Commit Graph

1685 Commits

Author SHA1 Message Date
Olli Pettay
63fd91b1fa Bug 534226 - Remove support for multiple presshells, r=bz, sr=roc 2010-01-07 12:36:11 +02:00
Robert O'Callahan
639decd49d Bug 527864. Test that we flush reflows before painting. 2010-01-07 14:12:21 +13:00
Robert O'Callahan
d3d23e846e Bug 532732. When rendering a background-attachment:fixed background to the screen, limit the background painting area to the viewport, since this is the only area where it will actually be visible and in many cases this means we don't need to think about tiling it. r=dbaron 2010-01-07 13:55:09 +13:00
Michael Ventnor
d1c7f0fda1 Bug 537471 - Detect empty shadow rects earlier. r=roc
--HG--
extra : rebase_source : 3d1a20314c9a771e987d9f82df0620d9d29720a6
2010-01-06 19:02:11 +01:00
Jacek Caban
e787295d9f Bug 530448 - Make TestPoisonArea.cpp compile on mingw, r=zwol 2010-01-05 18:29:08 -05:00
Simon Montagu
77681c9d20 Call AdjustOffsetsForBidi on a new continuation frame immediately after creating it. Bug 536963, r=roc 2010-01-03 00:46:19 -08:00
Timothy Nikkel
2178cc013a Bug 537562. Add crashtest. 2010-01-02 18:57:40 -06:00
Timothy Nikkel
97bc36029b Bug 499841. Fix a failed attempt at a merge. 2010-01-02 18:56:29 -06:00
Timothy Nikkel
c4d4fe1182 Bug 537059. Add crashtest.
--HG--
extra : rebase_source : 7a22e75f66a58c08d9ff25f264617ac31aea65b4
2010-01-02 02:22:22 -06:00
Timothy Nikkel
c581600f0c Bug 396367. Make the tests reset the textzoom after changing it.
--HG--
extra : rebase_source : 5afddd159099314576ad6623d11649266a7b4e77
2009-12-26 21:36:12 -06:00
Timothy Nikkel
bd15675d9d Bug 396367. Make sure that we have flushed content notifications anytime we process restyles that might cause frame construction. r=bzbarsky
--HG--
extra : rebase_source : c6aa412247ab70f5bf3a4688f262edcccd2d6fdd
2009-12-22 18:09:10 -06:00
Timothy Nikkel
b3a04ada9c Bug 499841. Add crashtest. 2010-01-01 20:49:24 -06:00
Timothy Nikkel
1172cc4cd0 Backed out changeset 63d4a49fbec1 (bug 396367) 2010-01-01 20:34:05 -06:00
Timothy Nikkel
6a84e052f2 Bug 396367. Make sure that we have flushed content notifications anytime we process restyles that might cause frame construction. r=bzbarsky
--HG--
extra : rebase_source : 67d99a197a4974684f57f23873033b85c276ef86
2009-12-22 18:09:10 -06:00
Timothy Nikkel
ad77d4e1d7 Bug 484400. When removing a positioned element use the placeholder frame to get the relevant containing block for first-letter processing. r=bzbarsky
--HG--
extra : rebase_source : 9a327e8f44541ac822bc7be52caa1ceab825ccd1
2009-12-22 17:47:25 -06:00
Timothy Nikkel
5f6a955352 Bug 499841. Pass the block frame that contains the first letter style to RemoveFirstLetterFrames so that the NS_BLOCK_HAS_FIRST_LETTER_CHILD bit can be unset on it. Always set the NS_BLOCK_HAS_FIRST_LETTER_CHILD bit on the first continuation only. r=bzbarsky
--HG--
extra : rebase_source : 411297490d6d18244fa7caa3194facfedc28ea86
2009-12-22 17:44:35 -06:00
Boris Zbarsky
cba4b340ad Trying to fix orange due to landing of bug 528306 by fixing assumption in test. 2009-12-31 17:12:43 -05:00
Boris Zbarsky
06b277fcc7 Bug 528306 part 3. Hook up restyle processing to nsRefreshDriver. r=dbaron 2009-12-31 14:07:57 -05:00
Boris Zbarsky
