git command to find a git revision corresponding to lkgr.

git command to find a git revision corresponding to lkgr.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=158892

[Jeffrey Yasskin, 2012-09-21 23:22:12]

I tried this and got:

$ git lkgr
fatal: ambiguous argument 'Partial-rebuilding': unknown revision or path not in
the working tree.
Use '--' to separate paths from revisions
fatal: bad revision '^1'
I am thoroughly confused.  Please send a bug report to
chrome-infrastructure-team.


$ git log
commit b329abd153855db3d2fd099a4b5cee60f2c663fa
    SVN changes up to revision 158098
...
$ curl -s http://chromium-status.appspot.com/lkgr
158017


The second run of `git lkgr` returns 79003d4eeb49f67b18255355878b6031f6e80c28.

[szager1, 2012-09-21 23:45:08]

On 2012/09/21 23:22:12, Jeffrey Yasskin wrote:
> I tried this and got:
> 
> $ git lkgr
> fatal: ambiguous argument 'Partial-rebuilding': unknown revision or path not
in
> the working tree.
> Use '--' to separate paths from revisions
> fatal: bad revision '^1'
> I am thoroughly confused.  Please send a bug report to
> chrome-infrastructure-team.
> 
> 
> $ git log
> commit b329abd153855db3d2fd099a4b5cee60f2c663fa
>     SVN changes up to revision 158098
> ...
> $ curl -s http://chromium-status.appspot.com/lkgr
> 158017
> 
> 
> The second run of `git lkgr` returns 79003d4eeb49f67b18255355878b6031f6e80c28.

That's weird.  It's possible that the git-svn cache was out-of-date, or
something like that.  I modified the script a bit for better error detection; I
*think* it would have caught your case, but it's hard to know for certain.

[Jeffrey Yasskin, 2012-09-22 00:16:06]

Thanks for writing this, btw.

https://codereview.chromium.org/10958056/diff/3001/git-lkgr
File git-lkgr (right):

https://codereview.chromium.org/10958056/diff/3001/git-lkgr#newcode9
git-lkgr:9: git_lkgr=`git svn find-rev r${svn_lkgr}`
I'm guessing a `git svn fetch` before this line would prevent the "Partial
rebuilding" error and make it more likely that the revision is found. You should
probably do a `git fetch` before that to make it run quickly. Alternately, do
those inside this `if` to speed up the common case.

https://codereview.chromium.org/10958056/diff/3001/git-lkgr#newcode15
git-lkgr:15: closest_commit=`git rev-list --ancestry-path --grep='SVN changes up
to revision [0-9]*' ${git_lkgr}..refs/remotes/origin/master | tail -1`
set pipefail to make the exit status of this command match the `git` exit status
instead of the `tail` exit status.

[szager1, 2012-09-22 00:54:47]

https://codereview.chromium.org/10958056/diff/3001/git-lkgr
File git-lkgr (right):

https://codereview.chromium.org/10958056/diff/3001/git-lkgr#newcode9
git-lkgr:9: git_lkgr=`git svn find-rev r${svn_lkgr}`
On 2012/09/22 00:16:06, Jeffrey Yasskin wrote:
> I'm guessing a `git svn fetch` before this line would prevent the "Partial
> rebuilding" error and make it more likely that the revision is found. You
should
> probably do a `git fetch` before that to make it run quickly. Alternately, do
> those inside this `if` to speed up the common case.

I'd prefer not run any `fetch` operations, to prevent any unintended side
effects.  I enhanced the error message to suggest running `git fetch`.  `git svn
fetch` will never do anything helpful for this script.

https://codereview.chromium.org/10958056/diff/3001/git-lkgr#newcode15
git-lkgr:15: closest_commit=`git rev-list --ancestry-path --grep='SVN changes up
to revision [0-9]*' ${git_lkgr}..refs/remotes/origin/master | tail -1`
On 2012/09/22 00:16:06, Jeffrey Yasskin wrote:
> set pipefail to make the exit status of this command match the `git` exit
status
> instead of the `tail` exit status.

Done.

[Jeffrey Yasskin, 2012-09-22 09:01:29]

lgtm

http://codereview.chromium.org/10958056/diff/8001/git-lkgr
File git-lkgr (right):

http://codereview.chromium.org/10958056/diff/8001/git-lkgr#newcode16
git-lkgr:16: set -o pipefail yes
Just "set -o pipefail".

