On 3/1/07, Ville Skyttä <ville.skytta@xxxxxx> wrote:
On Thursday 01 March 2007, Chris Weyl wrote:
> @@ -49,7 +49,7 @@
>
> %check
> # these will cause problems otherwise...
> -rm debug*list
> +rm -f debug*list
>
> ./Build test
>
> @@ -63,6 +63,9 @@
> %{_mandir}/man3/*
>
> %changelog
> +* Thu Mar 01 2007 Chris Weyl <cweyl@xxxxxxxxxxxxxxx> 1.26-3
> +- cause rm to not fail on non-existance of debug*list in check.
Looks like this was done in order to appease Module::Signature checks.
However, it's better to not run Module::Signature tests at all if they're
sanely avoidable, or if that fails, run them with special settings in order
to avoid it silently importing GPG keys into the build user's keyring. More
info (also linked to from the Perl SIG Wiki page):
http://koti.welho.com/vskytta/packagers-handbook/packagers-handbook.html#guidelines-perl-cpansign
Yep, a good diagnosis :)
Disabling the signature test doesn't seem to be easy with this package so the
network access problem remains, but the attached patch would fix the problem
with build user GPG keyring trashing. (The disabling debuginfo vs rm -f
debug*.list part is only cosmetic.)
Thanks for the patch -- I'll apply whenever I have a minute free later on today.
-Chis
--
Chris Weyl
Ex astris, scientia
--
fedora-extras-list mailing list
fedora-extras-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-extras-list