ef3cef478b Bug 528306 part 2. Make nsCSSFrameConstructor reference-counted. r=dbaron 2009-12-31 14:07:57 -05:00
Boris Zbarsky
5e1384376d Bug 528306 part 1. Don't stop the refresh driver timer on observer removal; instead just stop it if it fires when there are no observers. r=dbaron 2009-12-31 14:07:57 -05:00
Boris Zbarsky
35c346dcdf Bug 536623 and bug 537141. Don't attempt to maybe-create textframes for kids of parents that have xbl childlists on insert and append: we know we didn't skip creating any frames that didn't need skipping unconditionally in that situation, and trying to create them can put them in the wrong place. r=roc 2009-12-31 14:07:57 -05:00
Boris Zbarsky
a99f0f4f96 Bug 536716. Make BuildScrollFrame not set the primary frame on the content, fix letter frames to reset primary frames to 0 before resetting them to the new value as needed, fix things so <area> never gets a frame of its own, and reenable the SetPrimaryFrame assertion when the primary frame is being changed from one non-null frame to another. r=roc 2009-12-31 14:07:56 -05:00
L. David Baron
61bd9cfc56 Revert bulk of changeset 67ddbe030ab6 (originally for bug 457809) now that bug 253354 is fixed. 2009-12-31 12:49:22 -05:00
Neil Deakin
989819833a Bug 536486, remove IsDestroyingFrames check now that frames are available from content nodes, fixes possible leak, r=roc 2009-12-31 12:36:56 -05:00
L. David Baron
8483573833 Add nsPresContext::EnsureSafeToHandOutCSSRules. (Bug 536379) r=bzbarsky 2009-12-31 10:56:33 -05:00
Nochum Sossonko
1aa92b8e9b merge for backout 2009-12-29 17:34:15 -05:00
Nochum Sossonko
616050ef52 Backed out changeset 58a672dcb695 due to test failures 2009-12-29 17:33:09 -05:00
Boris Zbarsky
c3626d1503 Bug 536718. Clean up NotifyDestroyingFrame. r=roc 2009-12-29 15:13:54 -05:00
Boris Zbarsky
b472db2cb1 Bug 508473 followup. Remove some more dead code. 2009-12-29 15:13:53 -05:00
Jacek Caban
a736eeeb19 Bug 530448 - Make TestPoisonArea.cpp compile on new mingw, r=zwol 2009-12-29 14:24:43 -05:00
Boris Zbarsky
e76b6c5c78 Bug 500882 followup. Restore bit that shouldn't have gone away. 2009-12-27 08:41:08 -05:00
Daniel Holbert
178307b6c0 Bug 535004: Check if we've fired an unload event before calling OnPageShow, in DocumentViewerImpl::LoadComplete. (Now without NS_ABORT_IF_FALSE, since other things can trigger it.) r=smaug 2009-12-26 15:11:04 -08:00
Boris Zbarsky
8519610b49 Bug 500882 part 7. Remove the old GetPrimaryFrameFor API. r=roc, sr=dbaron 2009-12-24 16:20:06 -05:00
Boris Zbarsky
b3883855ac Bug 500882 part 5. Switch layout module to using the new GetPrimaryFrame API. r=roc 2009-12-24 16:20:06 -05:00
Boris Zbarsky
df21dcfefd Bug 500882 part 3. Switch accessibility module to using the new GetPrimaryFrame API. r=surkov, sr=roc 2009-12-24 16:20:05 -05:00
Boris Zbarsky
b2e4054198 Bug 500882 part 2. Switch setting of primary frames to the new API; switch nsFrameManager::GetPrimaryFrameFor to the new API; remove resulting gobs of dead code. r=roc 2009-12-24 16:20:05 -05:00
fantasai
1e880721fb Bug 508473 Part II: Remove DeletingFrameSubtree r=bz sr=roc 2009-12-24 00:20:41 -05:00
fantasai
714c1668f7 Bug 508473 Part I: Remove nsIFrame::RemovedAsPrimary r=bz sr=roc 2009-12-24 00:20:20 -05:00
L. David Baron
