Bug 1686106 - Allow the character map record created for unicode-range to be shared among FontFaces with the same descriptor. r=heycam This will reduce memory footprint when there are multiple @font-face rules with the same unicode-range descriptor, which is common when, for example, a site loads several styles of a family, all with the same character repertoire. Differential Revision: https://phabricator.services.mozilla.com/D101411

Bug 1685675 - Ship :autofill / :-webkit-autofill. r=heycam Spec PR at https://github.com/whatwg/html/pull/6273 Tests at https://github.com/web-platform-tests/wpt/pull/27098 Differential Revision: https://phabricator.services.mozilla.com/D101148

Bug 1685950 - update cssparser. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D101300

Bug 1685917 - Remove layout.css.unknown-webkit-pseudo-element. r=heycam Enabled since bug 1486325 (Firefox 64). Depends on D101290 Differential Revision: https://phabricator.services.mozilla.com/D101291

Bug 1685917 - Remove layout.css.file-selector-button.enabled pref. r=heycam On by default since bug 1662478 (Firefox 82) Depends on D101289 Differential Revision: https://phabricator.services.mozilla.com/D101290

Bug 1685917 - Remove the moz-focus-outer code. r=heycam This pseudo does nothing and we successfully removed it in bug 1655859 (firefox 81). Differential Revision: https://phabricator.services.mozilla.com/D101289

Bug 1684497 - Hoist ResetGlyphRuns out of the conditional statement. r=heycam This matches how it's done at the other callsite. (Also add an assertion that would've caught the omission sooner.) Differential Revision: https://phabricator.services.mozilla.com/D100779

Bug 1683295 - Make next_in_preorder generate slightly better code. r=heycam This avoids the panic code in release builds. Differential Revision: https://phabricator.services.mozilla.com/D100094

Bug 1682882 - Don't crash when trying to figure out the parent style of a node about to be reframed. r=heycam This is basically the situation described here: https://searchfox.org/mozilla-central/rev/8d722de75886d6bffc116772a1db8854e34ee6a7/layout/generic/nsIFrame.cpp#9961-9964 Differential Revision: https://phabricator.services.mozilla.com/D100097

Bug 1682988 - Use UTF8String for querySelector and similar OM methods. r=heycam This has no observable behavior change because we were converting to UTF-8 in ParseSelectorList. Differential Revision: https://phabricator.services.mozilla.com/D99961

Bug 1682003 - Avoid UTF-8 -> UTF-16 conversion during CSSOM serialization. r=heycam This lifts a bunch of string conversions higher up the stack, but allows us to make the servo code use utf-8 unconditionally, and seemed faster in my benchmarking (see comment 0). It should also make a bunch of attribute setters faster too (like setting .cssText), now that we use UTF8String for them (we couldn't because we couldn't specify different string types for the getter and setters). Differential Revision: https://phabricator.services.mozilla.com/D99590

Bug 1577530 - Cache theming values from nsTextControlFrame correctly. r=heycam The issue is that nsTextControlFrame might have a different used padding then the editing root block and thus the editing root's scrolled content, etc, because its padding doesn't match the style. This causes TextOverflow code to miscompute the content area size, which is sadness and causes the overflow markers to go over the padding box. The REFLOW_ROOT condition was introduced in bug 157846 and only partially takes care of the issue. This fix supersedes it, so we can remove it. Differential Revision: https://phabricator.services.mozilla.com/D99838

Bug 1470983 - Prelude: Remove use of gtk_widget_get_settings in LookAndFeel. r=heycam The GTK nsLookAndFeel implementation uses `gtk_widget_get_settings` on temporary widgets that aren't attached to anything, which the documentation says not to do. Empirically, this seems to return the same settings object as `gtk_settings_get_default`. Explicitly using `gtk_settings_get_default` is useful for remote look-and-feel, so that the parent process can register for property changes on that object to know when to update the child processes. Differential Revision: https://phabricator.services.mozilla.com/D94535

Bug 1603455 - Remove layout.css.xul-tree-pseudos.content.enabled. r=heycam We unshipped these a while ago and left the pref just for testing purposes. But now all the reftests using it were conveniently migrated to chrome:// tests, so we no longer need it. Differential Revision: https://phabricator.services.mozilla.com/D56950

Bug 1680084 - Use a more precise rule start for sanitization. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D98710

Bug 1680558 - Update cssparser to get ParserState in rule parser. r=heycam The changes should be trivial. The third_party changes are up for review in https://github.com/servo/rust-cssparser/pull/277 (and of course I'll land with a bump to 0.28 rather than the override after that gets r+'d). The basic idea is that with this we have the actual start offset of the rule, so we wouldn't include html comments or other invalid stuff we discard during sanitization in bug 1680084. But that's a separate change. Differential Revision: https://phabricator.services.mozilla.com/D98677

Bug 1680558 - Update cssparser to get ParserState in rule parser. r=heycam The changes should be trivial. The third_party changes are up for review in https://github.com/servo/rust-cssparser/pull/277 (and of course I'll land with a bump to 0.28 rather than the override after that gets r+'d). The basic idea is that with this we have the actual start offset of the rule, so we wouldn't include html comments or other invalid stuff we discard during sanitization in bug 1680084. But that's a separate change. Differential Revision: https://phabricator.services.mozilla.com/D98677

