action #30121
closedcombine bugzilla back links and bsc#xxx tagging
0%
Description
As I learned today from coolo, there's actually a back linking feature in openQA that looks at referrers coming from bugzilla (was partially broken on o3 until this morning). So if eg bug 123 has a link to openqa test 789 and someone clicks on that link, openQA would link 789 and bug 123 internally so test 789 doesn't get cleaned up.
Then there is a second mechanism in openQA. If someone writes bsc#123 into comment of test 789 openQA tags 789 with bsc#123. That is used eg. by totest manager and the badges.
Looks like both mechanisms are unrelated. What about making the back linking feature to also cause tagging? That would remove the need for a manual tagging step. Esp eg when using the bug file link openqa provides. Tagging would be just a matter of following back the link from bugzilla then
Updated by coolo almost 7 years ago
It should be noted though that this only works for the first time the bug is hit - you don't necessarly want to mention every openqa test hitting it afterwards :)
Updated by okurz over 4 years ago
- Priority changed from Normal to Low
- Target version set to future
Updated by okurz over 2 years ago
- Status changed from New to Rejected
- Assignee set to okurz
lnussel wrote:
Looks like both mechanisms are unrelated. What about making the back linking feature to also cause tagging? That would remove the need for a manual tagging step. Esp eg when using the bug file link openqa provides. Tagging would be just a matter of following back the link from bugzilla then
I guess you mean "labelling" instead of "tagging". We only have the concept of "build tags", e.g. as visible on job group pages like https://openqa.opensuse.org/group_overview/1 where we have a "published" tag on builds. For jobs there is "labelling". Every URL-referred openQA job already receives a label "label:linked" which makes the job to be considered important and hence is kept around for longer. That already happens automatically.
What we shouldn't really do is that label tests automatically as "ignorable" only because they are mentioned in a link coming from the outside. I don't see what else coudl be added hence rejecting the ticket. Feel free to respond with additional specific proposals or clarifications if I didn't get it.