RESOLVED FIXED 39097
webkit-patch barfed on upload with a new image test result
https://bugs.webkit.org/show_bug.cgi?id=39097
Summary webkit-patch barfed on upload with a new image test result
Simon Fraser (smfr)
Reported 2010-05-13 18:07:25 PDT
Created attachment 56040 [details] webkit-patch output I ran 'webkit-patch upload NNN' with a new image result added to my git index. In the diff file (attached), I see: Exception raised during decoding git binary patch: /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/open3.rb:67:in `exec': No such file or directory - /opt/local/bin/git apply --directory=/var/folders/4I/4IFxjUcVHIip-GD1W3OsEE+++TM/-Tmp- (Errno::ENOENT) from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/open3.rb:67:in `popen3' from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/open3.rb:53:in `fork' from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/open3.rb:53:in `popen3' from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/open3.rb:51:in `fork' from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/open3.rb:51:in `popen3' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:452:in `extract_contents_from_git_binary_chunk' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:373:in `initialize' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:372:in `collect' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:372:in `initialize' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:424:in `new' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:424:in `parse' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:424:in `collect' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:424:in `parse' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:15:in `prettify' from /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/prettify.rb:26 /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:458:in `extract_contents_from_git_binary_chunk' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:373:in `initialize' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:372:in `collect' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:372:in `initialize' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:424:in `new' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:424:in `parse' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:424:in `collect' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:424:in `parse' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/PrettyPatch.rb:15:in `prettify' /Volumes/InternalData/Development/webkit/OpenSource/BugsSite/PrettyPatch/prettify.rb:26
Attachments
webkit-patch output (30.89 KB, text/html)
2010-05-13 18:07 PDT, Simon Fraser (smfr)
no flags
Patch (1.05 KB, patch)
2010-08-17 11:26 PDT, Tony Chang
no flags
Patch (985 bytes, patch)
2010-08-19 10:08 PDT, Tony Chang
no flags
Simon Fraser (smfr)
Comment 1 2010-05-13 18:07:46 PDT
On SnowLeopard, WebKit r59396
Simon Fraser (smfr)
Comment 2 2010-07-13 21:51:24 PDT
No such file or directory - /opt/local/bin/git apply My git is in /usr/local/bin/git.
Simon Fraser (smfr)
Comment 3 2010-07-13 21:55:57 PDT
BugsSite/PrettyPatch/PrettyPatch.rb has GIT_PATH = "/opt/local/bin/git" which is bogus.
Eric Seidel (no email)
Comment 4 2010-07-13 22:18:14 PDT
PrettyPatch is thanks to senior Roben, although others (including myself) have hacked on it. Seems like an easy fix. We should probably use 'which git' if available.
Eric Seidel (no email)
Comment 5 2010-07-13 22:18:36 PDT
Thank you for the report, btw.
Tony Chang
Comment 6 2010-08-17 11:26:58 PDT
Tony Chang
Comment 7 2010-08-17 11:27:47 PDT
Comment on attachment 64610 [details] Patch It looks like other scripts also expect git to be in your path (e.g., prepare-ChangeLog, resolve-ChangeLog, VCSUtils.pm).
WebKit Review Bot
Comment 8 2010-08-17 11:30:07 PDT
Attachment 64610 [details] did not pass style-queue: Failed to run "['WebKitTools/Scripts/check-webkit-style']" exit_code: 1 BugsSite/ChangeLog:1: One or more unexpected \r (^M) found; better to use only a \n [whitespace/carriage_return] [1] Suppressing further [whitespace/carriage_return] reports for this file. Total errors found: 9 in 2 files If any of these errors are false positives, please file a bug against check-webkit-style.
Simon Fraser (smfr)
Comment 9 2010-08-17 12:58:27 PDT
prepare-ChangeLog works fine for me on the same machine.
Eric Seidel (no email)
Comment 10 2010-08-17 17:20:11 PDT
Comment on attachment 64610 [details] Patch OK. Makes sense.