683aa648c1 Backed out changeset 2752efeb2fdd (Bug 535004: Check if we've fired an unload event before calling OnPageShow, in DocumentViewerImpl::LoadComplete) because the NS_ABORT_IF_FALSE that it added is firing in reftest and crashtest, at least on Linux. 2009-12-23 23:05:15 -05:00
Daniel Holbert
43ca8b575b Bug 535004: Check if we've fired an unload event before calling OnPageShow, in DocumentViewerImpl::LoadComplete. r=smaug 2009-12-23 16:19:26 -08:00
Boris Zbarsky
976c983e86 Bug 535721. Deal with MoveChildrenTo on non-siblings. r=roc 2009-12-23 16:17:23 -05:00
Boris Zbarsky
32cfd3d5a7 Bug 535911. Silence the bogus assertion. r=roc 2009-12-23 16:17:23 -05:00
L. David Baron
b4840861ab Make nsRefreshDriver hold a reference to its observers while notifying them. (Bug 531585) r=bzbarsky 2009-12-23 14:10:31 -05:00
L. David Baron
179ee34b4d Make the transition manager reference-counted. (Bug 531585) r=bzbarsky 2009-12-23 14:10:31 -05:00
L. David Baron
8ee9665a58 Make nsRefreshDriver reference-counted instead of being a sub-object of the pres context. (Bug 531585) r=bzbarsky 2009-12-23 14:10:30 -05:00
Benjamin Smedberg
c2bbf9c677 Back out all the pieces of bug 531585 (transitioned event for end of CSS transitions) because of timeouts in test_transitions_events.html 2009-12-23 10:15:26 -05:00
L. David Baron
026f0e2d0b Make nsRefreshDriver hold a reference to its observers while notifying them. (Bug 531585) r=bzbarsky 2009-12-23 06:35:45 -05:00
L. David Baron
6294ba95b7 Make the transition manager reference-counted. (Bug 531585) r=bzbarsky 2009-12-23 06:35:45 -05:00
L. David Baron
262ab8b4d1 Make nsRefreshDriver reference-counted instead of being a sub-object of the pres context. (Bug 531585) r=bzbarsky 2009-12-23 06:35:45 -05:00
L. David Baron
eae023b6a3 Backout revisions fa5326c011b8, 8b22441911b0, and cfa10b01b1f6 (bug 531585) on suspicion of causing random orange bug 536382. 2009-12-22 15:46:23 -05:00
L. David Baron
dadb7490da Make nsRefreshDriver hold a reference to its observers while notifying them. (Bug 531585) r=bzbarsky 2009-12-21 16:46:25 -05:00
L. David Baron
d0934f9f1f Make the transition manager reference-counted. (Bug 531585) r=bzbarsky 2009-12-21 16:46:25 -05:00
L. David Baron
c4c186dd89 Make nsRefreshDriver reference-counted instead of being a sub-object of the pres context. (Bug 531585) r=bzbarsky 2009-12-21 16:46:24 -05:00
Olli Pettay
bf7962ffed Bug 535851 - noscript is rendered in print preview, r=roc 2009-12-19 12:35:54 +02:00
Masayuki Nakano
68361828eb Bug 532422 - n900: Cannot select any characters from sym map. r=enndeakin. 2009-12-17 10:42:02 -08:00
Boris Zbarsky
bdc41dba60 Bug 531148 followup. Actually return a value here! 2009-12-17 02:38:12 -05:00
Robert O'Callahan
8e327b88bf Bug 516885 / Bug 525375: Fix (harmless) assertions due to nsDisplayTransform passing non-null aVisibleRegionBeforeMove. r=dbaron 2009-12-12 13:27:15 -08:00
Robert Longson
819b6997d7 Bug 523576 - Fix compilation with disable-svg. r=dbaron 2009-12-12 19:43:34 +00:00
Robert O'Callahan
19da2f1872 Bug 530686. Only create an nsDisplayTableBorderBackground when it's going to have something to paint, since having one can disable some scrolling optimizations. r=dbaron 2009-12-12 10:50:30 -08:00
