Bug 36302 - EWS bots should process r+ bugs as well as r?
Summary: EWS bots should process r+ bugs as well as r?
Status: RESOLVED DUPLICATE of bug 35460
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: 528+ (Nightly build)
Hardware: PC OS X 10.5
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-03-18 10:37 PDT by Eric Seidel (no email)
Modified: 2010-04-16 17:34 PDT (History)
2 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Eric Seidel (no email) 2010-03-18 10:37:09 PDT
EWS bots should process r+ bugs as well as r?

We care just as much (or more) about r+ bugs not breaking anything (or knowing sooner that they will)!

Should be a pretty easy change.  May involve adding a new fetch_ method to BugzillaQueries:
http://trac.webkit.org/browser/trunk/WebKitTools/Scripts/webkitpy/bugzilla.py

fetch_potential_patch_ids:
http://trac.webkit.org/browser/trunk/WebKitTools/Scripts/webkitpy/commands/queues.py#L255

is the method that would need to be overridden in:
http://trac.webkit.org/browser/trunk/WebKitTools/Scripts/webkitpy/commands/early_warning_system.py

We might have to change the hierarchy some too since the EWSes inherit from AbstractReviewQueue.
Comment 1 Ojan Vafai 2010-03-18 10:55:36 PDT
The case I care about is when something gets r+'ed with comments. I need a way to make the changes and then send it through the EWS again. Are you saying I should post the patch and r+ it myself even though I'm not a reviewer?

Would be nice if there were a way to send something to the EWS without tying it to review state, e.g. if I could put up an attachment and then paste the URL to the attachment on a web form somewhere, that would be sufficient. Then we could add a "send to EWS" link in bugzilla.
Comment 2 Eric Seidel (no email) 2010-03-18 10:57:07 PDT
That sounds like try-bots. :)
Comment 3 Ojan Vafai 2010-03-18 11:05:16 PDT
Well, it's try bots, except it actually posts your bug to bugzilla and emails the relevant people. But, the point is, what's there now is pretty darn close.
Comment 4 Adam Barth 2010-03-18 12:23:03 PDT
If we have extra capacity, we can run on commit-queue? patches too.
Comment 5 Ojan Vafai 2010-03-18 12:38:57 PDT
(In reply to comment #4)
> If we have extra capacity, we can run on commit-queue? patches too.

But that would mean someone might come by and cq+ my change that I'm waiting on the EWS results for, no?
Comment 6 Adam Barth 2010-03-18 13:50:53 PDT
It sounds like you want try servers, not the EWS.  If something fails EWS, it notices various stakeholders.  If you're uploading patches you have no interest in committing, then we shouldn't be emailing folks on failure.

In the case you mentioned, that you have an R+ and you want to run it through some validation before landing, marking commit-queue? is appropriate.  If someone marks your patch commit-queue+, then it will get landed and go through our usual commit-queue (and, in the future, sheriff-bot) validation.

The more that I think about this, the more I think you should just use "webkit-patch land-safely".  That will upload your patch to bugs.webkit.org and run it through the commit-queue machinery.  We should make the commit-queue machinery robust enough so you can do that without worrying too much about breaking things.
Comment 7 Eric Seidel (no email) 2010-04-16 17:34:07 PDT

*** This bug has been marked as a duplicate of bug 35460 ***