Re-re-land gclient deletion of mismatching checkouts again

Re-re-land gclient deletion of mismatching checkouts again

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=253616

[borenet, 2014-02-18 21:56:58]

Trying this again.  Patch set 1 is the patch as it landed last, with buildslave
hostname fixes included.  The subsequent patch sets move the deletion logic into
gclient_scm.  Note that the deletion is *not* guarded in this patch.  I can put
in a guard, but it seemed like it did more harm than good with the last attempt.

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py
File gclient_scm.py (left):

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#oldcode1153
gclient_scm.py:1153: return
I removed this check altogether. Seems like we'd just want to delete the
directory and re-sync in this case.

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode1
gclient_scm.py:1: # Copyright (c) 2012 The Chromium Authors. All rights
reserved.
I moved around some logic.  I think it shouldn't have any effect in most cases,
but it makes the deletion easier.

[borenet, 2014-02-18 21:57:55]

On 2014/02/18 21:56:58, borenet wrote:
> Trying this again.  Patch set 1 is the patch as it landed last, with
buildslave
> hostname fixes included.  The subsequent patch sets move the deletion logic
into
> gclient_scm.  Note that the deletion is *not* guarded in this patch.  I can
put
> in a guard, but it seemed like it did more harm than good with the last
attempt.
> 
> https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py
> File gclient_scm.py (left):
> 
>
https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#oldcode1153
> gclient_scm.py:1153: return
> I removed this check altogether. Seems like we'd just want to delete the
> directory and re-sync in this case.
> 
> https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py
> File gclient_scm.py (right):
> 
> https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode1
> gclient_scm.py:1: # Copyright (c) 2012 The Chromium Authors. All rights
> reserved.
> I moved around some logic.  I think it shouldn't have any effect in most
cases,
> but it makes the deletion easier.

I will probably upload a different CL tomorrow which adds a bunch of tests to
verify that the upstream switching behaves the same.

[iannucci, 2014-02-21 02:17:00]

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode391
gclient_scm.py:391: not os.path.exists(os.path.join(self.checkout_path,
'.git')))):
I think this logic is more complex than it needs to be at this point, right?

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode1199
gclient_scm.py:1199: return # TODO(borenet): Get the svn revision number?
This is weird... I don't understand why this clause is necessary. IIUC, this
would happen when you:
  * have DEPS make the repo git
  * manually set up git-svn
  * have DEPS make the repo svn

Is this expected to happen? The only case I can think of is blink in the
chromium tree...

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode1208
gclient_scm.py:1208: if from_info and from_info['URL'] != base_url:
if not needs_delete and ....

?

[borenet, 2014-02-21 13:38:17]

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode391
gclient_scm.py:391: not os.path.exists(os.path.join(self.checkout_path,
'.git')))):
On 2014/02/21 02:17:00, iannucci wrote:
> I think this logic is more complex than it needs to be at this point, right?

Changed to just "os.path.exists".  This is more strict, since it will no longer
check out into a directory already containing a checkout, although the removal
logic above should prevent there being a checkout there anyway.

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode1199
gclient_scm.py:1199: return # TODO(borenet): Get the svn revision number?
On 2014/02/21 02:17:00, iannucci wrote:
> This is weird... I don't understand why this clause is necessary. IIUC, this
> would happen when you:
>   * have DEPS make the repo git
>   * manually set up git-svn
>   * have DEPS make the repo svn
> 
> Is this expected to happen? The only case I can think of is blink in the
> chromium tree...

One of my previous submissions broke a project whose workflow was something
like:

git svn clone svn://repo
gclient config svn://repo
gclient sync

They were depending on the functionality in SVNWrapper.update() which noticed
the .git directory and returned early.  I added this slightly more sophisticated
version because we don't want to delete the checkout if it's actually the right
repo.

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode1208
gclient_scm.py:1208: if from_info and from_info['URL'] != base_url:
On 2014/02/21 02:17:00, iannucci wrote:
> if not needs_delete and ....
> 
> ?

Not sure it's needed but added anyway.

[iannucci, 2014-02-24 21:18:19]

lgtm.... Pawel (cc'd) is trooper today

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/70001/gclient_scm.py#newcode1199
gclient_scm.py:1199: return # TODO(borenet): Get the svn revision number?
On 2014/02/21 13:38:17, borenet wrote:
> On 2014/02/21 02:17:00, iannucci wrote:
> > This is weird... I don't understand why this clause is necessary. IIUC, this
> > would happen when you:
> >   * have DEPS make the repo git
> >   * manually set up git-svn
> >   * have DEPS make the repo svn
> > 
> > Is this expected to happen? The only case I can think of is blink in the
> > chromium tree...
> 
> One of my previous submissions broke a project whose workflow was something
> like:
> 
> git svn clone svn://repo
> gclient config svn://repo
> gclient sync

:( That is so sad in so many ways :(

> 
> They were depending on the functionality in SVNWrapper.update() which noticed
> the .git directory and returned early.  I added this slightly more
sophisticated
> version because we don't want to delete the checkout if it's actually the
right
> repo.

[borenet, 2014-02-25 20:11:55]

As discussed, put the deletion logic behind a guard.  Added an exception for
deleting unversioned SVN checkouts to preserve the old behavior with the guard
in place.

[iannucci, 2014-02-25 22:32:03]

lgtm

Looks as good to me as any of the other attempts :/

[szager1, 2014-02-25 23:32:55]

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode366
gclient_scm.py:366: if subprocess2.capture(
self._Capture ?

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode368
gclient_scm.py:368: cwd=self.checkout_path).strip() != 'False':
Check for more standard boolean strings, e.g., '0', '[Nn]o', '[Ff]alse'

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode377
gclient_scm.py:377: needs_delete = True
Let me make sure I understand this...

If the url has changed, and I have 'remote.origin.gclient-auto-fix-url = False'
in my .git/config, then my checkout gets deleted and re-cloned from the new url?
 That seems a bit weird.

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode1190
gclient_scm.py:1190: if scm.GIT.IsGitSvn(self.checkout_path):
I read the previous comments on this, I understand it, and I am throwing up in
my mouth a little bit.  No offense, not your fault; I just thought you should
know.

https://codereview.chromium.org/164053003/diff/240001/scm.py
File scm.py (right):

https://codereview.chromium.org/164053003/diff/240001/scm.py#newcode172
scm.py:172: if not os.path.isdir(os.path.join(checkout_root, '.git')):
In some cases, .git will be a file, not a directory.

[iannucci, 2014-02-25 23:42:11]

On 2014/02/25 23:32:55, szager1 wrote:
> https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py
> File gclient_scm.py (right):
> 
>
https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode366
> gclient_scm.py:366: if subprocess2.capture(
> self._Capture ?
> 
>
https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode368
> gclient_scm.py:368: cwd=self.checkout_path).strip() != 'False':
> Check for more standard boolean strings, e.g., '0', '[Nn]o', '[Ff]alse'
> 
>
https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode377
> gclient_scm.py:377: needs_delete = True
> Let me make sure I understand this...
> 
> If the url has changed, and I have 'remote.origin.gclient-auto-fix-url =
False'
> in my .git/config, then my checkout gets deleted and re-cloned from the new
url?
>  That seems a bit weird.
> 
>
https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode1190
> gclient_scm.py:1190: if scm.GIT.IsGitSvn(self.checkout_path):
> I read the previous comments on this, I understand it, and I am throwing up in
> my mouth a little bit.  No offense, not your fault; I just thought you should
> know.
> 
> https://codereview.chromium.org/164053003/diff/240001/scm.py
> File scm.py (right):
> 
> https://codereview.chromium.org/164053003/diff/240001/scm.py#newcode172
> scm.py:172: if not os.path.isdir(os.path.join(checkout_root, '.git')):
> In some cases, .git will be a file, not a directory.

I don't think we have any tools which generate those style git checkouts, do we?
(but yes, it can be a application-level 'symlink' instead of a dir)

[borenet, 2014-02-26 14:26:35]

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode366
gclient_scm.py:366: if subprocess2.capture(
On 2014/02/25 23:32:56, szager1 wrote:
> self._Capture ?

This block is just moved and slightly modified from lines 371-391.  I'd prefer
to modify it as little as possible.

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode368
gclient_scm.py:368: cwd=self.checkout_path).strip() != 'False':
On 2014/02/25 23:32:56, szager1 wrote:
> Check for more standard boolean strings, e.g., '0', '[Nn]o', '[Ff]alse'

Again, this is just moved from below.  I can change it as you suggest if you
feel strongly, but would rather leave it alone.

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode377
gclient_scm.py:377: needs_delete = True
On 2014/02/25 23:32:56, szager1 wrote:
> Let me make sure I understand this...
> 
> If the url has changed, and I have 'remote.origin.gclient-auto-fix-url =
False'
> in my .git/config, then my checkout gets deleted and re-cloned from the new
url?
>  That seems a bit weird.

