Postmortem MWPS Bug Left Untouched for 3 weeks
We had an issue https://gitlab.com/gitlab-org/gitlab-ce/issues/29568 which was reported 4 weeks ago and took around 3 weeks to be fixed. The issue itself should have been fixed within a day. The series of events are as follows:
- @grzesiek noticed this bug first.
- Report on March 16th 2017 (my birthday).
- @filipa was the first one from the FE to notice the bug.
- @filipa did not have time to work on the issue so she asked @fatihacet to take a look.
- @fatihacet probably never noticed this bug because he was hard at work on the MR Widget redesign.
- @annabeldunstone noticed the bug.
- @smcgivern relays that no one is working on the bug but it is the highest priority.
- I request that @jivanvl take a look at the issue.
- April 7th 2017: @jivanvl was unable to reproduce.
- @jivanvl got together with @stanhu and was able to reproduce with some pair coding.
- @jivanvl and @stanhu discovered that a simple class name was at fault.
- This line in
app/views/projects/merge_requests/widget/open/_accept.html.haml
= link_to "#", class: "merge_when_pipeline_succeeds" do
- Was changed to
= link_to "#", class: "merge-when-pipeline-succeeds" do
- A fix was submitted by @jivanvl on April 10th 2017
- The fix was merged on April 11th 2017 by @stanhu
This issue was marked Next Patch Release
, meaning it should be fixed ASAP. Usually I like to get these fixes out in under 1 day, under a few hours if possible. Unfortunatly no one picked this issue up and the buck was passed around until no one did anything and @stanhu and I got involved.
The issue should have been picked up by the first available Frontend Engineer.
Also the issue was easily to reproduce, but some Frontend Engineers did not have the CI multi runner installed, which was needed in order to reproduce. Debugging questions were not asked and the issue continued to exist for 3 weeks.