Граф коммитов

22051 Коммитов

Автор SHA1 Сообщение Дата
L. David Baron 8f0a832c01 Bug 1371157 - Cache results of Windows system font lookups. r=jimm
I haven't really tested that this fixes the performance problem observed
in a profile without the patch because the steps to use the Gecko
Profiler on local builds on Windows are rather complicated.

MozReview-Commit-ID: FmGFs2Cvquv

--HG--
extra : transplant_source : %CB%AD%1E%2A%8F%DB%0D%9E%25%08%A82%13fP%BFS%82%BF%FC
2017-06-09 08:59:12 -07:00
L. David Baron f37c8a671f Bug 1371157 - Use override keyword in windows nsLookAndFeel header. r=jimm
MozReview-Commit-ID: 8bwmXmBFZB4

--HG--
extra : transplant_source : %EB%3A-%D7%3E%5E%22%A7%24%CAL%7D%A0%BB%E6%A9p%3E%86J
2017-06-09 08:59:12 -07:00
Chris Pearce f6c7ea61dd Bug 1346620 - Resolve symlinks and junction points in path to GMP dir when loading GMP process. r=bobowen
The sandbox blocks loading of GMPs when the GMP resides in a directory stored
in a path which contains a symlink or junction point. So resolve GMP paths
fully before instantiating the GMP process.

MozReview-Commit-ID: EvPCpNIDNwg

--HG--
extra : rebase_source : 7df8236c9988a674ae128faf63b949fd711ab4e0
2017-06-09 15:29:46 +12:00
Masayuki Nakano 24bfe124f4 Bug 1369695 Creating ITfInputProcessorProfiles and TSFStaticSink when they are used at first time r=m_kato
ITfProcessorProfiles are used by a debug method TSFTextStore::CurrentKeyboardLayoutHasIME() and TSFStaticSink (when it's initialized).  However, TSFStaticSink isn't necessary until when TSFTextStore needs to hack something for specific IME or notifying IMEHandler of active TIP change.  So, we can put off to create the instance of ITfInputProcessorProfiles and TSFStaticSink.

MozReview-Commit-ID: KcrqUbqz1do

--HG--
extra : rebase_source : f1821b782c6cd316a8f234a17ee3c92114547041
2017-06-08 23:55:57 +09:00
Masayuki Nakano e25479050a Bug 1369697 Create ITfDisplayAttributeMgr and ITfCategoryMgr when TSFTextStore::GetDisplayAttribute() needs to use them at first time r=m_kato
Creating them may be expensive due to allocating them in the heap.  So, let's put off to create them when first use.

MozReview-Commit-ID: HDgijJo7brU

--HG--
extra : rebase_source : 9e4e68bd048185fe38fd98bef9b5711e86f68999
2017-06-08 20:51:50 +09:00
Masayuki Nakano 696e4cb85a Bug 1250823 part 2 - IMEContentObserver should cache adding ranges as a range during document change r=smaug
Between nsIDocumentObserver::BeginUpdate() and nsIDocumentObserver::EndUpdate(), IMEContentObserver can cache added nodes as a range if they are consecutive nodes.  Even if a node is removed, a data node is changed or attribute is changed unexpectedly, IMEContentObserver can post text change of the added node range and handle it normally.

MozReview-Commit-ID: IttDHBkr92Y

--HG--
extra : rebase_source : f0849d5fab0b28bdfa311cf833a216d43b9215d2
2017-06-08 11:24:58 +09:00
Masayuki Nakano f6dd08693e Bug 1367692 - Make IMEHandler not restore default IMC unless legacy IMM-IME is active. r=m_kato
TIPs (and normal keyboard layouts) don't need IMC on focused window.  So, in most environment, it's not necessary to restore default IMC of focused window.

Therefore, this patch makes IMEHandler not restore default IMC unless legacy IMM-IME is active and disassociate IMC from focused window when IMM-IME isn't active.

However, this is risky change.  Therefore, the new behavior is disabled in default settings.  On the other hand, we need the new behavior only when MS-IME for Japanese is active on Win10.  Therefore, this patch adds a pref to enable/disable the hack and make it true in the default settings.

MozReview-Commit-ID: KAVxVT9CrsW
2017-06-07 11:42:27 +09:00
L. David Baron 88370e02d1 Bug 1367576 - Make IsWidgetStateSafe not truncate the bits that it needs to test. r=karlt
This fix a mistake that goes back to the original code from bug 174585
(gecko-dev 9611b23530, 2005-08-20).

(This makes me wonder how important the code is in the first place if it
didn't work correctly.)

