjamin Smedberg 995c9cbf9a Bug 582012 - register identifiers in the ID maps when they are created by the browser, so that a plugin never sees two identifiers for the same string/integer, r=bent 2010-07-28 13:20:33 -04:00
Tero Koskinen 8e16bba239 Bug 563737 - [QT] Detect when Qt plugins enter a nested event loop. r=cjones 2010-06-01 23:58:00 -04:00
Jim Mathies 9f80071784 Bug 568129 - Don't reset the delayed flash reply message flag, triggering the call at some point during fullscreen tear down. r=bent. 2010-05-26 20:07:25 -05:00
Jim Mathies 0879ff6221 Bug 558986 - detect and resolve races between MS-COM RPC and IPDL RPC by implementing IMessageFilter and processing pending IPDL RPC calls while waiting for MS-COM RPC responses. Original patch by bsmedberg, r=cjones/jimm/ehsan with followup fixes by Jim Mathies, r=bsmedberg
--HG--
extra : rebase_source : 214029d3188f8aa00948e0dcdc3c11cd279d1569
2010-05-21 11:48:34 -04:00
Oleg Romashin 3c481f45ca Bug 566845 - Make desktop flash plugin works on Fennec QT. r=dougt. 2010-05-21 13:19:57 -04:00
Benoit Girard 125ee2ceb5 Fixed PluginInstanceParent::AnswerNPN_ConvertPoint definition. b=565365 r=josh 2010-05-20 15:22:54 -04:00
Jim Mathies 8e6808699a Bug 566610 - Undo the nuclear reply message option, keep the set focus fix. r=bent. 2010-05-18 13:17:10 -05:00
Jim Mathies eb4761537a Bug 563377 - Flash fullscreen window refuses to close. r=bent. 2010-05-18 11:43:45 -05:00
Jim Mathies e9cf1d1dad Bug 558986 - Backout of fix for crashes in TSF with Flash/Silverlight on tablets. CLOSED TREE 2010-05-17 14:36:43 -05:00
Benjamin Smedberg 695ab55828 Bug 558986 - Fix for crashes in TSF with Flash/Silverlight on tablets (OLE/COM nested event loops). r=cjones, jimm, ehsan. 2010-05-17 11:33:34 -05:00
Benoit Girard 94d6258a5b Bug 528146: Run plug-in code on the thread that starts in main(). r=cjones
--HG--
rename : dom/plugins/PluginThreadChild.cpp => dom/plugins/PluginProcessChild.cpp
rename : dom/plugins/PluginThreadChild.h => dom/plugins/PluginProcessChild.h
rename : ipc/ipdl/test/cxx/IPDLUnitTestThreadChild.cpp => ipc/ipdl/test/cxx/IPDLUnitTestProcessChild.cpp
rename : ipc/ipdl/test/cxx/IPDLUnitTestThreadChild.h => ipc/ipdl/test/cxx/IPDLUnitTestProcessChild.h
2010-05-10 23:18:00 -05:00
Jim Mathies 0dfb101754 Bug 563377 - [OOPP] Add centralized ReplyMessage handler in PluginModuleChild for all child windows (various fullscreen hang fixs). r=bent. 2010-05-11 10:19:14 -05:00
Benjamin Smedberg 19727cf3ac Bug 564861 - NPN_IntFromIdentifier always returns INT_MIN when running in OOPP mode because of a backwards conditional check. Also, int identifiers of -1 don't work correctly when the parent first creates the identifier, because we were misusing a sentinel value. r=bent 2010-05-10 17:11:25 -04:00
Benjamin Smedberg 624e1bc012 Bug 561817 part C - restore the NestableTasksAllowed munging that was removed in patch B: I thought it was vestigial, but it's necessary, r=bent 2010-04-30 14:43:59 -04:00
Benjamin Smedberg b670251bc9 Bug 561817 part B - Install the nested event loop tracking mechanism for all RPC calls to plugins, not just specific handle-event calls, r=jimm 2010-04-30 14:42:51 -04:00
Karl Tomlinson 0ed80ea60e b=561308,539138 Realize GtkPlugs when associated with a socket for consistency with in-process behavior r=cjones
--HG--
extra : rebase_source : 96d433c381d6d560ddd566572cfc143d91113453
2010-04-29 22:43:15 +12:00
Karl Tomlinson 6eff65702c b=561690 NULL check aNPP in _geturlnotify for nspluginwrapper r=bsmedberg
--HG--
extra : rebase_source : 67c36002a3ef071b6ff6f31dcfe1954e3e61ca9e
2010-04-29 09:08:07 +12:00
Josh Aas 146d6023c8 Remove unimplemented status from NPN_ForceRedraw for IPC. We are ignoring calls on purpose and do not plan to implement it in any way at this point. b=555511 r=cjones 2010-03-30 23:53:53 -04:00
Chris Jones 8ef4352202 Bug 555309, part 3: Grab the plugin thread's ID on startup and blame it in hang dumps. r=bsmedberg 2010-03-30 19:24:47 -05:00
Oleg Romashin a6a9c5bdf6 Bug 554911 - mozilla-runtime process hang and not closed properly after closing fennec UI. r=dougt
--HG--
extra : rebase_source : 5490b2acfe1980e372b7918cf36e8e771b2a0747
2010-03-30 07:34:37 -04:00
Benjamin Smedberg 8cbfc7ec64 Bug 542700 - Async calls aren't cancelled when the instance is destroyed, and the list of pending calls isn't protected against simultaneous access, r=cjones
--HG--
extra : rebase_source : b0949fd4921d1dbb558e84e191387964af5767fc
2010-03-25 14:25:38 -04:00
Josh Aas 5baa1625d5 [OOP] Add support for NPN_ConvertPoint via IPC. b=555250 r=cjones 2010-03-26 16:55:56 -04:00
Josh Aas 342aa128ae [OOPP] Fix ordering of NP_Initialize and NP_GetEntryPoints calls for Mac OS X. The incorrect order causes Flash 10.1 to fail to load. b=555208 r=bsmedberg 2010-03-26 13:03:58 -04:00
Karl Tomlinson 9b9ff3a789 b=545297 don't add an extra reference to GtkPlug::socket_window with versions of GTK that already add a reference r=cjones 2010-03-25 16:18:25 +13:00
Karl Tomlinson 9828764f3e b=551875 handle g_main_depth() != 1 when events are dispatched r=cjones 2010-03-26 14:52:16 +13:00
Karl Tomlinson eee596b947 b=526393 propagate scroll events from GtkPlug to the embedder r=roc (relanded) 2010-03-24 16:42:06 +13:00
Karl Tomlinson 90e98e6604 backout 60fb978130f5 due to wrong bug no in the commit message 2010-03-26 14:35:56 +13:00
Josh Aas b5a2f793a8 Fix NPAPI NPN_PushPopupsEnabledState and NPN_PopPopupsEnabledState signature mismatches, statically initialize browser-side NPAPI function table with no casting. b=554170 r=bent 2010-03-25 19:45:33 -04:00
Ben Turner 8271154a90 Bug 521118 - '[OOPP] Need NPAPI threadsafety checks in plugin process'. r=cjones, a=dbaron for CLOSED TREE 2010-03-24 12:30:45 -07:00
Karl Tomlinson 54bc536db5 b=544211 propagate scroll events from GtkPlug to the embedder r=roc 2010-03-24 16:42:06 +13:00
Ben Turner 5c68b4c136 Bug 547359 - '[OOPP] Silverlight plugin takes 6X longer to load with IPC plugins than without - make NPIdentifier handler smarter/faster'. r=bsmedberg. 2010-03-23 07:30:51 -07:00
Josh Aas 36aed1ebfa Allow Gecko to build and run on Mac OS X with MOZ_IPC defined. b=552862 r=cjones 2010-03-17 14:38:28 -04:00
Benjamin Smedberg bb9f80b36c Bug 551049 (was bug 532208) part 2 - Delay delivery of NPP_DestroyStream until stream data is delivered, and make sure that data delivery doesn't re-enter, r=bent 2010-03-06 16:02:31 -05:00
Benjamin Smedberg a854facd28 Bug 551049 (was bug 532208) part 1 - Make PBrowserStream destructor uni-directional by delivering all data asynchronously and doing two-phase (acknowledged) destruction, r=bent 2010-02-23 16:45:00 -05:00
