On Thursday, 19. July 2007 19:51:41 Chris Evans wrote:
> Joost van Baal sent the following  at 19/07/2007 09:19:
> > Op Sun 15 Jul 2007 om 01:55:57 +0200 schreef Joost van Baal:
> >> Op Tue 10 Jul 2007 om 07:07:48 +0200 schreef Joost van Baal:
> >>> Op Sat  7 Jul 2007 om 07:47:55 +0100 schreef Chris Evans:
> >>>> There are some debs that rapidly get so their files don't match their
> >>>> debsums because they are updated by a route other than dpkg.  (The
> >>>> libraries/packages that come with the R statistical system for me.) 
> >>>> It would be great if it were possible to exclude these from the
> >>>> debsums check in systraq. I think the real feature request would be
> >>>> for debsums though, not for systraq though perhaps there would be a
> >>>> way you could grep the debsums output to ignore a set of debs we could
> >>>> specify?
> >>
> >> This could be done by changing /etc/systraq/systraq.d/nn-debsums.
> >>
> >> Perhaps the output of "dpkg -l" could be fed to debsums, after some
> >> modifications.  Your idea is interesting, for sure.
> >
> > Otoh, this might be a bug in the R packages.  Perhaps such a package
> > should exclude changing files from their md5sum list.  I am not quite
> > sure...

The R packages finally got fixed :-)

> I think that would be a step too far but perhaps the package debsums
> system could be modified to allow some packages to be marked as rapidly
> changing and an option to debsums allow for these to be ignored.

And I plan to hunt down all the packages that misbehave w.r.t. .md5sums and 
debsums for jessie. Lets fix the buggy packages. No need for working around 
this by maintaining blacklists in other packages ....

Andreas


-- 
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Reply via email to