MozReview-Commit-ID: B6q0o5n5hDw
2017-06-06 22:27:18 -07:00
L. David Baron e04a5a47c6 Bug 1367576 - Shrink existing caches to the size that's actually needed. r=karlt
Now that, thanks to bug 1367577, we have the theme constants in an enum,
we can make these arrays smaller rather than assuming that the constants
might use any valid uint8_t value.

MozReview-Commit-ID: A6GjTarVurc
2017-06-06 22:27:18 -07:00
L. David Baron d1ddf5e48e Bug 1367576 - Cache results of getting GTK widget borders. r=karlt
See comments in the header file.

This also clears out mSafeWidgetStates in ThemeChanged since that seems
like a good thing to do, and marks nsNativeThemeGTK as final.

MozReview-Commit-ID: 5Zne4eGbGlh
2017-06-06 22:27:18 -07:00
L. David Baron ee1674ecc5 Bug 1367576 - Refactor to allow for caching of some gtk widget padding/border results. r=karlt
This refactors the two nearly-identical callsites into a method so that
I can do caching in that method in the next patch.

Note that there was a slight difference between them in that the
aWidgetFlags parameter to GetGtkWidgetAndState was only passed from one
callsite.  However, given that the aState parameter is null, this
doesn't cause any behavior differences.  (Some controls in
GetGtkWidgetAndState null-check aWidgetFlags and some don't!)

Note also that this makes it always assign a result (often zero).  This
is fine for both callsites; GetWidgetPadding previously assigned zero
right before the call, and GetWidgetBorder did so at the start of the
function (and wasn't modified in between, since it was immediately
before the switch that the modified code is a case in).

MozReview-Commit-ID: IKurwry3UTi
2017-06-06 22:27:17 -07:00
L. David Baron 06ad92ffc6 Bug 1367576 - Remove unused ishtml parameter to moz_gtk_get_widget_border. r=karlt
This was needed for the (now-unused) GTK2 version of the code.

MozReview-Commit-ID: GocgC4OZ76p
2017-06-06 22:27:17 -07:00
Chris H-C c57d39227d bug 546387 - Don't try to set the GTK clipboard with null items r=karlt
It causes an assert failure in gtk which prints to the console.

MozReview-Commit-ID: A4106Z4rT36

--HG--
extra : rebase_source : 642c94a95cfa3939bc475e9139ee63caa78e3005
2017-05-26 11:50:13 -04:00
Masayuki Nakano 94170bb269 Bug 1370198 Back out the patch for bug 1369419 because the patch couldn't fix bug 1361132 and causes new crash bugs r=jimm
MozReview-Commit-ID: Eq7Dkk6Ghwc

--HG--
extra : rebase_source : edc2d10f4e0359152927830a269f74b75daf5883
2017-06-05 21:52:16 +09:00
Sebastian Hengst 5e4755df68 merge mozilla-inbound to mozilla-central. r=merge a=merge
MozReview-Commit-ID: IdRgoZ9hiXq
2017-06-03 20:15:27 +02:00
Jim Chen 4230e5eacd Bug 1369107 - 4. Remove GeckoAppShell.ContextGetter; r=snorp
Remove ContextGetter and all implementations now that it's no longer
being used.

MozReview-Commit-ID: G5jWFsQPfWw
2017-06-02 16:13:42 -04:00
Jim Chen c611475410 Bug 1369107 - 3. Don't use GeckoAppShell.getContext for plugin code; r=snorp
Use getApplicationContext or GeckoApp.getPluginContext instead of
getContext so we can get rid of ContextGetter.

MozReview-Commit-ID: B8tTPeZ5gvt
2017-06-02 16:13:41 -04:00
Masayuki Nakano d3789d2d6b Bug 1369419 GetMessage() and PeekMessage() shouldn't be used directly as far as possible r=jimm
In TSF mode, application should retrieve messages with ITfMessagePump::GetMessage() or ITfMessagePump::PeekMessage() since TSF/TIP may handle the message before or after the host application handles it.

This patch rewrites the API users with WinUtils::(Get|Peek)Message() which use ITfMessagePump if it's available.

