Re: The PCMCIA and kernel packages
Hi,
>>"Brian" == Brian Mays <bem5r@virginia.edu> writes:
>
Brian> This is good; things are much clearer now. So dchanges should
Brian> be run with the -n (no edit) option and perhaps a cs="New
Brian> kernel version $(KVERS)." override?
Right. Hmmm. You also need to know who is going to sign the
modules produced (and is unlikely to be the primary maintainer for
the module, espescially since all the modules have different
maintainers anyway ;-)
So, I need to send in the name of the person running the
kernel build, to be used in the dchanges pgp=XXX option.
Also, (this is a user interface issue), where should the
finished module files be produced by the target kdist* in the
module's debian.rules?
(1) /usr/src/module-X-$(KVERS)_$(version)_$(arch).deb ?
or in
(2) $(KSRC)/../module-X-$(KVERS)_$(version)_$(arch).deb?
My /usr/src is on a tiny, ancient partition; all my free space
happens to reside on /var, so I put my kernel sources there, and
would like the built things to be produced there as well.
Also, all the resulting packages will be in one place if we
use option (2) above, and allows the user to dictate where the files
are ceated, to an extent. In the default case, KSRC=/usr/src/linux,
and everything *will* be created into /usr/src/, but we are not
locked into that directory.
So, This is what I have now:
...
...
# only generate module image packages modules_image: for
$$module in $(MODULE_LOC)/* ; do \ (cd $$module; \
./debian.rules KVERS=$(version) KSRC=$(SRCTOP) \
KMAINT=$(maintainer) kdist_image; )\
done
# generate the modules package and sign them with dchanges
modules: for $$module in $(MODULE_LOC)/* ; do \ (cd $$module; \
./debian.rules KVERS=$(version) KSRC=$(SRCTOP) \
KMAINT=$(maintainer) kdist; ) \
done
Have I missed any other information that might be needed by
the modules?
Hmm. This proposal should probably be gathered together and put
into a coherent document, destined ultimately to be put into
/usr/doc/debian/package-developer/ or something like that, after its
blessed by Brian/Simon/Bruce/Ian. (I've saved the last few messages,
thankfully).
manoj
--
"Standards committees are not the best ways to create a standard.
Standards meetings and standards themselves are horribly political
things. One thing that people forget is that many standards are made
by rather small groups of people. A few good people can really save
the day, and a few idiots can really make it miserable for years to
come." Dennis Ritchie, coinventor of Unix
Manoj Srivastava Systems Research Programmer, Project Pilgrim,
Phone: (413) 545-3918 A143B Lederle Graduate Research Center,
Fax: (413) 545-1249 University of Massachusetts, Amherst, MA 01003
<srivasta@pilgrim.umass.edu> <URL:http://www.pilgrim.umass.edu/%7Esrivasta/>
Reply to: