Subject: | |
From: | |
Reply To: | Alan J. Flavell |
Date: | Thu, 7 Apr 2005 10:48:08 +0100 |
Content-Type: | TEXT/PLAIN |
Parts/Attachments: |
|
|
OK, so overnight there was a new ERRATA bundle.
This morning, on one of our SL 3.0.3 installations which is manually
updated, I tried running yum update manually and got this:
[...]
librsvg2 100 % done 5/22
/usr/bin/update-gdk-pixbuf-loaders: line 27:
/etc/gtk-2.0/i386-redhat-linux-gnu/gdk-pixbuf.loaders: No such file or
directory
error: %post(librsvg2-2.2.3-6) scriptlet failed, exit status 1
[...]
This *could* however be a consequence of something I had done on that
node, in earlier attempts to recover from the errors which we had been
discussing.
On those of our installations which get updated automatically by cron,
on the other hand, I see that the reports include mention of a
gdk-pixbuf update done, without any error reports.
So maybe things are now, for practical purposes, back in synch, for
those who hadn't tried to recover things manually in the meantime?
A colleague writes to me in private mail (where "latest" refers to
the situation a day or two back now):
___
/
The latest gtk2 spec file has lines
# librsvg2-2.2.3-3 contains the adjustments for the gtk2 biarch changes
Conflicts: librsvg2 < 2.2.3-3
This can also be revealed by the obscure rpm option
$ rpm -q --conflicts gtk2
librsvg2 < 2.2.3-3
The "rpm -q --conflicts" command is really not easy to stumble across - I
guessed. It is mentioned if you say "rpm" on its own but is not explained.
I conclude that both apt-get and yum are not 100% cautious in their use of
rpm and can create situations where conflicts exist and probably
requirements not satisfied.
\___
I pass this on in the hope it may be helpful.
|
|
|