What's the expected behavior in that case?  It seems to me that, if what you
have locally doesn't match what you want (as specified in DEPS), and it can't be
auto-fixed, the right behavior (with --force anyway) is to delete and re-clone.

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode1190
gclient_scm.py:1190: if scm.GIT.IsGitSvn(self.checkout_path):
On 2014/02/25 23:32:56, szager1 wrote:
> I read the previous comments on this, I understand it, and I am throwing up in
> my mouth a little bit.  No offense, not your fault; I just thought you should
> know.

Yep.

https://codereview.chromium.org/164053003/diff/240001/scm.py
File scm.py (right):

https://codereview.chromium.org/164053003/diff/240001/scm.py#newcode172
scm.py:172: if not os.path.isdir(os.path.join(checkout_root, '.git')):
On 2014/02/25 23:32:56, szager1 wrote:
> In some cases, .git will be a file, not a directory.

Changed to exists()

[szager1, 2014-02-26 20:33:14]

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode366
gclient_scm.py:366: if subprocess2.capture(
On 2014/02/26 14:26:35, borenet wrote:
> On 2014/02/25 23:32:56, szager1 wrote:
> > self._Capture ?
> 
> This block is just moved and slightly modified from lines 371-391.  I'd prefer
> to modify it as little as possible.

Actually, I think this is the ideal time to improve the code health in this way.

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode368
gclient_scm.py:368: cwd=self.checkout_path).strip() != 'False':
On 2014/02/26 14:26:35, borenet wrote:
> On 2014/02/25 23:32:56, szager1 wrote:
> > Check for more standard boolean strings, e.g., '0', '[Nn]o', '[Ff]alse'
> 
> Again, this is just moved from below.  I can change it as you suggest if you
> feel strongly, but would rather leave it alone.

Same comment, it should be easy to make the behavior more intuitive:

if self._Capture(...) not in ['0', 'false', 'False', 'no', 'No']:

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode377
gclient_scm.py:377: needs_delete = True
On 2014/02/26 14:26:35, borenet wrote:
> On 2014/02/25 23:32:56, szager1 wrote:
> > Let me make sure I understand this...
> > 
> > If the url has changed, and I have 'remote.origin.gclient-auto-fix-url =
> False'
> > in my .git/config, then my checkout gets deleted and re-cloned from the new
> url?
> >  That seems a bit weird.
> 
> What's the expected behavior in that case?  It seems to me that, if what you
> have locally doesn't match what you want (as specified in DEPS), and it can't
be
> auto-fixed, the right behavior (with --force anyway) is to delete and
re-clone.

I don't know what the expected behavior is; but setting
remote.origin.gclient-auto-fix-url=False *sounds* like the intended behavior is
"hands off my checkout, gclient!"

I guess I'd like to hear from maruel, but my gut tells me we shouldn't blow away
the checkout when that config is set.

[borenet, 2014-02-26 21:18:19]

Uploaded patch set 10.

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode366
gclient_scm.py:366: if subprocess2.capture(
On 2014/02/26 20:33:15, szager1 wrote:
> On 2014/02/26 14:26:35, borenet wrote:
> > On 2014/02/25 23:32:56, szager1 wrote:
> > > self._Capture ?
> > 
> > This block is just moved and slightly modified from lines 371-391.  I'd
prefer
> > to modify it as little as possible.
> 
> Actually, I think this is the ideal time to improve the code health in this
way.

Ok.  I was trying to be non-intrusive.

https://codereview.chromium.org/164053003/diff/240001/gclient_scm.py#newcode377
gclient_scm.py:377: needs_delete = True
On 2014/02/26 20:33:15, szager1 wrote:
> On 2014/02/26 14:26:35, borenet wrote:
> > On 2014/02/25 23:32:56, szager1 wrote:
> > > Let me make sure I understand this...
> > > 
> > > If the url has changed, and I have 'remote.origin.gclient-auto-fix-url =
> > False'
> > > in my .git/config, then my checkout gets deleted and re-cloned from the
new
> > url?
> > >  That seems a bit weird.
> > 
> > What's the expected behavior in that case?  It seems to me that, if what you
> > have locally doesn't match what you want (as specified in DEPS), and it
can't
> be
> > auto-fixed, the right behavior (with --force anyway) is to delete and
> re-clone.
> 
> I don't know what the expected behavior is; but setting
> remote.origin.gclient-auto-fix-url=False *sounds* like the intended behavior
is
> "hands off my checkout, gclient!"
> 
> I guess I'd like to hear from maruel, but my gut tells me we shouldn't blow
away
> the checkout when that config is set.

Changed this a bit: if gclient-auto-fix-url is not set, an exception will be
thrown.  In that case, delete the checkout.  If no exception is thrown, auto-fix
if it's something true-like, or just move on otherwise.  That's close to the
original behavior.

[szager1, 2014-02-26 21:34:07]

https://codereview.chromium.org/164053003/diff/280001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/280001/gclient_scm.py#newcode365
gclient_scm.py:365: auto_fix = False
Shouldn't this be:

auto_fix = 'true'

?

[borenet, 2014-02-26 21:52:18]

https://codereview.chromium.org/164053003/diff/280001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/280001/gclient_scm.py#newcode365
gclient_scm.py:365: auto_fix = False
On 2014/02/26 21:34:07, szager1 wrote:
> Shouldn't this be:
> 
> auto_fix = 'true'
> 
> ?

I *think* the behavior we want is to delete the checkout if gclient-auto-fix-url
is not set.  If it's 'true' or 'false' or any variant thereof, we'll auto-fix or
ignore.

[szager1, 2014-02-26 22:16:52]

https://codereview.chromium.org/164053003/diff/280001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/280001/gclient_scm.py#newcode365
gclient_scm.py:365: auto_fix = False
On 2014/02/26 21:52:18, borenet wrote:
> On 2014/02/26 21:34:07, szager1 wrote:
> > Shouldn't this be:
> > 
> > auto_fix = 'true'
> > 
> > ?
> 
> I *think* the behavior we want is to delete the checkout if
gclient-auto-fix-url
> is not set.  If it's 'true' or 'false' or any variant thereof, we'll auto-fix
or
> ignore.

Ah, I see.  I'm not trying to torture you with this review, really I'm not :)

It should be OK to update the URL without deleting the checkout if
remote.origin.gclient-auto-fix-url is unset.  Basically, I think the logic
should be:

If remote.gclient-auto-fix-url is unset or true:
  if the url can be updated:
    update url
  else:
    delete checkout and re-clone
else:
  pass

[borenet, 2014-02-26 22:49:52]

https://codereview.chromium.org/164053003/diff/280001/gclient_scm.py
File gclient_scm.py (right):

https://codereview.chromium.org/164053003/diff/280001/gclient_scm.py#newcode365
gclient_scm.py:365: auto_fix = False
On 2014/02/26 22:16:53, szager1 wrote:
> On 2014/02/26 21:52:18, borenet wrote:
> > On 2014/02/26 21:34:07, szager1 wrote:
> > > Shouldn't this be:
> > > 
> > > auto_fix = 'true'
> > > 
> > > ?
> > 
> > I *think* the behavior we want is to delete the checkout if
> gclient-auto-fix-url
> > is not set.  If it's 'true' or 'false' or any variant thereof, we'll
auto-fix
> or
> > ignore.
> 
> Ah, I see.  I'm not trying to torture you with this review, really I'm not :)
> 
> It should be OK to update the URL without deleting the checkout if
> remote.origin.gclient-auto-fix-url is unset.  Basically, I think the logic
> should be:
> 
> If remote.gclient-auto-fix-url is unset or true:
>   if the url can be updated:
>     update url
>   else:
>     delete checkout and re-clone
> else:
>   pass

Got it.  Changed to switch the checkout rather than deleting if
gclient-auto-fix-url is true or not set.  I don't think there's a "url can't be
updated" case, since we _CheckClean, FetchAndReset, right?

[szager1, 2014-02-26 22:53:32]

lgtm, thanks!  You may fire when ready.

[borenet, 2014-02-26 23:07:40]

On 2014/02/26 22:53:32, szager1 wrote:
> lgtm, thanks!  You may fire when ready.

Thanks for the review.

