Closed
Bug 515447
Opened 15 years ago
Closed 6 years ago
textrun / text reflow causes 10 seconds of beachball loading tinderbox logs
Categories
(Core :: Layout: Text and Fonts, defect)
Tracking
()
RESOLVED
WORKSFORME
Performance Impact | ? |
People
(Reporter: sayrer, Unassigned)
References
Details
Attachments
(3 files)
It's hard to open a big tinderbox log without spending a long time with a beachball on OS X, at least.
Reporter | ||
Comment 1•15 years ago
|
||
Reporter | ||
Comment 2•15 years ago
|
||
Comment 3•15 years ago
|
||
Do you have the HTML5 parser enabled?
Reporter | ||
Comment 4•15 years ago
|
||
http://blog.mozilla.com/rob-sayre/2009/09/09/spincontrol/
Reporter | ||
Comment 5•15 years ago
|
||
(In reply to comment #3) > Do you have the HTML5 parser enabled? Yes!
Comment 6•15 years ago
|
||
Does this go away if you disable the HTML5 parser? The HTML5 parser disabled the chunking of text nodes; we used to create separate text nodes for each 8K chunk of text; now it's all one huge text node. I think there are some existing bugs on the resulting performance regressions.
Reporter | ||
Comment 7•15 years ago
|
||
(In reply to comment #6) > Does this go away if you disable the HTML5 parser? > > The HTML5 parser disabled the chunking of text nodes; we used to create > separate text nodes for each 8K chunk of text; now it's all one huge text node. It does go away. At least we got a new tool out of it. :)
Related to bug 502568.
Comment 9•6 years ago
|
||
Another sample.
Comment 10•6 years ago
|
||
According to catlee, this still affects us, all these years later.
Whiteboard: [qf]
Updated•6 years ago
|
Attachment #9018049 -
Attachment mime type: application/octet-stream → text/plain
Comment 12•6 years ago
|
||
(In reply to Mike Conley (:mconley) (:⚙️) from comment #10) > According to catlee, this still affects us, all these years later. Can you/he post a testcase where this is bad? The first two attachments here load near instantly for me, on Linux nightly. (And for now Firefox prompts me to download the 3rd attachment rather than directly rendering it, so I'm not sure if that one is bad or good.)
Flags: needinfo?(catlee)
Updated•6 years ago
|
Flags: needinfo?(mconley)
Updated•6 years ago
|
Attachment #9018049 -
Attachment description: live_backing.log → live_backing.txt
Attachment #9018049 -
Attachment filename: live_backing.log → live_backing.txt
Comment 13•6 years ago
|
||
https://taskcluster-artifacts.net/XZmhivqwQUKrk15pEu7R9Q/0/public/logs/live_backing.log
Flags: needinfo?(mconley)
Updated•6 years ago
|
Attachment #9018049 -
Attachment description: live_backing.txt → live_backing.txt.gz
Attachment #9018049 -
Attachment mime type: text/plain → application/octet-stream
Updated•6 years ago
|
Attachment #9018049 -
Attachment filename: live_backing.txt → live_backing.txt.gz
Comment 14•6 years ago
|
||
That file is a 27 meg text file (orders of magnitude larger than the testcases we had here, which now load instantly). I spun off bug 1501415 for that issue, and I'm going to close this old bug as WORKSFORME, since its original testcases are fine now.
Status: NEW → RESOLVED
Closed: 6 years ago
Flags: needinfo?(catlee)
Resolution: --- → WORKSFORME
Updated•2 years ago
|
Performance Impact: --- → ?
Whiteboard: [qf]
You need to log in
before you can comment on or make changes to this bug.
Description
•