diversions!? arrrgghhh!!!!
Hi,
I'm dealing with a bug in wmaker and it's becoming quite
frustrating. First some background info:
* Back in the old days of wmaker 0.6.3, it included
/usr/X11R6/bin/asclock; the afterstep package also included it.
The previous mantainer of wmaker added a diversion for that file,
/usr/X11R6/bin/asclok -> /usr/X11R6/bin/asclock.afterstep
* In 0.15.0 (the first non-hamm release) I removed asclock from
wmaker and made it a standalone package, asclock, which wmaker now
recommends because of a bug (something or another about not
inducing trauma on users because the dammed thing is no longer
available)
* Because of the two previous points, 0.15.0's preinst tries to
remove the diversion.
Now I'm having a nasty problem because of all this:
Preparing to replace wmaker 0.20.3-1 (using wmaker_0.20.3-3.deb) ...
Removing diversion of asclock.1x to asclock.afterstep.1x...
dpkg-divert: rename involves overwriting
`/usr/X11R6/man/man1/asclock.1x.gz' with
different file `/usr/X11R6/man/man1/asclock.afterstep.1x.gz', not allowed
dpkg: error processing wmaker_0.20.3-3.deb (--unpack):
subprocess pre-installation script returned error exit status 2
The offending command in preinst:
dpkg-divert --quiet --package wmaker \
--divert $old_asclock_man_divert_to \
--remove $asclock_man
Please note it doesn't contain --rename
Now, WHAT THE HECK IS dpkg-divert TALKING ABOUT THEN!!!??!??!?!
Beleive me, I'd LOVE to go RTFM... but there's no FM for dpkg-divert.
Because of a bug in previous versions of wmaker it seems the
diversion was never removed as intended, and I THINK asclock's version
was diverted...
now, asclock conflicts with afterstep (<= 1.4-6) -- the one in hamm
*and* slink, last time I checked; so there's NO WAY afterstep's
version is installed if asclock is installed.
OTOH, if afterstep is installed and asclock is not, where does the
second copy of asclock.1x.gz come from?!? wmaker doesn't provide it!
Just to complicate things further, /usr/X11R6/bin/asclock is not a
file but an alternative:
/usr/X11R6/bin/asclock-24bpp
/usr/X11R6/bin/asclock-4bpp
/usr/X11R6/bin/asclock-8bpp
/usr/X11R6/man/man1/asclock.1x.gz is a real file and
/usr/X11R6/man/man1/asclock-24bpp.1x.gz
/usr/X11R6/man/man1/asclock-4bpp.1x.gz
/usr/X11R6/man/man1/asclock-8bpp.1x.gz
are symlinks to it.
And to make the picture even prettier, I CAN'T REPRODUCE THIS BUG
because all them systems I have here, either don't have this mess up
going on on them or don't have the diversion for some reason or
another (all of them are hamm or bo->hamm systems, and in all of them
the diversion is simply gone) That's why I uploaded -3 and claimed
the bug was fixed, because I thought I had reproduced the scenario
that triggers it, but it's obvious that I didn't. I asked Adam Di
Carlo to test wmaker 0.20.3-4, and he tells me that it fixes zippo.
That's *all* I know about this. I'm waiting for someone to reply
with some more info.
Any ideas about how I can handle this?
Marcelo
Reply to: