gclient: Actually move or delete mismatched checkouts

gclient: Actually move or delete mismatched checkouts

Followup to https://codereview.chromium.org/189913020/ ("gclient: print a warning if a dep would get deleted or moved in the future")

BUG=skia:1638
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=268944

[iannucci, 2014-04-07 18:43:12]

Preemptive lgtm

[Paweł Hajdan Jr., 2014-04-15 13:01:47]

Please make sure to test with "fetch chromium", see
https://goto.google.com/zhiyk (Googlers only)

[DaleCurtis, 2014-04-24 17:32:40]

Hey Eric, what's the plan this CL?  I thought it would be landing before "flag
day".  I was planning to roll over the FFmpeg repo before then.

[borenet, 2014-04-24 17:47:53]

On 2014/04/24 17:32:40, DaleCurtis wrote:
> Hey Eric, what's the plan this CL?  I thought it would be landing before "flag
> day".  I was planning to roll over the FFmpeg repo before then.

Dale, I'm still fixing bugs and dealing with workflows which trigger the
warning.  If this were submitted now, either those workflows would fail to
checkout or, worse, checkouts would get deleted.

Believe me, we want to get Skia switched over before "flag day" as well.

[scottmg, 2014-04-30 17:19:49]

Hi, I just got a warning pointing to this CL when I Ctrl-C'd a gclient sync (I
think maybe the output is normally swallowed, which might be concerning for this
not getting reported).

Warnings:
WARNING: Upcoming change in https://codereview.chromium.org/225403015 would
cause D:\src\cr3\src\chrome\app\theme\default_100_percent\google_chrome to be
deleted or moved to the side. This is intended to ease changes to DEPS in the
future. If you are seeing this warning and haven't changed the DEPS file, please
contact borenet@ immediately.
WARNING: Upcoming change in https://codereview.chromium.org/225403015 would
cause D:\src\cr3\src\pdf to be deleted or moved to the side. This is intended to
ease changes to DEPS in the future. If you are seeing this warning and haven't
changed the DEPS file, please contact borenet@ immediately.

At the moment I have a "regular" src.git for chromium set up via "fetch
chromium", and two other svn entries for src-internal and pdf. Is this because
I'm using svn for those deps and I need to switch to git?

[scottmg, 2014-04-30 17:22:14]

On 2014/04/30 17:19:49, scottmg wrote:
> Hi, I just got a warning pointing to this CL when I Ctrl-C'd a gclient sync (I
> think maybe the output is normally swallowed, which might be concerning for
this
> not getting reported).
> 
> Warnings:
> WARNING: Upcoming change in https://codereview.chromium.org/225403015 would
> cause D:\src\cr3\src\chrome\app\theme\default_100_percent\google_chrome to be
> deleted or moved to the side. This is intended to ease changes to DEPS in the
> future. If you are seeing this warning and haven't changed the DEPS file,
please
> contact borenet@ immediately.
> WARNING: Upcoming change in https://codereview.chromium.org/225403015 would
> cause D:\src\cr3\src\pdf to be deleted or moved to the side. This is intended
to
> ease changes to DEPS in the future. If you are seeing this warning and haven't
> changed the DEPS file, please contact borenet@ immediately.
> 
> At the moment I have a "regular" src.git for chromium set up via "fetch
> chromium", and two other svn entries for src-internal and pdf. Is this because
> I'm using svn for those deps and I need to switch to git?

Hm, actually, running gclient sync -v, I don't see any warnings, so I think
perhaps this is just a minor issue with how Ctrl-C gets handled in gclient (or
something).

[borenet, 2014-04-30 17:22:45]

On 2014/04/30 17:19:49, scottmg wrote:
> Hi, I just got a warning pointing to this CL when I Ctrl-C'd a gclient sync (I
> think maybe the output is normally swallowed, which might be concerning for
this
> not getting reported).
> 
> Warnings:
> WARNING: Upcoming change in https://codereview.chromium.org/225403015 would
> cause D:\src\cr3\src\chrome\app\theme\default_100_percent\google_chrome to be
> deleted or moved to the side. This is intended to ease changes to DEPS in the
> future. If you are seeing this warning and haven't changed the DEPS file,
please
> contact borenet@ immediately.
> WARNING: Upcoming change in https://codereview.chromium.org/225403015 would
> cause D:\src\cr3\src\pdf to be deleted or moved to the side. This is intended
to
> ease changes to DEPS in the future. If you are seeing this warning and haven't
> changed the DEPS file, please contact borenet@ immediately.
> 
> At the moment I have a "regular" src.git for chromium set up via "fetch
> chromium", and two other svn entries for src-internal and pdf. Is this because
> I'm using svn for those deps and I need to switch to git?

