Bug 1551385 - part3 : add test 'test_webvtt_overlapping_time.html' and 'vtt_overlapping_time.html'. r=heycam Adding tests to ensure that when cues with overlapping times, the one with earlier end timestamp should disappear when the media time reaches its end time. In this test, we have two cues with overlapping time, when the video starts, both cues should be displayed. When the time passes 1 seconds, the first cue should disappear and the second cues should be still displayed. Differential Revision: https://phabricator.services.mozilla.com/D31172

Bug 1551385 - part2 : recompute display state when display cues amount is different from the one we rendered last time. r=heycam If the amount of cues which are going to be displayed is different from the one we displayed last time, we have to compute cues' display state again because cue's position might be affected by other cues. Differential Revision: https://phabricator.services.mozilla.com/D31170

Bug 1551385 - part1 : let 'processCue()' handle cleaning cues div. r=heycam We can actually let `processCue()` to handle rendering cues or cleaning displayed cues, no need to use another way to clear the cue. The advantages is to make the code cleaner and easier to read, now we just need to know JS side would handle all rendering stuffs for us. We don't need to have different behavior when there is no showing cue. The way we clear displayed cues are intuitive, we would remove all child nodes under the overlay, which are used to display cues. Differential Revision: https://phabricator.services.mozilla.com/D31171

Bug 1544242 - Add a test for a bug that this inadvertently fixes. r=heycam Always generating the slot assignment combinator means that we check the current host correctly. Differential Revision: https://phabricator.services.mozilla.com/D32405

Bug 1544242 - Cleanup selector-matching for nested pseudo-elements, match ::slotted correctly when there's no selector before it, and add tests. r=heycam,mats D29542 fixed the bogus checks that was making nested pseudo-elements match author rules. This adds tests and ends up being just a cleanup, though as it turns out we it also fixes an issue with ::slotted() matched from Element.matches. Differential Revision: https://phabricator.services.mozilla.com/D27529

Bug 1551385 - part3 : add test 'test_webvtt_overlapping_time.html' and 'vtt_overlapping_time.html'. r=heycam Adding tests to ensure that when cues with overlapping times, the one with earlier end timestamp should disappear when the media time reaches its end time. In this test, we have two cues with overlapping time, when the video starts, both cues should be displayed. When the time passes 1 seconds, the first cue should disappear and the second cues should be still displayed. Differential Revision: https://phabricator.services.mozilla.com/D31172

Bug 1551385 - part2 : recompute display state when display cues amount is different from the one we rendered last time. r=heycam If the amount of cues which are going to be displayed is different from the one we displayed last time, we have to compute cues' display state again because cue's position might be affected by other cues. Differential Revision: https://phabricator.services.mozilla.com/D31170

Bug 1551385 - part1 : let 'processCue()' handle cleaning cues div. r=heycam We can actually let `processCue()` to handle rendering cues or cleaning displayed cues, no need to use another way to clear the cue. The advantages is to make the code cleaner and easier to read, now we just need to know JS side would handle all rendering stuffs for us. We don't need to have different behavior when there is no showing cue. The way we clear displayed cues are intuitive, we would remove all child nodes under the overlay, which are used to display cues. Differential Revision: https://phabricator.services.mozilla.com/D31171

Bug 1548342 - Fix nested-pseudo-elements matching author rules better. r=heycam This makes *|*::marker do the intended thing in UA sheets, so I think it's better, and it's a bit less special-casey. We may want to re-introduce the changes for pseudo-elements at some point, depending on the WG decision, but this patch makes each_non_document_style_data consistent with the rule collector. The changes of each_non_document_style_data on their own should fix some bugs, but it doesn't because: * This is only hit for pseudos that allow user-action-state pseudo-classes. * The containing shadow check worked for them anyway. * We don't allow any pseudo after ::slotted() or that isn't tree-abiding per the CSS specs (we should maybe enable one of the moz-range stuff to be tree-abiding). So ::placeholder is the only one that right now fits the bill to trigger the bugs this would fix, but it doesn't since I couldn't make ::placeholder match :hover / :active / :focus anyhow (inside or outside a shadow tree). I've left the ProbeMarkerPseudoStyle changes for now since they are technically a bit more consistent than what was there before, but we could revert those if they cause trouble, we could rely on the UA rule matching, but we would need to change that if we allow the ::foo::before and such to match. Differential Revision: https://phabricator.services.mozilla.com/D29542

Bug 1546255 - When an @import stylesheet is cloned, we should also make sure to restyle. r=heycam We need to notify the whole parent chain (using NOTIFY) rather than just mStyleSets, since mStyleSets is empty for @import-ed sheets. Differential Revision: https://phabricator.services.mozilla.com/D31787

Bug 1550554 - Use OwnedSlice for will-change. r=heycam We could use ArcSlice if wanted I guess, your call. Though will change is not supposed to be used very frequently. Differential Revision: https://phabricator.services.mozilla.com/D30548

Bug 1550554 - Add bindings for box shadows, and remove nsCSSShadowArray and friends. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D30547

Bug 1550554 - Use ArcSlice for quotes. r=heycam This saves the intermediate allocation. Differential Revision: https://phabricator.services.mozilla.com/D30546

Bug 1550554 - Share computed and specified value representation of -moz-context-properties. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D30545

Bug 1550554 - Use an ArcSlice as the computed value representation of inherited list properties. r=heycam This adds a bit of complexity, which I think will pay off in the end. Removals incoming. Differential Revision: https://phabricator.services.mozilla.com/D30544

Bug 1550554 - Implement ArcSlice::default(). r=heycam Share a singleton to avoid allocating for empty lists. Differential Revision: https://phabricator.services.mozilla.com/D30543

Bug 1549593 - Use OwnedSlice in the specified and computed values of most vector properties. r=heycam This is just a refactor in the right direction. Eventual goal is: * All inherited properties use ArcSlice<>. * All reset properties use OwnedSlice<> (or ThinVec<>). No conversion happens at all, so we can remove all that glue, and also compute_iter and co. Of course there's work to do, but this is a step towards that. Differential Revision: https://phabricator.services.mozilla.com/D30127

Bug 1550554 - Use OwnedSlice for will-change. r=heycam We could use ArcSlice if wanted I guess, your call. Though will change is not supposed to be used very frequently. Differential Revision: https://phabricator.services.mozilla.com/D30548

Bug 1550554 - Add bindings for box shadows, and remove nsCSSShadowArray and friends. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D30547

Bug 1550554 - Use ArcSlice for quotes. r=heycam This saves the intermediate allocation. Differential Revision: https://phabricator.services.mozilla.com/D30546

Bug 1550554 - Share computed and specified value representation of -moz-context-properties. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D30545

Bug 1550554 - Use an ArcSlice as the computed value representation of inherited list properties. r=heycam This adds a bit of complexity, which I think will pay off in the end. Removals incoming. Differential Revision: https://phabricator.services.mozilla.com/D30544

Bug 1550554 - Implement ArcSlice::default(). r=heycam Share a singleton to avoid allocating for empty lists. Differential Revision: https://phabricator.services.mozilla.com/D30543

Bug 1549593 - Use OwnedSlice in the specified and computed values of most vector properties. r=heycam This is just a refactor in the right direction. Eventual goal is: * All inherited properties use ArcSlice<>. * All reset properties use OwnedSlice<> (or ThinVec<>). No conversion happens at all, so we can remove all that glue, and also compute_iter and co. Of course there's work to do, but this is a step towards that. Differential Revision: https://phabricator.services.mozilla.com/D30127

Bug 1551652. Stop using [array] in nsILayoutHistoryState. r=heycam I read through <https://searchfox.org/mozilla-central/search?q=%5B%5EA-Za-z_%5D%5BGg%5DetKeys%5B%5EA-Za-z_%5D&case=true&regexp=true&path=> and as far as I can tell the only JS consumer is SessionHistory.jsm, which was not passing the optional "count" arg already. There are no C++ consumers. Differential Revision: https://phabricator.services.mozilla.com/D31113

Bug 1551065 - Properly discard empty lines for -webkit-line-clamp. r=heycam That being said, this code is clearly not getting hit, are we 100% sure that this is interoperable with what WebKit / Blink do? Differential Revision: https://phabricator.services.mozilla.com/D30830

Bug 1551045 - add debug log in `vtt.jsm`. r=heycam We should add some debug logs, which should be easily enable and disable, in vtt.jsm in order to increase the efficiency of debugging. Differential Revision: https://phabricator.services.mozilla.com/D30798

Bug 1536762 - part5 : adjustment should only happen when box doesn't go over the boundary along the moving direction. r=heycam When adjusting box's position, we would choose a axis first and move the box along this axis to see whether we can place this box without overlapping with other boxes and fully inside the rendering area. If the box has been over the boundary along the moving direction, we should move box back to the original position and change the moving direction to see whether we can find another best place for the box. Although the adjustment can run without timeout now, it still doesn't match the result of the reference, so change the state to `fail`. Differential Revision: https://phabricator.services.mozilla.com/D29765

Bug 1536762 - part4 : disable tests on Android and Windows. r=heycam On Android, sometime we can't get the first frame, which would cause intermittent fail. See bug1546128. On Windows, sometime we would get incorrect text color rendering, which would also cause intermittent fail. See bug1546133. It seems that we won't get these two bug fixed recently, so I'm going to disable them on these two platforms. Differential Revision: https://phabricator.services.mozilla.com/D28359

Bug 1536762 - part3 : enable wpt tests and add 'fuzzy' comparision. r=heycam Sometime there are some color rendering differences between the test and the reference, which is small enough to be ignored because human can't be aware of that difference. Differential Revision: https://phabricator.services.mozilla.com/D27179

Bug 1536762 - part2 : use first line box's size as positioning basic unit. r=heycam,baku According to the spec [1] 7.2.10.2, we should use the first line box's height or width as positioning unit to adjust box's position. We will also use this value to adjust box when `snap-to-line` is false. There, we implement a new chrome-only API to acquire this information, which would return the first line box's size in the `block` frame. [1] https://www.w3.org/TR/webvtt1/#ref-for-webvtt-cue-snap-to-lines-flag-12 Differential Revision: https://phabricator.services.mozilla.com/D29745

Bug 1536762 - part1 : use unscaled bounding box'size as a cue box's size. r=baku,heycam When adjusting cue's position, the goal is to place cue boxes fully inside the rendering area, and place in the right place. In order to check whether the box is correctly inside the rendering area, we have to know the correct size of the cue box. Therefore, we implement the new chrome-only APIs to acquire this information (box's height and width), and this returned value won't be affected by CSS transformation, which is good for positioning. Differential Revision: https://phabricator.services.mozilla.com/D27178

Bug 1549559 - Add bindings for ArcSlice and ThinArc, and use them to reduce copies of SVG path data. r=heycam As I said over bug 1549593, the eventual goal is to use ArcSlice in all inherited properties. But this seemed like a good first candidate that doesn't require me to move around a lot more code, since we were already using cbindgen for the path commands. Differential Revision: https://phabricator.services.mozilla.com/D30134

Bug 1549559 - Introduce ArcSlice, a small wrapper over ThinArc but without an explicit header. r=heycam We could make the header PhantomData or something, but then we wouldn't be able to bind to C++, since C++ doesn't have ZSTs. So add a canary instead to add a runtime check of stuff being sane. Differential Revision: https://phabricator.services.mozilla.com/D30133

Bug 1549593 - Move OwnedSlice to style_traits. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D30126

Bug 1548691 - Use the owned slice type for basic shape polygon coordinates. r=TYLin,heycam This enables destructors for tagged unions in cbindgen, implemented in: * https://github.com/eqrion/cbindgen/pull/333 Which allow us to properly generate a destructor for the cbindgen-generated StyleBasicShape (which now contains an OwnedSlice). For now, we still use the glue code to go from Box<BasicShape> to UniquePtr<BasicShape>. But that will change in the future when we generate even more stuff and remove all the glue. I could add support for copy-constructor generation to cbindgen for tagged enums, but I'm not sure if it'll end up being needed, and copy-constructing unions in C++ is always very tricky. Differential Revision: https://phabricator.services.mozilla.com/D29769

Bug 1548691 - Add an owned slice type which cbindgen can understand. r=heycam Passing these by value won't be ok of course, but that's fine. I plan to combine this with https://github.com/eqrion/cbindgen/pull/333 to actually be able to share representation for ~all the things, this is just the first bit. Box<T>, Atom and Arc<T> will be much easier since cbindgen can understand them without issues. It's boxed slices the only ones I should need something like this. I could avoid it if I rely on Rust's internal representation, which we can per [1], but then I need to teach cbindgen all about slices, which is generally hard, I think. [1]: https://github.com/rust-lang/unsafe-code-guidelines/blob/master/reference/src/layout/pointers.md Differential Revision: https://phabricator.services.mozilla.com/D29768

Bug 1545177 - Add crashtest. r=heycam Depends on D27980 Differential Revision: https://phabricator.services.mozilla.com/D27982

Bug 1545177 - Add WPT test for modifications to FontFace descriptors after initial creation. r=heycam Depends on D27980 Differential Revision: https://phabricator.services.mozilla.com/D29257

Bug 1545177 - If descriptors of a FontFace are modified after creation, update the associated font entry so that face selection will respect the new values, and mark font sets as dirty. r=heycam Differential Revision: https://phabricator.services.mozilla.com/D27980

Bug 1545430 - Implement selector-matching for ::part(). r=heycam Also fairly straight-forward. This may get more complicated when we do part forwarding, if any. I've opened https://github.com/w3c/csswg-drafts/issues/3841 in what I think would be a cleaner model for forwarding. Differential Revision: https://phabricator.services.mozilla.com/D28063

