Doug Turner
|
e342d43d13
|
Bug 550545 - The NPAPI ImageExpose event should be disabled when building for Qt. r=vlad
--HG--
extra : rebase_source : b66a81ba5a019e7da387eb81d4fd2454be196fe7
|
2010-03-10 15:01:46 -08:00 |
|
Benjamin Smedberg
|
0ac333446e
|
Bug 550944 - If a stream is alive at instance destruction, make sure we fire NPP_DestroyStream and NPP_URLNotify correctly. Also make the new-stream functions (NPN_GetURL etc) fail while we are destroying a plugin instance, r=josh
--HG--
extra : rebase_source : e94bc08bcc04c2ead28ac14bce7e3320a49c9d5a
|
2010-03-08 16:54:04 -05:00 |
|
Julian Seward
|
d06bc6c5d0
|
Bug 548441. Don't write to ndata after we delete the object it's in. r=bzbarsky
|
2010-03-09 14:38:13 -05:00 |
|
Benjamin Smedberg
|
ced61e4b02
|
Followup to bug 548217 - Instead of using a bizarre dual-refcounting scheme, explicitly track when nsPluginStreamListenerPeer.mLocalCachedFile is shared and should be deleted, r=josh
--HG--
extra : rebase_source : 97f6bebfdb99f0c0cbd8cd033e62492fcb755d09
|
2010-03-06 14:38:54 -05:00 |
|
Benjamin Smedberg
|
2a5f3518de
|
1;2cBug 548217 - Make NP_SEEK streams work properly: they shouldn't crash if NPN_RequestRead is called after all the stream data is delivered, and they should call NPP_DestroyStream when NPN_DestroyStream is called, but not before. To this end, the classes nsPluginStreamInfo and nsPluginStreamListenerPeer have been unified: the nsNPAPIPluginStreamListener holds the stream listener peer alive so that additional range requests can be made on it. In addition the nsNPAPIPluginStreamListener holds *itself* alive for NP_SEEK streams, and only calls releases itself/calls CleanUpStream when the plugin explicitly calls NPN_DestroyStream or when the associated plugin instance is being destroyed. r=josh
--HG--
extra : rebase_source : 372df117d2b0c955605f44ee0e2fa512f0ad972b
|
2010-02-23 21:58:27 -05:00 |
|
Chris Jones
|
f0560d9e8e
|
Bug 441324, part 1: Shuffle around includes to avoid problems with #define malloc et al. r=bsmedberg
|
2010-03-03 23:02:55 -06:00 |
|
Benjamin Smedberg
|
0b629a5939
|
Backed out changeset f829f942873d - bug 548217 because of topcrash bug 549112
|
2010-02-27 17:36:54 -05:00 |
|
Benjamin Smedberg
|
958042217d
|
Bug 548217 - Make NP_SEEK streams work properly: they shouldn't crash if NPN_RequestRead is called after all the stream data is delivered, and they should call NPP_DestroyStream when NPN_DestroyStream is called, but not before. To this end, the classes nsPluginStreamInfo and nsPluginStreamListenerPeer have been unified: the nsNPAPIPluginStreamListener holds the stream listener peer alive so that additional range requests can be made on it. In addition the nsNPAPIPluginStreamListener holds *itself* alive for NP_SEEK streams, and only calls releases itself/calls CleanUpStream when the plugin explicitly calls NPN_DestroyStream or when the associated plugin instance is being destroyed. r=josh/jst
|
2010-02-23 21:58:27 -05:00 |
|
Benjamin Smedberg
|
7985a1dcb1
|
Backed out changeset 77dc38d8196e - bug 548217 because even though this patch is correct, it exposes a bug in the OOPP code which got backed out.
|
2010-02-25 03:57:57 -08:00 |
|
Benjamin Smedberg
|
4a49a68e09
|
1;2cBug 548217 - Make NP_SEEK streams work properly: they shouldn't crash if NPN_RequestRead is called after all the stream data is delivered, and they should call NPP_DestroyStream when NPN_DestroyStream is called, but not before. To this end, the classes nsPluginStreamInfo and nsPluginStreamListenerPeer have been unified: the nsNPAPIPluginStreamListener holds the stream listener peer alive so that additional range requests can be made on it. In addition the nsNPAPIPluginStreamListener holds *itself* alive for NP_SEEK streams, and only calls releases itself/calls CleanUpStream when the plugin explicitly calls NPN_DestroyStream or when the associated plugin instance is being destroyed. r=josh
--HG--
extra : rebase_source : fba343acbfcaa785c7405552d42f4558424b485c
|
2010-02-23 21:58:27 -05:00 |
|
Martin Stransky
|
b3a1889d8e
|
Bug 528900: Null check at _[pop|push]popupsenabledstate to avoid crashing r=jst
|
2010-02-20 08:53:10 -05:00 |
|
Brad Lassey
|
23156d33b8
|
Winmo bustage fix from bug 545080 r=dougt
|
2010-02-20 00:53:38 -05:00 |
|
Benjamin Smedberg
|
171347def1
|
Followup to bug 545080, convert additional functions to the explicit "W" variants so that even when we're not building with -DUNICODE we always use unicode paths, r=Thunderbird bustage
|
2010-02-19 15:31:28 -05:00 |
|
Benjamin Smedberg
|
a408137a65
|
Bug 545080 and bug 537344 - When UNICODE is defined in modules/plugin/base/src, we follow codepaths that aren't right, and the fields of nsPluginInfo are double-converted to UTF8, which is bad, r=josh
|
2010-02-13 16:59:39 -05:00 |
|
Karl Tomlinson
|
0cd7c3af25
|
b=540114 process reparent requests before removing the X error handler r=cjones
|
2010-02-18 11:05:24 +13:00 |
|
Karl Tomlinson
|
048a190877
|
b=540114 reparent foreign children of the socket window before it gets destroyed r=cjones
|
2010-02-16 10:40:46 +13:00 |
|
Benjamin Smedberg
|
256648ae37
|
Bug 543977 - NPClass.invoke should be optional when a class just wants to implement NPClass.invokedefault, r=jst
--HG--
extra : rebase_source : 2e1fb9896bcbf12bc6d2772c02a1cc9557efe6eb
|
2010-02-17 13:30:54 -05:00 |
|
Karl Tomlinson
|
a4cd333dde
|
b=543901 only call NPPluginFuncs::getvalue when non-NULL and remove NPP_GetValue NPPVpluginWindowBool NPPVpluginTransparentBool r=cjones
|
2010-02-04 13:02:48 +13:00 |
|
Oleg Romashin
|
afd6f98795
|
bug 544193 - Qt version for plugins embedding functionality required. r=mozbugz.
|
2010-02-10 22:48:20 +02:00 |
|
Ted Mielczarek
|
d40733d453
|
Bug 541076 - add minidump id and plugin name to PluginCrashed event. p=ted+dolske, r=jst, r=ted
|
2010-02-09 17:05:31 -08:00 |
|
Doug Turner
|
c5fb72ab64
|
Bug 543976 - Clean up Maemo defines. r=ted
|
2010-02-09 11:02:11 -08:00 |
|
Josh Aas
|
c04a923d42
|
Fix our plugin library loading code so that it works in 64-bit builds. 64-bit NSPR does not support bundle loading. b=544579 r=smichaud
|
2010-02-08 13:22:25 -06:00 |
|
Josh Aas
|
e0d1ee7974
|
More complete plugin architecture check for Mac OS X. Support 64-bit-only plugins and stop assuming that any universal binary plugin has the correct architecture. Fixes bug 544433 and bug 544437. r=smichaud
|
2010-02-06 23:29:58 -06:00 |
|
Josh Aas
|
20a6128dae
|
Make all scriptable methods in nsIPluginHost appear before the C++ code. b=542712 r=jst
|
2010-02-06 16:07:13 -06:00 |
|
Benjamin Smedberg
|
151a3c2903
|
Bug 544016 - If two plugins are available for a MIME type, prefer the enabled one to the disabled/blocklisted one, r=josh
--HG--
extra : rebase_source : 8c4382719dd8886634a8f89a85d89af1e96f6c4d
|
2010-02-03 16:52:07 -05:00 |
|
Benjamin Smedberg
|
6db43fb19a
|
revert changeset f61e06c31086 (bug 542971) since it caused bug 543376.
|
2010-02-02 17:18:37 -08:00 |
|
Benjamin Smedberg
|
fb8f40b46e
|
Bug 542248 - ASSERTION: Shouldn't return null and NPERR_NO_ERROR (and then crash) navigation from one silverlight page to another, r=josh
|
2010-02-02 09:46:25 -05:00 |
|
Ben Turner
|
dac5a3fee3
|
Bug 542915 - 'Protect DelayedReleaseGCCallback from reentering and double-freeing NPObjects' r=jst+josh.
--HG--
extra : transplant_source : %5B%C1%C3QH%22%22%24%9D%A2%CE%E9%907%941%8F%F5%07%19
|
2010-02-01 12:35:48 -08:00 |
|
Josh Aas
|
4a93f5fe9f
|
Clean up a bunch of things in nsNPAPIPlugin. b=543339
|
2010-01-31 16:41:26 -05:00 |
|
Daniel Holbert
|
32a1fb0c00
|
Bug 542959: Add 'const' to NP_GetMIMEDescription's parameter, to fix "deprecated conversion from string constant to ‘char*’" warning. r=cjones
|
2010-01-29 12:18:50 -08:00 |
|
Josh Aas
|
46d060dccd
|
Get rid of nsPluginInstanceTag. Clarifies the plugin object hierarchy and makes it easier to get from one object to another. We avoid unnecessary null checks and strong references due to clarified lifetimes dependencies between objects. Makes all of this code less error-prone. b=542971 r=jst
|
2010-01-29 14:42:54 -05:00 |
|
Ben Turner
|
33a8aae576
|
Bug 542337 - '[OOPP] hang in test_plugin_clipping2.xhtml' Followup fix, r=roc.
