Sanitize git progress message output.

Sanitize git progress message output.

Committed as revision 103787.

[M-A Ruel, 2011-09-28 00:20:05]

I'm fine with the change but Chase or Nicolas can you confirm it works on
msysgit?

[cmp, 2011-09-28 01:35:10]

I like this change, it solves a real problem we want solved.  As for testing it
on Windows, somebody should try it first.  I don't have cycles right now to do
that.  Maybe in the next couple days, but can't be sure.  Maybe you or nsylvain
will have cycles before me?

http://codereview.chromium.org/8052025/diff/1/gclient_utils.py
File gclient_utils.py (right):

http://codereview.chromium.org/8052025/diff/1/gclient_utils.py#newcode363
gclient_utils.py:363: in_line = ''
FYI changing this can have wide and deep effects.  I worry about places that use
this method and pass a filter_fn which assumes it's only called on a newline
where with this change "all of those" methods will be called on carriage return.

"All of those" in quotes because I don't have an accounting of how many places
that is.  Hmm.

[M-A Ruel, 2011-09-28 12:52:09]

I'm fine with the shotgun approach. Commit and I'll revert if I see angry mobs
or stack traces.

lgtm

[szager, 2011-09-29 17:39:48]

On 2011/09/28 01:35:10, cmp wrote:
> I like this change, it solves a real problem we want solved.  As for testing
it
> on Windows, somebody should try it first.  I don't have cycles right now to do
> that.  Maybe in the next couple days, but can't be sure.  Maybe you or
nsylvain
> will have cycles before me?

I ran this on my Windows machine.  Some of the initial output seems to get
buffered ('Counting remote objects...'), but then it catches up and produces the
correct (unbuffered) output.

Committing...

[commit-bot: I haz the power, 2011-09-29 17:40:08]

CQ is trying tha patch. Follow status at
https://chromium-status.appspot.com/cq/szager@google.com/8052025/6001

[M-A Ruel, 2011-10-04 16:58:33]

This broke all the git gclient_scm_test.py. Please fix ASAP.

[szager, 2011-10-04 17:50:37]

Reverted with revision 103935.

Can you suggest a non-hard way to update the expectations of the test?

Thanks,

Stefan

On Tue, Oct 4, 2011 at 9:58 AM,  <maruel@chromium.org> wrote:
> This broke all the git gclient_scm_test.py. Please fix ASAP.
>
> http://codereview.chromium.org/8052025/
>

[M-A Ruel, 2011-10-04 18:02:36]

On 2011/10/04 17:50:37, szager wrote:
> Reverted with revision 103935.
> 
> Can you suggest a non-hard way to update the expectations of the test?

You could _Run() to filter out --progress, but still expect the flag when it is
expected to be provided. With that, none(?) of the expectation will need to be
modified. Should work fine enough. That said, I don't know what msysgit does on
windows with that.