Closed
Bug 1415496
Opened 7 years ago
Closed 6 years ago
CSS performance regression due to temporary instrumentation code to debug a crash
Categories
(Core :: CSS Parsing and Computation, defect, P3)
Tracking
()
RESOLVED
WORKSFORME
People
(Reporter: grekovdmitry16, Unassigned)
References
Details
Attachments
(1 file)
|
41.18 KB,
text/html
|
Details |
User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:58.0) Gecko/20100101 Firefox/58.0
Build ID: 20171103003834
Steps to reproduce:
I have written table row and column hover highlighting in CSS only using pseido-elements (See HTML file in attachments)
Actual results:
New Quantum CSS engine (FF version 58.0b1 (64-bit)) is actually slower than old engine (FF version 56.0.2 (64-bit)). Try to move mouse cursor over table fast and see that columns and rows are highlighted too late. Old engine is performing a lot faster, difference is easily seen by unarmed eye.
Expected results:
I expect that new engine should be at least as fast as old one (or better - faster).
|
Reporter | |
Comment 1•7 years ago
|
||
I looked also at performance tab in Dev tools, and it seems, that graphics become slower, not styling.
|
||
Updated•7 years ago
|
Component: Untriaged → CSS Parsing and Computation
Product: Firefox → Core
|
||
Comment 2•7 years ago
|
||
I can confirm that this is a regression.
The regression points to : https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=1498abcee79d232125fb44846849160ec3c1fa51&tochange=0efa71a9a60ac4373222b58d5750b453e7bd7318
with specific : Bug 1406727. r=mstange,froydnj a=abillings
Flags: needinfo?(mstange)
|
|
||
Comment 3•7 years ago
|
||
and here is a profile: https://perfht.ml/2ymeSP2
|
||
Comment 4•7 years ago
|
||
The slowdown is caused by temporary instrumentation code that we added in order to track down the cause of a crash. Once the cause has been found, the instrumentation code will be removed and things should be fast again.
Flags: needinfo?(mstange)
|
|
Reporter | |
Comment 5•7 years ago
|
||
Ok, thanks for support!
Summary: Quantum CSS engine is actually slower than old engine in some cases → CSS performance regression due to temporary instrumentation code to debug a crash
|
|
Reporter | |
Comment 6•6 years ago
|
||
Seems fixed now.
|
|
||
Comment 7•6 years ago
|
||
Is there anything actionable, or can this bug be closed?
Here is a profile with WR on latest nightly: http://bit.ly/2JRXLz6
Flags: needinfo?(mstange)
|
|
Reporter | |
Comment 8•6 years ago
|
||
You can close. Thank you.
|
|
||
Comment 9•6 years ago
|
||
I agree, looks fixed to me. I don't know which bug removed the crash instrumentation.
Status: UNCONFIRMED → RESOLVED
Closed: 6 years ago
Flags: needinfo?(mstange)
Resolution: --- → WORKSFORME
You need to log in
before you can comment on or make changes to this bug.
Description
•