WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
57662
new-run-webkit-tests: fix feature detection, skipped platform lists on mac
https://bugs.webkit.org/show_bug.cgi?id=57662
Summary
new-run-webkit-tests: fix feature detection, skipped platform lists on mac
Dirk Pranke
Reported
2011-04-01 14:23:38 PDT
new-run-webkit-tests: fix feature detection, skipped platform lists on mac
Attachments
Patch
(5.19 KB, patch)
2011-04-01 15:07 PDT
,
Dirk Pranke
no flags
Details
Formatted Diff
Diff
update w/ feedback from abarth
(5.31 KB, patch)
2011-04-01 16:41 PDT
,
Dirk Pranke
abarth
: review+
abarth
: commit-queue-
Details
Formatted Diff
Diff
Show Obsolete
(1)
View All
Add attachment
proposed patch, testcase, etc.
Dirk Pranke
Comment 1
2011-04-01 15:07:19 PDT
Created
attachment 87922
[details]
Patch
Adam Barth
Comment 2
2011-04-01 16:32:39 PDT
Comment on
attachment 87922
[details]
Patch View in context:
https://bugs.webkit.org/attachment.cgi?id=87922&action=review
> Tools/Scripts/webkitpy/layout_tests/port/mac.py:134 > + def _path_to_webcore_library(self): > + return self._build_path('WebCore.framework/Versions/A/WebCore')
This method does not appear to be used in this patch.
> Tools/Scripts/webkitpy/layout_tests/port/webkit.py:186 > + entries = self._filesystem.glob(self._webkit_baseline_path('*'))
Doesn't this return a super huge list? This doesn't seem like the right algorithm. You want to enumerate the directories and then ask whether they're in the baseline search path, not enumerate all the things in those directories.
> Tools/Scripts/webkitpy/layout_tests/port/webkit.py:187 > + platform_prefix = self.name().split('-', 1)[0]
This should be a helpful function instead of munging the string here.
Dirk Pranke
Comment 3
2011-04-01 16:36:46 PDT
(In reply to
comment #2
)
> (From update of
attachment 87922
[details]
) > View in context:
https://bugs.webkit.org/attachment.cgi?id=87922&action=review
> > > Tools/Scripts/webkitpy/layout_tests/port/mac.py:134 > > + def _path_to_webcore_library(self): > > + return self._build_path('WebCore.framework/Versions/A/WebCore') > > This method does not appear to be used in this patch. > > > Tools/Scripts/webkitpy/layout_tests/port/webkit.py:186 > > + entries = self._filesystem.glob(self._webkit_baseline_path('*')) > > Doesn't this return a super huge list? This doesn't seem like the right algorithm. You want to enumerate the directories and then ask whether they're in the baseline search path, not enumerate all the things in those directories. >
The glob() returns LayoutTests/platform/*, nor platform/*/*. It's a small list and the isdir call filters for the directories.
> > Tools/Scripts/webkitpy/layout_tests/port/webkit.py:187 > > + platform_prefix = self.name().split('-', 1)[0] > > This should be a helpful function instead of munging the string here.
Will add.
Dirk Pranke
Comment 4
2011-04-01 16:38:44 PDT
(In reply to
comment #3
)
> (In reply to
comment #2
) > > (From update of
attachment 87922
[details]
[details]) > > View in context:
https://bugs.webkit.org/attachment.cgi?id=87922&action=review
> > > > > Tools/Scripts/webkitpy/layout_tests/port/mac.py:134 > > > + def _path_to_webcore_library(self): > > > + return self._build_path('WebCore.framework/Versions/A/WebCore') > > > > This method does not appear to be used in this patch. > > > > > Tools/Scripts/webkitpy/layout_tests/port/webkit.py:186 > > > + entries = self._filesystem.glob(self._webkit_baseline_path('*')) > > > > Doesn't this return a super huge list? This doesn't seem like the right algorithm. You want to enumerate the directories and then ask whether they're in the baseline search path, not enumerate all the things in those directories. > > > > The glob() returns LayoutTests/platform/*, nor platform/*/*. It's a small list and the isdir call filters for the directories. > > > > Tools/Scripts/webkitpy/layout_tests/port/webkit.py:187 > > > + platform_prefix = self.name().split('-', 1)[0] > > > > This should be a helpful function instead of munging the string here. > > Will add.
Er, actually, this was left over from an earlier, incorrect patch and isn't even needed. Deleted.
Dirk Pranke
Comment 5
2011-04-01 16:41:06 PDT
Created
attachment 87941
[details]
update w/ feedback from abarth
Adam Barth
Comment 6
2011-04-01 17:42:22 PDT
Comment on
attachment 87941
[details]
update w/ feedback from abarth View in context:
https://bugs.webkit.org/attachment.cgi?id=87941&action=review
> Tools/Scripts/webkitpy/layout_tests/port/webkit.py:186 > + basename = self._filesystem.basename(entry)
Shouldn't this be inside the if since the variable is only used inside the if?
Adam Barth
Comment 7
2011-04-01 17:45:59 PDT
Comment on
attachment 87941
[details]
update w/ feedback from abarth View in context:
https://bugs.webkit.org/attachment.cgi?id=87941&action=review
> Tools/Scripts/webkitpy/layout_tests/port/mac.py:134 > + def _path_to_webcore_library(self): > + return self._build_path('WebCore.framework/Versions/A/WebCore')
I would have made this change in a separate patch, but i think it's fine to add it to this patch.
Dirk Pranke
Comment 8
2011-04-01 17:51:03 PDT
(In reply to
comment #6
)
> (From update of
attachment 87941
[details]
) > View in context:
https://bugs.webkit.org/attachment.cgi?id=87941&action=review
> > > Tools/Scripts/webkitpy/layout_tests/port/webkit.py:186 > > + basename = self._filesystem.basename(entry) > > Shouldn't this be inside the if since the variable is only used inside the if?
Sure, it can be. I will fix this when I land it. Thanks for the review.
Dirk Pranke
Comment 9
2011-04-01 17:55:32 PDT
Committed
r82753
: <
http://trac.webkit.org/changeset/82753
>
WebKit Review Bot
Comment 10
2011-04-01 20:16:38 PDT
http://trac.webkit.org/changeset/82759
might have broken SnowLeopard Intel Release (WebKit2 Tests) The following tests are not passing: inspector/styles/styles-add-blank-property.html plugins/windowless_plugin_paint_test.html
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