On 2025-08-28 20:46:47, '[email protected]' via sage-devel wrote: > Lines annotated with "known bug" are always skipped. What do you think > about the following: > 1. Introduce a new "random failure" tag and use that to annotate doctests > that are known to fail sometimes (either with segfaults, timeouts, or > differing output) > 2. Normally on CI, skip the execution of such tests. > 3. In one special run after a release (or perhaps also for PRs?), only > execute the "random failure" tests. Since there will not be that many such > annotated tests, we could even run it a couple of times and only report an > error if none of the runs was successful.
I think this would be counter-productive because it would completely hide these failures from developers, but leave them there for users to hit. Not tested / known bug would be better in that regard, since it hides them from everyone. But I really don't think we need a special solution for this. These tests usually aren't that hard to fix. In the time it will take to come up with an elaborate way to ignore them, most of them could be fixed. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/sage-devel/aLHtzkwZqNcX53Kb%40mertle.
