>>>> On Sat, 24 Feb 2001 03:47:56 +0900, "Stephen J.
Turnbull" <turnbull(a)sk.tsukuba.ac.jp> said:
>>>> "Dres" == James LewisMoss
Dres> OK. I hate to bring this up, but
looking at it the person who
Dres> reported this bug to me looks to have a point.
Stephen> I don't see one, just a hyperactive imagination.
That's kewl. The only point I saw is that there are Makefiles and
such in cvs that aren't generally available. (No source packages cvs
Stephen> All .debs suffer from the same "bug." None of them have
Stephen> debian/rules in them, but many have scripts which are surely
Stephen> source code. The GPL doesn't say that just because you
Stephen> provide some sources in a particular medium you have to
Stephen> provide them all that way.
I'm not sure what you are saying here.
Stephen> Nor does it say anything about replicating the binary you
Stephen> have, only about access to the sources used to build the
Stephen> binary. Debian doesn't keep even one past copy of .diff.gz
Stephen> on its servers, let alone all of them for three years!! Do
Stephen> you have a global CVS repository? If not, you guys are all
Stephen> in big trouble by that argument....
Nor does this have anything to do with the argument (though
technically I can provide every bit of source, diff.gz etc for all my
packages for three years but you are right Debian in general is in
trouble in this department, but again other violations of the GPL have
nothing to do with this issue).
Stephen> 1. We're basically covered in spirit under GPL 3(b); every
Stephen> XEmacs installation that handles packages (>=20.4)
Stephen> should have instructions for how to get to
. So maybe we have to add a README to
Stephen> every package.
This seems a reasonable way of doing it. ANd I agree in spirit it's
not a problem. Just in a very literal reading people could read it
Stephen> 2. In the case of .el -> .elc, emacs itself is the
Stephen> Makefile! For an individual .el, M-x byte-compile-file
Stephen> is all that's necessary in most cases. M-x
Stephen> byte-recompile-directory for whole packages. Where
Stephen> something more complex is needed, as with VM and I
Stephen> believe Gnus, we do supply the upstream Makefile. If we
Stephen> don't it's a bug, but only in the particular package.
Stephen> The scripts and so on that exist in the xemacs-packages
Stephen> hierarchy are mostly concerned with building the
Stephen> distribution, not with .el -> .elc compilation. AFAIK,
Stephen> the GPL does not require that you provide the tools to
Stephen> recreate the distribution tarballs, merely the software.
Stephen> To the extent that the clean environment affects the
Stephen> build, the user would have to check out the same version
Stephen> of XEmacs, build it the same way, and check out all the
Stephen> packages (in compilation, a (require ...) basically
Stephen> includes the macros in the required file, and they get
Stephen> inlined). Nobody in free software provides that level
Stephen> of replicability, I don't see why we should be burdened
Stephen> with it.
No one ever argued this. Bringing it up is a waste of time.
＠James LewisMoss <dres(a)debian.org> | Blessed Be!
| Linux is kewl!
＠"Argue for your limitations and sure enough, they're yours." Bach