Add tryserver.api.gerrit_change_info that returns detailed info about the
current Gerrit CL. It makes a Gerrit RPC in the background on the first read.
This step will be present in all tryjobs.
This mechanism can be used for extra details about the current CL.
This CL starts to use this in bot_update for gerrit_ref, and stops using
patch_ref property, and also update bot_update to use when the destination
branch is needed.
Also this CL removes no longer used properties in bot_update,
namely most of patch_* properties.
build.git manual roll: https://chromium-review.googlesource.com/c/chromium/tools/build/+/1217626
Recipe-Manual-Change: build
Recipe-Nontrivial-Roll: infra
Recipe-Nontrivial-Roll: skia
Recipe-Nontrivial-Roll: skiabuildbot
Recipe-Nontrivial-Roll: build_limited_scripts_slave
Recipe-Nontrivial-Roll: release_scripts
R=tandrii@chromium.org
Bug: 694348
Change-Id: I945d061b74278c3d1d227578abe16a64cdaca647
Reviewed-on: https://chromium-review.googlesource.com/1213557
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
CrOS chromium-pfq builder uses bot_update, but checks out Chromium on its own,
effectively ignoring input commit in bot_update.
To resolve a P0, add ignore_input_commit parameter.
Long term, bot_update API must be redesigned.
TBR=tandrii@chromium.org
Bug: 882094
Change-Id: I4f43348e3d5f81bf603dd2c110829bce58acad16
Reviewed-on: https://chromium-review.googlesource.com/1216787
Reviewed-by: Lann Martin <lannm@chromium.org>
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
tryserver module's purpose is to provide high level API for patches.
Update it to use buildbucket.build.input.gerrit_changes[0].
Add helper properties for gerrit repo URL and gerrit ref.
This will be used by both bot_update and gclient.
R=hinoka@chromium.org, tandrii@chromium.org
Bug: 694348
Change-Id: I7eb416e75737310404c4354fa680fd3499eddaa7
Reviewed-on: https://chromium-review.googlesource.com/1211653
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Ryan Tseng <hinoka@chromium.org>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
Use self.m.buildbucket.build.input.gitiles_commit.id instead.
It encapsulates revision and parent_got_revision properties.
Also, honor host and project of the specified commit.
This enables overriding non-first gclient solution at the build-level
for arbitrary buildbucket builds.
Bug: 877161
Change-Id: I3f92f04ca84a1d30cbbb2670f62bea2bd9ee5932
Reviewed-on: https://chromium-review.googlesource.com/1200066
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
Use repo_path_map in all places where patch_projects is currently used.
Add gclient.get_repo_path function to resolve path to a repo given its URL.
infra.git requires a non-trivial roll because of incorrect expectations.
Recipe-Nontrivial-Roll: build
Recipe-Nontrivial-Roll: infra
Bug: 877161
Change-Id: Iaea0a3f14609bf0c18d4ecdf8564153b8052584d
Reviewed-on: https://chromium-review.googlesource.com/1208363
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
When checking if buildbucket.build_input.gitiles_commit exists, use
HasField protobuf method, as opposed to truth check. A protobuf message
field is always true.
Bug: 877161
Change-Id: Ie4a7e22b985776441a74283b09646e8837c326f1
Reviewed-on: https://chromium-review.googlesource.com/1196930
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
This is in order to use built-in presubit recipe.
R=maruel
Change-Id: I3e50af7dc3a9ad83f0f358285395794ea599a4e8
Reviewed-on: https://chromium-review.googlesource.com/1176650
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
Commit-Queue: Andrii Shyshkalov <tandrii@chromium.org>
Also add shortcuts, so we don't have to look through all the branches
if the patch is based on a common branch (i.e. master, infra/config, lkgr)
Bug: 870279
Change-Id: I625a8481dccac9a475b096b926e6fab7efe676b0
Reviewed-on: https://chromium-review.googlesource.com/1161094
Commit-Queue: Edward Lesmes <ehmaldonado@chromium.org>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
This is an automated CL created by the recipe roller. This CL rolls recipe
changes from upstream projects (e.g. depot_tools) into downstream projects
(e.g. tools/build).
Please review the expectation changes, and LGTM+CQ.
More info is at https://goo.gl/zkKdpD. Use https://goo.gl/noib3a to file a bug.
recipe_engine:
https://crrev.com/bd5111dec043a367d69d85b81289e2207aa26f6e [cipd] Clean docs, pylint errors, add pkg-fetch, pkg-deploy. (iannucci@chromium.org)
R=tandrii@chromium.org
Recipe-Tryjob-Bypass-Reason: Autoroller
Bugdroid-Send-Email: False
Change-Id: I19d44529f7caa41c9577e57f6a9a35815627c4c3
Reviewed-on: https://chromium-review.googlesource.com/1170243
Commit-Queue: Recipe Roller <recipe-roller@chromium.org>
Commit-Queue: Robbie Iannucci <iannucci@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
When syncing remote branches, checkout origin/<branch> instead of
<branch>, since there might be a local branch with that name.
Bug: 859703
Change-Id: I9a49b940f53f75a97e5458fdcf578c3b52b26de2
Reviewed-on: https://chromium-review.googlesource.com/1128271
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
Commit-Queue: Edward Lesmes <ehmaldonado@chromium.org>
Today, whenever LUCI scheduler sets gitiles_commit, a revision property
is also always set, thus this CL will have no effect on today's production.
The goal of CL is to pave the way to avoid this duplication. For starters,
one could start real writing tests with
+ api.buildbucket.ci_build(..., revision='sha1')
without the need to duplicate the same revision as a property. All
recipes which WAI with such input are thus ready to work w/o 'revision'
property. After that, we'll be able to stop the duplication in LUCI scheduler.
R=hinoka
Recipe-Nontrivial-Roll: infra
Recipe-Nontrivial-Roll: build
Change-Id: I1773b1c3d2a98a9dbd1cfdd9ac2b4fb99745ee99
Reviewed-on: https://chromium-review.googlesource.com/1123713
Commit-Queue: Andrii Shyshkalov <tandrii@chromium.org>
Reviewed-by: Ryan Tseng <hinoka@chromium.org>
It is not possible to use cwd argument, since none of the methods called by this
module define cwd explicitly and low-level modules like api.step do not support
it anymore, resulting in a TypeError exception thrown due to unexpected keyword
argument, e.g. see https://ci.chromium.org/swarming/task/3e74a52905a71110. Given
that any users using cwd argument would be broken anyway, it should be safe to
make this change.
Planned usage: https://crrev.com/c/1122874
R=tandrii@chromium.org, iannucci@chromium.org
Bug: 793687
Change-Id: Iadf6bf91d8d41dbc6751a848bb69d7922a206ade
Reviewed-on: https://chromium-review.googlesource.com/1122873
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Commit-Queue: Sergiy Byelozyorov <sergiyb@chromium.org>
gclient should be the one syncing the solutions to the appropriate
revisions.
Bug: 850812, 853032
Change-Id: Ieefc5661627d4864deb0d4e7053168a99da29d29
Reviewed-on: https://chromium-review.googlesource.com/1102833
Reviewed-by: Aaron Gable <agable@chromium.org>
Commit-Queue: Edward Lesmes <ehmaldonado@chromium.org>
shallow mode is a special git mode that only checks out a shallow copy
of git so the full history is not downloaded. It is not well supported
by googlesource.com servers because git has to do a giant negotiation
with the server and transmit every single commit that the local checkout
has or doesn't have (instead of just the latest commit of each ref),
which ends up taking more than 10x longer than a non-shallow checkout.
R=ehmaldonado, hinoka, iannucci
Recipe-Nontrivial-Roll: infra
Recipe-Nontrivial-Roll: build
Recipe-Nontrivial-Roll: build_limited_scripts_slave
Bug: 855137
Change-Id: I5f9e31f8b2730cf80b79bfd0e08201c33eb942f5
Reviewed-on: https://chromium-review.googlesource.com/1110450
Commit-Queue: Andrii Shyshkalov <tandrii@chromium.org>
Reviewed-by: Ryan Tseng <hinoka@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Allows easily getting at more than just the commit message.
I intend to look at the author to whitelist bots from binary size
checks (initially).
Bug: 702625
Change-Id: I684fdfacc88635ec4849216c74183f63f9d54810
Reviewed-on: https://chromium-review.googlesource.com/1070483
Commit-Queue: agrieve <agrieve@chromium.org>
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
It fails when both the revision and the url are None.
Bug: 841936
Change-Id: Idef45a015624a92226d4ccd38ed5b978bf786993
Reviewed-on: https://chromium-review.googlesource.com/1053996
Commit-Queue: Edward Lesmes <ehmaldonado@chromium.org>
Reviewed-by: Ryan Tseng <hinoka@chromium.org>
Nodir Turakulov
Nodir Turakulov
Corentin Wallez
Edward Lemur
John Budorick
Ben Pastene
Andrii Shyshkalov
recipe-roller
Robert Iannucci
Ryan Tseng
Michael Achenbach
Sergiy Byelozyorov
Andrew Grieve