Use author as determined from scm if we can not get it from rietveld

Use author as determined from scm if we can not get it from rietveld

Until an issue is uploaded to Rietveld, we don't know the official
email address to use for an owners check. There are three ways to fix
this: we could attempt to log in to rietveld prior to doing the check
and extract the address to use, or we could use ~/.last_codereview_email_address,
or we can use the email address we can determine from the checkout.

All three options have flaws; the first is particularly awkward since there
doesn't seem to be a good way to fetch the email without posting an issue.
The second is flawed if we use different addresses for different repos,
and the third is flawed if the checkout's email address is different from
the rietveld address, or if it is anonymous.

However, since this is only being used for owners checks (in this case),
anonymous checkouts probably don't matter, and hopefully the cases where
the email addresses differ are rare.

R=maruel@chromium.org
BUG=118388, 150049
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=186259

[Dirk Pranke, 2013-02-28 21:12:42]

WDYT of this? I'm not entirely happy with it, but I think it'll fix 80-90% of
the cases.

[M-A Ruel, 2013-03-03 01:58:57]

I think I'd prefer to auth before running the test but it has the disadvantage
of contacting the server to be able to run git cl presubmit.

It would be possible to auth to Rietveld with a "dummy" get request.

https://codereview.chromium.org/12377023/diff/1/presubmit_canned_checks.py
File presubmit_canned_checks.py (right):

https://codereview.chromium.org/12377023/diff/1/presubmit_canned_checks.py#ne...
presubmit_canned_checks.py:796: if not owner_email:
owner_email = owner_email or input_api.change.author_email

https://codereview.chromium.org/12377023/diff/1/presubmit_canned_checks.py#ne...
presubmit_canned_checks.py:799: message = ''
this line doesn't seem useful anymore.

[M-A Ruel, 2013-03-03 01:59:23]

On 2013/03/03 01:58:57, Marc-Antoine Ruel wrote:
> I think I'd prefer to auth before running the test but it has the disadvantage
> of contacting the server to be able to run git cl presubmit.
> 
> It would be possible to auth to Rietveld with a "dummy" get request.

To clarify, I don't know what's best either.

[Dirk Pranke, 2013-03-04 21:41:12]

> I think I'd prefer to auth before running the test but it has the 
> disadvantage of contacting the server to be able to run 
> git cl presubmit.
>
> It would be possible to auth to Rietveld with a "dummy" get request.

Yeah, I didn't like having to do the round trip. Also, from what I can tell, the
cookies we get back from Rietveld are encrypted, so we'd have to find a dummy
request that actually returned the email, and from looking at the api, I didn't
see such a thing. Adding a new request seemed like it was more trouble than it
was worth.

Given that it's not clear what the "right" fix is, but that this'll probably fix
most of the issues, do you think it's okay to land this and see how it goes?

https://codereview.chromium.org/12377023/diff/1/presubmit_canned_checks.py
File presubmit_canned_checks.py (right):

https://codereview.chromium.org/12377023/diff/1/presubmit_canned_checks.py#ne...
presubmit_canned_checks.py:796: if not owner_email:
On 2013/03/03 01:58:57, Marc-Antoine Ruel wrote:
> owner_email = owner_email or input_api.change.author_email

Done.

https://codereview.chromium.org/12377023/diff/1/presubmit_canned_checks.py#ne...
presubmit_canned_checks.py:799: message = ''
On 2013/03/03 01:58:57, Marc-Antoine Ruel wrote:
> this line doesn't seem useful anymore.

You're right. Removed the reference, below, as well.

[Marc-Antoine Ruel (Google), 2013-03-05 02:05:20]

lgtm

[commit-bot: I haz the power, 2013-03-05 21:03:14]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dpranke@chromium.org/12377023/5002

[commit-bot: I haz the power, 2013-03-05 21:06:04]

Change committed as 186259