| Summary: | dev-util/git-1.6.4.4 test suite fails on alpha | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Tobias Klausmann (RETIRED) <klausman> |
| Component: | New packages | Assignee: | Robin Johnson <robbat2> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | alpha, ricmm |
| Priority: | High | ||
| Version: | 2008.0 | ||
| Hardware: | Alpha | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | |||
| Bug Blocks: | 287031 | ||
| Attachments: |
build log
t3700 result file #1 t3700 result file #2 |
||
|
Description
Tobias Klausmann (RETIRED)
2009-10-11 16:03:07 UTC
Created attachment 206778 [details]
build log
Created attachment 206780 [details]
t3700 result file #1
Created attachment 206782 [details]
t3700 result file #2
please retest with FEATURES=userpriv With FEATURES=userpriv, the test suite passes. Maybe the ebuild should check&warn if it isn't set? The way distcc's pump mode works, all people who use it must currently set -userpriv (and -usersandbox). I've talekd to zmedico about that but a fix/feature for that won't be around for a while now. Feel free to close as fixed if you think the test/warn isn't necessary. I'll stabilize dev-util/git-1.6.4.4 in a few moments. We do try to disable tests that break under FEATURES=-userpriv, and there IS a warning that is already displayed (it's in your build-log), this is just one that got missed. Is t3700 the only testcase that fails for you? (In reply to comment #6) > We do try to disable tests that break under FEATURES=-userpriv, and there IS a > warning that is already displayed (it's in your build-log), this is just one > that got missed. Whoops, I should've seen that. My bad, then. > Is t3700 the only testcase that fails for you? The test suite terminates right after that, so I don't know if it's the only one that breaks with -userpriv. With user privileges, the whole test suite passes. if you look at the src_test block, there's two tests already disabled for -userpriv: t0004, t1004 just add in t3700 there for yourself. it doesn't fail here, but this isn't the first time i've seen an arch-specific test. I've done some more (and deeper) testing today and foudn that not only test 3700 has the uid-0 problem, but test 7300 (sic), too. So I've added both to the corresponding section in the ebuild. Feel free to revert if I've SNAFU'd anything. Closing since the original issue is dealt with. thanks. |