New issue
Advanced search Search tips

Issue 738231 link

Starred by 2 users

Issue metadata

Status: WontFix
Owner: ----
Closed: Oct 31
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 3
Type: Feature



Sign in to add a comment

CL labelled as Trybot-Ready +1 can't be applied in pre-cq.

Project Member Reported by pho...@chromium.org, Jun 29 2017

Issue description

I have CL stack as follows:

crosreview.com/#/c/540269
crosreview.com/#/c/556455
crosreview.com/#/c/539030
crosreview.com/#/c/539029

All CLs were marked Trybot-Ready=+1, but since CL 556455 was new, it had not been reviewed yet. When the pre-cq tried to apply 556455, it borked:

"15:25:58: INFO: Failed creating transaction for phobbs:540269:2f55b5c2: CL:540269 depends on CL:556465, which is not marked Code-Review=+2."

See the logs here: https://luci-logdog.appspot.com/v/?s=chromeos%2Fbb%2Fchromeos%2Fpre-cq-launcher%2F9475%2F%2B%2Frecipes%2Fsteps%2FPreCQLauncher%2F0%2Fstdout

Why does launching a Pre-CQ trybot require Code-Review=+2 on every CL in the stack?
 
Labels: -Type-Bug Type-Feature
Owner: dgarr...@chromium.org
Status: Assigned (was: Untriaged)
I think this is more-or-less intended behavior.

dgarrett@ - is this a duplicate of an known bug?  Should
we close it as WontFix?

Comment 2 by pho...@chromium.org, Jun 30 2017

Why is this intended?
Owner: nxia@chromium.org

Comment 4 by nxia@chromium.org, May 18 2018

Cc: -akes...@chromium.org -nxia@chromium.org
Components: -Infra>Client>ChromeOS Infra>Client>ChromeOS>CI
Owner: ----
Status: Available (was: Assigned)
Oof, I'll second comment #2. This really doesn't sound like (good) intentional behavior.

Anecdote: I'm pretty sure some developers are working around this by just +2'ing their own stuff sometimes (I mean, the PreCQ is almost telling them to...). See this CL (and its CQ-DEPEND):

https://chromium-review.googlesource.com/c/chromiumos/platform2/+/1260426

and some of the discussion, particularly:

"self reviewed otherwise pre-cq will complain the depended cl not CR+2."
Cc: lannm@google.com
Status: WontFix (was: Available)
PreCQ will be going away as we roll out parallel CQ in Q1. As such, this report will be obsoleted then. As a side effect of rolling out PCQ, all CL's will no longer need to be CR+2 to do a test run. (CQ+1 will be like launching a tryjob/PreCQ; CQ+2 will be for doing a real CQ run.)
#6 "As a side effect of rolling out PCQ, all CL's will no longer need to be CR+2 to do a test run." wouldn't this mean anyone with a chromium account can run any arbitrary code on our builders without approval? -cracks knuckles and makes a mine-bitcoin@chromium.org account-
We will prevent launching CQ+1 if the CQ+1 flagger doesn't have commit access. That will be orthogonal to code review.

#8 Is that how pre-CQ works today? That seems like a decent solution.
Today, Pre-CQ only looks at CQ+2 labels. Please add your comments to this DD so we don't lose them: https://docs.google.com/document/d/1AI8xuDHyG1nzuzDgpRjm_jVyDS1cBF1ujsB3Gl49QLM/edit

Sign in to add a comment