MozReview-Commit-ID: LwHIgp7SxLH

--HG--
extra : rebase_source : aa5750af9812f9b107c29546cbee6f9eede6ebfa
2017-06-02 12:02:35 +09:00
James Willcox bd21ee4811 Bug 1367287 - Reference count GeckoSurfaceTexture r=jchen
MozReview-Commit-ID: 1JJVzCmANyH
2017-06-02 10:14:20 -05:00
Bob Owen 595abc49e3 Bug 1369386 Part 2: Follow-up #if to fix non-Windows build failures. r=me 2017-06-02 12:07:31 +01:00
Bob Owen 508a716403 Bug 1369386: Fix old invalid millimeter page size prefs using print_paper_size_type as trigger. r=jimm
These old invalid prefs were caused by us not clearing out old inch page size prefs when the
user switched to a standard mm page size (for example A4).
This fix uses the fact that an old print_paper_size_type pref was removed at the same time
to trigger a one off sanitization.
2017-06-02 11:58:44 +01:00
Karl Tomlinson 94ab0b4d5c Bug 1368597 label CompositorWidget and WindowSurface as Graphics BUG_COMPONENT r=rhunt
--HG--
extra : rebase_source : 5f2cf496030a5736424e66be1a3b8a53188d5e22
2017-05-30 14:29:22 +12:00
Carsten "Tomcat" Book f69c1db21b Merge mozilla-central to autoland 2017-06-01 13:57:22 +02:00
Carsten "Tomcat" Book 972b7a5149 merge mozilla-inbound to mozilla-central a=merge 2017-06-01 13:50:56 +02:00
Ryan VanderMeulen 984edc2be4 Merge m-c to inbound. a=merge 2017-05-31 20:26:16 -04:00
Ryan VanderMeulen b2f44e5181 Merge inbound to m-c. a=merge 2017-05-31 20:25:52 -04:00
Martin Stransky 5f9fc84d1a Bug 1364355 - Implement UpdateOpaqueRegion() by gdk_window_set_opaque_region(), r=karlt
MozReview-Commit-ID: 50N99ZlvFUh

--HG--
extra : rebase_source : d095f93a858271c8a622f127724b06246e9f4bf9
2017-05-26 21:27:32 +02:00
Martin Stransky 0d7ee0c159 Bug 1364355 - Add support for using ARGB windows with OMTC when an X11 compositor is active. r=karlt
Authored by Andrew Comminos <andrew@comminos.com>

MozReview-Commit-ID: FIQBHSXgjMh

--HG--
extra : rebase_source : 1ae73bed3b2b933d11803ae3b93afbf3d1e1f570
2017-06-01 11:28:50 +02:00
Andrew Comminos 302c93427f Bug 1364355 - Enable argb visual for GTK window behind hidden preference mozilla.widget.use-argb-visuals, r=karlt
This preference is default to false and allows to experiment with transparent widgets on Gtk.
Original patch autor is Andrew Comminos <andrew@comminos.com>.

MozReview-Commit-ID: JZkCjBWny3m

--HG--
extra : rebase_source : 116c4977d7d7f3927e6a4df203584d8b9c9ad57b
2017-05-25 15:45:22 +02:00
Botond Ballo 4810f9b6a9 Bug 1369074 - Store the layers id in a variable of the proper type (uint64_t) in nsBaseWidget::StartAsyncScrollbarDrag(). r=kats
MozReview-Commit-ID: App50MV57JJ

--HG--
extra : rebase_source : 5892eea589227df0511e63bc260d99ece9cbf9a4
2017-05-31 13:48:01 -04:00
Brendan Dahl 9541cf8db0 Bug 1367227 - Part 1 - Match headless fullscreen behavior to other widget backends. r=jrmuizel
Restore size mode after leaving fullscreen. Also, trigger fullscreen
and size mode changed events during MakeFullScreen.
2017-05-31 16:29:10 -07:00
Andrea Marchesini f9b3992646 Bug 1368941 - Remove gonk specific code from ProcessPriorityManager, r=gsvelto 2017-05-31 21:23:52 +02:00
Carsten "Tomcat" Book 1d032e6328 Merge mozilla-central to autoland 2017-05-31 12:35:56 +02:00
Jorg K 09557f6df0 Bug 1353625 - detect zero (0) as invalid resolution. r=bobowen a=tomcat 2017-05-30 23:44:00 +02:00
Ryan VanderMeulen bd262b8cd1 Merge inbound to m-c. a=merge 2017-05-30 16:09:14 -04:00
Kilik Kuo 444875671e Bug 1350253 - Provide GeckoHlsDemuxerWrapper & corresponding generated JNI files as the glue for HLSDemuxer and GeckoHlsPlayer. r=jolin,jya
MozReview-Commit-ID: 62cnGQsmdNS

--HG--
extra : rebase_source : 4431d9e55f4ee38faeae50d03b46fbb57629fb8e
2017-05-25 20:48:03 +08:00
James Cheng ce2fbd2836 Bug 1350250 - Add a GeckoHlsResourceWrapper java glue. r=jolin,jya
MozReview-Commit-ID: KMAqkJSk2tT

