Hello I just checked the configuration of the old (which is now in stable) vs the new (which is in testing and unstable) debian pkg: the old one had --with-file-coding on the new has it off. Can somebody comment on this, for I want to report the debian manager about the issue. Another point is the strange option: '--package-path=~/.xemacs:~/.xemacs/packages:~/.xemacs/xemacs-packages: :/usr/share/xemacs21/xemacs-packages:/usr/share/xemacs21/site-packages' I remember David Kastrup complaining about it, since it shadows 3rd party pkg which are put in say site-lisp. Does anybody has a comment about that issue as well? Thanks Uwe Brauer

Uwe> I just checked the configuration of the old (which is now in Uwe> stable) vs the new (which is in testing and unstable) debian Uwe> pkg: the old one had --with-file-coding on the new has it Uwe> off. Open a Debian bug, that's just dumb, as it introduces gratuitous incompatibilities with Mule XEmacs and GNU Emacs. I'll check whether it's default in 21.4 and if not, I'll submit a patch and recommendation to make it so for the next release. It's a user-visible change, so we've been conservative about it, but I think its time has come. Uwe> Another point is the strange option: Uwe> '--package-path=~/.xemacs:~/.xemacs/packages:~/.xemacs/xemacs-packages: Uwe> :/usr/share/xemacs21/xemacs-packages:/usr/share/xemacs21/site-packages' Uwe> I remember David Kastrup complaining about it, since it Uwe> shadows 3rd party pkg which are put in say site-lisp. Does Uwe> anybody has a comment about that issue as well? As Adrian pointed out, site-packages should come ahead of xemacs-packages. Packages for XEmacs use should be put in one of the site-packages hierarchies. site-lisp is deprecated, but an option is provided to configure it (I forget whether it defaults on or off now). You should check for that option in the Debian build. -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.

On Thu, 29 Sep 2005, oub(a)mat.ucm.es wrote: The default is no file coding on unix, at least when building without mule. --with-file-coding should default to yes. --with-file-coding (Bool) Support detection and translation of newline conventions. Default: `yes' on Windows, `no' on Unix. file-coding comes along for the ride with mule, so the comment isn't exactly correct. You can't build mule without getting the effect of file coding: if test "$with_file_coding" = "yes" && test "$with_mule" = "no"; then AC_DEFINE(FILE_CODING) XE_ADD_OBJS(file-coding.o) fi if test "$with_mule" = "yes" ; then AC_CHECKING(for Mule-related features) AC_DEFINE(MULE) AC_DEFINE(FILE_CODING) XE_ADD_OBJS(mule.o mule-ccl.o mule-charset.o file-coding.o) ... fi The configure script should probably whine if -with-file-coding=no in the with_mule if block. -jeff

On 9/29/05, Uwe Brauer <oub(a)mat.ucm.es&gt; wrote: Hi Uwe, In answer to your question, no, I'm not yet sure that the default configure option should be changed for 21.4. I'm fine with changing the 21.5 default, however. - Vin -- Whoever you are, no matter how lonely, the world offers itself to your imagination, calls to you like the wild geese, harsh and exciting-- over and over announcing your place in the family of things. Mary Oliver

On 9/29/05, Uwe Brauer <oub(a)mat.ucm.es&gt; wrote: Yes, I think each of those issues should be addressed in the Debian package. Thanks, Vin -- Whoever you are, no matter how lonely, the world offers itself to your imagination, calls to you like the wild geese, harsh and exciting-- over and over announcing your place in the family of things. Mary Oliver

On 9/29/05, Uwe Brauer <oub(a)mat.ucm.es&gt; wrote: Personally, I think arguing about defaults (particularly build-time defaults) is largely a waste of time. The /instantiation/ of a particular default in a package, on the other hand, can make an enormous difference, because nowadays relatively few users compile from source. - Vin -- Whoever you are, no matter how lonely, the world offers itself to your imagination, calls to you like the wild geese, harsh and exciting-- over and over announcing your place in the family of things. Mary Oliver

Glynn> I suspect that defaulting to "off" on Unix is just the Glynn> usual backward-compatibility issue, where version N+1 Glynn> behaves like version N unless you explicitly ask for the Glynn> new features. That's right, and that's basically Vin's point. However, Python now defaults "universal newlines" (which has basically the same effect as our --enable-file-coding --disable-mule) on. I bet Perl does the same. That's why I'm lukewarm in favor of changing the default now. Glynn> OTOH, it's almost certain that anyone using XEmacs on Glynn> Windows would want it enabled. Which it is. -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.

Uwe> Would it be best just not to set that variable [--site-lisp] Uwe> at all. I remember that was what I did when I compiled xemacs Uwe> for myself and I had never problems of installing 3rd party Uwe> packages. I think that's best, but it will depend on the rest of the Debian set up. -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.