Rules and their declarations are a single object as far as the CC is concerned. They have a single nsCycleCollectionISupports and they are represented by a single node in the CC graph. That single object has two nsWrapperCache instances in it that point to different JS objects, and we need to make sure that the ordering of the unlink operations for those nsWrapperCache instances is handled correctly.
Differential Revision: https://phabricator.services.mozilla.com/D38326
--HG--
extra : moz-landing-system : lando
When we unlink, we may null out the mFontFaceSet pointer, but still keep
mInFontFaceset as true. Which can cause null dereference bugs as described in
comment 0.
This is mostly a speculative fix since I think that's a plausible cause for a
bug like that, but it may not be the only cause for the crash.
Differential Revision: https://phabricator.services.mozilla.com/D38822
--HG--
extra : moz-landing-system : lando
Many of these features are non-functional these days, but I'll keep them
hooked up in case we decide to fix them.
Differential Revision: https://phabricator.services.mozilla.com/D38413
--HG--
extra : moz-landing-system : lando
They're wrong. When a property that affects the parent frame changes, we get a
hint for both frames. This fixes this bug.
Depends on D38598
Differential Revision: https://phabricator.services.mozilla.com/D38599
--HG--
extra : moz-landing-system : lando
This is IMO the right RestyleManager change for what bug 1527210 tried to fix.
We need to apply the animation hints to the primary frame, not the style frame.
The other non-RestyleManager bits of that bug still apply and look fine to me.
Differential Revision: https://phabricator.services.mozilla.com/D38598
--HG--
extra : moz-landing-system : lando
Seems we can leave this node alive for too long if the user scrolls between
domcontentloaded (where GoToAnchor is called) and onload (where ScrollToAnchor()
is called).
Though it seems we can leave it for too long if we don't end up calling
ScrollToAnchor(), the documentation of the method claims that it's cleared
unconditionally.
Differential Revision: https://phabricator.services.mozilla.com/D38398
--HG--
extra : moz-landing-system : lando
This is a potential fix that I thought it was worth doing rather than
implementing Blink's platform-dependent silliness. This ensures that the display
frame always has enough space to display itself.
Note that it may still get clipped, if there's no room for both the display
frame and the button.
Differential Revision: https://phabricator.services.mozilla.com/D37922
--HG--
extra : moz-landing-system : lando
Unmasking is an optional style of showing password. Therefore, if callers of
`nsIEditor::Unmask()` specify middle of surrogate pair(s), it may mean that
they want to expand the unmask range from shorter range which does not include
the high and/or low surrogate. Therefore, one of the surrogates is in unmasked
range, we unmask the surrogate pair. However, we handle this in a lot of
places, i..e., we have duplicated code. This can get rid of these duplicates
with making `nsIEditor::Unmask()` expand the range automatically.
Differential Revision: https://phabricator.services.mozilla.com/D38432
--HG--
extra : moz-landing-system : lando
In some cases, other tools may show selected content in their UI. E.g.,
character palette of macOS. Therefore, we shouldn't allow them to show
masked password.
Differential Revision: https://phabricator.services.mozilla.com/D38430
--HG--
extra : moz-landing-system : lando
Double click, long tap, moving selection with keyboard and deleting text
scan word boundary. With these changes, the text node has raw password
value even if it's masked visually. For making safer,
`nsTextFrame::PeekOffsetWord()` should search word boundary in masked text.
Then, we can hide word boundary at masked range, but keep allowing to look
for word boundary only in unmasked range.
Differential Revision: https://phabricator.services.mozilla.com/D38013
--HG--
extra : moz-landing-system : lando
Anonymous text node in password field has `NS_MAYBE_MASKED` flag and
`TextEditor` managing the node has range of unmasking. This patch makes
`BuildTextRunsScanner::BuildTextRunForFrames()` treat text frame in
password field as `text-transform` applied. For recording the unmask range,
this patch creates `nsTransformedCharStyle::mMaskPassword` and treats unmask
range start/end edges as different style boundary. Therefore,
`nsCaseTransformTextRunFactory::TransformString()` can replace each character
with a password mask character with the information.
Note that we need to kill bidi algorithm in password fields with forms.css
because caret position will tell everybody whether the typing character is
an RTL or an LTR character.
Differential Revision: https://phabricator.services.mozilla.com/D38009
--HG--
extra : moz-landing-system : lando
With the previous patches, editor has stopped masking characters in password
field. Instead, layout code needs to handle it. However, layout code
especially around `nsTextFrame` is performance critical area. Therefore,
layout code requires a quick way to check whether a text node in password
field or not.
This patch creates new flag for `CharacterData` node and marks all text nodes
whose characters should be masked with the flag when `EditorBase` or
`nsTextControlFrame` creates them.
Differential Revision: https://phabricator.services.mozilla.com/D38006
--HG--
extra : moz-landing-system : lando
Now, `TextEditRules` should use the new editor API to mask/unmask characters
if it's for a password editor. With this change, it does not need to manage
masked characters and unmasked characters separately since the anonymous
text node always have unmasked characters and `nsTextFrame` will mask the
characters at painting time.
Differential Revision: https://phabricator.services.mozilla.com/D38005
--HG--
extra : moz-landing-system : lando
Currently it's completely unclear at use sites that the getters for `once`
static prefs return the pref value from startup, rather than the current pref
value. (Bugs have been caused by this.) This commit improves things by changing
the getter name to make it clear that the pref value obtained is from startup.
This required changing things within libpref so it distinguishes between the
"base id" (`foo_bar`) and the "full id" (`foo_bar` or
`foo_bar_DoNotUseDirectly` or `foo_bar_AtStartup` or
`foo_bar_AtStartup_DoNotUseDirectly`; the name used depends on the `mirror` and
`do_not_use_directly` values in the YAML definition.) The "full id" is used in
most places, while the "base id" is used for the `GetPrefName_*` and
`GetPrefDefault_*` functions.
(This is a nice demonstration of the benefits of the YAML file, BTW. Making
this change with the old code would have involved adding an entry to every
single pref in StaticPrefList.h.)
The patch also rejigs the comment at the top of StaticPrefList.yaml, to clarify
some things.
Differential Revision: https://phabricator.services.mozilla.com/D38604
--HG--
extra : moz-landing-system : lando
Delete `return false` at the end of the if-statement block that handling
the multicol subtree reframing, and let it fall though the bottom of
WipeContainingBlock() where there is a complete logic for ib-split
reframing.
Differential Revision: https://phabricator.services.mozilla.com/D38548
--HG--
extra : moz-landing-system : lando
CSSStyleSheet.prototype.deleteRule can throw an IndexSizeError. This change
makes it so that the exception thrown has the max valid index for that function.
Differential Revision: https://phabricator.services.mozilla.com/D38519
--HG--
extra : moz-landing-system : lando
Emilio Cobos Álvarez
Boris Zbarsky
harry
Ciure Andrei
Coroiu Cristina
Cameron McCormack
Bogdan Tara
Edwin Gao
Hiroyuki Ikezoe
Masayuki Nakano
Nicholas Nethercote
Andreea Pavel
Mihai Alexandru Michis
Jonathan Kew
longsonr
Charlie Marlow
Gurzau Raul
Ting-Yu Lin
Nicolas Silva
Blito
Andrea Marchesini
RuchikaGosain