deps2git: Improve support for WebRTC and its deps.

deps2git: Improve support for WebRTC and its deps.

* Remove old unused WebRTC URLs (stable branch was removed
  about a year ago).
* Make the WebRTC matching more generic, so it can match both
  trunk and deps. This will add support for:
  - deps/third_party/junit 
  - deps/third_party/openmax  
  - deps/third_party/winsdk_samples_v71
  and thus the special case for openmax can be removed.

BUG=419516
TESTED=Passing build with:
scripts/tools/run_recipe.py webrtc/standalone buildername="Linux64 Debug" mastername=client.webrtc slavename=x revision='"7414"' use_mirror=True
when combined with https://codereview.chromium.org/472963003/

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=292399

[kjellander_chromium, 2014-10-02 20:09:27]

kjellander@chromium.org changed reviewers:
+ hinoka@chromium.org

[kjellander_chromium, 2014-10-02 20:09:27]

This is based on the changes we made when we sat down looking at this in STO +
some additional cleanup.

[kjellander_chromium, 2014-10-09 09:27:51]

On 2014/10/02 20:09:27, kjellander wrote:
> This is based on the changes we made when we sat down looking at this in STO +
> some additional cleanup.

ping!

[Ryan Tseng, 2014-10-10 05:02:00]

hinoka@google.com changed reviewers:
+ agable@chromium.org, hinoka@google.com

[Ryan Tseng, 2014-10-10 05:02:01]

lgtm % comment.  Make sure to inform chrome-troopers and Aaron when you land
this.

https://codereview.chromium.org/627493002/diff/1/svn_to_git_public.py
File svn_to_git_public.py (right):

https://codereview.chromium.org/627493002/diff/1/svn_to_git_public.py#newcode68
svn_to_git_public.py:68: return (path, GIT_HOST + 'external/webrtc/%s/%s' %
(match.group(1), repo),
I think you still need to special case /trunk/src, because that points to
external/webrtc/src instead of external/webrtc/trunk/src

[kjellander_chromium, 2014-10-10 08:46:20]

agable: can you take a look at the unittest?

https://codereview.chromium.org/627493002/diff/1/svn_to_git_public.py
File svn_to_git_public.py (right):

https://codereview.chromium.org/627493002/diff/1/svn_to_git_public.py#newcode68
svn_to_git_public.py:68: return (path, GIT_HOST + 'external/webrtc/%s/%s' %
(match.group(1), repo),
On 2014/10/10 05:02:00, Ryan Tseng wrote:
> I think you still need to special case /trunk/src, because that points to
> external/webrtc/src instead of external/webrtc/trunk/src

No, the src subdirectory of trunk is long gone and no longer used. See
https://chromium.googlesource.com/external/webrtc/src (the latest commit is 2
years old). I think that directory was renamed to webrtc and thus is reflected
by the https://chromium.googlesource.com/external/webrtc/trunk/webrtc/ mirror
instead.
I filed http://crbug.com/422217 to get that mirror removed.

So the cases we have is (SVN -> mirror):
svn/deps/*        -> external/webrtc/deps/third_party/junit	
                     external/webrtc/deps/third_party/openmax	
                     external/webrtc/deps/third_party/winsdk_samples_v71	
                     external/webrtc/deps/tools/android
                     (all created in http://crbug.com/390681)
svn/trunk         -> external/webrtc/
svn/trunk/talk    -> external/webrtc/trunk/talk/
svn/trunk/webrtc  -> external/webrtc/trunk/webrtc/

I added a unittest just to make sure I didn't miss anything.

[agable, 2014-10-10 08:57:51]

LGTM % comments on the test.

https://codereview.chromium.org/627493002/diff/40001/svn_to_git_public_unitte...
File svn_to_git_public_unittest.py (right):

https://codereview.chromium.org/627493002/diff/40001/svn_to_git_public_unitte...
svn_to_git_public_unittest.py:14: 'deps/third_party/junit':
'/deps/third_party/junit',
Not a huge fan of adding *real* urls to this test. There should be urls of all
"shapes" that webrtc urls take (e.g. "deps/third_party/foo", "trunk",
"trunk/bar"), but not a bunch of redundant ones.

https://codereview.chromium.org/627493002/diff/40001/svn_to_git_public_unitte...
svn_to_git_public_unittest.py:24: git_base_url = svn_to_git_public.GIT_HOST +
'external/webrtc'
Also, this kind of logic in a test is unfortunate. Better to have
'external/webrtc/...' in all of the expectations above.

[agable, 2014-10-10 08:58:36]

Oh, also, is there a PRESUBMIT.py for this repo? If so, does it automatically
find this test?

P.S. Please name it "svn_to_git_public_test.py", to match our other PRESUBMIT
regexes.

[kjellander_chromium, 2014-10-10 09:03:53]

Done and renamed. I named it _unittest.py because there was another test in
deps2git/ with that suffix, just assuming that was correct.

https://codereview.chromium.org/627493002/diff/40001/svn_to_git_public_unitte...
File svn_to_git_public_unittest.py (right):

https://codereview.chromium.org/627493002/diff/40001/svn_to_git_public_unitte...
svn_to_git_public_unittest.py:14: 'deps/third_party/junit':
'/deps/third_party/junit',
On 2014/10/10 08:57:51, agable wrote:
> Not a huge fan of adding *real* urls to this test. There should be urls of all
> "shapes" that webrtc urls take (e.g. "deps/third_party/foo", "trunk",
> "trunk/bar"), but not a bunch of redundant ones.

I mostly wanted them for my own sake to doublecheck all the cases. I wiped that
now and put in a more generic set now.

https://codereview.chromium.org/627493002/diff/40001/svn_to_git_public_unitte...
svn_to_git_public_unittest.py:24: git_base_url = svn_to_git_public.GIT_HOST +
'external/webrtc'
On 2014/10/10 08:57:51, agable wrote:
> Also, this kind of logic in a test is unfortunate. Better to have
> 'external/webrtc/...' in all of the expectations above.

Done.

[kjellander_chromium, 2014-10-10 09:09:47]

The CQ bit was checked by kjellander@chromium.org

[commit-bot: I haz the power, 2014-10-10 09:09:55]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/627493002/180001

[commit-bot: I haz the power, 2014-10-10 09:10:09]

Committed patchset #5 (id:180001) as 292399

[kjellander_chromium, 2014-10-13 06:23:38]

On 2014/10/10 09:10:09, I haz the power (commit-bot) wrote:
> Committed patchset #5 (id:180001) as 292399

This was reverted in https://codereview.chromium.org/651833002/ due to breaking
bots in http://build.chromium.org/p/chromium.webrtc/waterfall
due to deps in
http://chromegw.corp.google.com/viewvc/chrome/trunk/deps/third_party/webrtc/w...
not getting the right mirrors.