David Kastrup writes: > Franz Haeuslschmid <[EMAIL PROTECTED]> writes: > >> David Kastrup writes: >> >>> I have just thrown out the references to TEXIPWD completely. I >>> have my doubts that this makes a difference though. Could you >>> check again? >> >> Works for me. And yes, it doesn't make any difference. > > Uh, what? Ah, ok, you were not the one reporting the problem in the > first place.
So many names... >> Don't get me wrong. The remark on `texi2dvi' versions was meant as >> a hint to Uwe, the original poster. In my opinion, it is in the >> responsibility of the developer to provide a sane environment for >> the build of AUCTeX. However, it may be a task of the documentation >> to describe such a development environment. > > It is quite clearly stated that a recent version of makeinfo is > required _if_ you are going to build from CVS _or_ are touching source > files. I'm sometimes too focused on making things work again. If my idea or guessing was appropriate, you _cannot blame_ AUCTeX's build system or the developer. This case could turn out to be subtle: Cygwin's `texi2dvi' is a shell script which wraps calls to TeX and others. And that is the place where notorious path clashes between Unix and MS-Windows applications happen, I think. As far as I can see, at least MiKTeX provides its own version of makeinfo. In order to avoid those path clashes, wouldn't it be reasonable to point out possible problems, when using (and mixing) MiKTeX and texinfo (the latter provided by Cygwin or MSYS)? > If texi2dvi gets called in the process of a normal build straight from > the tarball, this is a bug, I'd say. In the untouched tarball, for > normal targets makeinfo should not be required. That is sensible. Makefiles will have to be remade. Franz. _______________________________________________ auctex mailing list [email protected] http://lists.gnu.org/mailman/listinfo/auctex
