Git has been using version 2 of their wire protocol by default since
2.26, released in 2020. Stop setting it explicitely.
Change-Id: Ibc93a672b41c5a8183bc3200c506006f6e08af94
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6070782
Commit-Queue: Josip Sokcevic <sokcevic@chromium.org>
Reviewed-by: Josip Sokcevic <sokcevic@chromium.org>
May help user to diagnose/debug some issues.
Bug: b/372552465
Change-Id: Ia9fc4fd49c8ee80bccd98442ed5589324d0b23b9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6068966
Reviewed-by: Gavin Mak <gavinmak@google.com>
Auto-Submit: Allen Li <ayatane@chromium.org>
Commit-Queue: Allen Li <ayatane@chromium.org>
Modifying the system-level Git config may require admin privileges,
depending on where Git was installed.
This CL makes the bootstrap.py script target a user's Git config
at the global level, instead of system level.
Checks the global Git config for postprocess version and explicit
enabling of global config changing. Logs a warning if out of date
and not enabled so the user knows how to agree to global config
changes.
Bug: b/379888465, b/360206460
Change-Id: I9be395b64d5429f85514397918f4a6140dfe2a1a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6034204
Reviewed-by: Allen Li <ayatane@chromium.org>
Reviewed-by: Josip Sokcevic <sokcevic@chromium.org>
Commit-Queue: Anne Redulla <aredulla@google.com>
Gerrit allows multiline footers with indents.
e.g.,
Test: something
looks great
This CL fixes the bug such that add_footer correctly inserts
a given message after multiline footers.
Bug: 379923433
Change-Id: I9b3f793095b63b0586123543a2f8d49f0503fca0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6064963
Reviewed-by: Josip Sokcevic <sokcevic@chromium.org>
Commit-Queue: Scott Lee <ddoman@chromium.org>
The param before_key used to be used in presubmit_support.py, but
it was removed in 2018, and has never been used anywhere since.
- https://crrev.com/c/1351509
- http://shortn/_YQUh4yC9Tc
This CL removes it to simplify the code before the next CL.
It hasn't been used for several years and will unlikely be used
in the future.
Bug: 379923433
Change-Id: Ib0da340235a991b6d8edca041eae811540d4aa60
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6064306
Commit-Queue: Scott Lee <ddoman@chromium.org>
Reviewed-by: Josip Sokcevic <sokcevic@chromium.org>
This CL adds a hint on how to log into GCS to the 401 error message
displayed to the user.
Bug: 381634957
Change-Id: Id72b5030c018b28a7297b7ff3f279485c6da4d9e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6056844
Reviewed-by: Gavin Mak <gavinmak@google.com>
Commit-Queue: Alan Cutter <alancutter@chromium.org>
crrev.com/c/6049353 updated bot_update.ensure_checkout to comput the
output commit even when not setting the output commit so that callers
could do custom logic for setting output commits while still having
access to the one computed by bot_update. Unfortunately, this changed
behavior for some builds that have set_output_commit set false. When
computing the output commit, if none of the conditions that set the ref
on the output commit were true then an AssertionError would be raised
indicating that the caller should call
buildbucket.set_output_gitiles_commit. With crrev.com/c/6049353, the
assertion can be raised even if set_output_commit is false.
This changes it so that the constraint on the ref field is only enforced
when set_output_commit is true.
Bug: 366409421
Change-Id: Ifb3a39784d75a3c47a7d4a6e2ac9957c29a09938
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6064719
Auto-Submit: Garrett Beaty <gbeaty@google.com>
Reviewed-by: Yiwei Zhang <yiwzhang@google.com>
Commit-Queue: Yiwei Zhang <yiwzhang@google.com>
Getting a lot of "(401) Unauthorized" responses for these now. This at
least makes the script not crash.
Change-Id: Ib42ccbdb3326a0918015e9d5e1cc9392a9c1a469
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6049877
Reviewed-by: Gavin Mak <gavinmak@google.com>
Commit-Queue: Andrew Grieve <agrieve@chromium.org>
Auto-Submit: Andrew Grieve <agrieve@chromium.org>
These were introduced in crrev.com/c/6024739 and currently break
autoninja for non-Googlers who manually enable Siso by setting the
use_siso=true GN arg.
$ gn gen --args='use_siso=true' out/Debug
$ autoninja -C out/Debug base
Traceback (most recent call last):
File "/home/philwo/depot_tools/autoninja.py", line 504, in <module>
sys.exit(main(sys.argv))
^^^^^^^^^^^^^^
File "/home/philwo/depot_tools/autoninja.py", line 493, in main
exit_code = _main_inner(input_args, build_id, should_collect_logs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/philwo/depot_tools/autoninja.py", line 249, in _main_inner
assert use_siso != False
^^^^^^^^^^^^^^^^^
AssertionError
R=ukai@google.com
Change-Id: I6536576d39b4ef094daa4e7ec80f3d743ecefd85
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6058992
Auto-Submit: Philipp Wollermann <philwo@google.com>
Reviewed-by: Fumitoshi Ukai <ukai@google.com>
Commit-Queue: Fumitoshi Ukai <ukai@google.com>
If use_reclient_value.py is not found and args.gn only has
use_remotexec=true, the autoninja fails to start reproxy.
Change-Id: I6608ea48b66864bf2ab89176b6b01a79e07b7564
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6059221
Reviewed-by: Fumitoshi Ukai <ukai@google.com>
Commit-Queue: Aleksei Khoroshilov <akhoroshilov@brave.com>
https://crrev.com/c/6048393 provides build/toolchain/use_reclient_value.py
to get value for use_reclient.
Use it in autoninja if it exists.
Bug: 341167943
Change-Id: Ia6207ea044cc109fa156deff3c3357475368576f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6024739
Commit-Queue: Fumitoshi Ukai <ukai@google.com>
Auto-Submit: Fumitoshi Ukai <ukai@google.com>
Reviewed-by: Takuto Ikuta <tikuta@chromium.org>
Reviewed-by: Junji Watanabe <jwata@google.com>
out_commit is parsed from got_revision even when set_output_commit is False. This allows for the calling code to implement its own logic for setting output commit, but still use the commit from bot_update if needed.
Example of usage: https://crrev.com/c/6049557.
Bug: 366409421
Change-Id: I49ce89730fe780d7561077f82797b30fad87e65e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6049353
Reviewed-by: Garrett Beaty <gbeaty@google.com>
Commit-Queue: Alex Kravchuk <alexanderkr@google.com>
Reviewed-by: Yiwei Zhang <yiwzhang@google.com>
This reverts commit 66b3972fc5.
Reason for revert: This won't work because chromium_checkout relies on the position being set in the output commit in order to set information for RDB. We'll have to come up with a different solution.
Original change's description:
> Add parse_commit_position parameter to BotUpdateApi.ensure_checkout.
>
> When true and got_revision_cp is present, output commit ref and position are set from got_revision_cp. Default value is true.
>
> For Chrome builds that build from tags we will set parse_commit_position to false to show tags instead of branches on Buildbucket UI, e.g. refs/tags/132.0.6824.0 instead of refs/branch-heads/6824@{#1}.
>
> Bug: 366409421
> Change-Id: I77d01615edb6b791445a06469f80c673c97ad8d6
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6000539
> Reviewed-by: Garrett Beaty <gbeaty@google.com>
> Reviewed-by: Scott Lee <ddoman@chromium.org>
> Commit-Queue: Alex Kravchuk <alexanderkr@google.com>
Bug: 366409421
Change-Id: I21a5fc6e011a64fffd1d2ef0689d5a3f3843482c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6037603
Reviewed-by: Garrett Beaty <gbeaty@google.com>
Commit-Queue: Alex Kravchuk <alexanderkr@google.com>
Reviewed-by: Scott Lee <ddoman@chromium.org>
gn finds // by searching .gn in parent directories.
To support out dir is not 2 directories from root (i.e. out/Default),
search .gn in parent directory for '//' location.
tools/licenses/licenses.py runs 'gn gen' in out/Default/$tmp.
for such case, '//' should be '../../..', or fail to import '//path'
Change-Id: I6e5ccbe93cb5e51704f31d4eb558c03560286865
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6051636
Reviewed-by: Philipp Wollermann <philwo@google.com>
Commit-Queue: Fumitoshi Ukai <ukai@google.com>
Reviewed-by: Takuto Ikuta <tikuta@chromium.org>
This reverts commit 9dd0755b69.
Reason for revert: 'Supplement:Apache-2.0' is considered to be allowed, but is not technically a valid SPDX identifier.
Original change's description:
> Adding Supplement:Apache-2.0 to chromium license allowlist.
>
> The 'Supplement:Apache-2.0' contains the full license text and is the preferred version over just 'Apache-2.0'.