Robert O'Callahan
97c78b1258 Bug 531461. Avoid creating an nsDisplayTableBorderBackground if every part of the table is hidden. r=dbaron 2009-12-12 10:50:25 -08:00
Boris Zbarsky
4e9f77e7b1 Bug 531148. Fix appends of a block to the trailing inline of an {ib} split which has an inline parent. r=roc 2009-12-11 22:36:34 -08:00
Boris Zbarsky
af198139a9 Bug 526596. Make sure appends to an {ib} split coalesce blocks correctly if the appended frames start with a block and the trailing inline of the split is empty. r=roc 2009-12-11 22:36:33 -08:00
Olli Pettay
2f4c5ea761 Bug 487667 - Clone documents for printing, r=roc+jst+joe+dbaron
--HG--
extra : rebase_source : dbab242066ba97c0d7f34a2fb701a1c465f26ec4
2009-12-10 20:02:13 -08:00
Boris Zbarsky
99027aa785 Bug 525608 part 3. Change all style set consumers that ask for a pseudo style context to indicate what they're actually asking for (pseudo-element, anonymous box, or xul tree thing). r=dbaron 2009-12-11 02:37:40 -05:00
Boris Zbarsky
e9f228c7d5 Backing out bug 525608 and bug 525952 2009-12-10 18:28:08 -08:00
Boris Zbarsky
0a82b262db Bug 525608 part 3. Change all style set consumers that ask for a pseudo style context to indicate what they're actually asking for (pseudo-element, anonymous box, or xul tree thing). r=dbaron 2009-12-10 14:36:05 -08:00
Boris Zbarsky
cc6eb188e6 Bug 523294 part 3. Get rid of the now-unused aStateMask argument to AttributeChanged. r=dbaron 2009-12-10 14:36:04 -08:00
Boris Zbarsky
129197fdc7 Bug 523294 part 2. Get rid of the aAttribute argument to SelectorMatches; handle attribute changes by checking for matches on selectors including the given attribute both before and after the change. r=dbaron 2009-12-10 14:36:03 -08:00
Olli Pettay
ec076a94d3 Bug 528829 - If mouse button is held down during page transition, Firefox stops responding to mouse clicks, r=roc 2009-12-10 10:58:33 -08:00
Timothy Nikkel
d16beb2417 Bug 525751. When using a subdocument frame as a fallback for drawing canvas background color override the canvas frame check. r=roc a=roc
--HG--
extra : rebase_source : 6719372b7404a598a4ce1c1f5190a4f86d0d896b
2009-12-01 02:49:53 -06:00
Zack Weinberg
e125669860 Bug 531847: Include address chosen for frame poisoning in breakpad reports. r=ted a192=roc 2009-12-01 14:48:23 -08:00
Robert O'Callahan
d866e87ce4 Bug 531371. Correctly propagate the backround from a <body style="display: table"> to the viewport as needed. r=bzbarsky 2009-12-01 12:21:00 -05:00
Robert O'Callahan
1eb1c0bb6b Bug 530472. Make sure to check across document boundaries to see whether the clipFrame is an ancestor of the moving frame. r=dbaron 2009-11-30 18:14:04 -05:00
Robert O'Callahan
6780628363 Bug 530686. Don't redraw box-shadows over the bounding rect of the dirty region; just redraw them over a (somewhat simplified) dirty region. r=dbaron 2009-11-30 18:14:04 -05:00
Zack Weinberg
cf67dd36c8 Bug 530598: If we can't get the preferred address for a poison page, take whatever the OS will give us, rather than giving up 2009-11-30 08:51:07 -08:00
Serge Gautherie
2a485f24e8 Bug 381467 - Switch application/x-javascript to application/javascript; (Ev1) /layout/base/crashtests/123946-1.html, EOL/whitespace fix (only).
r=dbaron.
