Bug 46682 – Patch locks should expire if a patch is marked for retry

• Component: New Bugs
• Status: RESOLVED FIXED

[Eric Seidel (no email)]

Patch locks should expire if a patch is marked for retry

[Eric Seidel (no email)]

Created attachment 69000 [details]
Patch

[Adam Barth]

Comment on attachment 69000 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=69000&action=review

> WebKitTools/QueueStatusServer/handlers/updatestatus.py:68
> +        active_items.expire_item(queue_status.active_patch_id)
> +        active_items.put()

This is a read-modify-write.  Does it need to be in a transaction to ensure atomicity?

[Eric Seidel (no email)]

(In reply to comment #2)
> (From update of attachment 69000 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=69000&action=review
> 
> > WebKitTools/QueueStatusServer/handlers/updatestatus.py:68
> > +        active_items.expire_item(queue_status.active_patch_id)
> > +        active_items.put()
> 
> This is a read-modify-write.  Does it need to be in a transaction to ensure atomicity?

It would, yes.  I guess I shoudl do that. :)  Originally I thougtht a transaction wouldn't be needed here, but since we're replacing the whole list, that could cause badness in the multiple-writer case.

[Eric Seidel (no email)]

Created attachment 69005 [details]
Patch

[Eric Seidel (no email)]

Now with transactiony goodness.

[Adam Barth]

Comment on attachment 69005 [details]
Patch

yay transactions

[WebKit Commit Bot]

Comment on attachment 69005 [details]
Patch

Clearing flags on attachment: 69005

Committed r68474: <http://trac.webkit.org/changeset/68474>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.