I've found over the last couple of years as we have made more use of the
exclude lists (as defined in ProblemList.txt in the repo) that at least in client-libs we have had many (well I'd say at least 10) cases where there are 2 entries in the problem list referencing the same test, often under different bug ids, referencing a different platform.
Would it make sense for jtreg to balk at such cases so that we don't add duplicates ? That could quickly shut down testing if someone does that without a jtreg run and it couldn't be pushed unless the problem list is clean so it isn't without difficulties and consequences but I think such a fail-fast mechanism would be highly effective in keeping such duplicates from hanging around.
An alternative would be a github bot that spots this .. I believe we already have that for detecting that a closed/resolved bug is being added to the ProblemList.
So if you think that would be better, maybe close this and link to a skara JBS issue
exclude lists (as defined in ProblemList.txt in the repo) that at least in client-libs we have had many (well I'd say at least 10) cases where there are 2 entries in the problem list referencing the same test, often under different bug ids, referencing a different platform.
Would it make sense for jtreg to balk at such cases so that we don't add duplicates ? That could quickly shut down testing if someone does that without a jtreg run and it couldn't be pushed unless the problem list is clean so it isn't without difficulties and consequences but I think such a fail-fast mechanism would be highly effective in keeping such duplicates from hanging around.
An alternative would be a github bot that spots this .. I believe we already have that for detecting that a closed/resolved bug is being added to the ProblemList.
So if you think that would be better, maybe close this and link to a skara JBS issue