Commit Graph

22 Commits

Author SHA1 Message Date
Mike Conley
a8482b31c5 Bug 1146955 - Have PluginContent and PluginCrashReporter work with GMPs and pluginIDs. r=felipe.
--HG--
extra : rebase_source : 44b0737c88442b5afbb5f9f2e9458c6321442e8e
extra : source : 8da0582cde7ed894e195f9886d701976a5f25976
extra : histedit_source : b0f8cae0da36ef1d3e2b70955d03e032143b2ae7
2015-04-27 15:10:54 -04:00
Mike Conley
6a2fda6f73 Bug 1110887 - Add test to exercise both cases of some non-determinism when dealing with plugin crashes in e10s. r=felipe.
--HG--
extra : source : 71310d7e7e2cc37d7da0131ad8e68d4db0788415
extra : amend_source : 5c0178440a40c2f198ad69944dfd600816804914
2015-03-27 16:32:24 -04:00
Mike Conley
a077dba15e Bug 1110887 - Make the plugin crash reporter work with e10s. r=felipe.
This patch does the following:
* Has the gPluginHandler tell content processes when a NPAPI plugin crashes
* Introduces PluginCrashReporter, and renames TabCrashReporter.jsm to ContentCrashReporters.jsm.
* Makes gPluginHandler use PluginCrashReporter to submit plugin crashes.
* If a plugin crash report is submitted, puts all visible instances into the submitted state.
* Makes the plugin crashed notification bar work with run IDs.
* Removes event and message listeners in PluginContent when uninitting.

--HG--
extra : source : 503877a91767a0920fcf0c092c726240d94c8db4
2015-03-27 23:34:15 -04:00
Carsten "Tomcat" Book
3630ba09bb Backed out changeset 503877a91767 (bug 1110887) for bc1 test failures
--HG--
extra : rebase_source : 0f4e369314de68f9c68276c4d28f9b240101c2ce
2015-04-09 13:11:32 +02:00
Carsten "Tomcat" Book
a808046b2f Backed out changeset 71310d7e7e2c (bug 1110887)
--HG--
extra : rebase_source : e89fe430036926c1cc4077154958339c6dbfb3ca
2015-04-09 13:05:55 +02:00
Mike Conley
952ec40084 Bug 1110887 - Add test to exercise both cases of some non-determinism when dealing with plugin crashes in e10s. r=felipe.
--HG--
extra : rebase_source : 9542bb8f9177680a45deb2056c1a66af59b180d8
2015-03-27 16:32:24 -04:00
Mike Conley
a3893685f8 Bug 1110887 - Make the plugin crash reporter work with e10s. r=felipe.
This patch does the following:
* Has the gPluginHandler tell content processes when a NPAPI plugin crashes
* Introduces PluginCrashReporter, and renames TabCrashReporter.jsm to ContentCrashReporters.jsm.
* Makes gPluginHandler use PluginCrashReporter to submit plugin crashes.
* If a plugin crash report is submitted, puts all visible instances into the submitted state.
* Makes the plugin crashed notification bar work with run IDs.
* Removes event and message listeners in PluginContent when uninitting.

--HG--
extra : rebase_source : 04f1f5c755bd4b8e16b7f73cf88955760944b358
2015-03-27 23:34:15 -04:00
Alessio Placitelli
d19436f3f5 Bug 1134279 - Change Telemetry data producers to use the correct recording flags. r=vladan 2015-03-23 04:39:00 +01:00
Bill McCloskey
e46e199ff2 Bug 1142542 - Use AppConstants in browser/modules (r=gavin) 2015-03-13 08:19:09 -07:00
Yury Delendik
0aeec913e7 Bug 1130032 - Missing plugin UI is broken. r=mconley
--HG--
extra : rebase_source : dbd952962fb2023ca4e61a4dbd4184c53860edec
2015-02-06 15:15:25 -06:00
Connor
6ac643be3c Bug 1045100 - This is 1 of 2 patches. This patch implements on the front end the new "PluginCrashedEvent" WebIDL binding. r=gfritzsche,mconley 2014-10-22 00:41:32 -04:00
Bill McCloskey
0d32d7e902 Bug 1085710 - Stop trying to send nsIPluginTag over IPC (r=mconley) 2014-10-29 08:04:51 -07:00
Yury Delendik
b35ef8b853 Bug 1075088 - Add telemetry probes for Flash content. r=gfritzsche,mconley
--HG--
extra : rebase_source : 2afac68cc01c045c4d4c85d663e0e76bb4e828ae
2014-10-09 16:05:20 -05:00
Mike Quested
85aabf627f Bug 1074976 - Move function makeNicePluginName() into BroswerUtils.jsm module. r=felipe 2014-10-06 11:40:00 -04:00
Mike Quested
9ff43809eb Bug 847072 - Regex test for Java plugin name, output "Java" instead of whole version string. r=mconley 2014-10-01 04:09:00 +02:00
Mike Conley
97117148cf Bug 1069075 - Send location along with CTP messages. r=gfritzsche.
Before, we were comparing Principals sent up from PluginContent against the
Principal of the selected browser - this was in order to account for races
(and intermittent oranges) where the browser has moved away from a page before
a CTP message has been received by the parent.

Comparing Principals works beautifully, except in the case of Data URIs, since
Data URIs inherit the Principals of the page they were linked from. That means
that if we're on a page with a link to a Data URI that embeds a plugin, if we
quickly browse to that page and back before a CTP message reaches the parent,
when the CTP message finally arrives, the Principals will match. We solve this
by sending the location as well as the Principal, to do a final comparison if
the Principals do match.

--HG--
extra : rebase_source : 095706306d241a254f62b28527a96a427d7e55c3
extra : histedit_source : 3afe814271166f70afce6a423bbb7809c5a983c7
2014-09-28 11:01:29 -04:00
Mike Conley
c4006a26e4 Bug 1069075 - Fix race condition in flushing pluginData cache in PluginContent.jsm. r=gfritzsche.
--HG--
extra : rebase_source : 0a0c801e0a8802e0a8ae04ac34e9977576f64e1e
extra : histedit_source : 03138083b2ac0892c2d66898db918528dda01fe0
2014-09-26 17:17:36 -04:00
Mike Conley
4a1e8018d2 Bug 1070053 - Avoid spurious hidden-plugin notifications when changing locations by doing a Principal comparison. r=gfritzsche
When the PluginRemoved event is fired when changing locations, it's fired
asynchronously such that the document that the plugin belongs to has already
been unloaded. This was causing the hidden plugin notification to appear in
some cases when users browsed away from documents that had hidden plugins
in them. Now we pass the Principal for the unloading document back to the
parent and do a comparison with the current browser Principal to ensure
that they match before showing the hidden plugin notification.

--HG--
rename : browser/base/content/test/plugins/plugin_small.html => browser/base/content/test/plugins/plugin_small_2.html
extra : rebase_source : e748e3b09de77cc7796b1a78f8e39a23af64049a
2014-09-24 10:30:18 -04:00
Justin Dolske
c80af2ab0d Bug 836415 - Kill PFS. r=bsmedberg 2014-09-18 10:01:23 -07:00
Mike Conley
6aecdba124 Bug 899347 - [e10s] Make click-to-play work with e10s. Originally written by mbrubeck. r=gfritzsche,felipe,Unfocused.
We now use message passing and content scripts in order to implement click-to-play. This allows
us to use click-to-play in e10s windows.

--HG--
extra : rebase_source : 6239a4b49e0af89800c5618e4589f8aef03f58d4
2014-09-17 16:06:58 -04:00
Ryan VanderMeulen
e84550c5e3 Backed out changeset 5747034f2bcc (bug 899347) for mochitest-bc failures.
CLOSED TREE
2014-09-15 15:30:39 -04:00
Mike Conley
91e7e2f96f Bug 899347 - [e10s] Make click-to-play work with e10s. Originally written by mbrubeck. r=gfritzsche,felipe,Unfocused.
We now use message passing and content scripts in order to implement click-to-play. This allows
us to use click-to-play in e10s windows.

--HG--
extra : rebase_source : 719afe669c638ba4883f76f18aee38df6b60b4ad
2014-09-15 13:33:50 -04:00