WebKit Commit Bot
Comment 11 2010-08-17 18:05:38 PDT
Comment on attachment 64610 [details] Patch Rejecting patch 64610 from commit-queue. Unexpected failure when processing patch! Please file a bug against webkit-patch. Failed to run "['./WebKitTools/Scripts/webkit-patch', '--status-host=queues.webkit.org', 'land-attachment', '--force-clean', '--build', '--non-interactive', '--ignore-builders', '--build-style=both', '--quiet', 64610, '--test', '--parent-command=commit-queue', '--no-update']" exit_code: 1 Last 500 characters of output: webkitpy/tool/commands/stepsequence.py", line 64, in _run step(tool, options).run(state) File "/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/webkitpy/tool/steps/validatereviewer.py", line 66, in run if self._has_valid_reviewer(changelog_entry): File "/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/webkitpy/tool/steps/validatereviewer.py", line 48, in _has_valid_reviewer if changelog_entry.reviewer(): AttributeError: 'NoneType' object has no attribute 'reviewer'
Eric Seidel (no email)
Comment 12 2010-08-17 19:36:59 PDT
Comment on attachment 64610 [details] Patch I'm confused by that commit-queue error. Lets try again. :)
WebKit Commit Bot
Comment 13 2010-08-17 20:09:20 PDT
Comment on attachment 64610 [details] Patch Rejecting patch 64610 from commit-queue. Unexpected failure when processing patch! Please file a bug against webkit-patch. Failed to run "['./WebKitTools/Scripts/webkit-patch', '--status-host=queues.webkit.org', 'land-attachment', '--force-clean', '--build', '--non-interactive', '--ignore-builders', '--build-style=both', '--quiet', 64610, '--test', '--parent-command=commit-queue', '--no-update']" exit_code: 1 Last 500 characters of output: webkitpy/tool/commands/stepsequence.py", line 64, in _run step(tool, options).run(state) File "/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/webkitpy/tool/steps/validatereviewer.py", line 66, in run if self._has_valid_reviewer(changelog_entry): File "/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/webkitpy/tool/steps/validatereviewer.py", line 48, in _has_valid_reviewer if changelog_entry.reviewer(): AttributeError: 'NoneType' object has no attribute 'reviewer'
Eric Seidel (no email)
Comment 14 2010-08-17 21:47:02 PDT
My guess is its failing to parse the changelog. Just not sure why.
Adam Barth
Comment 15 2010-08-17 21:52:47 PDT
Likely because of the \r complained about in https://bugs.webkit.org/show_bug.cgi?id=39097#c8
Eric Seidel (no email)
Comment 16 2010-08-18 09:16:45 PDT
Comment on attachment 64610 [details] Patch Seems like we should update webkit-patch to detect \r in a ChangeLog and barf nicely. This can't be cq+'d as is. And ideally the \r would be removed from the ChangeLog anyway.
Tony Chang
Comment 17 2010-08-19 09:50:45 PDT
*** Bug 38101 has been marked as a duplicate of this bug. ***
Tony Chang
Comment 18 2010-08-19 10:08:57 PDT
Tony Chang
Comment 19 2010-08-19 10:09:38 PDT
Comment on attachment 64867 [details] Patch I removed the \r's in ChangeLog (the whole file had \r's) and set the svn:eol-style to native.
Adam Barth
Comment 20 2010-08-19 11:00:05 PDT
Comment on attachment 64867 [details] Patch Okiedokes. Thanks for fixing the ChangeLog.
WebKit Commit Bot
Comment 21 2010-08-19 11:29:05 PDT
Comment on attachment 64867 [details] Patch Clearing flags on attachment: 64867 Committed r65686: <http://trac.webkit.org/changeset/65686>
WebKit Commit Bot
Comment 22 2010-08-19 11:29:11 PDT
All reviewed patches have been landed. Closing bug.
Tony Chang
Comment 23 2010-08-19 13:03:04 PDT
It turns out that whatever computer runs PrettyPatch.rb when you click "Formatted Diff" on the bots depended on the full git path. For example, the following patch fails even though it includes the git binary information in the diff. https://bugs.webkit.org/attachment.cgi?id=64890&action=prettypatch Where does this run? Can we fix the path to git on the bots?
Eric Seidel (no email)
Comment 24 2010-08-19 13:08:32 PDT
wms is our man! if he can't do it, no one can!
William Siegrist
Comment 25 2010-08-19 13:29:51 PDT
Dave, looks like we need to adjust the PATH for the open2() in attachment.cgi? http://trac.webkit.org/browser/trunk/BugsSite/attachment.cgi#L395
Eric Seidel (no email)
Comment 26 2010-08-19 13:34:45 PDT
jparent is also on rather intimate terms with bugzilla...
Tony Chang
Comment 27 2010-08-19 13:56:51 PDT
(In reply to comment #25) > Dave, looks like we need to adjust the PATH for the open2() in attachment.cgi? > > http://trac.webkit.org/browser/trunk/BugsSite/attachment.cgi#L395 Alternately, maybe we can just symlink /usr/bin/git to /opt/local/bin/git on bugs.webkit.org? I could also change PrettyPatch.rb to use /usr/bin/env git.
William Siegrist
Comment 28 2010-08-26 08:57:39 PDT
Bugs' PrettyPatch has been fixed in <http://trac.webkit.org/changeset/66104>.
David Kilzer (:ddkilzer)
Comment 29 2010-08-26 09:53:53 PDT
(In reply to comment #28) > Bugs' PrettyPatch has been fixed in <http://trac.webkit.org/changeset/66104>. Inadvertent changes backed out in r66105. <http://trac.webkit.org/changeset/66105>
Note You need to log in before you can comment on or make changes to this bug.