!!! Digest verification failed: !!! /usr/portage/distfiles/libksba-1.0.3.tar.bz2 !!! Reason: Failed on RMD160 verification !!! Got: 91423921eec58b053853b04a9565f208a6041ed4 !!! Expected: c47f576627aca28aac3365c65846b641c8c5ed6f
I downloaded the file again and it looks correct... I digested this again and "cvs diff" shows no change.
Well, that must be because mirror://gnupg is made of fail... There's tens of mirrors that are missing the tarball completely or plain fail in weird ways and return some junk instead of the proper tarball. Sounds like the gnupg stuff in $PORTDIR/profiles/thirdpartymirrors could do with some cleanup.
I never understood which mirror sync when... Maybe the best way is to drop the gpg mirrors and use only master site. What do you think?
> I never understood which mirror sync when... You mean the gentoo ones, or... ? Beyond initial mirroring, the SRC_URI is only used when you set GENTOO_MIRRORS="" on command-line or in make.conf, or as a fallback when the tarball is not found on any of the mirror://gentoo servers (which was my case here) - unless you of course set RESTRICT="mirror" or RESTRICT="primaryuri" in the ebuild ;) > Maybe the best way is to drop the gpg mirrors and use only master site. I didn't check how widely it mirror://gnupg used in the tree and how much work would that mean, but seeing how many broken ones mirrors have in the list, I'd say it indeed would be better to refer just to a known-to-be-working location in SRC_URI (or get an updated/fixed list from upstream if any?)
I don't mean gentoo ones... The problem is gnupg mirrors. If I will have one more issue with these I will drop the usage of gnupg mirrors, and refer only to the master gnupg site. Thanks!
Ah, OK... I'll file another bug or reopen this one if I hit the gnugp mirrors suckage too often. :)
