Igor Oliveira authored 7 years ago and Pier Angelo Vendrame committed 1 year ago

Many fonts have issues with their vertical metrics. they
are used to influence the height of ascenders and depth
of descenders. Gecko uses it to calculate the line height
(font height + ascender + descender), however because of
that idiosyncratic behavior across multiple operating
systems, it can be used to identify the user's OS.

The solution proposed in the patch uses a default factor
to be multiplied with the font size, simulating the concept
of ascender and descender. This way all operating
systems will have the same line height.

System fonts are an enormous fingerprinting vector.
Even with font allow lists and with our custom configuration on Linux,
which counter metrics measurements, getComputedStyle leaks several
details.
This patch counters both these kinds of attacks.

Bug 1855345 - Make `AutoCopyListener` ignore any selection changes caused by Selection API r=karlt,smaug, a=dsmith

Even though there is a reasonable usabe like bug 1567160, we should
ensure that `AutoCopyListener` never update clipboard when the selection
change is caused by a Selection API call.

Differential Revision: https://phabricator.services.mozilla.com/D191585

This patch doesn't change behavior; it just removes some redundant checks.

(Note that the tight-scoping around nsAutoMicroTask/ReduceAnimations is
important for the RAII object to be torn down at the intended time; but it
doesn't need its own if-check.)


Original Revision: https://phabricator.services.mozilla.com/D191911

Differential Revision: https://phabricator.services.mozilla.com/D191987

This patch doesn't change behavior.

This is a bit gross, but it helps us make better use of the backtrace to figure
out what's going on in a given crash report.

Original Revision: https://phabricator.services.mozilla.com/D191909

Differential Revision: https://phabricator.services.mozilla.com/D191986

Bug 1861259 part 1: Refactor out a helper function for loop body in nsRefreshDriver::Tick.  a=diannaS

This doesn't affect behavior; it's just a pure refactoring, moving code from
being inline to being in a helper function.

Original Revision: https://phabricator.services.mozilla.com/D191908

Differential Revision: https://phabricator.services.mozilla.com/D191985

These telemetry probes have long since expired, so remove them. This
avoids taking unnecessary timestamps in these functions, which we have
seen can be expensive on certain devices.

Original Revision: https://phabricator.services.mozilla.com/D185007

Differential Revision: https://phabricator.services.mozilla.com/D192116

Bug 1856497 - Allow user input events to bypass the delay handling check if this is a non-visible document r=smaug, a=dmeehan

Differential Revision: https://phabricator.services.mozilla.com/D190133

Bug 1856514 - Make sure nsRefreshDriver will tick enough times for user input handling r=smaug, a=dmeehan

Differential Revision: https://phabricator.services.mozilla.com/D190108

Differential Revision: https://phabricator.services.mozilla.com/D186053

In some edge cases, we may start handling user input events that
are not supposed to be handled by the current document.

This patch introduces some delay to handle user input events.

Differential Revision: https://phabricator.services.mozilla.com/D180976

As expected because the NS_ASSERTION was removed in the patch.

Make the errors more visible in viewsource/XMLPrettyPrint by adding a
semitransparent red background and switching from a dotted underline to
a wavy underline. Both styles were chosen for consistency with the
devtools debugger's error styles. Same background color and same
underline appearance.

Differential Revision: https://phabricator.services.mozilla.com/D188284

Bug 1809492. Clear pointer to nsTreeBodyFrame on XULTreeElement after any possible calls that can set it. r=emilio, a=RyanVM

Differential Revision: https://phabricator.services.mozilla.com/D184318

Bug 1837070 - Ensure we flush OffscreenCanvas when the display list is recreated. r=lsalzman, a=RyanVM

When we recreate the display list, e.g. after tabbing away, we need to
ensure we flush the OffscreenCanvas because our resource handles will be
invalidated. This requires the OffscreenCanvasDisplayHelper to know
which OffscreenCanvas it belongs to, and a worker reference if necessary
to do the dispatching for the flush. Otherwise we would require the
canvas to be updated manually by the web application to repaint.