[borenet, 2014-02-26 23:07:46]

The CQ bit was checked by borenet@google.com

[commit-bot: I haz the power, 2014-02-26 23:07:53]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/borenet@google.com/164053003/300001

[commit-bot: I haz the power, 2014-02-26 23:09:35]

Change committed as 253616

[mattm, 2014-02-26 23:54:43]

On 2014/02/26 23:09:35, I haz the power (commit-bot) wrote:
> Change committed as 253616

This change broke fetching a new checkout:

mmueller2 15:48:34 /y/src/chrome$ fetch --nohooks chromium
Running: svn ls --non-interactive svn://svn.chromium.org/chrome
PRESUBMIT.py
branches/
codereview.settings
releases/
trunk/
Running: git config --global core.deltaBaseCacheLimit 1G
Running: gclient config --spec 'solutions = [{u'"'"'managed'"'"': False,
u'"'"'name'"'"': u'"'"'src'"'"', u'"'"'url'"'"':
u'"'"'https://chromium.googlesource.com/chromium/src.git'"'"',
u'"'"'custom_deps'"'"': {}, u'"'"'deps_file'"'"': u'"'"'.DEPS.git'"'"',
u'"'"'safesync_url'"'"': u'"'"''"'"'}]'
Running: gclient sync --nohooks
Traceback (most recent call last):
  File "/y/src/depot_tools/gclient.py", line 1926, in <module>
    sys.exit(Main(sys.argv[1:]))
  File "/y/src/depot_tools/gclient.py", line 1916, in Main
    return dispatcher.execute(OptionParser(), argv)
  File "/y/src/depot_tools/subcommand.py", line 245, in execute
    return command(parser, args[1:])
  File "/y/src/depot_tools/gclient.py", line 1697, in CMDsync
    ret = client.RunOnDeps('update', args)
  File "/y/src/depot_tools/gclient.py", line 1239, in RunOnDeps
    work_queue.flush(revision_overrides, command, args, options=self._options)
  File "/y/src/depot_tools/gclient_utils.py", line 843, in run
    self.item.run(*self.args, **self.kwargs)
  File "/y/src/depot_tools/gclient.py", line 672, in run
    file_list)
  File "/y/src/depot_tools/gclient_scm.py", line 142, in RunCommand
    return getattr(self, command)(options, args, file_list)
  File "/y/src/depot_tools/gclient_scm.py", line 307, in update
    self.UpdateSubmoduleConfig()
  File "/y/src/depot_tools/gclient_scm.py", line 218, in UpdateSubmoduleConfig
    gclient_utils.CheckCallAndFilter(cmd, **kwargs)
  File "/y/src/depot_tools/gclient_utils.py", line 453, in CheckCallAndFilter
    **kwargs)
  File "/y/src/depot_tools/subprocess2.py", line 237, in __init__
    super(Popen, self).__init__(args, **kwargs)
  File "/usr/lib/python2.7/subprocess.py", line 679, in __init__
    errread, errwrite)
  File "/usr/lib/python2.7/subprocess.py", line 1249, in _execute_child
    raise child_exception
OSError: [Errno 2] No such file or directory: '/y/src/chrome/src'
Sending crash report ...
  args: ['/y/src/depot_tools/gclient.py', 'sync', '--nohooks']
  cwd: /y/src/chrome
  exception: [Errno 2] No such file or directory: '/y/src/chrom
  host: mmueller2.mtv.corp.google.com
  stack:   File "/y/src/depot_tools/gclient.py", line 1926,
  user: mattm
  version: 2.7.3 (default, Sep 26 2013, 20:03:06)
[GCC 4.6.3
A stack trace has been sent to the maintainers.
Traceback (most recent call last):
  File "/y/src/depot_tools/fetch.py", line 299, in <module>
    sys.exit(main())
  File "/y/src/depot_tools/fetch.py", line 295, in main
    return run(options, spec, root)
  File "/y/src/depot_tools/fetch.py", line 289, in run
    return checkout.init()
  File "/y/src/depot_tools/fetch.py", line 152, in init
    super(GclientGitSvnCheckout, self).init()
  File "/y/src/depot_tools/fetch.py", line 118, in init
    self.run_gclient('sync', '--nohooks')
  File "/y/src/depot_tools/fetch.py", line 75, in run_gclient
    return self.run(cmd_prefix + cmd, **kwargs)
  File "/y/src/depot_tools/fetch.py", line 65, in run
    return subprocess.check_call(cmd, **kwargs)
  File "/usr/lib/python2.7/subprocess.py", line 511, in check_call
    raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '('gclient', 'sync', '--nohooks')'
returned non-zero exit status 1



I reverted it in my depot_tools checkout and now am fetching successfully.

[szager1, 2014-02-26 23:57:42]

Thanks for the heads up, I reverted the change.