--HG--
extra : rebase_source : efe73b1f70e5bde6c6d9fd7ad36ee98ad21eaf7c
2017-05-17 17:05:42 +08:00
Andrea Marchesini c1c91c120d Bug 1368666 - Get rid of screen-state-changed observer in ProcessPriorityManager, r=gsvelto 2017-05-30 16:27:12 +02:00
Carsten "Tomcat" Book b318c7dca7 merge mozilla-inbound to mozilla-central + UPGRADE_NSS_RELEASE a=merge
--HG--
rename : toolkit/components/extensions/test/xpcshell/xpcshell.ini => toolkit/components/extensions/test/xpcshell/xpcshell-common.ini
extra : amend_source : 458fd54fe8070ca3034ac441267ff7025adb5251
2017-05-30 11:37:46 +02:00
Kris Maglione a176402b70 Bug 1365660: Part 5b - Enable compositing for popups with remote content on Windows. r=bas
MozReview-Commit-ID: GzpCBUnVkdY

--HG--
extra : rebase_source : bc2a6459e64847b0153e09e392173bac00d97b4e
2017-05-29 14:58:48 -07:00
Kilik Kuo 7f124232fa Bug 1350241 -Part1: Provide data structure of {Audio,Video}Info and demuxed sample for the use of HLS on Fennec. r=jolin,jya
MozReview-Commit-ID: 11triLS84Gq

--HG--
extra : rebase_source : a1c0d04a476560156ddf4919ab4fc277833b4db1
2017-05-25 12:46:06 +08:00
Manish Goregaokar 616a37d6b7 Bug 1367275 - Part 2: stylo: Support -moz-window-shadow; r=xidorn
MozReview-Commit-ID: LiYlfmaCfHH
2017-05-27 13:21:44 -07:00
Phil Ringnalda c3d89ab945 Backed out 3 changesets (bug 1364355) for timeouts in test_panel_anchoradjust.xul
Backed out changeset 7cc2790574dc (bug 1364355)
Backed out changeset 2ee6598089a3 (bug 1364355)
Backed out changeset c4c2645ca8b7 (bug 1364355)

MozReview-Commit-ID: 5JEGvHeCWIv
2017-05-26 20:30:45 -07:00
Martin Stransky eb881749c2 Bug 1364355 - Implement UpdateOpaqueRegion() by gdk_window_set_opaque_region(), r=karlt
MozReview-Commit-ID: 50N99ZlvFUh

--HG--
extra : rebase_source : 63bcdeaa5f7129a9f0d171d9826f33b8dc6bbe34
2017-05-26 21:27:32 +02:00
Martin Stransky 3d05b54851 Bug 1364355 - Add support for using ARGB windows with OMTC when an X11 compositor is active. r=karlt
Authored by Andrew Comminos <andrew@comminos.com>

MozReview-Commit-ID: FIQBHSXgjMh

--HG--
extra : rebase_source : 22885a2846ec0bc718196d6e16bacb37b0410b25
2017-05-25 15:55:20 +02:00
Andrew Comminos f2d7adfe22 Bug 1364355 - Enable argb visual for GTK window behind hidden preference mozilla.widget.use-argb-visuals, r=karlt
This preference is default to false and allows to experiment with transparent widgets on Gtk.
Original patch autor is Andrew Comminos <andrew@comminos.com>.

MozReview-Commit-ID: JZkCjBWny3m

--HG--
extra : rebase_source : 18a3dd1ecb2ef87910307e3085a343a4d75403af
2017-05-25 15:45:22 +02:00
Stephen A Pohl 352c559fa6 Bug 1368077: Strip page titles from URLs on the pasteboard when dragging on OSX/macOS. r=mstange 2017-05-26 11:45:28 -04:00
Ryan VanderMeulen 8543438674 Merge inbound to m-c. a=merge 2017-05-25 16:34:16 -04:00
Stephen A Pohl b633564ae8 Bug 1367720: Restore ability to drag & drop files from Finder to Firefox. r=mstange 2017-05-25 10:54:52 -04:00
Farmer Tseng f5f7370562 Bug 1359713 - Stop loading the contents of PDF files into memory in PDFViaEMFPrintHelper r=jwatt
This change changes PDFViaEMFPrintHelper so that instead of loading the entire
contents of a PDF document into memory to pass to PDFium, it now simply passes
PDFium a file system path to load the PDF from. This should help us reduce
memory use (and potentially avoid running out of memory) when larger PDFs are
printed.


MozReview-Commit-ID: GISolZYC2GJ

--HG--
extra : rebase_source : a59925542b6c075cd14c67577653163cc2d49c0f
2017-05-12 12:42:20 +08:00