Differential Revision: https://phabricator.services.mozilla.com/D184324

The caller can already run script, it's just that other script-running
functions are already marked as _BOUNDARY (incorrectly). Will fix
properly in a follow-up bug.

MANUAL PUSH: Bustage fix CLOSED TREE

Bug 1843700 - Hide popups synchronously if needed, even if already hiding async. r=NeilDeakin a=pascalc

Differential Revision: https://phabricator.services.mozilla.com/D186126

Differential Revision: https://phabricator.services.mozilla.com/D184410

Differential Revision: https://phabricator.services.mozilla.com/D184848

Just like we rollup menus.

This is needed because tooltips otherwise get hidden on mousedown and
don't flush the view manager.

This matches the similar hunk for GTK in bug 1569439.

Differential Revision: https://phabricator.services.mozilla.com/D183489

ThinArc is way more complex than it needs to be, and miri complains
about various things when using the selectors crate with it.

This doesn't fully fix miri, but it gets much closer. ThinArc becomes
just an alias for Arc<HeaderSlice<>>. This allows to simplify the
bindings to ArcSlice too, since now the existing Arc<> code can just be
used.

Differential Revision: https://phabricator.services.mozilla.com/D183011

Bug 1843847. Clear nsDisplayListBuilder::mCurrentContainerASR at the end of a paint. r=mstange, a=dsmith

Differential Revision: https://phabricator.services.mozilla.com/D183741

Bug 1843086: Line participants use block start ascent, instead of not-yet stored baseline. r=emilio, a=RyanVM

Differential Revision: https://phabricator.services.mozilla.com/D183432

Bug 1842325. Ensure that WebRenderLayerManager::mTarget is always cleared after a transaction. r=gfx-reviewers,lsalzman, a=dsmith

The scope exit in WebRenderLayerManager::EndTransactionWithoutLayer isn't necessary because there are no early returns in that function, but I added it just in case an early return is added in the future.

The assert in nsDisplayList::PaintRoot currently holds because of how the current code is, but these changes will mean EndEmptyTransaction will handle that case as far as clearing mTarget is concerned.

Differential Revision: https://phabricator.services.mozilla.com/D183069

Differential Revision: https://phabricator.services.mozilla.com/D182618

Bug 1840804 - Make `TextEditor::OnBlur` stop finalizing `Selection` if new element already gets focus r=m_kato, a=dsmith

Before bug 1770874, `EditorBase::OnBlur` checked that for both `TextEditor`
and `HTMLEditor`.  However, accidentally, I removed the check from `TextEditor`.
Therefore, a call of `EditorBase::FinalizeSelection()` will hide the caret
even after another editor gets focus.

Therefore, this patch just take it back into `TextEditor::OnBlur`.

Note that I don't think the design mode handling is required there because
`TextEditor`s shouldn't be created in the design mode document.

Differential Revision: https://phabricator.services.mozilla.com/D182468

Bug 1840700 - Make `Selection::AddRangeAndSelectFramesAndNotifyListenersInternal()` update interline position even if it does nothing r=jjaschke, a=dsmith

It always set to "start of next line" when new range was added.  However,
after bug 1815802, it does nothing when adding range is one of its ranges.
Therefore, if interline position is set to "end of line" by our internal
code, like editor, `getSelection().addRange(getSelection().getRangeAt(0))`
keeps the last interline position.

Differential Revision: https://phabricator.services.mozilla.com/D182300

Was introduced in bug 1831148, regresses the behaviour fixed by
bug 1419225.

Differential Revision: https://phabricator.services.mozilla.com/D182346

This actually fixes the bug. The issue is that we end up with
nsITreeView.setTree(null) then nsITreeView.setTree(<tree>).

This code deals with the first call:

  https://searchfox.org/mozilla-central/rev/c936f47f3a629ae49a4d528d3366bf29f2d4e4a7/browser/components/places/content/treeView.js#1651

But the later call doesn't restore the state properly and a bunch of nodes end
up unparented. That might be worth fixing on its own...

This is probably a long/forever-standing bug that was uncovered by different
amount of reflow calls.

Depends on D181379

Differential Revision: https://phabricator.services.mozilla.com/D181380

Backed out changeset 47f0f49b4036 (bug 1834786) from beta, the regressor (bug 1834081) is being backed out

Bug 1826206 - Require nsISerialEventTarget for RetargetDeliveryTo, r=necko-reviewers,valentin, a=dmeehan

This avoids potential issues where multiple OnDataAvailable callbacks or
similar could theoretically be called concurrently on different
StreamTransportService threads when targeting the STS - these cases will
now target a TaskQueue on the STS instead, structurally ensuring serial
execution.

Differential Revision: https://phabricator.services.mozilla.com/D179984

Bug 1836670 - Use NonNull in cssparser::CowRcStr. r=layout-reviewers,glandium,dlrobertson, a=dmeehan

This prevents the compiler from misoptimizing this code due to undefined
behavior.

Upstream PR: https://github.com/servo/rust-cssparser/pull/345

Differential Revision: https://phabricator.services.mozilla.com/D179921

The test changes from the previous patch uncover an issue on shippable
builds, see referenced bug.

MANUAL PUSH: Orange fix CLOSED TREE

See https://github.com/w3c/csswg-drafts/issues/7961#issuecomment-1549874958

In general I agree with Tab that media blocks should parse the same
regardless of whether they are nested.

We should add a test that covers this to WPT, but for now tweak the
local reftest.

MANUAL PUSH: Trivial orange fix CLOSED TREE

Bug 1835066 - Tweak test_parser_diagnostics_unprintables.html to keep testing unexpected token errors. r=dholbert

Without nesting, the parser error is the unexpected token error, emitted
from cssparser.

With nesting enabled these behave the same as other strings outside of a
block (so an "unexpected end of input" error instead).

Tweak the test a bit to keep testing what this test wants to test (that
unexpected token errors are escaped).

Differential Revision: https://phabricator.services.mozilla.com/D179830

Backed out changeset 7bc8c25b2935 (bug 1833244)
Backed out changeset 4576af83a4ec (bug 1833244)
Backed out changeset 90a5bbba7b9c (bug 1833244)

Bug 1833244 p3. Implement sheet orientation switching/rotation for `page-orientation`. r=AlaskanEmily,dholbert

Where supported (print preview and print-to-PDF), this implements changing the
orientation and/or rotation of print sheets, as appropriate, in response to CSS
`page-orientation`. When supported we:

- in the single page-per-sheet case, rotate the sheet in order to implement
  any `page-orientation` rotation on the sheet. Rotating the sheet is necessary
  so that the pages in the PDF files that we output are correct.

- in the multiple pages-per-sheet case, we already rotate individual pages in
  their grid cell. This change keeps such pages rotated, as appropriate, but
  augments that behavior by switching the orientation of the sheet (based on
  the first page on the sheet) if necessary to best place the page to make
  maximum use of the space.

Depends on D179423

Differential Revision: https://phabricator.services.mozilla.com/D179448

OS print drivers/devices know nothing about page dimensions unless we tell
them. Previously, the physical page dimensions (including orientation) have
always been the same, so communicating their dimensions once at the start of
a print has been enough. In preparation for supporting different "physical"
page dimensions (in the immediate future only different page orientations) when
we save to PDF, we need to have the infrastructure to pass dimensions through
on a page-by-page basis. This patch adds that.

None of the PrintTarget subclasses do anything with this extra information yet,
but in a follow-up patch PrintTargetPDF will use this information to create
PDFs with mixed page orientations.

Differential Revision: https://phabricator.services.mozilla.com/D179423