Bug 1679208 - Fix pre-existing unit confusion in gtk scrollbar painting. r=heycam This is very visible with the previous changes in the bug. Differential Revision: https://phabricator.services.mozilla.com/D98240

Bug 1679208 - Use more strongly typed rects in nsNativeBasicTheme. r=heycam As per review comments. Depends on D97937 Differential Revision: https://phabricator.services.mozilla.com/D98099

Bug 1679208 - Improve non-native theme with fractional scaling. r=heycam This greatly improves non-native theme on Windows in presence of fractional DPI (see bug 1657191). This matches what the Windows native theme does, to my knowledge, see ScaleForFrameDPI in nsNativeThemeWin.h. However I think the right fix is to not make these LayoutDeviceInt* things but LayoutDevice*. That's a bigger fix, and fits well as a follow-up to this bug, so let's land this in the meantime. Differential Revision: https://phabricator.services.mozilla.com/D97937

Bug 1675015: Add Windows high contrast support to non-native theme. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D96803

Bug 1677187 - ignore x values for vertical and y values for horizontal text in textPaths r=heycam Differential Revision: https://phabricator.services.mozilla.com/D97086

Bug 1676843 - Make innerWidth/innerHeight return the actual CSS size rather than a rounded version of it. r=heycam On Nightly, for now. This allows websites to get the real viewport size. The rounded size has caused problems in the past (see bug 1648298 / bug 1648265), and changing it would be ideal. I told the CSSWG that we can experiment with this on Nightly to alleviate (or prove) compat concerns, in which case we'd need to use a different solution, see https://github.com/w3c/csswg-drafts/issues/5260. Differential Revision: https://phabricator.services.mozilla.com/D96826

Bug 1676543 - Remove some invalid assertions. r=heycam We might clear this flag in PostRebuildAllStyleData before actually flushing style, so if the timing is wrong you can hit this assertion. This doesn't impact correctness in any way that I can see. Differential Revision: https://phabricator.services.mozilla.com/D96685

Bug 1634437 - Dispatch NativeInit as an idle runnable on Linux r=heycam EnsureInit takes a long time to run on Linux, in addition to it's timing, it also runs within a critical path of MSG_ConstructBrowser. Scheduling it as an idle runnable gives it a chance to run before MSG_ConstructBrowser. Differential Revision: https://phabricator.services.mozilla.com/D83230

Bug 1675853 - FontFace.load() doesn't need to flush the user fontset. r=heycam This one is even more subtle. This could cause a call from SetStatus() to OnFontFaceStatusChange() to not be done as part of the load() function (if the face is about to get added to the set or such). However, when it's actually added via UpdateRules, we'd properly invalidate the relevant state as well, so I think this is also fine, and our existing tests also agree. Depends on D96342 Differential Revision: https://phabricator.services.mozilla.com/D96343

Bug 1675950 - Check for explicit restyles posted to pseudo-elements to check whether we need a style flush in getComputedStyle. r=heycam Font loading invalidation may invalidate only pseudo-elements, like it happens on this test-case. Once we get to testing first-line / first-letter in the test-case (css/css-values/ch-pseudo-recalc-on-font-load.html), we end up flushing because those end up posting the restyle to the parent element correctly. Differential Revision: https://phabricator.services.mozilla.com/D96373

Bug 1675950 - Check for explicit restyles posted to pseudo-elements to check whether we need a style flush in getComputedStyle. r=heycam Font loading invalidation may invalidate only pseudo-elements, like it happens on this test-case. Once we get to testing first-line / first-letter in the test-case (css/css-values/ch-pseudo-recalc-on-font-load.html), we end up flushing because those end up posting the restyle to the parent element correctly. Differential Revision: https://phabricator.services.mozilla.com/D96373

Bug 1675853 - FontFace.load() doesn't need to flush the user fontset. r=heycam This one is even more subtle. This could cause a call from SetStatus() to OnFontFaceStatusChange() to not be done as part of the load() function (if the face is about to get added to the set or such). However, when it's actually added via UpdateRules, we'd properly invalidate the relevant state as well, so I think this is also fine, and our existing tests also agree. Depends on D96342 Differential Revision: https://phabricator.services.mozilla.com/D96343

Bug 1676083 - Use font-variant-numeric in the video control duration. r=heycam If you keep the controls showing on a video, depending on your fonts you can see the progress bar growing and shrinking depending on the current duration value. This is more noticeable at high video speed. We can use font-variant-numeric: tabular-nums; to make Gecko choose the right glyphs if available so that the size doesn't change over time. Differential Revision: https://phabricator.services.mozilla.com/D96370

Bug 1675853 - FontFace getters don't need to flush the fontset. r=heycam There's no way these need to do a flush as they just return data from the descriptors etc. Similarly .loaded also doesn't rely on the font face set. Differential Revision: https://phabricator.services.mozilla.com/D96341

[css-grid-3] Bug 1665152 - Add more editors to the CSS Masonry draft spec. r=heycam Per CSSWG meeting at TPAC: https://www.w3.org/2020/10/20-css-irc NPOTB DONTBUILD Differential Revision: https://phabricator.services.mozilla.com/D94204

Bug 1668875 - Invalidate images for media query changes more granularly. r=heycam The SVG code invalidating too much causes hangs with pages that have lots of SVGs. Differential Revision: https://phabricator.services.mozilla.com/D94186

Bug 1671863 - Skip reftest on mingw platform. r=heycam The RDD doesn't start on MinGW and so the AV1 decoder isn't available. Differential Revision: https://phabricator.services.mozilla.com/D93909

Bug 1490969 - Remove intermittent annotations on mask-image tests, r=heycam,mats,TYLin,emilio I couldn't reproduce this bug with a try push containing 20 retriggers, so I think it's safe to remove the annotations. If this comes up again as an intermittent, we can revert this patch and investigate further. Differential Revision: https://phabricator.services.mozilla.com/D93126

Bug 1484928 - Add an use counter for when people mix content: url() and <img src>. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D93071

Bug 1668253 - Convert hyphenation from ipc::SharedMemoryBasic to base::SharedMemory, and freeze blocks before sharing to content processes. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D91945

Bug 1669162 - Update mapped_hyph to 0.4.2 so that .dic parse errors are non-fatal. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D92435

Bug 1669459 - Introduce InspectorUtils.supports(). r=rcaliman,heycam Which allows to run CSS.supports in chrome/user-agent context. We should probably add a couple more tweaks for stuff like quirks or what not (the `ParsingMode` enum), but that seem lower priority. I only added the one-value version of CSS.supports because it should be enough and supporting the two value version required a bit of awkward code due to when we parse the CSS property name right now. Differential Revision: https://phabricator.services.mozilla.com/D92585

Bug 1669459 - Clean-up CSS.supports. r=heycam Use UTF8String rather than converting on the callee (which can avoid copies), and also it doesn't really throw, so don't claim it does. Differential Revision: https://phabricator.services.mozilla.com/D92583

Bug 1669600 - Properly distinguish between empty children and no children in rule iterator. r=heycam So that skip_children(), which just pops the stack and is used by the dynamic media query evaluation code, works as it should. Differential Revision: https://phabricator.services.mozilla.com/D92717

Bug 1665773 - Ensure CompressedGlyph::SetComplex() always clears the glyphCount to zero; let gfxShapedText::SetDetailedGlyphs() own the setting of this field. r=heycam We also rearrange bits in the CompressedGlyph so that the advance field doesn't need to be shifted, and remove unnecessary glyphCount parameters from various setter methods. Differential Revision: https://phabricator.services.mozilla.com/D90678

Bug 1667192 - patch 3 - For Android, package the uncompiled versions of hyphenation files. r=heycam This reduces the APK size by about 1.4MB. Differential Revision: https://phabricator.services.mozilla.com/D91411

Bug 1667192 - patch 2 - Update nsHyphenator to support loading either precompiled or uncompiled data. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D91410

Bug 1667192 - patch 1.1 - Run `mach vendor rust` to update in-tree mapped_hyph source. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D91409

Bug 1667192 - patch 1 - Update to mapped_hyph v0.4.0, to get FFI access to hyphenation table compilation. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D91408

Bug 1658342 - Enable the remoteObjectEmbed pref by default, r=heycam,valentin This also disables the cross-origin resource timing test when Fission is enabled, as resource timing is is broken for cross-origin frames (bug 1658097). Differential Revision: https://phabricator.services.mozilla.com/D91326

Bug 1658342 - Enable the remoteObjectEmbed pref by default, r=heycam,valentin This also disables the cross-origin resource timing test when Fission is enabled, as resource timing is is broken for cross-origin frames (bug 1658097). Differential Revision: https://phabricator.services.mozilla.com/D91326

Bug 1659227 - Handle SVG{Inner,Outer}Frame as the clip frame for partial pre-render. r=heycam,botond In SVG frame tree SVG{Inner,Outer}Frame is normally not scrollable but clips descendants, thus nsLayoutUtils::GetNearestScrollableFrame is not able to find the clip/scroll frame. So in this change, we introduce GetNearestOverflowClipFrame which is a similar function to GetNearestScrollableFrame but it also checks whether the frame is SVGInnerFrame or SVGOuterFrame which is clipping the given nsIFrame during walking up the frame tree. In this change we also change the logic to get the clip/scroll frame and the scroll id as follows; 1) Get the nearest overflow clip frame in the same document 2) Use the root frame if there is no clip frame 3) Then try to find the scroll id if the clip frame is asynchronously scrollable so that it becomes clearer before. It's still messy though. Differential Revision: https://phabricator.services.mozilla.com/D90647

Bug 1665258 - Shrink `CSSEquivTable`. r=heycam Better field packing reduces it from 48 bytes to 40 bytes on 64-bit. This reduces the memory used by each process by 336 bytes on 64-bit. Differential Revision: https://phabricator.services.mozilla.com/D90344