Removed an assert in presubmit_shim.py

Removed an assert in presubmit_shim.py

The removed assert prevents calling presubmit_support with an empty
password. But an empty password is required to connect to a locally-
running rietveld instance.

BUG=330923
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=242939

[pgervais, 2013-12-28 00:35:04]

This modification can be risky, since some code might rely on this assert being
present to catch problems. Feel free to suggest addition reviewers if required.

[Paweł Hajdan Jr., 2014-01-02 12:25:30]

https://codereview.chromium.org/117513003/diff/1/verification/presubmit_shim.py
File verification/presubmit_shim.py (right):

https://codereview.chromium.org/117513003/diff/1/verification/presubmit_shim....
verification/presubmit_shim.py:18: import presubmit_support  # pylint:
disable=F0401
I don't remember hitting pylint errors on importing modules when running CQ
presubmit. Please figure this out locally and don't add suppressions in this CL
- they shouldn't be needed.

[pgervais, 2014-01-02 16:10:21]

On 2014/01/02 12:25:30, Paweł Hajdan Jr. wrote:
>
https://codereview.chromium.org/117513003/diff/1/verification/presubmit_shim.py
> File verification/presubmit_shim.py (right):
> 
>
https://codereview.chromium.org/117513003/diff/1/verification/presubmit_shim....
> verification/presubmit_shim.py:18: import presubmit_support  # pylint:
> disable=F0401
> I don't remember hitting pylint errors on importing modules when running CQ
> presubmit. Please figure this out locally and don't add suppressions in this
CL
> - they shouldn't be needed.

The problem happens when running pylint alone, not during presubmit. I run
pylint on the fly in my editor, and it does hit these errors, because it doesn't
know about the crazy changes to sys.path.
And yes, running pylint on the fly is useful because it avoids creating one
local git commit per style fix when trying to upload the CL. 

However, if you have a way to tell pylint about the full path needed to import
everything, feel free to tell me. I agree that added suppressions is
sub-optimal.

[Paweł Hajdan Jr., 2014-01-02 16:45:13]

On 2014/01/02 16:10:21, pgervais wrote:
> The problem happens when running pylint alone, not during presubmit. I run
> pylint on the fly in my editor, and it does hit these errors, because it
doesn't
> know about the crazy changes to sys.path.
> And yes, running pylint on the fly is useful because it avoids creating one
> local git commit per style fix when trying to upload the CL. 
> 
> However, if you have a way to tell pylint about the full path needed to import
> everything, feel free to tell me. I agree that added suppressions is
> sub-optimal.

Note: you can reply to specific comments in Rietveld, and it's slightly easier
to respond to them that way (proper threading, especially if there are multiple
comments).

Please remove the pylint overrides from this CL. They are not related, and I
don't think we're going to add them to all files (CQ is not the only affected
place).

Feel free to start a discussion on one of the MLs about how to get standalone
pylint to work. Maybe just looking at what the presubmit check does can lead so
some solutions. Good luck.

[pgervais, 2014-01-03 20:21:22]

I have removed the pylint comments.

On 2014/01/02 16:45:13, Paweł Hajdan Jr. wrote:
> 
> Feel free to start a discussion on one of the MLs about how to get standalone
> pylint to work. Maybe just looking at what the presubmit check does can lead
so
> some solutions. Good luck.

The paths added to sys.path are often hard-coded into PRESUBMIT.py or any other
file. 
Having them in a reusable form will take a lot of refactoring, so for now I'm to
copy-pasting them in one of my personal scripts.
One long-term solution could be something similar to the runit.py script in
build/scripts/tools/

+stip, who is involved in runit.py

[ghost stip (do not use), 2014-01-03 20:26:49]

I'm going to side with Paweł on this one, in that we should attempt to move
toward a unified sys.path style.

One thing you can do is run your editor under runit.py itself (`runit.py vim` or
`runit.py emacs`).

[pgervais, 2014-01-03 21:51:55]

On 2014/01/03 20:26:49, stip wrote:
> I'm going to side with Paweł on this one, in that we should attempt to move
> toward a unified sys.path style.
> 
I agree. We should talk about that some day.

> One thing you can do is run your editor under runit.py itself (`runit.py vim`
or
> `runit.py emacs`).

In the present case (commit-queue), there's no runit.py.
Even using 'runit.py emacs' doesn't work when you're working on several build/
at the same time (that happens).

Now that I have removed those pylint comments, can I have your opinion on the
essential modification (assert removal)?

[ghost stip (do not use), 2014-01-03 21:55:17]

lgtm

[commit-bot: I haz the power, 2014-01-03 22:16:23]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/pgervais@chromium.org/117513003/80001

[commit-bot: I haz the power, 2014-01-03 22:16:52]

Change committed as 242939