>
> Change-Id: Ie4566ffdcd20771a0170190270d8d45e1b77980d
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6040532
> Reviewed-by: Jordan Brown <rop@google.com>
> Reviewed-by: Rachael Newitt <renewitt@google.com>
> Commit-Queue: Rachael Newitt <renewitt@google.com>
> Auto-Submit: Jordan Brown <rop@google.com>
Change-Id: Ic6529514304b826dbc8840a468bab47321c30471
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6040754
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Auto-Submit: Jordan Brown <rop@google.com>
The 'Supplement:Apache-2.0' contains the full license text and is the preferred version over just 'Apache-2.0'.
Change-Id: Ie4566ffdcd20771a0170190270d8d45e1b77980d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6040532
Reviewed-by: Jordan Brown <rop@google.com>
Reviewed-by: Rachael Newitt <renewitt@google.com>
Commit-Queue: Rachael Newitt <renewitt@google.com>
Auto-Submit: Jordan Brown <rop@google.com>
This expands the names of OSes in a .gclient files target_os list that
gclient will recognize. This should prevent someone from putting
"windows" in their .gclient file's target_os list with no effect. (This
keeps happening to me.)
Bug: None
Change-Id: I7051caacb7fbe769bbc2d1c25c179d6335aa924e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6033012
Reviewed-by: Josip Sokcevic <sokcevic@chromium.org>
Reviewed-by: Dirk Pranke <dpranke@google.com>
Reviewed-by: Junji Watanabe <jwata@google.com>
Commit-Queue: Ben Pastene <bpastene@chromium.org>
By default it's false and the bundled git is used.
Bug: b/379888465
Change-Id: I7d3d91da2ade6534d5b80479a050d6f608c16b2e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6033861
Reviewed-by: Josip Sokcevic <sokcevic@chromium.org>
Auto-Submit: Anne Redulla <aredulla@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@chromium.org>
Rename parse_commit_position parameter to parse_commit_position_for_tags. If input commit ref is not a tag, or parse_commit_position_for_tags is True, execute the old code path. This makes the changes backward compatible with the existing code.
Bug: 366409421
Change-Id: I58405325e406a82c2e255fe5d3d4a2883c98a84a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6035030
Reviewed-by: Scott Lee <ddoman@chromium.org>
Commit-Queue: Alex Kravchuk <alexanderkr@google.com>
Reviewed-by: Garrett Beaty <gbeaty@google.com>
The flag to filter for branches was renamed in git 2.46.0 [1] from
`-h` to `-b`. This causes previous versions of git to fail when running
ls-remote.
To be compatible across a version upgrade, a second git call would be
required to retrieve the version and determine the correct flag.
However, the filter feature is not used yet and can also be solved by
specifying the absolute ref path within the recipe. Therefore, this
change removes ls-remote's filter flags.
[1] https://git-scm.com/docs/git-ls-remote/2.42.0
R=machenbach@chromium.org, gavinmak@google.com
Bug: none
Change-Id: I840fb3ac56daf47c798d5c32f00de841b8db2771
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6028418
Auto-Submit: Alexander Schulze <alexschulze@chromium.org>
Reviewed-by: Gavin Mak <gavinmak@google.com>
Commit-Queue: Gavin Mak <gavinmak@google.com>
Call `git ls-remote` to determine the head revision for a branch, or
the revision for a tag. The method is useful for bots which do not
require a checkout.
Bug: none
Change-Id: Ibc01d1fb9c2714bd4d719159dbdaf3d7684ea1e9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6012563
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Alexander Schulze <alexschulze@chromium.org>
Reviewed-by: Gavin Mak <gavinmak@google.com>
When true and got_revision_cp is present, output commit ref and position are set from got_revision_cp. Default value is true.
For Chrome builds that build from tags we will set parse_commit_position to false to show tags instead of branches on Buildbucket UI, e.g. refs/tags/132.0.6824.0 instead of refs/branch-heads/6824@{#1}.
Bug: 366409421
Change-Id: I77d01615edb6b791445a06469f80c673c97ad8d6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/6000539
Reviewed-by: Garrett Beaty <gbeaty@google.com>
Reviewed-by: Scott Lee <ddoman@chromium.org>
Commit-Queue: Alex Kravchuk <alexanderkr@google.com>
Henrique Ferreiro
recipe-roller
Allen Li
Anne Redulla
Scott Lee
Alan Cutter
Garrett Beaty
Will Yeager
Yiwei Zhang
Andrew Grieve
Philipp Wollermann
Aleksey Khoroshilov
Fumitoshi Ukai
Gary Tong
Alex Kravchuk
Gavin Mak
Jordan Brown
Ben Pastene
Junji Watanabe
Congbin Guo
Alex Schulze