When doing a 'git cl archive' if there are no branches other then master the
command will crash with a 'ValueError: Number of processes must be at least 1'
when attempting to create the thread pool.
This Cl modifies the code to use the maximum of the changes and 1 as the number
of threads to create so we have at least one thread to work with.
Review-Url: https://codereview.chromium.org/2228723003
if when execute command "git rev-parse --show-toplevel" in a directory
that is not a subdirectory of a git repo, CheckCallAndFilter will not
throw exception, so the top_dir may be updated with a invalid path.
BUG=633971
Signed-off-by: hanpfei <hanpfei@gmail.com>
Review-Url: https://codereview.chromium.org/2200193003
"gn format" changed to remove the --in-place argument. This updates the
instructions for formatting and fixes existing usage.
git cl format --diff never worked, it just printed the full formatted output
unconditionally. This change improves things slightly by indicating there's a
diff for a file. A proper diff can be hooked up in a later stage if somebody
so desires.
Review-Url: https://codereview.chromium.org/2220123002
It might be the case for some repositories that there is no merge base
between the current git HEAD and origin/master, which causes git
merge-base to exit with a non-zero code and cause calls to gclient
status/diff/pack to fail, as well as gclient sync if the repository in
question has been removed from DEPS.
This is true for the external/webrtc/trunk/talk repository, for example.
Its recent release branches (branch-heads/45 all the way to /53 at
least) have no ancestry shared with its master branch, so gclient
sync'ing from a Chromium M51 checkout to an M52 one (where it's no
longer in DEPS) fails because of the failed git merge-base calls.
We now ignore failures and just don't specify a merge base when calling
"git diff".
BUG=633962
R=iannucci@chromium.org,agable@chromium.org,maruel@chromium.org
Review-Url: https://codereview.chromium.org/2215673002
"git retry" assumes that depot_tools is in PATH; in some cases it isn't,
and there's currently no way to add it.
BUG=
Recipe-Tryjob-Bypass-Reason: Expectations known to change
Review-Url: https://codereview.chromium.org/2192163002
This is a workaround for the fact that 'gclient revert' calls
'gclient update' with the same `options` object, despite not
actually having the same set of command line options.
R=iannucci@chromium.org, kevcheng@chromium.org
BUG=631220
Review-Url: https://codereview.chromium.org/2184493002
If an exception occurs during a try block which prevents a step from actually executing, then the finally block's attempt to access step.active_result will fail, and the original exception will be hidden behind a new exception from the recipe engine. This CL changes this behavior so that the original exception will always fire.
Review-Url: https://codereview.chromium.org/2175103002
add infra_swarmbucket config for infra_paths module to override
git_cache. The value of path_config property value will be specified by
kitchen.
This will still delete cache every time. The goal of this change is to
fix build.git swarmbucket builds.
TBR=dnj@chromium.org
BUG=
Review-Url: https://codereview.chromium.org/2176913002
Reason for revert:
this is wrong. if directory is not empty, git clone fails
Original issue's description:
> bot_update: fix git dir check
>
> existance of a dir does not mean it is a git repo. Check .git dir
> presence instead.
>
> This is the reason why swarmbucket build.git builds fail.
>
> TBR=hinoka@chromium.org
> BUG=
>
> Committed: cdac80c9fbTBR=hinoka@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review-Url: https://codereview.chromium.org/2173683003
Since different projects can have different configurations for what
the maximum value of the "Code-Review" label is in Gerrit, this teaches
presubmit_support to inspect the maximum configured value and see who
has granted it (the same behavior as the Submit button in Gerrit itself).
R=andybons@chromium.org, martiniss@chromium.org, tandrii@chromium.org
BUG=630738
Review-Url: https://codereview.chromium.org/2178673002
That recipe no longer exists, as of
e0cc9e4949
Referring to it in the tutorial is misleading. It seems that the docs scripts
haven't been run in awhile, so there are some unrelated changes as well.
Review-Url: https://codereview.chromium.org/2162223003
Traceback (most recent call last):
...
File "...\\git_map_branches.py", line 303, in main
mapper.start()
File "...\\git_map_branches.py", line 169, in start
self.__append_branch(root)
File "...\\git_map_branches.py", line 272, in __append_branch
self.__append_branch(child, depth=depth + 1)
File "...\\git_map_branches.py", line 267, in __append_branch
line.append(run('log', '-n1', '--format=%s', branch))
File "...\git_common.py", line 599, in run
return run_with_stderr(*cmd, **kwargs)[0]
File "...\git_common.py", line 668, in run_with_stderr
raise subprocess2.CalledProcessError(retcode, cmd, os.getcwd(), ret, err)
subprocess2.CalledProcessError: Command '...\\git.bat -c color.ui=never log -n1 --format=%s readme' returned non-zero exit status 128 in ...
fatal: ambiguous argument 'readme': both revision and filename
Use '--' to separate paths from revisions, like this:
'git <command> [<revision>...] -- [<file>...]'
BUG=
Review-Url: https://codereview.chromium.org/2153803002
Chrome on iOS downstream repository tracks Chromium via DEPS and wants
to use cpplint canned presubmit check but the cpplint.py errors due at
the include guard as it stops at the inner most git repository.
Add a new parameter --project_root to cpplint.py. If set, it overrides
the automatic detection of the project root that searches the root of
the version control repository.
The --root parameter cannot be used as it is used after the automatic
resolution is performed (and allow chopping the head of the relative
path while the need is to expend the path to include ios_internal/).
BUG=598090
Review-Url: https://codereview.chromium.org/2036773002