New issue
Advanced search Search tips

Issue 736767 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner:
Closed: Sep 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Bug



Sign in to add a comment

WPT import deleted tests that were supposed to be exported

Project Member Reported by cbiesin...@chromium.org, Jun 26 2017

Issue description

In this CL: https://chromium-review.googlesource.com/c/547019/

A number of tests in css-flexbox-1 were deleted. But those tests were added in https://chromium-review.googlesource.com/544318, supposed to be exported to web-platform-tests. Why did that not happen?
 
Cc: qyears...@chromium.org
Owner: jeffcarp@chromium.org
Looking into this.
Status: Started (was: Untriaged)
Status: Assigned (was: Started)
The reason the change wasn't exported was a patch failure:

error: css/css-flexbox-1/percentage-heights-001.html: already exists in working directory

This was probably due to the CL being able to land in Chromium when external/wpt wasn't completely up to sync. I can make a manual PR upstream based on the patch sometime this week (or if you're willing, that's great too). Thank you for bringing this up and I'm sorry this happened.
Ah I see. I can make a manual pull request right now.

(But that failure mode isn't great... nobody would've noticed if I hadn't watched this manually)
https://github.com/w3c/web-platform-tests/pull/6345 -- now I just need a reviewer! (or do I? can I bypass that?)
Ok thanks! I just approved the PR. I'll make a follow-up bug to see if we can make the patch failure mode more obvious.
Bug:  crbug.com/738502 
Components: -Blink>Infra Blink>Infra>Ecosystem
Status: WontFix (was: Assigned)
I believe this bug is no longer relevant. The immediate problem was fixed. Furthermore, both the importer and exporter have been improved since then and there is no known "data loss" issue (like reverting unexported changes in this case) at the moment.

Or in fact, I'd prefer reopening the more generic  issue 738502  which was merged into this CL to discuss if we should turn the exporter red (or send some other kind of signals) when an in-flight CL cannot apply cleanly to upstream.

Sign in to add a comment