Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=875150

--- Comment #18 from Honza Horak <[email protected]> ---
(In reply to comment #17)
> (In reply to comment #16)
> > Tom, do you remember if there is a reason to run the tests in %build 
> > section?
> 
> The reason I've historically run mysql's regression tests (and also
> postgresql's) in the %build part is that %check is misdesigned: it runs the
> checks only after the %install section, so that a lot of work is wasted if
> the regression test fails.

Well, since test-suite is not interrupted in case some tests fail and comparing
duration of testing to duration of %install section, I don't see the wasting
time too much important here.

> I might be willing to tolerate that and use %check if it actually did
> anything useful, like say if rpmbuild had an option to control whether to
> run the %check part or not. Since it doesn't, and we have to roll our own
> support for that anyhow (cf %runselftest in these specfiles), I find %check
> to be completely useless and best ignored.

Actually, rpmbuild allows to skip %check section with --nocheck now, but I'm
not sure if it can be used in wrappers like fedpkg, if people don't use
rpmbuild directly.

Anyway, I can imagine someone can benefit from moving the tests to %check
section, so I'd vote for the change.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug 
https://bugzilla.redhat.com/token.cgi?t=yVO55OV3sL&a=cc_unsubscribe
_______________________________________________
package-review mailing list
[email protected]
https://admin.fedoraproject.org/mailman/listinfo/package-review

Reply via email to