--HG--
extra : transplant_source : ._%92%D6%40%FE%1E%92%01%11%7F%BA%2A%EF%24Y%84%DC6%0E
|
2010-01-28 12:23:07 -08:00 |
|
Ben Turner
|
453fbad6ed
|
Bug 542337 - '[OOPP] hang in test_plugin_clipping2.xhtml' r=jimm.
|
2010-01-28 11:32:41 -08:00 |
|
Benjamin Smedberg
|
4245b947f3
|
Bug 539063 - Implement a whitelist/blacklist for OOPP, r=josh
|
2010-01-28 10:31:53 -05:00 |
|
Oleg Romashin
|
2d419b5929
|
Bug 534644 - implement Qt ipc/chromium backend, first part. r=benjamin, r=jones.chris.g
|
2010-01-28 07:22:41 +02:00 |
|
Chris Jones
|
c22de50bdf
|
Bug 542356: Initialize the NPError outparam passed to NPP_New() so that plugins must set it to NPERR_NO_ERROR (and so as to work around a valgrind UMR false-positive for OOPP). r=joshmoz
|
2010-01-27 17:45:34 -06:00 |
|
Jim Mathies
|
757c1b3d63
|
Bug 540052 - [OOPP] Test failure in test_plugin_focus.html. r=bent.
--HG--
extra : rebase_source : bdb40f27e098519ecd71abd75a80e8e612f888f8
|
2010-01-26 15:40:15 -06:00 |
|
Josh Aas
|
2e7df95914
|
Get rid of nsPluginInstanceTagList. Make this code easier to deal with and less error-prone. b=535643 r=jst
|
2010-01-26 20:30:58 -05:00 |
|
Gavin Sharp
|
937dce11b7
|
Bug 189378: add pref to disable plugins in the browser, r=josh
--HG--
extra : rebase_source : 7f70e2585d6b2840fa801969e5af50c5e81caca8
|
2009-10-05 05:00:57 -04:00 |
|
Ben Turner
|
471d1a3e1b
|
Bug 521377 - 'NPRuntime: Segfault when NPP_GetValue_NPPVpluginScriptableNPObject returns a null actor'. r=bsmedberg+valgrind.
--HG--
extra : transplant_source : %F8%7C%B6%C49%24%9A%E0%96%EA%10%E9%C7%C0%EF%EA%D6%7FT%9A
|
2010-01-15 12:35:57 -08:00 |
|
Peter Van der Beken
|
076d23ead6
|
Backout 76cdc8296409 and 9baa220b27c0 (Bug 521377 - 'NPRuntime: Segfault when NPP_GetValue_NPPVpluginScriptableNPObject returns a null actor') to try fo fix orange.
|
2010-01-19 13:00:58 +01:00 |
|
Ben Turner
|
37ff4409a5
|
Bug 521377 - 'NPRuntime: Segfault when NPP_GetValue_NPPVpluginScriptableNPObject returns a null actor'. r=bsmedberg+valgrind.
|
2010-01-15 12:35:57 -08:00 |
|
Chris Jones
|
80b004dd05
|
Bug 538532: When a plugin crashes, make sure all its instances are "stopped" ASAP. r=bsmedberg r=josh
--HG--
extra : rebase_source : 4355e574e5cf74da76ef5773e16fa3bd9a09225e
|
2010-01-12 00:45:24 -06:00 |
|
Chris Jones
|
2d1e4c2722
|
Bug 538532: Add the helper method nsPluginHost::FindTagForPlugin(). r=josh
--HG--
extra : rebase_source : c7192818d9f6e42510e681286a4dda774df8a94c
|
2010-01-12 00:45:23 -06:00 |
|
Josh Aas
|
9f1683fecf
|
Sent "PluginCrashed" notification when a plugin crashes. b=519541 r=jst r=bsmedberg
|
2010-01-13 11:42:41 -05:00 |
|
Jim Mathies
|
bd95331a3a
|
Bug 536303 - OOPP: disable oopp for Windows 2000. r=benjamin.
|
2010-01-10 17:50:36 -06:00 |
|
Benjamin Smedberg
|
86980c504e
|
Bug 532246 - Plugins should not be able to obtain an NPObject* belonging to a different NPP, because that instance may go away before this one and the NPObject* may be invalidated and destroyed. Double-wrap in these cases. r=josh sr=jst
|
2009-12-02 12:02:41 -05:00 |
|
Josh Aas
|
0487ff44e4
|
Minor cleanup, no bug. a=bsmedberg
|
2010-01-05 13:30:00 -05:00 |
|
Michael Wu
|
27fd1e717b
|
Bug 273025 - "bad logic results in potential leak xor crash based on flow" (Improve GetURL/PostURL code, v4 (2/2)) [r+sr=jst]
|
2009-12-27 14:26:00 -06:00 |
|
Oleg Romashin
|
bd7430f326
|
b=532627; nsPluginHostImpl::WritePluginInfo should call PR_Sync before PR_Close, r=jst
|
2009-12-23 13:58:54 +02:00 |
|