Stefan

On Wed, Feb 26, 2014 at 3:54 PM,  <mattm@chromium.org> wrote:
> On 2014/02/26 23:09:35, I haz the power (commit-bot) wrote:
>>
>> Change committed as 253616
>
>
> This change broke fetching a new checkout:
>
> mmueller2 15:48:34 /y/src/chrome$ fetch --nohooks chromium
> Running: svn ls --non-interactive svn://svn.chromium.org/chrome
> PRESUBMIT.py
> branches/
> codereview.settings
> releases/
> trunk/
> Running: git config --global core.deltaBaseCacheLimit 1G
> Running: gclient config --spec 'solutions = [{u'"'"'managed'"'"': False,
> u'"'"'name'"'"': u'"'"'src'"'"', u'"'"'url'"'"':
> u'"'"'https://chromium.googlesource.com/chromium/src.git'"'"',
> u'"'"'custom_deps'"'"': {}, u'"'"'deps_file'"'"': u'"'"'.DEPS.git'"'"',
> u'"'"'safesync_url'"'"': u'"'"''"'"'}]'
> Running: gclient sync --nohooks
> Traceback (most recent call last):
>   File "/y/src/depot_tools/gclient.py", line 1926, in <module>
>     sys.exit(Main(sys.argv[1:]))
>   File "/y/src/depot_tools/gclient.py", line 1916, in Main
>     return dispatcher.execute(OptionParser(), argv)
>   File "/y/src/depot_tools/subcommand.py", line 245, in execute
>     return command(parser, args[1:])
>   File "/y/src/depot_tools/gclient.py", line 1697, in CMDsync
>     ret = client.RunOnDeps('update', args)
>   File "/y/src/depot_tools/gclient.py", line 1239, in RunOnDeps
>     work_queue.flush(revision_overrides, command, args,
> options=self._options)
>   File "/y/src/depot_tools/gclient_utils.py", line 843, in run
>     self.item.run(*self.args, **self.kwargs)
>   File "/y/src/depot_tools/gclient.py", line 672, in run
>     file_list)
>   File "/y/src/depot_tools/gclient_scm.py", line 142, in RunCommand
>     return getattr(self, command)(options, args, file_list)
>   File "/y/src/depot_tools/gclient_scm.py", line 307, in update
>     self.UpdateSubmoduleConfig()
>   File "/y/src/depot_tools/gclient_scm.py", line 218, in
> UpdateSubmoduleConfig
>     gclient_utils.CheckCallAndFilter(cmd, **kwargs)
>   File "/y/src/depot_tools/gclient_utils.py", line 453, in
> CheckCallAndFilter
>     **kwargs)
>   File "/y/src/depot_tools/subprocess2.py", line 237, in __init__
>     super(Popen, self).__init__(args, **kwargs)
>   File "/usr/lib/python2.7/subprocess.py", line 679, in __init__
>     errread, errwrite)
>   File "/usr/lib/python2.7/subprocess.py", line 1249, in _execute_child
>     raise child_exception
> OSError: [Errno 2] No such file or directory: '/y/src/chrome/src'
> Sending crash report ...
>   args: ['/y/src/depot_tools/gclient.py', 'sync', '--nohooks']
>   cwd: /y/src/chrome
>   exception: [Errno 2] No such file or directory: '/y/src/chrom
>   host: mmueller2.mtv.corp.google.com
>   stack:   File "/y/src/depot_tools/gclient.py", line 1926,
>   user: mattm
>   version: 2.7.3 (default, Sep 26 2013, 20:03:06)
> [GCC 4.6.3
> A stack trace has been sent to the maintainers.
> Traceback (most recent call last):
>   File "/y/src/depot_tools/fetch.py", line 299, in <module>
>     sys.exit(main())
>   File "/y/src/depot_tools/fetch.py", line 295, in main
>     return run(options, spec, root)
>   File "/y/src/depot_tools/fetch.py", line 289, in run
>     return checkout.init()
>   File "/y/src/depot_tools/fetch.py", line 152, in init
>     super(GclientGitSvnCheckout, self).init()
>   File "/y/src/depot_tools/fetch.py", line 118, in init
>     self.run_gclient('sync', '--nohooks')
>   File "/y/src/depot_tools/fetch.py", line 75, in run_gclient
>     return self.run(cmd_prefix + cmd, **kwargs)
>   File "/y/src/depot_tools/fetch.py", line 65, in run
>     return subprocess.check_call(cmd, **kwargs)
>   File "/usr/lib/python2.7/subprocess.py", line 511, in check_call
>     raise CalledProcessError(retcode, cmd)
> subprocess.CalledProcessError: Command '('gclient', 'sync', '--nohooks')'
> returned non-zero exit status 1
>
>
>
> I reverted it in my depot_tools checkout and now am fetching successfully.
>
>
> https://codereview.chromium.org/164053003/

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.