Michael,
> The problem is that one can never cover most of the possibilities,
> and even worse some possibilities are interpreted differently by
> vendors providing the pristine sources.
In the case of OpenPKG packages which include an explicit datestamp in the
filename, that would seem the more obvious string to parse than the
name-$MAJOR.$MINOR{.$BUGFIX}.
I realise that the package information is not the same as the filename,
though it does include the datestamp in the "Release" field. Of course it
is conceivable that a later release of an earlier version may exist.
It's certainly a nontrivial problem to solve, but at least it's not a
significant problem.
> I think that the logic you described is a part of RPM rather than OpenPKG
I don't know how this fits in the RPM/OpenPKG picture, but on the face of
it, it would make sense for this datestamp comparison to "work" in
OpenPKG.
> In any case, OpenPKG package writers should do their best to avoid this
> problem by not packaging beta grade software at all.
Well, some folks' idea of "beta" or 0.x is leagues ahead of what other
folks will sell you as version 9.5. I don't think that as packagers we can
do much other than accept name and version from the distribution.
That is of course all opinion and supposition.
-Andrew-
--
_________________________________________________________
| -Andrew J. Caines- 703-886-2689 Unix Systems Engineer |
| <[EMAIL PROTECTED]> WorldCom Web Hosting |
______________________________________________________________________
The OpenPKG Project www.openpkg.org
User Communication List [EMAIL PROTECTED]