Benjamin Smedberg 961a9e2c1e Revert the RPC-style race resolution from bug 532208 in order to use the asynchronous delivery patches. 2010-03-06 11:08:19 -05:00
Chris Jones ee351fc953 Bug 532208: Defer deletion of browser streams to later tasks. r=bsmedberg 2010-02-26 13:07:56 -06:00
Karl Tomlinson a0842f75ce b=548434 NULL check aNPP in _invalidaterect for nspluginwrapper r=cjones
--HG--
extra : rebase_source : 6af217c167b313f034461ec63d7dd45cb5bd5225
2010-02-24 16:19:51 -08:00
Benjamin Smedberg 5839dc62ac Merge backout of bug 532208 due to test failures, crashes, and overall badness. 2010-02-25 03:03:12 -08:00
Benjamin Smedberg 11b1d27433 Backed out changeset e4a7ea0bb90f bug 532208 - asynchronous stream delivery because write events were being dispatched in odd re-entrant states and NPP_DestroyStream was being delivered before all the stream data was delivered. 2010-02-25 03:00:56 -08:00
Karl Tomlinson 30883fb3f1 b=517133 Get crash stats for X errors r=bsmedberg
--HG--
extra : rebase_source : e9b2319a566ecb5a8a3cd6ba322df97e1ff3ffc8
2010-02-24 14:25:16 -08:00
Benjamin Smedberg fbec9f6e59 Bug 532208 - Make PBrowserStream destructor uni-directional by delivering all data asynchronously and doing two-phase (acknowledged) destruction, r=bent
--HG--
extra : rebase_source : 1155cb3c0d548712b274c81e7cf597eaf091d220
2010-02-23 16:45:00 -05:00
Jim Mathies 39cc8f63c3 Bug 547353 - [OOPP] Mouse pointer coordinates misaligned with winless Silverlight. r=bent. 2010-02-24 16:13:48 -08:00
Chris Jones 2a273c0016 Bug 544945, part 2: Periodically unblock the parent to allow it to process events while the plugin subprocess is in a nested event loop. r=karlt
--HG--
extra : rebase_source : bf9a2df9d7d6b420fb93d11dc5abaeb4be1214db
2010-02-16 12:44:25 -06:00
Chris Jones 2e0e523879 Bug 544945, part 1: Detect nested glib event loops in the plugin subprocess. r=karlt
--HG--
extra : rebase_source : 454553501de351b66e328a7005e6779088f2a2fa
2010-02-16 12:44:24 -06:00
Benjamin Smedberg 8a8ef6a58f Bug 544074 - Crashes at mozilla::plugins::PluginModuleChild::UnregisterActorForNPObject - the hash of scriptable objects is mutating while we enumerate the actors being destroyed after a plugin instance is destroyed. Save the list of actors off into a separate non-mutating table, r=bent 2010-02-24 16:14:13 -05:00
Benjamin Smedberg a91ddf5a32 Bug 547894 - Allow null in NPN_GetStringIdentifier(s) to fix a topcrash with the divx plugin, r=bent 2010-02-24 16:34:17 -05:00
Oleg Romashin 12fceeb0cd bug 544193 - Qt version for plugins embedding functionality required. r=mozbugz. 2010-02-10 22:48:20 +02:00
Benjamin Smedberg c0e27c0256 Bug 535120 - NPN_GetValue(NPNVToolkit) doesn't work with null npp, r=bent
--HG--
extra : rebase_source : 67d03285f34260c02c49579ef8ff664e04bc4a84
2010-02-03 17:17:41 -05:00
Benjamin Smedberg d5eb4d3947 Bug 516515 - Don't initialize XPCOM in plugin processes, r=cjones
--HG--
extra : rebase_source : cab356b862a66fb7f5955b1f936944f2395bba3c
2010-02-03 17:17:09 -05:00
Chris Jones 66af1ab007 Bug 535612: Set GDK_NATIVE_WINDOWS in plugin processes to work around plugins that don't interact well with client-side GDK windows. r=karlt r=roc 2010-02-02 00:28:28 -06:00