I just made a bunch of branches disappear from `git map-branches` by
creating a big cycle (none of them were a root, thus none was listed at
all). This CL makes `git map-branches` check for such cycles and report
them as warnings.
R=iannucci@chromium.org
Change-Id: I723b327282a1c90bc9da5438777653e48f8bfd1d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1528985
Auto-Submit: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Commit-Queue: Robbie Iannucci <iannucci@chromium.org>
Using just @{u} to navigate upstream breaks navigating back down when
the upstream reaches a non-local branch, e.g. origin/master. This is
because git branch walks the reflog for detached heads, and sees @{u}
instead of e.g. origin/master, and so tries to calculate HEAD@{u}
which fails. See also the message send to the git mailing list:
https://public-inbox.org/git/CAGRskv__Qe-F-grBtq5v-uhHE+BQAcAcOQtX-GKB+MNOrxs0hQ@mail.gmail.com/
This patch works around this by manually specifying the branch from
which we are calculating upstream, e.g. for branch foo we do
git checkout foo@{u} instead of just git checkout @{u}. This ensures
that the reflog has a valid entry (albeit still a relative one).
Change-Id: I339d4e4288a1e104b9d7d3d568c45b2b156f0689
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1534217
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Commit-Queue: Robbie Iannucci <iannucci@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Writing to ninjalog.cfg comes with the risk of failure (Ctrl+C at the
wrong time) which can lead to an empty or corrupted file. This will
cause errors on subsequent builds when loading the config file, and
will lose the users opt-in/out settings.
This change ensures that the config file is only written when something
has changed, thus essentially avoiding the risk.
Bug: 941810
Change-Id: If266f07536569c9ac91ee37c28797c8b2ccc6fdc
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1522291
Reviewed-by: Takuto Ikuta <tikuta@chromium.org>
Reviewed-by: Fumitoshi Ukai <ukai@chromium.org>
Commit-Queue: Bruce Dawson <brucedawson@chromium.org>
This reverts commit 2d29261e94.
Reason for revert: Both jbudorick and I asked you not to land this today :(.
Original change's description:
> Remove exception for relative CIPD
>
> This seems to no longer be necessary, as relative paths 'just work' for
> CIPD deps now.
>
> Change-Id: I3843548dc74d29af90e0c402f0459dbe60934423
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1529132
> Reviewed-by: Dirk Pranke <dpranke@chromium.org>
> Reviewed-by: John Budorick <jbudorick@chromium.org>
> Commit-Queue: Brandon Tolsch <btolsch@chromium.org>
TBR=dpranke@chromium.org,jbudorick@chromium.org,btolsch@chromium.org
Change-Id: I1d9a771ef34dae0dec15499d7cf28e4ba5daf2ff
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1531193
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Dirk Pranke <dpranke@chromium.org>
This seems to no longer be necessary, as relative paths 'just work' for
CIPD deps now.
Change-Id: I3843548dc74d29af90e0c402f0459dbe60934423
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1529132
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Reviewed-by: John Budorick <jbudorick@chromium.org>
Commit-Queue: Brandon Tolsch <btolsch@chromium.org>
The scripts still work with Python 2.
There are no intended behaviour changes.
Bug: 939847
Change-Id: Icada60c5b2cf351d62aead26b7364fcef2c2a3e5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1524486
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Raul Tambre <raul@tambre.ee>
Needed by modules using DIA SDK, which is located in "[windows_sdk]/DIA SDK",
not under "[windows_sdk]/win_sdk".
Change-Id: I1ca42a38ef8d9d1cbfde5e63e786b7d792504f3c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1526453
Commit-Queue: Robbie Iannucci <iannucci@chromium.org>
Auto-Submit: Veranika Liaukevich <veranika@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Instead of creating an empty dict and setting all the required git variables
there, add those variables to os.environ so that we do not drop other
required variables -- in my case, git was not in /usr/bin so invoking 'git
commit' in _add_schema_commit fails with "no such file or directory".
Change-Id: I8c9b14e1f534efc8f40d21b312a5396b2f65f4a1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1528290
Auto-Submit: Raphael Kubo da Costa <raphael.kubo.da.costa@intel.com>
Commit-Queue: Robbie Iannucci <iannucci@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
It breaks gclient for some users depending on locale settings (see bug).
The feature just doesn't seem worth the complexity.
Bug: 934279
Change-Id: I2c6e2b77ab764381db9f2c4e63ff8a84526769bb
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1520646
Reviewed-by: Elliott Friedman <friedman@chromium.org>
Reviewed-by: Ryan Tseng <hinoka@chromium.org>
Commit-Queue: Ryan Tseng <hinoka@chromium.org>
Auto-Submit: Hans Wennborg <hans@chromium.org>
The function we use with gclient_utils has moved to gclient_paths.
However, I am not sure gclient_utils is used by others or not.
Additional change:
- sorted testMembersChanged members for ease of understanding what is
different.
