Stefan Siegel <[EMAIL PROTECTED]> writes:
> It seems to work, but urpmi still has a bug:
>
> ------------------------------------------------------------------------------
> # urpmi fetchmail
> --14:22:45--
>
>ftp://ftp.uni-bayreuth.de/pub/linux/Mandrake/updates/8.0/RPMS/fetchmail-5.7.4-5.1mdk.i586.rpm
> => `/var/cache/urpmi/rpms/.listing'
> Verbindungsaufbau zu ftp.uni-bayreuth.de:21... verbunden!
> Anmelden als anonymous ... Angemeldet!
> ==> TYPE I ... fertig. ==> CWD pub/linux/Mandrake/updates/8.0/RPMS ... fertig.
> ==> PASV ... fertig. ==> LIST ... fertig.
>
> 0K -> ....
>
> 14:22:53 (7.81 KB/s) - /var/cache/urpmi/rpms/.listing gespeichert [4837]
>
> /var/cache/urpmi/rpms/.listing gelscht.
> --14:22:53--
>
>ftp://ftp.uni-bayreuth.de/pub/linux/Mandrake/updates/8.0/RPMS/fetchmail-5.7.4-5.1mdk.i586.rpm
> => `/var/cache/urpmi/rpms/fetchmail-5.7.4-5.1mdk.i586.rpm' ==> CWD
> nicht erforderlich.
> ==> PASV ... fertig. ==> RETR fetchmail-5.7.4-5.1mdk.i586.rpm ... fertig.
> Länge: 349,361
>
> 0K -> .......... .......... .......... .......... .......... [ 14%]
> 50K -> .......... .......... .......... .......... .......... [ 29%]
> 100K -> .......... .......... .......... .......... .......... [ 43%]
> 150K -> .......... .......... .......... .......... .......... [ 58%]
> 200K -> .......... .......... .......... .......... .......... [ 73%]
> 250K -> .......... .......... .......... .......... .......... [ 87%]
> 300K -> .......... .......... .......... .......... . [100%]
>
> 14:22:54 (356.13 KB/s) - /var/cache/urpmi/rpms/fetchmail-5.7.4-5.1mdk.i586.rpm
> gespeichert [349361]
>
> installing /var/cache/urpmi/rpms/fetchmail-5.7.4-5.1mdk.i586.rpm
> Preparing...
>
>###################################################################################################
> [100%]
> package fetchmail-5.7.4-5mdk (which is newer than fetchmail-5.7.4-5.1mdk) is
> already installed
> Installation failed
> Try installation without checking dependencies? (y/N)
This is "easy", release numbering is not consistent with rpm vision of sorting
release (or version).
I checked rpmtools version_compare (which do the same as rpmvercmp of rpmlib)
and it return the first argument as always greater than the second (which is
absurd). But rpmlib says the second argument is always greater than the first
one :-)
Use for example "4.1m" and "4m" for comparison. I tested it on rpm 4.0 only, not
checked for 4.0.3
François.
