Track the performance of adding sources after page load in the debugger
Categories
(DevTools :: Debugger, enhancement)
Tracking
(firefox116 fixed)
Tracking | Status | |
---|---|---|
firefox116 | --- | fixed |
People
(Reporter: ochameau, Assigned: ochameau)
References
(Blocks 1 open bug)
Details
Attachments
(5 files, 2 obsolete files)
While investigating bug 1651849, I realized that many slowness were about sources being added after the toolbox opening. Or sources simply not throttled in the massive first bucket.
These late sources, throttled in a couple of subsequent throlling bucket would actually cause reducer and components updates, leading to slow reflows.
Unfortunately, this doesn't seem to be correctly tracked by DAMP as we do not add sources after the page load, so that they are all notified almost all at once.
It would be nice to add a test adding many sources, causing many throttle buckets. This would verify that incremental reducer/selectors/components updates is not extremely costly.
Assignee | ||
Comment 1•2 years ago
|
||
This helps verify that adding new sources incrementaly wouldn't cause
massive slowness in reducer/components/reflows.
Updated•2 years ago
|
Assignee | ||
Comment 2•2 years ago
|
||
Also remove useless dbg
argument to evalInContent.
Comment 3•2 years ago
|
||
Comment on attachment 9321829 [details]
Bug 1778275 - [devtools] Create debugger context only once in DAMP tests.
Revision D171983 was moved to bug 1835775. Setting attachment 9321829 [details] to obsolete.
Assignee | ||
Comment 4•2 years ago
|
||
We weren't printing the exception message.
Error mesage and its related stacktrace were also off.
Allow exception to be a string.
Assignee | ||
Comment 5•2 years ago
|
||
Now that reducers are cleared per target/thread, listening to NAVIGATE
for the full clearance of reducer data is wrong.
We should rather wait for the REMOVE_THREAD.
Also add an explicit waiting code for the selected source.
Comment 6•2 years ago
|
||
Comment on attachment 9336583 [details]
Bug 1778275 - [devtools] Improve exception message in DAMP.
Revision D179459 was moved to bug 1836146. Setting attachment 9336583 [details] to obsolete.
Assignee | ||
Comment 7•2 years ago
|
||
Assignee | ||
Comment 8•2 years ago
|
||
(Instead of forcing to select original or generated based on shouldSelectOriginalLocation flag).
This also add an await for the selectLocation action in order to better wait for all async work.
This required to tweak the minified test as we don't want to wait for full selection.
Assignee | ||
Comment 9•2 years ago
|
||
Assignee | ||
Updated•1 years ago
|
Comment 10•1 years ago
|
||
Comment 11•1 years ago
|
||
Comment 12•1 years ago
|
||
Comment 13•1 years ago
|
||
bugherder |
Comment 14•1 years ago
|
||
Comment 15•1 years ago
|
||
bugherder |
Assignee | ||
Updated•1 years ago
|
Assignee | ||
Updated•1 years ago
|
Comment 16•1 years ago
|
||
== Change summary for alert #38599 (as of Wed, 07 Jun 2023 22:56:55 GMT) ==
Regressions:
Ratio | Test | Platform | Options | Absolute values (old vs new) |
---|---|---|---|---|
160% | damp custom.jsdebugger.project-search.DAMP | linux1804-64-shippable-qr | e10s fission stylo webrender | 1,039.22 -> 2,705.13 |
156% | damp custom.jsdebugger.project-search.DAMP | windows10-64-shippable-qr | e10s fission stylo webrender-sw | 914.66 -> 2,338.51 |
154% | damp custom.jsdebugger.project-search.DAMP | windows10-64-shippable-qr | e10s fission stylo webrender | 919.93 -> 2,340.97 |
153% | damp custom.jsdebugger.project-search.DAMP | linux1804-64-shippable-qr | e10s fission stylo webrender-sw | 976.68 -> 2,473.37 |
130% | damp custom.jsdebugger.project-search.DAMP | macosx1015-64-shippable-qr | e10s fission stylo webrender-sw | 954.98 -> 2,197.40 |
129% | damp custom.jsdebugger.project-search.DAMP | macosx1015-64-shippable-qr | e10s fission stylo webrender | 951.12 -> 2,177.61 |
41% | damp custom.jsdebugger.open.DAMP | windows10-64-shippable-qr | e10s fission stylo webrender-sw | 632.11 -> 890.01 |
38% | damp custom.jsdebugger.open.DAMP | linux1804-64-shippable-qr | e10s fission stylo webrender | 813.86 -> 1,121.56 |
38% | damp custom.jsdebugger.open.DAMP | macosx1015-64-shippable-qr | e10s fission stylo webrender | 581.99 -> 800.48 |
37% | damp custom.jsdebugger.open.DAMP | linux1804-64-shippable-qr | e10s fission stylo webrender-sw | 799.99 -> 1,094.55 |
... | ... | ... | ... | ... |
7% | damp custom.jsdebugger.close.DAMP | windows10-64-shippable-qr | e10s fission stylo webrender | 29.68 -> 31.68 |
6% | damp custom.jsdebugger.stepOver.DAMP | linux1804-64-shippable-qr | e10s fission stylo webrender | 441.97 -> 466.39 |
3% | damp custom.jsdebugger.pretty-print.DAMP | windows10-64-shippable-qr | e10s fission stylo webrender | 282.95 -> 291.64 |
3% | damp custom.jsdebugger.pause.DAMP | windows10-64-shippable-qr | e10s fission stylo webrender-sw | 242.34 -> 249.58 |
2% | damp custom.jsdebugger.pretty-print.DAMP | windows10-64-shippable-qr | e10s fission stylo webrender-sw | 284.16 -> 290.06 |
For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=38599
Assignee | ||
Comment 17•1 years ago
|
||
Oh wait I actually did not land the very last patch...
Comment 18•1 years ago
|
||
Updated•1 years ago
|
Comment 19•1 years ago
|
||
bugherder |
Description
•