Bug 1545425 - Add parsing support for ::part(). r=heycam Disabled for now of course. This should be pretty uncontroversial I'd think. Differential Revision: https://phabricator.services.mozilla.com/D28060

Bug 1544242 - Fix selector-matching for nested pseudo-elements. r=heycam,mats We always include the combinator for pseudo-elements now (not including it was just an optimization) in order to not match when nested pseudo-elements are involved. We could add a more generic check in `matches_simple_selector` like: ``` if element.is_pseudo_element() { match *selector { Component::PseudoElement(..) | Component::NonTSPseudoClass(..) => {}, _ => return false, } } ``` But even that wouldn't be enough to make selectors like `:hover::marker` not match on the `::before::marker` pseudo-element, plus that code is really hot. So for now do the check on the `next_element_for_combinator` function. It's a bit hacky but it's the best I could came up with... While at it, simplify some checks to use is_pseudo_element() instead of implemented_pseudo_element() directly. Only the Rust patch as-is would make markers for ::before and ::after on list items not show up, so we also need to switch ::marker to use ProbeMarkerStyle() rather than ProbePseudoElementStyle(), since the marker should exist even if it matches no rules. Differential Revision: https://phabricator.services.mozilla.com/D27529

Bug 1544242 - Check iterator length in SelectorIter::is_featureless_host_selector. r=heycam I'm going to unconditionally generate the PseudoElement combinator, and this causes issues since we'll put the raw `::pseudo` selectors in the host bucket, which is obviously wrong. Differential Revision: https://phabricator.services.mozilla.com/D27528

Bug 1544170 Part 1 - Notify record/replay system about parsed style sheets, r=heycam. Differential Revision: https://phabricator.services.mozilla.com/D27408

Bug 1544546 - Un-fill the styleset and drop the cascade data when removing a pres shell. r=heycam This should keep memory usage a bit more under control when loosing a pres shell. Differential Revision: https://phabricator.services.mozilla.com/D27571

Bug 1544548 - Insert quirks.css later so that we hit the cascade data cache more often. r=heycam Bug 1535788 moved the quirks.css insertion earlier, so the order will not match when the commonish case where the document gets a shell and then knows the compat mode, which means that we hit the cascade data cache a bit less often. I'll also fix the order of the preference sheet vs. the quirks sheet in a followup, right now whatever gets first ends ups earlier on the list. Differential Revision: https://phabricator.services.mozilla.com/D27578

Bug 1542935 - Document::CompatibilityModeChanged should call EnsureStyleFlush. r=heycam Following the reasoning from the previous commit. Selectors may become case sensitive or case insensitive as a result of this operation, something I forgot about when writing the patch for bug 1535788. Now Document::CompatibilityModeChanged is actually only called if the compatibility mode actually changes, and thus we always need to restyle. The Stylist already takes care of fully invalidating the document when it changes (via mark_origins_dirty and co.), so we technically just need to guarantee that a style flush will happen. In practice I doubt this call will have any effect in practice given how early in the document's lifetime the compatibility mode can change, but still I think it's worth landing. Depends on D27416 Differential Revision: https://phabricator.services.mozilla.com/D27417

Bug 1542935 - Re-introduce a call to EnsureStyleFlush from PresShell::Init to fix an Android scrolling regression. r=heycam Sorry for the vague commit message, but I haven't dug yet on why or how the Android code is depending on this. This call used to be part of nsPresContext::CompatibilityModeChanged, which unconditionally called PresShell::EnsureStyleFlush. This was not (in theory, at least) always necessary. There's there's no point in ensuring a style flush is going to happen if styles haven't changed, and CompatibilityModeChanged() didn't actually guarantee that the compat mode was different at all before my patch. Styles only change if the compat mode actually changes (since then selectors become case-sensitive or case-insensitive), or more obviously when you insert or remove the quirks.css stylesheet, and in that case ApplicableStylesChanged makes sure that the flush happens. Yet here we are, and not having that early call to EnsureStyleFlush, even in the case there's no quirks mode or quirks sheet change or anything of that sort (this happens even on XUL docs, which are always FullStandards) makes the first (and only the first) browsing session in Geckoview have terrible scrolling performance. I'm calling it a day for today, will investigate as time permits in bug 1544185. Differential Revision: https://phabricator.services.mozilla.com/D27416

Bug 1543304 - Add MOZ_DBG support for nsCSSPropertyID and nsCSSPropertyIDSet; r=heycam Differential Revision: https://phabricator.services.mozilla.com/D26856