Bug#703436: Multi-arch builds uses wrong UDEB_EXCLUDE
On Tue, Mar 19, 2013 at 05:35:24PM +0200, Robert Spencer wrote:
>Package: debian-cd
>Version: 3.1.11
>
>When building multi-arch ISOs I noted a error that repeated twice.
>
>WARNING: Unable to read UDEB_EXCLUDE file
>/home/idms/debian-cd/data/squeeze/amd64_netinst_udeb_exclude
>WARNING: Unable to read UDEB_EXCLUDE file
>/home/idms/debian-cd/data/squeeze/amd64_netinst_udeb_exclude
>
>The fact that it repeated exactly looked suspicious. Further
>investigation revealed that the same exclude file was used despite
>the arch.
>
>WARNING: Unable to read UDEB_EXCLUDE file
>/home/idms/debian-cd/data/squeeze/amd64_netinst_udeb_exclude [amd64]
>WARNING: Unable to read UDEB_EXCLUDE file
>/home/idms/debian-cd/data/squeeze/amd64_netinst_udeb_exclude [i386]
Ack, good catch.
>Looking at the code showed that it made false assumptions and
>therefore the amd64 run tainted the i386 one.
>
>I've assumed that the code is supposed to be checking for a settings
>in CONF.sh and added a check for that.
>
>I'm also assuming that UDEB_INCLUDE, being in the same area as
>UDEB_EXCLUDE, suffers from the same problem.
>
>Attached please find a patch file that fixes the error.
Ish. In fact, there's a deeper bug here - the udeb include/exclude
code is actually worse than you think. At the moment, we get away with
things only because the amd64 and i386 files provided with debian-cd
are identical. The code here just doesn't work properly with
multi-arch builds as there is no way to specify different files in
CONF.sh for the different arches. Equally, d-i only looks for its
include and exclude lists in one place on an install CD regardless of
architecture so there's currently no way of passing different config
for the different arches *anyway*.
As you might guess, this piece of the code hasn't been played with for
a while!
I'm thinking a better way to handle this would be to pick up on the
data files for all arches rather than just the first one, and merge
them. What do you think?
--
Steve McIntyre, Cambridge, UK. steve@einval.com
"Further comment on how I feel about IBM will appear once I've worked out
whether they're being malicious or incompetent. Capital letters are forecast."
Matthew Garrett, http://www.livejournal.com/users/mjg59/30675.html
Reply to: