Closed
Bug 1356655
Opened 7 years ago
Closed 7 years ago
adjustTabstrip is slow, due to flushing layout
Categories
(Firefox :: Tabbed Browser, enhancement, P1)
Firefox
Tabbed Browser
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox55 | --- | fixed |
People
(Reporter: florian, Assigned: dao)
References
(Blocks 2 open bugs)
Details
(Whiteboard: [photon-performance])
Attachments
(1 file)
On a fast last generation macbook, I just had adjustTabstrip cause 37ms of jank (within 700ms of jank while detaching a mozreview tab to a new window). These 37ms are spent flushing layout. I'm almost sure this layout flush isn't needed, because on one window I had several hundred tabs, so removing one won't change anything to what we do with close tab icons, and the new window had only one tab so we could guess easily without flushing that there was plenty of room. Could we somehow use overflow/underflow event to replace this code? Or would get use getBoundsWithoutFlushing to get a rought estimate, and if it's close to the threshold do an actual flush?
|
Assignee | |
Updated•7 years ago
|
Flags: needinfo?(dao+bmo)
|
||
Updated•7 years ago
|
Flags: qe-verify?
Priority: -- → P2
|
|
Assignee | |
Comment 2•7 years ago
|
||
(In reply to Florian Quèze [:florian] [:flo] from comment #0) > Could we somehow use overflow/underflow event to replace this code? The tab clip width is intentionally larger than the tab minimum width, so we can't replace this code, but we can return early if we're overflowing.
Assignee: nobody → dao+bmo
Status: NEW → ASSIGNED
Flags: needinfo?(dao+bmo)
|
|
Assignee | |
Updated•7 years ago
|
Flags: qe-verify? → qe-verify-
|
|
||
Updated•7 years ago
|
Priority: P2 → P1
| Comment hidden (mozreview-request) |
|
Reporter | |
Comment 4•7 years ago
|
||
| mozreview-review | ||
Comment on attachment 8858644 [details] Bug 1356655 - Let adjustTabstrip return early if the tab strip is overflowing. https://reviewboard.mozilla.org/r/130612/#review133212 Thanks for looking into this! :-) ::: browser/base/content/tabbrowser.xml:5934 (Diff revision 1) > > <method name="adjustTabstrip"> > <body><![CDATA[ > + // If we're overflowing, tab widths don't change anymore, so we can > + // return early to avoid flushing layout. > + if (this.getAttribute("overflow") == "true") { I started looking at the code to try to find the answer to this question, but you probably know it or can find it faster than me: will adjustTabstrip be called again after we remove the 'overflow' attribute? I assume if we are removing it because we closed just one tab it won't matter, but if we removed plenty of tabs (eg. we triggered removeTabsToTheEndFrom from the third tab), it may change whether we need to display the close buttons.
|
|
Assignee | |
Comment 5•7 years ago
|
||
(In reply to Florian Quèze [:florian] [:flo] from comment #4) > I started looking at the code to try to find the answer to this question, > but you probably know it or can find it faster than me: will adjustTabstrip > be called again after we remove the 'overflow' attribute? > I assume if we are removing it because we closed just one tab it won't > matter, but if we removed plenty of tabs (eg. we triggered > removeTabsToTheEndFrom from the third tab), it may change whether we need to > display the close buttons. We call adjustTabstrip in _endRemoveTab.
|
|
Reporter | |
Comment 6•7 years ago
|
||
| mozreview-review | ||
Comment on attachment 8858644 [details] Bug 1356655 - Let adjustTabstrip return early if the tab strip is overflowing. https://reviewboard.mozilla.org/r/130612/#review133888
Attachment #8858644 -
Flags: review?(florian) → review+
Pushed by dgottwald@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/7640ebe4a0e2 Let adjustTabstrip return early if the tab strip is overflowing. r=florian
|
||
Comment 8•7 years ago
|
||
| bugherder | ||
https://hg.mozilla.org/mozilla-central/rev/7640ebe4a0e2
Status: ASSIGNED → RESOLVED
Closed: 7 years ago
status-firefox55:
--- → fixed
Resolution: --- → FIXED
Target Milestone: --- → Firefox 55
|
|
||
Updated•7 years ago
|
Iteration: --- → 55.4 - May 1
|
|
Assignee | |
Updated•7 years ago
|
|
|
Reporter | |
Updated•7 years ago
|
No longer blocks: photon-performance-triage
|
|
Reporter | |
Updated•7 years ago
|
Blocks: photon-performance-triage
You need to log in
before you can comment on or make changes to this bug.
Description
•