Re: GNU Autoconf test version 2.59d available

[Andreas Büning]

Hello!

Ralf Wildenhues wrote:
> 
> GNU Autoconf test version 2.59d is now available.
> 
> This is a beta release, intended to be largely identical to 2.60,
> to be released very soon, if no unexpected issues turn up.  So test it
> now, use it with your code, and report any remaining issues, please!

I tried it on OS/2 and I found the following (minor) issues:

- The code that searches for the best shell (_AS_DETECT_BETTER_SHELL,
  around line 234 in lib/m4sugar/m4sh.m4) doesn't use $ac_executable_extensions,
  so for systems which use .exe extensions no shell can be found here.
  Please, correct this. If you like I'll write the necessary one or two lines
  to patch this.

  Btw, ac_executable_extension seems to be undocumented. If I don't set this
  variable, no executables like grep or perl can be found at all. I wonder
  whether this is also a problem on Windows or other systems having .exe
  files. Would it be possible to document this or any similar variable, please?

- When I recreated the autoconf 2.59d configuration files by calling
  aclocal/automake 1.9.6 and an (temporarily) installed autoconf 2.59d,
  I got the following warnings:


configure: creating cache /dev/null

[...]

checking for a sed that does not truncate output... E:/bin/sed.exe
configure: not updating unwritable cache /dev/null
configure: creating ./config.status
config.status: creating config/Makefile
config.status: WARNING:  config/Makefile.in seems to ignore the --datarootdir 
setting
config.status: creating tests/Makefile
config.status: WARNING:  tests/Makefile.in seems to ignore the --datarootdir 
setting
config.status: creating tests/atlocal
config.status: creating man/Makefile
config.status: WARNING:  man/Makefile.in seems to ignore the --datarootdir 
setting
config.status: creating lib/emacs/Makefile
config.status: WARNING:  lib/emacs/Makefile.in seems to ignore the 
--datarootdir setting
config.status: creating Makefile
config.status: WARNING:  Makefile.in seems to ignore the --datarootdir setting
config.status: creating doc/Makefile
config.status: WARNING:  doc/Makefile.in seems to ignore the --datarootdir 
setting
config.status: creating lib/Makefile
config.status: WARNING:  lib/Makefile.in seems to ignore the --datarootdir 
setting
config.status: creating lib/Autom4te/Makefile
config.status: WARNING:  lib/Autom4te/Makefile.in seems to ignore the 
--datarootdir setting
config.status: creating lib/autoscan/Makefile
config.status: WARNING:  lib/autoscan/Makefile.in seems to ignore the 
--datarootdir setting
config.status: creating lib/m4sugar/Makefile
config.status: WARNING:  lib/m4sugar/Makefile.in seems to ignore the 
--datarootdir setting
config.status: creating lib/autoconf/Makefile
config.status: WARNING:  lib/autoconf/Makefile.in seems to ignore the 
--datarootdir setting
config.status: creating lib/autotest/Makefile
config.status: WARNING:  lib/autotest/Makefile.in seems to ignore the 
--datarootdir setting
config.status: creating bin/Makefile
config.status: WARNING:  bin/Makefile.in seems to ignore the --datarootdir 
setting
config.status: executing tests/atconfig commands

This happened appearently because all Makefile.in had no datarootdir,
htmldir and all other new entries. I guess, this is the case because automake
internally called autoconf (or autom4te?) but I still had 2.59 installed in my
$PATH (which didn't know datarootdir) while 2.59d was only in a test directory
(not in $PATH). I would call this at least unexpected behaviour.
It could easily happen if someone has autoconf 2.59 / automake 1.9.6 in 
/usr/local
but a separate installation of autoconf 2.60 somewhere else.


Regards,
Andreas Büning