>>>>> "Dan" == Dan Gohman <gohmandj_at_mrs.umn.edu> writes:

    Dan> I'm using the Debian source files as found on the coda ftp
    Dan> site. These work quite well (even a debconf setup for
    Dan> coda-client!). So far, I've noticed the following problems:

I've got a few problems, comments, and suggestions, too.

It would be nice if the debian subdirectory could be in the CVS tree.
I would suggest a PMS subdirectory to configs, with "debian",
"redhat", "SuSE" (and/or just one "RPM" for the RPM-using distros),
"FreeBSD", etc subdirs.  Then add some code to bootstrap.sh to copy
the data into the appropriate place.

I successfully built all packages using --srcdir.  I had to hack the
debian/rules file in a minor way.  I plan to put together some patches
(including the debian portion of the bootstrap.sh patch suggested
above).  Remind me if you're interested, otherwise I probably won't
get to it until I next have a "build Coda frenzy."

I noticed the debconf stuff being there, but it didn't seem to get
triggered using dpkg -i ... is that only for dselect?

venus-setup still refers to /etc/conf.modules; on my system that file is
ignored.  It should check for /etc/modules.conf and/or /etc/modutils.

There's a lot of cruft in venus-setup that doesn't get into the real
/etc/coda/venus.conf.  There should be a comment explaining that these
are just fallbacks.

venus-setup (apparently) did not get run as part of the dpkg -i
process for some reason (previously installed coda detected, maybe?);
the installed venus.conf has rootserver=testserver.coda.cs.cmu.edu in
it, oops.

vice-setup also did not request any user input; I don't know if it ran
or not.

What is the cachesize parameter at the bottom of venus.conf?

The coda-client and coda-server packages are insufficiently careful
about common files.  I inadvertantly installed coda-server on a
notebook with an overfull /usr partition, so I removed it---which cost
me my codaconfedit and coda-setup-ports scripts at least.  coda-common
package, anyone?

In venus-setup, an instance of @sysconfdir@ didn't get substituted.

    Dan> Having /usr/coda/tmp is suboptimal;

Agreed, this looks like an oversight.

-- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."