Bug: 939959
Change-Id: I8b822e26afc73b39d2a33c5e07779c9c14cbd394
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1521850
Commit-Queue: Yoshisato Yanagisawa <yyanagisawa@chromium.org>
Reviewed-by: Nico Weber <thakis@chromium.org>
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Reviewed-by: Fumitoshi Ukai <ukai@chromium.org>
Warning message appears when trying to cherry-pick into a Chromium branch,
but doesn't actually mean anything right now.
"""
About to land on 3163. Continue (y/n)? y
It seems this repository has a Commit Queue, which can test and land changes for you.
Are you sure you wish to bypass it?
Press Enter to bypass CQ, or Ctrl+C to abort
"""
We want to disable until CQ for branches is fully developed.
Bug: 753213
Change-Id: I1cdfa084a6bad1be56b85a56e94e2dd27a099b2a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1518291
Commit-Queue: Karen Qian <karenqian@google.com>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
I've been using this recipe module for a new internal builder
(http://shortn/_wvjqgrpIGb) and it seems odd to have the package
name and instance ID in the step text rather than a link directly
to the site. (Perhaps the module predates the CIPD UI rewrite?)
Recipe-Nontrivial-Roll: infra
Change-Id: Iab6080c3c6110ac2205cb3de397055b30d881d62
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1496953
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Commit-Queue: John Budorick <jbudorick@chromium.org>
sys takes essentially no time to import (I think possibly because
it's a built-in?), and importing it as-needed is fragile. Make
it a top-level import to gclient_paths to prevent future breakages.
BUG=939959
Change-Id: I1ebe1260afd927355d668acf3cab5067ff0a2e03
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1515669
Reviewed-by: Nico Weber <thakis@chromium.org>
Commit-Queue: Dirk Pranke <dpranke@chromium.org>
Trying to get "host_cpu" in DEPS file may fail on AIX.
Getting the "machine ID number" using "uname -m" on an AIX box, in an LPAR
environment will not generate a unique machine identifier and instead might
return a string such as "00f9445f4c00". "platform.processor()" will be used
as a fall back on AIX.
Change-Id: I7fada10059e29066f5a13d6135b01eaeaccb769d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1504554
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Commit-Queue: Dirk Pranke <dpranke@chromium.org>
gclient_utils.py is a kitchen sink and is for that reason expensive
to import. Move the comparatively cheap and simple path routines
to a new gclient_paths module and use that in gn.py, clang_format.py,
dart_format.py.
(To be able to move FindGclientRoot() to gclient_paths.py,
make it use io.open() instead of FileRead(). FileRead() tries
to paper over invalid utf-8, but that was added for presubmits,
not for .gclient files, so this is hopefully fine.)
Cuts gn.py overhead in half (on my Windows laptop from 0.6s to 0.25s,
still high; on my Mac laptop from 0.1s to 0.05s), and probably helps
the other two too.
Completely remove PathDifference() since it's unused.
Bug: 939959
Change-Id: I6a70f6e4c16062b622fb2df8778e8a598d4cc956
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1512058
Commit-Queue: Nico Weber <thakis@chromium.org>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
If we already have the correct commit, it is not necessary to fetch. Any
fetches can takes minutes due to gerrit problems, and often fail
completely, so it is helpful to avoid them completely whenever we can.
Reland now that skia autoroller does its own fetch.
Bug: 938627
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1504102
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
Change-Id: I5a6877125c682ff0d98bd6d2ffe495245f74c513
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1512024
Auto-Submit: Michael Spang <spang@chromium.org>
Commit-Queue: Edward Lesmes <ehmaldonado@chromium.org>
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
We should pass --verify to tell rev-parse that we're just testing,
and use the ^{commit} dereference operator to avoid the hack of
removing a character from the commit hash to see if it's really
in the object database.
Bug: 938627
Change-Id: Ic6ea898b0a5a6a1a5d706c7586c7208ec8ca2ce2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1504104
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
This reverts commit 61a1b5f310.
Reason for revert:
It's blocking skia autoroller
https://skia-review.googlesource.com/c/buildbot/+/198700
Original change's description:
> Avoid unnecessary invocations of "git fetch"
>
> If we already have the correct commit, it is not necessary to fetch. Any
> fetches can takes minutes due to gerrit problems, and often fail
> completely, so it is helpful to avoid them completely whenever we can.
>
> Bug: 938627
>
> Change-Id: I54494cad9b9aa0090f51674b6b9f56bb3e842f12
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1504102
> Commit-Queue: Michael Spang <spang@chromium.org>
> Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
TBR=spang@chromium.org,agable@chromium.org,ehmaldonado@chromium.org
Change-Id: If8bd26d74bf84ff0ca69378dbfe6b9369b9988d3
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 938627
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/1508980
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
Commit-Queue: Edward Lesmes <ehmaldonado@chromium.org>
Clemens Hammacher
Leszek Swirski
btolsch
recipe-roller
Bruce Dawson
Vadim Shtayura
Dirk Pranke
Takuto Ikuta
Raul Tambre
Veranika Liaukevich
Raphael Kubo da Costa
Hans Wennborg
Yoshisato Yanagisawa
Karen Qian
John Budorick
Lutz Justen
recipe-roller
Milad Farazmand
Nico Weber
Michael Spang
Edward Lesmes