WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
Bug 78012
run-perf-tests doesn't recognize paths that start with PerformanceTests
https://bugs.webkit.org/show_bug.cgi?id=78012
Summary
run-perf-tests doesn't recognize paths that start with PerformanceTests
Ryosuke Niwa
Reported
2012-02-07 11:45:04 PST
run-perf-tests PerformanceTests/Parser runs exactly 0 tests because run-perf-tests recognizes that as PerformanceTests/PerformanceTests/Parser. It's annoying as hell.
Attachments
Fixes the bug
(4.88 KB, patch)
2012-02-07 15:18 PST
,
Ryosuke Niwa
no flags
Details
Formatted Diff
Diff
Updated per Adam's comment
(3.34 KB, patch)
2012-02-08 01:10 PST
,
Ryosuke Niwa
abarth
: review+
Details
Formatted Diff
Diff
Show Obsolete
(1)
View All
Add attachment
proposed patch, testcase, etc.
Adam Barth
Comment 1
2012-02-07 12:35:19 PST
run-webkit-tests has some code to remove the LayoutTests from command line arguments. Presumably you want something similar here.
Ryosuke Niwa
Comment 2
2012-02-07 13:49:12 PST
(In reply to
comment #1
)
> run-webkit-tests has some code to remove the LayoutTests from command line arguments. Presumably you want something similar here.
Something like that. I just looked at its implementation in manager.py and it seems way too complicated for run-perf-test. I'm just going to use filesystem.relpath instead.
Ryosuke Niwa
Comment 3
2012-02-07 15:18:29 PST
Created
attachment 125942
[details]
Fixes the bug
Ryosuke Niwa
Comment 4
2012-02-07 15:19:33 PST
I feel like we should be able to use the same approach in run-webkit-tests but there's some special treatment for / and \ there so not sure if that's possible.
Adam Barth
Comment 5
2012-02-07 16:14:59 PST
Comment on
attachment 125942
[details]
Fixes the bug View in context:
https://bugs.webkit.org/attachment.cgi?id=125942&action=review
> Tools/Scripts/webkitpy/performance_tests/perftestsrunner.py:111 > + if self._host.filesystem.exists(arg):
It's strange that this checks the real file system for existence. I would expect this to be a syntactic transformation that doesn't query the local file system.
Ryosuke Niwa
Comment 6
2012-02-07 16:28:31 PST
Comment on
attachment 125942
[details]
Fixes the bug View in context:
https://bugs.webkit.org/attachment.cgi?id=125942&action=review
>> Tools/Scripts/webkitpy/performance_tests/perftestsrunner.py:111 >> + if self._host.filesystem.exists(arg): > > It's strange that this checks the real file system for existence. I would expect this to be a syntactic transformation that doesn't query the local file system.
Otherwise how would I figure out whether we should use relpath or not? An alternative is to add both since find_files.find only finds existing files.
Adam Barth
Comment 7
2012-02-07 16:33:01 PST
> Otherwise how would I figure out whether we should use relpath or not? An alternative is to add both since find_files.find only finds existing files.
That could work. You can also look whether the string PerformanceTests appears in the path.
Ryosuke Niwa
Comment 8
2012-02-07 16:35:31 PST
(In reply to
comment #7
)
> > Otherwise how would I figure out whether we should use relpath or not? An alternative is to add both since find_files.find only finds existing files. > > That could work. You can also look whether the string PerformanceTests appears in the path.
I want to avoid hard-coding "PerformanceTests" since I'm hoping that I can also replace the equivalent function in manager.py by this one as a followup.
Hajime Morrita
Comment 9
2012-02-07 21:47:12 PST
(In reply to
comment #8
)
> I want to avoid hard-coding "PerformanceTests" since I'm hoping that I can also replace the equivalent function in manager.py by this one as a followup.
You can just pass the "base directory name" as an argument to the (coming shared) function. Avoiding file system access is good for performance reason. In this case, startup speed. Anyway, I like this change. I'm tired to remove "PerformanceTest" part from my completed-by-the-shell path ;-)
Ryosuke Niwa
Comment 10
2012-02-08 01:10:41 PST
Created
attachment 126024
[details]
Updated per Adam's comment
Adam Barth
Comment 11
2012-02-08 01:15:56 PST
Comment on
attachment 126024
[details]
Updated per Adam's comment That certainly looks nicer. :)
Ryosuke Niwa
Comment 12
2012-02-08 01:51:25 PST
Committed
r107053
: <
http://trac.webkit.org/changeset/107053
>
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug