set_failure_hash assumes that (StepFailure is raised) => (some steps
were ran), which is not a correct implication.
Check that there was a step before.
R=iannucci@chromium.org
BUG=
Change-Id: I36ad187584cb42696676e3339f6a77fc77dfd3b1
Reviewed-on: https://chromium-review.googlesource.com/412994
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Currently if path_config is not specified, infra_paths
defaults to buildbot paths. We want to make this behavior explicit,
so we don't have to introduce a new "luci" path configuration which
tells infra_paths module to do nothing.
Buildbot now specifies path_config=buildbot since
https://chromium-review.googlesource.com/c/412029/ ,
so now we can remove this default.
This also affects all expectation files because this CL switches
the default path configuration from Buildbot to LUCI.
In particular, it changes default git cache dir from [GIT_CACHE]
to [CACHE]/git.
Also code paths that work only in Buildbot environment has to
have a test with path_config=buildbot.
R=iannucci@chromium.org, martiniss@chromium.org
BUG=662586
Change-Id: Ifeabe13e69810dd24a6deb4830699f8c9fafafae
Reviewed-on: https://chromium-review.googlesource.com/411989
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
We cannot easily move git cache location on all buildbot bots, but we
need to define a new value for LUCI.
Introduce infra_paths.default_git_cache_dir which must be used instead
of using path['git_cache'] directly.
R=iannucci@chromium.org, martiniss@chromium.org
BUG=662586
Change-Id: I5244d7a761da94858df798abbd423c85f55936af
Reviewed-on: https://chromium-review.googlesource.com/411993
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Update recipe modules that use path "build" to either check whether it
is defined or fail with an explanation if it is not defined.
This CL marks failure code paths as `# pragma: no cover`. This comment
will be removed in https://chromium-review.googlesource.com/c/411989/
R=iannucci@chromium.org, martiniss@chromium.org
BUG=662586
Change-Id: Ieb7637deefa6e366dfe4c30c7711d60daa06575a
Reviewed-on: https://chromium-review.googlesource.com/412225
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Use depot_tools module's methods or package_repo_resource function to
determine paths of tools that live in depot_tools
R=iannucci@chromium.org, martiniss@chromium.org
BUG=662586
Change-Id: Ic3172308c140f7f19e982ec6ec5a4b1b83261e1b
Reviewed-on: https://chromium-review.googlesource.com/411936
Reviewed-by: Erik Staab <estaab@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
slave_build path, defined by infra_paths module, is unnecessarily tied
to our buildbot setup which we are trying to deprecate in favor of LUCI.
In practice it is always == start_dir and that's what we want in LUCI too.
Update all code that uses slave_build path to use start_dir instead.
R=iannucci@chromium.org, martiniss@chromium.org
BUG=662586
Change-Id: Id52aa535d6a4f52670d581bf56aab65f670b1861
Reviewed-on: https://chromium-review.googlesource.com/412075
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Reviewed-by: Stephen Martinis <martiniss@chromium.org>
CURRENT_WORKING_DIR in path module was renamed to START_DIR. Update
infra_paths accordingly.
This includes recipe_engine roll.
R=iannucci@chromium.org, martiniss@chromium.org
BUG=662586
Change-Id: I913c16b1a185c5b753284ed934a5e60912887f0f
Reviewed-on: https://chromium-review.googlesource.com/412055
Reviewed-by: Stephen Martinis <martiniss@chromium.org>
Commit-Queue: Nodir Turakulov <nodir@chromium.org>
Add the "gsutil" recipe module to depot_tools. This module is copied
from "tools/build"'s recipe module set at:
d6205ab0c0e8e1920f4af1617b5b3767cd944492.
Two Chromium-specific functions have been removed:
- download_with_polling
- download_latest_file
It also adds a function, "gsutil_py_path", which mimics the function in
the "depot_tools" recipe module. Ultimately, the latter will be phased
out in favor of the module-specific function.
BUG=chromium:662895
TEST=expectations
Change-Id: Ie049d637593b3f02b9ff029fbd20fd3f7170faa6
Reviewed-on: https://chromium-review.googlesource.com/411970
Commit-Queue: Daniel Jacques <dnj@chromium.org>
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Remove nested retries on checkout. Since the outer loop handles retries
specially, we want to let it exclusively handle retries. This will
prevent the inner retry loop triggering a timeout condition (rather than
a failure condition) with the outer loop.
BUG=chromium:665116
TEST=None
Change-Id: I4bf0c08b75861bf1c95060452118c314a121f9f3
Reviewed-on: https://chromium-review.googlesource.com/411397
Reviewed-by: Ryan Tseng <hinoka@chromium.org>
Commit-Queue: Ryan Tseng <hinoka@chromium.org>
Currently, 'bot_update' uses the 'gclient' that is on the system path.
Now, it will use the 'gclient.py' that is in the same 'depot_tools'
checkout as the 'bot_update' recipe module.
Also don't ignore "git_cache" move errors.
BUG=664254,663990,663440
TEST=None
Review-Url: https://codereview.chromium.org/2492963002
Reason for revert:
Broke build CQ: https://luci-milo.appspot.com/swarming/task/324a5636a5cfdf10/steps/Uncaught_Exception/0/logs/exception/0
Original issue's description:
> fix swarmbucket infra path
>
> Do not rely on buildbot config which goes up 4 directories. Everything
> should happen in the workdir. Also exclude any project-specific paths.
>
> This fixes the problem that swarmbucket builds expect git_cache to be in
> /b/git_cache. It should be in the CWD/git_cache. This will fix Windows
> builders.
>
> R=vadimsh@chromium.org
> BUG=
>
> Committed: b8b5aaea92
TBR=vadimsh@chromium.org,nodir@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/2484433002
Do not rely on buildbot config which goes up 4 directories. Everything
should happen in the workdir. Also exclude any project-specific paths.
This fixes the problem that swarmbucket builds expect git_cache to be in
/b/git_cache. It should be in the CWD/git_cache. This will fix Windows
builders.
R=vadimsh@chromium.org
BUG=
Review-Url: https://codereview.chromium.org/2477033002
Hypothesis: Sometimes bot update fails because windows fails to delete
a lockfile associated with a git process.
Test: If this happens, let's delete that lockfile and try again.
BUG=651602
Review-Url: https://codereview.chromium.org/2382653005
Reason for revert:
Pointing to the wrong file :(
Original issue's description:
> bot_update: add --auth-refresh-token-json passthrough for apply_issue
>
> BUG=642150
>
> Committed: e465667e78
TBR=vadimsh@chromium.org,martiniss@chromium.org,tandrii@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=642150
Review-Url: https://codereview.chromium.org/2350363003
Reason for revert:
broke win_chromium_x64_rel_ng/builds/281022
because of apply_issue.bat calls local python.
Original issue's description:
> bot_update: ensure correct depot_tools checkout is used.
>
> R=vadimsh@chromium.org,martiniss@chromium.org
> BUG=635641,647446
>
> Committed: 972ac50401
TBR=martiniss@chromium.org,vadimsh@chromium.org,iannucci@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=635641,647446
Review-Url: https://codereview.chromium.org/2349673002
Reason for revert:
The "git rebase --abort" should only happen when there is a failure during rebase not when it succeeds.
Original issue's description:
> Skip apply_gerrit rebase when there is a rebase failure.
>
> Not doing so causes a failure in the 'git checkout $base_rev' step as visible here: https://build.chromium.org/p/client.skia.fyi/builders/Infra-PerCommit-Trybot/builds/4995
> This causes the checkout to remain in the "rebase in progress" state.
>
>
> BUG=chromium:645955
> BUG=skia:5749
>
> Committed: 9736fe8938
TBR=agable@chromium.org,andybons@chromium.org,tandrii@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:645955
Review-Url: https://codereview.chromium.org/2331103002
A while back, we removed the last pieces of code in
src/third_party/libjingle/source/talk and it is no longer synced
in Chromium's DEPS file. Clean this up to avoid confusion in the
buildbot recipes.
BUG=webrtc:4256
Review-Url: https://codereview.chromium.org/2325043002
This is the simplest change I can make here. The bot_update.py script already
ignores the value of force (and sets it to True in all cases). This is an
expectations-only change.
I'll refactor the content of the bot_update.py script to account for this
removed option in a followup CL as well.
R=agable@chromium.org, estaab@chromium.org, hinoka@chromium.org, martiniss@chromium.org
BUG=643885
Review-Url: https://codereview.chromium.org/2316323002
Nodir Turakulov
Dan Jacques
agable
martiniss
recipe-roller
tandrii
katthomas
borenet
hinoka
vadimsh
rmistry
iannucci
kjellander
bpastene
andybons