I don't know much about the output getting swallowed, but my experience with
interrupting a "gclient sync" is typically that I have to delete the checkout
and start over.  I can usually do some surgery like deleting only the
directories which are in a broken state.

[borenet, 2014-05-05 20:09:51]

Tested:

$ fetch chromium
(no errors, warnings, moves, or removals)
$ cd src
$ vi .DEPS.git
(hack Skia DEPS to check out the git repo in third_party/skia)
$ git add .DEPS.git && git commit
$ gclient sync --nohooks
Warnings:
Conflicting directory ...../src/third_party/skia moved to
...../_bad_scm/src/third_party/skiaoGzFUr.
$ git reset --hard HEAD
$ gclient sync
Warnings:
Conflicting directory ...../src/third_party/skia/include moved to
...../_bad_scm/src/third_party/skia/includeCT3pH5.
Conflicting directory ...../src/third_party/skia/gyp moved to
...../_bad_scm/src/third_party/skia/gyplEWFry.
Conflicting directory ...../src/third_party/skia/src moved to
...../_bad_scm/src/third_party/skia/srcOxX9QB.
$ vi .DEPS.git
(hack Skia DEPS to check out the git repo in third_party/skia)
$ git add .DEPS.git && git commit
$ gclient sync --nohooks
WARNING: 'src/third_party/skia/include' is no longer part of this client.  It is
recommended that you manually remove it.

WARNING: 'src/third_party/skia/src' is no longer part of this client.  It is
recommended that you manually remove it.

https://codereview.chromium.org/225403015/diff/80001/tests/gclient_scm_test.py
File tests/gclient_scm_test.py (right):

https://codereview.chromium.org/225403015/diff/80001/tests/gclient_scm_test.p...
tests/gclient_scm_test.py:683: gclient_scm.os.environ['CHROME_HEADLESS'] = '1'
This seems gross; is there a better way?

[iannucci, 2014-05-05 20:18:11]

lgtm, did you test with CHROME_HEADLESS=1 && --force?

https://codereview.chromium.org/225403015/diff/80001/tests/gclient_scm_test.py
File tests/gclient_scm_test.py (right):

https://codereview.chromium.org/225403015/diff/80001/tests/gclient_scm_test.p...
tests/gclient_scm_test.py:683: gclient_scm.os.environ['CHROME_HEADLESS'] = '1'
On 2014/05/05 20:09:52, borenet wrote:
> This seems gross; is there a better way?

Hm. I think mox could do this for you, but I'd have to dig into it.

[borenet, 2014-05-05 20:30:52]

On 2014/05/05 20:18:11, iannucci wrote:
> lgtm, did you test with CHROME_HEADLESS=1 && --force?

testUpdateGitForce does this, but just checked manually:

$ CHROME_HEADLESS=1 gclient sync --force --nohooks
...
Warnings:
Conflicting directory /usr/local/google/chromium/src/third_party/skia deleted.


> 
> https://codereview.chromium.org/225403015/diff/80001/tests/gclient_scm_test.py
> File tests/gclient_scm_test.py (right):
> 
>
https://codereview.chromium.org/225403015/diff/80001/tests/gclient_scm_test.p...
> tests/gclient_scm_test.py:683: gclient_scm.os.environ['CHROME_HEADLESS'] = '1'
> On 2014/05/05 20:09:52, borenet wrote:
> > This seems gross; is there a better way?
> 
> Hm. I think mox could do this for you, but I'd have to dig into it.

[borenet, 2014-05-07 20:54:50]

The CQ bit was checked by borenet@google.com

[commit-bot: I haz the power, 2014-05-07 20:55:41]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/borenet@google.com/225403015/80001

[commit-bot: I haz the power, 2014-05-07 20:57:29]

Change committed as 268944