2009-11-30 03:44:02 +01:00
Olli Pettay
2a7bca1c92 Bug 530880, r=bz, sr=roc
--HG--
extra : rebase_source : 4f1e95223fd0c167fa5ec882b46ae36f9edbb071
2009-11-26 11:12:20 +02:00
Ginn Chen
546bba6b66 Bug 529764 nsPresArena.cpp failed to compile on Solaris r=zweinberg 2009-11-20 14:11:42 +08:00
Robert O'Callahan
a3f1f75383 Bug 528493. Prevent script execution while initializing nsDocumentViewer. r=bzbarsky 2009-11-19 12:40:31 +13:00
Robert O'Callahan
a3558ce3da Backed out changeset 93a0acf68dd6 2009-11-19 15:15:44 +13:00
Robert O'Callahan
5fa533b23e Bug 528493. Prevent script execution while initializing nsDocumentViewer. r=bzbarsky 2009-11-19 12:40:31 +13:00
Boris Zbarsky
e38ec39351 Bug 505059 build bustage fix. 2009-11-18 08:58:44 -05:00
Boris Zbarsky
71f48241bd Bug 505059. Don't shut down the layout module while someone is holding a reference to the stylesheet service. r=dbaron 2009-11-18 08:25:02 -05:00
Boris Zbarsky
679d2edfd4 Bug 526375. Enforce always appending to the last continuation for {ib} splits. r=roc 2009-11-18 08:25:02 -05:00
Robert O'Callahan
54824bc7d1 Bug 528604. Don't add areas-to-repaint when scrolling for non-moving clip rects that aren't clipping any moving content. r=dbaron 2009-11-18 13:38:17 +13:00
Boris Zbarsky
2b37acc71b Bug 521542. Fix IsValidSibling to not lie when non-table-related siblings are involved. r=bernd 2009-11-17 17:50:04 -05:00
L. David Baron
e3330c4996 Null check mShell in nsPresContext::MediaFeatureValuesChanged. (Bug 528832) r=bzbarsky 2009-11-17 13:11:15 -08:00
Zack Weinberg
3c349b25a6 Bug 522088 and bug 507924: Ensure that the value used for frame poisoning is a pointer to an inaccessible page of virtual memory. 2009-11-17 11:17:20 -08:00
Boris Zbarsky
c1595ce69a Bug 518114 part 2. Use state bits to keep track of what a placeholder is a placeholder for. r=roc 2009-11-16 16:00:07 -05:00
Boris Zbarsky
ffab603a47 Bug 518114 part 1. Don't force a virtual GetType call in GetFloatFromPlaceholder. r=roc 2009-11-16 16:00:07 -05:00
David Bolter
ae4129ee61 Bug 525579 - topcrash [@ nsAccessibilityService::GetAccessible(nsIDOMNode*, nsIPresShell*, nsIWeakReference*, nsIFrame**, int*, nsIAccessible**)] (thanks dbaron) r=roc,surkov 2009-11-16 09:03:53 -05:00
Mark Finkle
13b8532386 Bug 507817: BorderImage should not call ExtractCurrentFrame each time it draws [r=dbaron r=joedrew] 2009-11-13 23:23:00 -05:00
L. David Baron
0d9c6b626c Make DEBUG_TRACEMALLOC_PRESARENA compile again by spelling PRUint32 correctly. 2009-11-12 17:40:05 -08:00
Boris Zbarsky
42cd75865c Bug 526178 followup. Remove now-unused AutoEnqueueBinding. 2009-11-12 14:27:33 -05:00
Boris Zbarsky
dcde9c8929 Bug 528002. InvalidateCanvasIfNeeded can do a lot less work. r=roc 2009-11-11 18:28:16 -05:00
Simon Montagu
df69e48755 Drill down through placeholder frames to the real frames in bidi reordering. Bug 525740, r=roc 2009-11-08 09:40:46 -08:00
Boris Zbarsky
f35446936d Bug 526178. Fix constructor firing order to match Gecko 1.9.1 again. r=roc 2009-11-04 17:58:13 -05:00
Boris Zbarsky
53f028f33b Backed out changeset 2fa27d8cd3d2 (bug 526178) to fix browser-chrome orange.
--HG--
extra : rebase_source : 6be60f6e6586e924f7607f135a3589d9e425a0ec
2009-11-04 20:41:06 -05:00
Boris Zbarsky
482ee7e084 Bug 526178. Fix constructor firing order to match Gecko 1.9.1 again. r=roc 2009-11-04 17:58:13 -05:00
Robert O'Callahan
7e04a64cff Bug 513082. Take advantage of our knowledge that nsDisplayClip, when used in a blitting scrolling frame, must snap to device pixels, to get more accurate blit regions. r=dbaron 2009-11-04 07:39:42 +13:00