[commit-bot: I haz the power, 2012-09-23 01:18:06]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/szager@chromium.org/10958056/4002

[commit-bot: I haz the power, 2012-09-23 01:20:11]

Presubmit check for 10958056-4002 failed and returned exit status 1.

Running presubmit commit checks ...
Checking out rietveld...
Running save-description-on-failure.sh
Running push-basic.sh
Running upstream.sh
Running submit-from-new-dir.sh
Running abandon.sh
Running submodule-merge-test.sh
Running upload-local-tracking-branch.sh
Running hooks.sh
Running post-dcommit-hook-test.sh
Running upload-stale.sh
Running patch.sh
Running basic.sh

** Presubmit ERRORS **
Missing LGTM from an OWNER for files in these directories:
    depot_tools

Presubmit checks took 124.0s to calculate.

[cmp, 2012-09-25 01:04:23]

http://codereview.chromium.org/10958056/diff/4002/git-lkgr
File git-lkgr (right):

http://codereview.chromium.org/10958056/diff/4002/git-lkgr#newcode1
git-lkgr:1: #!/bin/bash
need some lines here like:

# Copyright (c) 2012 The Chromium Authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.

http://codereview.chromium.org/10958056/diff/4002/git-lkgr#newcode17
git-lkgr:17: closest_commit=`git rev-list --ancestry-path --grep='SVN changes up
to revision [0-9]*' ${git_lkgr}..refs/remotes/origin/master | tail -1`
nit: wrap at 80 chars

http://codereview.chromium.org/10958056/diff/4002/git-lkgr#newcode19
git-lkgr:19: echo 'Could not find a blessed git commit (with accurate .DEPS.git
and submodules) after lkgr'
nit: wrap at 80 chars

http://codereview.chromium.org/10958056/diff/4002/git-lkgr#newcode25
git-lkgr:25: echo 'I am thoroughly confused.  Please send a bug report to
chrome-infrastructure-team.'
nit: wrap at 80 chars
let's use something like the following message:
  ... Please file a bug report at http://new.crbug.com/.

[cmp, 2012-09-25 01:05:52]

Admittedly I had something else in mind that involved setting up a src.git ref
on the server so that users could follow lkgr locally just like they follow
master.  It might still be worth setting up a lkgr ref (or a lkgr branch) but
just having this will do a lot to help almost all users.  Nice work.

lgtm with the nits I mentioned above.

[szager1, 2012-09-26 21:11:01]

http://codereview.chromium.org/10958056/diff/4002/git-lkgr
File git-lkgr (right):

http://codereview.chromium.org/10958056/diff/4002/git-lkgr#newcode1
git-lkgr:1: #!/bin/bash
On 2012/09/25 01:04:24, cmp wrote:
> need some lines here like:
> 
> # Copyright (c) 2012 The Chromium Authors. All rights reserved.
> # Use of this source code is governed by a BSD-style license that can be
> # found in the LICENSE file.

Done.

http://codereview.chromium.org/10958056/diff/4002/git-lkgr#newcode17
git-lkgr:17: closest_commit=`git rev-list --ancestry-path --grep='SVN changes up
to revision [0-9]*' ${git_lkgr}..refs/remotes/origin/master | tail -1`
On 2012/09/25 01:04:24, cmp wrote:
> nit: wrap at 80 chars

Done.

http://codereview.chromium.org/10958056/diff/4002/git-lkgr#newcode19
git-lkgr:19: echo 'Could not find a blessed git commit (with accurate .DEPS.git
and submodules) after lkgr'
On 2012/09/25 01:04:24, cmp wrote:
> nit: wrap at 80 chars

Done.

http://codereview.chromium.org/10958056/diff/4002/git-lkgr#newcode25
git-lkgr:25: echo 'I am thoroughly confused.  Please send a bug report to
chrome-infrastructure-team.'
On 2012/09/25 01:04:24, cmp wrote:
> nit: wrap at 80 chars
> let's use something like the following message:
>   ... Please file a bug report at http://new.crbug.com/.

Done.

[commit-bot: I haz the power, 2012-09-26 21:11:16]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/szager@chromium.org/10958056/14001

[commit-bot: I haz the power, 2012-09-26 21:16:24]

Change committed as 158892