On Fri, Oct 12, 2012 at 12:02:21 +0200, Didier 'OdyX' Raboud wrote: > Hi Julien, and thanks for your feedback. > > Le mercredi, 10 octobre 2012 21.34:01, Julien Cristau a écrit : > > […] > > this whole stuff smells like insanity but I guess it's no worse than it > > used to be > > Indeed. It's a whole mess. And don't get me started on bzr. :> > > > > === modified file 'debian/cups.postinst' > > > --- old/debian/cups.postinst 2012-08-02 16:37:24 +0000 > > > +++ new/debian/cups.postinst 2012-08-03 10:09:05 +0000 > > > @@ -175,8 +175,8 @@ > > > > > > fi > > > > > > fi > > > > > > - if dpkg --compare-versions "$2" lt "1.5.3"; then > > > - rmmod usblp || : > > > + if dpkg --compare-versions "$2" lt "1.5.3-2+wheezy0"; then > > > > > > + modprobe usblp >/dev/null 2>&1 || : > > > fi > > > > > > fi > > > > 1) why? > > As you can see from the diff, the previous postinst was rmmod'ing usblp, so a > live-system that kept being upgraded without reboots would not have its usblp > module reloaded after cups changed its mind. > > > 2) if modprobe needs to be there shouldn't it rather be modprobe -b? > > Indeed. > > As you might have seen, I have uploaded cups 1.5.3-2.1 and then 1.5.3-2.2 to > have it built on all release architectures. The mips FTBFS was due to timing > issues improperly addressed before for arm. Waiting as long as needed for all > the print jobs to finish before starting the subsequent tests helps there. > > What's your take there ? The above modprobe is currently in unstable but we > can either a) drop it completely; b) make it modprobe -b. > Either works for me, with a slight preference for option a.
Cheers, Julien
signature.asc
Description: Digital signature
