Summary: | sci-libs/umfpack-5.2.0 fails tests | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Markus Rothe (RETIRED) <corsair> |
Component: | Current packages | Assignee: | Gentoo Science Related Packages <sci> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | m.langer798 |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 211791 | ||
Attachments: | umfpack build.log |
Description
Markus Rothe (RETIRED)
2008-03-14 09:25:33 UTC
Created attachment 146094 [details]
umfpack build.log
Hi Markus, Thanks for the heads-up! Unfortunately, the tests fail on my opteron as well so these may actually be faulty test routines. We should probably contact upstream and ask them what's up with that. Best, Markus I contacted upstream. The difference you see in the logs are difference in round offs and are fine. So you can go ahead and keyword it. Now I will try to convince upstream to update the regression test to make them more robust. added ~ppc64 (In reply to comment #3) > I contacted upstream. The difference you see in the logs are difference in > round offs and are fine. So you can go ahead and keyword it. > > Now I will try to convince upstream to update the regression test to make them > more robust. > Why not RESTRICT="test" in the meantime? I'd rather remove the offending tests from the test suite since the rest seems to work fine. |