Re: backwards compatibility was Re: Uploaded kernel-package 3.61
Hi,
Well, about the libc6 and kernel headers and the proposal that
a libc6-kernel-headers package be created, there in one non-kludgy
way I can think of off hand.
a) I modify the kernel-package package, and add a target that shall
create a package called libc-kernel-header. This target shall not
be called by default for all builds, but shall need to be called
manually.
This target requires the co-operation of the lic6 developers
(for different architectures) to determine where to put the
headers in /usr/include, and has to be cordinated with a release
of libc6-dev that does not install the symlinks in /usr/include,
and depends on the new package libc-kernel-headers. Also, ths
requires the coperation of kernel package maintainers, who shall
have to manually create the kernel-version 2.0.32 packages, and
additionally the libc-kernel-headers package.
b) the libc developer creates a new libc6-dev package that depends on
libc-hernel-headers package
c) the kernel package maintainers releases the 2.0.32 packages again,
and also releases libc-kernel-headers package (this should be
released and installed on the ftp site first, or else libc6-dev
shall be unconfigurable.
d) repeat for all architectures
Whenever it is decided that libc be recompiled with new kernel
headers, the libc developer has to ask kernel package maintainers for
a new libc-kernel-headers package.
And then we have to deal with people who shall ignore all
warnings and create their own libc-kernel-headers packages with the
wrong darned kernel version, and stuff shall start breaking on their
machine (well, it may), and we shall never hear the end of how debian
is unstable.
Personally, I do not think this is worth it. but I am
unwilling to further "fight the good fight" on this issue. If I am
told what the consensus for the libc-kernel-headers package is, and
where to put the directories, I shall modify kernel-package to make
it easier for kernel-* package maintainers to create the darned
package
manoj
tired,
>>"Christian" == Christian Schwarz <schwarz@monet.m.isar.de> writes:
Christian> (Sorry to jump in so late.)
Christian> I know that we decided last time that the current solution
Christian> is sufficient. We thought that having the `kernel headers
Christian> FAQ' (is it package somewhere, already?) will stop these
Christian> discussions. At least, until now these discussion did not
Christian> stop.
Christian> However, since nearly every user (ok, perhaps only
Christian> developers--but these are a lot) will use libc6-dev and
Christian> kernel-headers, wouldn't it be better to choose the
Christian> solution with the little extra work (that's the solution
Christian> Remco suggested--it has been suggested a few times already)
Christian> if we can stop the confusion here?
Christian> BTW, I still doubt that it would be that much extra
Christian> work. AFAIK, the only `problem' is that the libc6-dev
Christian> maintainer and the kernel-headers maintainer need to
Christian> communicate before doing new upstream releases--but having
Christian> the maintainers of such important parts of the system
Christian> staying in contact with each other is probably a good thing
Christian> anyways.
--
"By the time they had diminished from 50 to 8, the other dwarves
began to suspect "Hungry." a Larson cartoon
Manoj Srivastava <srivasta@acm.org> <http://www.datasync.com/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: