Convert (nearly) all layout tests to unix line endings

Convert (nearly) all layout tests to unix line endings

Rietveld and the CQ can't apply patches against files with DOS line
endings, so convert them all to unix.  This skips mhtml tests, which
apparently need their carriage returns.

These files were found via git grep -I -l $'\r' | grep -v "gif$" | grep
-v mhtml (apparently some gif files are marked as text), and then
converted by opening them all in vim and doing a :bufdo set ff=unix.

The attached code review only includes a sample of two files, as
Rietveld is also not happy about attaching 1800 files to a code review.

BUG=357373

[enne (OOO), 2014-03-26 00:34:24]

This patch will need to be dcommitted.

If you really want to see the full patch, I've included a git diff -R --reverse
version of the patch here: (internal link only for 14mb patch, sorry):
http://www/~enne/patches/line_ending_layout_test_reverse.patch

(You can only turn on white space in the diff when you're looking at the patch
applied in reverse, sorry.  Thanks, git.)

[enne (OOO), 2014-03-27 19:14:55]

eseidel: what do you think of this? I could also land these in chunks of ~100 or
something to make it more sane than one monster patch.

[enne (OOO), 2014-03-27 20:29:23]

+ojan for opinions/review

[eseidel, 2014-03-27 21:35:24]

So long as we aren't braking intentional testing of \r, etc, i'm all for making
the repository more consistent and easy to work with.  We could also just fix
Rietveld to be better with these types of patches, which is something Jason may
be working on (you can ask him he sits not very far from you).

I'm surprised you need to do this by hand, and that SVN/Git doesn't just do this
for you when you set the line-ending types on a file.

[eseidel, 2014-03-27 21:37:02]

I would poke szager or jrobbins before committing (since they both sit near you)
and make sure this sounds like a san way to do this.  But yes, I'm supportive of
making Blink easier to work in by making things like line endings consistent.

I'm not concerned about breaking git blame.  git blame or similar tools will
eventually be taught how to ignore whitespace changes if they don't already know
how today.

rslgtm.

[szager1, 2014-03-27 21:47:59]

On 2014/03/27 21:37:02, eseidel wrote:
> I would poke szager or jrobbins before committing (since they both sit near
you)
> and make sure this sounds like a san way to do this.  But yes, I'm supportive
of
> making Blink easier to work in by making things like line endings consistent.
> 
> I'm not concerned about breaking git blame.  git blame or similar tools will
> eventually be taught how to ignore whitespace changes if they don't already
know
> how today.
> 
> rslgtm.

If it layout tests pass and it gets through the try bots, lg to me.

[Jason Robbins -- corp, 2014-03-27 22:10:49]

On 2014/03/27 21:35:24, eseidel wrote:
> Rietveld to be better with these types of patches, which is something Jason
may
> be working on (you can ask him he sits not very far from you).

I'm not planning any change to how Rietveld handles line endings.  
The chromium style guide says to use LF:
http://dev.chromium.org/developers/coding-style#TOC-Non-coding-miscellany

[enne (OOO), 2014-03-27 22:34:45]

Thanks.  :)

I'll go ahead and land these in chunks of ~100 or so and just TBR them all.