* David Kastrup (2005-05-24) writes: > All our help strings with the exception of this single one are > formatted manually. There is no point in having a single string > formatted differently, actually. Maybe we will just format everything > with AS_HELP_STRING at some point of time, but I don't think we should > do that right now.
Okay, with _this_ ratio of manually formatted strings to AS_HELP_STRING there probably is not much harm done. I thought we would use this more extensively. > It would also be impolite to demand more recent versions than what is > currently provided by MSYS, Cygwin and, well. Uhm. Woody. Ok, > people will need backported Emacsen for that to work with our sources, > anyway, and so a backported autoconf would not seem all that bad... I don't know if MSYS even comes with autoconf but as far as I understand the MinGW web site, MinGW includes autoconf 2.59 (<URL:http://sourceforge.net/project/showfiles.php?group_id=2435>). The same is true for Cygwin (<URL:http://cygwin.com/packages/autoconf/>). And for woody (you are considering this seriously?) there is a backport: <URL:http://www.backports.org/debian/dists/stable/autoconf/binary-i386/>. So outdated versions on these environments aren't really an argument. > By the way: does anyone object if I replace the initial #!/bin/sh line > in install-sh and mkinstalldirs with just : ? The scripts get called > from the Makefile (which requires a Bourne shell) anyway, and there is > so much that can go wrong with #!/bin/sh... Like there being no > executable /bin/sh, or there being no executable /bin/sh^M, or a > number of different things. Have we had problems with this? I briefly followed the discussion about line endings emacs-devel and must say that I find it ridiculous that line endings can be replaced during a CVS checkout. I mean, that's a problem of the CVS client, right? -- Ralf _______________________________________________ auctex-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/auctex-devel
