WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
107606
EWS creates 0 byte sized log files
https://bugs.webkit.org/show_bug.cgi?id=107606
Summary
EWS creates 0 byte sized log files
Lucas Forschler
Reported
2013-01-22 17:50:11 PST
Ever since
r138264
, EWS creates 0 length log files. The only logs which have content, are those with errors. It is strange to see a directory full of empty files. We should not be creating them unless something is to be put in them. We might not be able to detect this until the run is over... at which we could check for 0 length files, and prune them.
Attachments
Patch
(3.45 KB, patch)
2013-11-11 09:16 PST
,
László Langó
no flags
Details
Formatted Diff
Diff
Patch
(3.99 KB, patch)
2013-11-21 06:47 PST
,
László Langó
no flags
Details
Formatted Diff
Diff
Patch
(4.61 KB, patch)
2013-11-22 07:52 PST
,
László Langó
no flags
Details
Formatted Diff
Diff
Patch
(7.17 KB, patch)
2013-11-25 00:49 PST
,
László Langó
no flags
Details
Formatted Diff
Diff
Show Obsolete
(3)
View All
Add attachment
proposed patch, testcase, etc.
Eric Seidel (no email)
Comment 1
2013-01-22 17:57:37 PST
They should be smaller, but not zero sized. :( We can revert the change.
László Langó
Comment 2
2013-11-11 09:16:18 PST
Created
attachment 216578
[details]
Patch
Ryosuke Niwa
Comment 3
2013-11-21 03:59:25 PST
Comment on
attachment 216578
[details]
Patch View in context:
https://bugs.webkit.org/attachment.cgi?id=216578&action=review
> Tools/ChangeLog:13 > + > + * Scripts/webkitpy/tool/bot/queueengine.py: > + (QueueEngine.run): > + (QueueEngine._open_work_log): > + (QueueEngine._ensure_work_log_closed): > + * Scripts/webkitpy/tool/commands/earlywarningsystem.py: > + (AbstractEarlyWarningSystem.review_patch):
I'd like to see some description as to why this bug manifested and how you fixed it. There is a lot of code change here, and the correctness of code change is not self-evident.
László Langó
Comment 4
2013-11-21 05:53:59 PST
(In reply to
comment #3
)
> (From update of
attachment 216578
[details]
) > View in context:
https://bugs.webkit.org/attachment.cgi?id=216578&action=review
> > > Tools/ChangeLog:13 > > + > > + * Scripts/webkitpy/tool/bot/queueengine.py: > > + (QueueEngine.run): > > + (QueueEngine._open_work_log): > > + (QueueEngine._ensure_work_log_closed): > > + * Scripts/webkitpy/tool/commands/earlywarningsystem.py: > > + (AbstractEarlyWarningSystem.review_patch): > > I'd like to see some description as to why this bug manifested and how you fixed it. > There is a lot of code change here, and the correctness of code change is not self-evident.
There was a modification in
r138264
, that tried to make less log, because some of the messeges were duplicated. After this the EWS created the log file (with the same name as the bugID) but doesn't write anything into it, even if there were errors during the build. So i managed this. Now it will create the log file only if there is an error, and wont create if dont want to write anything to it. The main problem was that, the ScriptError was not raised and the QueueEngine never get that. Now it gets the error and handles it.
Ryosuke Niwa
Comment 5
2013-11-21 06:01:12 PST
Please add per-function/code change comment as to how you fixed that problem.
László Langó
Comment 6
2013-11-21 06:47:18 PST
Created
attachment 217559
[details]
Patch
Ryosuke Niwa
Comment 7
2013-11-21 19:23:04 PST
Comment on
attachment 217559
[details]
Patch View in context:
https://bugs.webkit.org/attachment.cgi?id=217559&action=review
> Tools/ChangeLog:12 > + Reviewed by NOBODY (OOPS!).
This line should appear before the long description but after the bug url.
> Tools/ChangeLog:15 > + (QueueEngine.run): Open the log file only if a ScriptError was raised.
Why?
> Tools/ChangeLog:16 > + (QueueEngine._open_work_log): Does not need to tee STDOUT to log file anymore.
Why?
> Tools/ChangeLog:17 > + (QueueEngine._ensure_work_log_closed): Close the logfile.
Why?
> Tools/ChangeLog:19 > + (AbstractEarlyWarningSystem.review_patch): Raise again the captured ScriptError.
Why?
Ryosuke Niwa
Comment 8
2013-11-21 19:24:11 PST
The current change log comments simply repeats the code change. I'm looking for is an explanation as to why each code change is made.
László Langó
Comment 9
2013-11-22 07:52:43 PST
Created
attachment 217684
[details]
Patch
Ryosuke Niwa
Comment 10
2013-11-22 07:56:59 PST
Comment on
attachment 217684
[details]
Patch Thanks.
WebKit Commit Bot
Comment 11
2013-11-22 08:24:04 PST
Comment on
attachment 217684
[details]
Patch Clearing flags on attachment: 217684 Committed
r159690
: <
http://trac.webkit.org/changeset/159690
>
WebKit Commit Bot
Comment 12
2013-11-22 08:24:07 PST
All reviewed patches have been landed. Closing bug.
WebKit Commit Bot
Comment 13
2013-11-22 09:39:37 PST
Re-opened since this is blocked by
bug 124782
Alexey Proskuryakov
Comment 14
2013-11-22 09:41:00 PST
This broke webkitpy tests, rolling out. It's not obvious whether these just need an update of results.
http://build.webkit.org/builders/Apple%20MountainLion%20Release%20WK1%20(Tests)/builds/15098/steps/webkitpy-test/logs/stdio
László Langó
Comment 15
2013-11-25 00:49:43 PST
Created
attachment 217777
[details]
Patch
László Langó
Comment 16
2013-11-25 00:58:39 PST
(In reply to
comment #14
)
> This broke webkitpy tests, rolling out. It's not obvious whether these just need an update of results. > >
http://build.webkit.org/builders/Apple%20MountainLion%20Release%20WK1%20(Tests)/builds/15098/steps/webkitpy-test/logs/stdio
Unfortunately i forgot to update the unit test. It really was not obvious, but this change is needed, because of the patch. I wrote some explanation into the Change log too.
WebKit Commit Bot
Comment 17
2013-11-25 07:45:37 PST
Comment on
attachment 217777
[details]
Patch Clearing flags on attachment: 217777 Committed
r159749
: <
http://trac.webkit.org/changeset/159749
>
WebKit Commit Bot
Comment 18
2013-11-25 07:45:41 PST
All reviewed patches have been landed. Closing bug.
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