From: Hans Ulrich Niedermann (vserver_at_n-dimensional.de)
Date: Fri 07 Jan 2005 - 03:53:23 GMT
Enrico Scholz <enrico.scholz_at_informatik.tu-chemnitz.de> writes:
> Hans Ulrich Niedermann <vserver_at_n-dimensional.de> writes:
>
>> The changes:
>> * Detect and require "sort"
>> * Use "sort -n" instead of multiplication magic to determine whether the
>> dietlibc version is sufficiently new. This test should be more robust.
>
> Sorry, I added some extra functionality in the meantime which was not
> submitted yet. So your patch can not be applied.
OK, so let's sort it out...
>> What they fix:
>> * handle a version number like 0.27-7
>
> Can you check if CVS ([1]) fixes your problem?
It does not.
For _dietlibc_ver_min="27-7", the line
_dietlibc_ver_min=${_dietlibc_ver_min%%[!0-9]*}
doesn't do what it is supposed to do.
>> * handle "diet -v" output like "diet version $VERSION" in addition to
>> the old "dietlibc-$VERSION"
>
> hmm... seems to be an error/incompatibility in your dietlibc. Its
> Makefile defines
>
> | VERSION=dietlibc-$(shell head -n 1 CHANGES|sed 's/://')
>
> (note the leading 'dietlibc-') which seems to be overridden by your
> distribution with 'VERSION=0.27-7'
Hmm... I'd say that VERSION=1.2.3.4 is nothing which should be
discounted in general.
Gruß,
Uli
_______________________________________________
Vserver mailing list
Vserver_at_list.linux-vserver.org
http://list.linux-vserver.org/mailman/listinfo/vserver