>>>> "Kyle" == Kyle Jones
Kyle> Once running in place fom a source tree became an option I
Kyle> never looked back. Uninstalling a source tree is easy.
Yeah, I also just noticed that something I semi-rely on (that
find-func finds the source library) doesn't work for autoloaded core
Lisp. It finds them in the installed Lisp directory, of course.
> library on the path. I would prefer those semantics (rather
> than defaulting to searching load-path) to reduce the incidence
> of the FAQ, but maybe I'm just a worry-wart.
Kyle> If "make install" is what we expect people to do, then the
Kyle> wiping load-history and searching load-path should be the
Kyle> default action.
OK on defaulting to searching load-path, then. As Andy points out, a
big hunk (and rapidly increasing) of our user population uses Windows
binaries, either the netinstaller or InstallShield. The Unix-y
solution (have the build process redump from the installed Lisp) isn't
failsafe for binary distributions.
I don't understand why wiping load-history is desirable, though.
Shouldn't that be accurate? And it shouldn't hurt for it to be
accurate. If we default to searching load-path, only those who know
where to look will ever notice that load-history contains references
to files that don't exist on their systems.
If we do default to munging load-history, should we try to warn users
that the libraries they're editing need to be explicitly loaded to get
XEmacs to use them?
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
My nostalgia for Icon makes me forget about any of the bad things. I don't
have much nostalgia for Perl, so its faults I remember. Scott Gilbert c.l.py