HasRTLChars() appears in profile of attachment 8848015 after landing the patches
for bug 1391538 because the landing made text in <input> or <textarea> always
treated as non-single byte characters. Therefore, HasRTLChars() is now called
by nsTextFragment::SetTo() a lot in editors.
HasRTLChar() checks if it's in an RTL character for each character until it
becomes true. However, if character is less than the minimum RTL character,
U+0590, it works faster at least for Western languages.
MozReview-Commit-ID: 4YecxQWUcmK
--HG--
extra : rebase_source : 172be670795f0e2155d4bd5044ee37f652367734
nsTextFragment::SetTo() may be hot path of performance test but it needs to be
careful for computing allocation size. However, nsTextFragment uses uint32_t
with only 29 bits to store its text length and using CheckedInt sometimes causes
appearing CheckedInt in profile. So, using CheckedInt in it doesn't make sense
(and also wrong).
This patch defines NS_MAX_TEXT_FRAGMENT_LENGTH and make SetTo() check the length
of new text by itself.
MozReview-Commit-ID: 5zHtU1Kk6X2
--HG--
extra : rebase_source : e9aecc6b721bb5211ccc1194e027343ef090e2c2
This adds some tests to ensure the hidden messages toolbar works
as expected. There are some for testing the component itself, as
well as some to test the store.
Some fixtures were modified as well to better represent the state
of the application.
MozReview-Commit-ID: 3Swqff1mbck
--HG--
extra : rebase_source : 48cff5c3c5e44e5aea9437d1243604ed69b19222
This adds a "filteredMessagesCount" property to the messages store,
which is updated at the same time as we update the "visibleMessages"
property.
The global number of hidden messages is then shown next to the text search
input, and a button can be clicked to Reset the filters to their original
values.
This only takes into account messages that are filtered-out because
the default filters changes. Which means that non-default filters, like
CSS, Network and XHR don't impact filteredMessagesCount.
We take this as an opportunity to rename the match*Filter functions to
pass*Filter, since it better represents what the function does (a network
message is not impacted by a level filter, but it can be misleading to say
that it matches level filters).
MozReview-Commit-ID: BLPmCFNtzEl
--HG--
extra : rebase_source : 8ea4abbc8f8847a306d3bac4fd87d07e0ccb8f89
We already support the actual functionality, but nobody added support
for the IDL property to the .webidl file.
Also added <style nonce> to the web-platform-tests reflection tests,
since nobody updated that to the current spec either.
This does not add support for .nonce to SVGScriptElement, because I
couldn't find any standard that specified it. I updated the wpt tests
to expect .nonce to work on HTMLScriptElement but not SVGScriptElement.
MozReview-Commit-ID: F1K7WMfMoDi
--HG--
extra : rebase_source : 247c63b63446dc0d60062bb9d9c61228c379b989
The GCC plugin used for hazard builds was built for GCC 4.9. Until a new
plugin is built by a toolchain job for the "default" version of GCC we
build Firefox with, the builds using the plugin need to stay with the
version the plugin was built for.
For the Obama wikipedia page, this covers about 85% of the unmeasured
ComputedValues structs. The about:memory output looks like this:
> +---2,443,648 B (02.41%) -- computed-values
> | +--1,088,272 B (01.07%) -- dom
> | +----945,744 B (00.93%) -- non-dom
> | +----409,632 B (00.40%) -- visited
I'm not sure why some CVs are still being missed.
MozReview-Commit-ID: 1bYWwSi4ihn
--HG--
extra : rebase_source : 14e4bd36a54bbbd8fd265f559704bec5a5e3b154
This corrects a bug introduced in 265873cf1388.
MozReview-Commit-ID: LkZlTVAM17E
--HG--
extra : rebase_source : bf442e620abd6b47adee3d4c56e6f0c19964aea1
More progress on unifying how Gecko and Servo track stylist dirtiness.
Source-Repo: https://github.com/servo/servo
Source-Revision: e60266a72302a0ed332f1dd18d335b6092c47da4
--HG--
extra : subtree_source : https%3A//hg.mozilla.org/projects/converted-servo-linear
extra : subtree_revision : fd8bbea82543a1a6a759806329e7ef2abe78bcc8
Although, we still have an issue that is leading of text in <input type="text">
element causes overflow and scrollable (bug 1378065), we have no other
compatibility issue reports and the issue is not so important nor easy to fix.
So, now, let's make the new Japanese default font settings ride the train.
MozReview-Commit-ID: J3SN9FtHoZg
--HG--
extra : rebase_source : 1ec0c11afc49b08bf38cc7b1a7e928a09f494e8b
The window rect's state is already tested a few lines down.
MozReview-Commit-ID: BLDufKrl8ey
--HG--
extra : rebase_source : 535872620741ffdd424f1794b355e1f0975ba76c
When we set the window rect to width/height 400, we only want to test
those qualities.
MozReview-Commit-ID: 76qqCOOwjXQ
--HG--
extra : rebase_source : a8ebbb4a848beba6de43c2176d9322806ae80ac1
Because the Maximize Window, Minimize Window, and Fullscreen Window
commands are now idempotent we will want to use the Set Window Rect to
bring them back to the normal window state.
MozReview-Commit-ID: BxtAJizfti6
--HG--
extra : rebase_source : c3110bb53ca2dc795e0e6756525435ba2ea61a8e
The x and y variables are not defined but needs to be extracted from
new_position.
MozReview-Commit-ID: Ds8cPlufjUa
--HG--
extra : rebase_source : 6edb0445ac89d0202a4ef037b0cd38af226c6507
The other window state manipulation commands are named minimizeWindow
and maximizeWindow.
MozReview-Commit-ID: IBVqJSRwG8x
--HG--
extra : rebase_source : a1efaf7cf2a5c068b62c60453b1f8814d34b043b
Tim Nguyen
Masayuki Nakano
Aryeh Gregor
Luca Greco
Nicolas Chevobbe
Anthony Ramine
Sebastian Hengst
Mike Hommey
Simon Sapin
Bevis Tseng
Nicholas Nethercote
sotaro
Wes Kocher
Gregory Szorc
ffxbld
Gabriel Luong
Simon Lindholm
Jan Beich
Emilio Cobos Álvarez
Masatoshi Kimura
Andreas Tolfsen