Re: Confusing/bogus message in config.log

[Derek R. Price]

Tim Van Holder wrote:

> True. But if you apply 'the program being run', then you need a seperate
> version for configure, as it is neither autoconf, nor the package,
> really.
> I guess that is what it boils down to: do we see 'configure' as a) a
> program in its own right, b) an inextricable part of a package, or
> c) part of autoconf (though included with many other packages).
>
> I suppose b) seems the most logical choice.
> So the shortest thing that has all the info and is compliant, would
> probably be
>
>         GNU configure 2.50, for CVS 1.11.1
>
> With the one drawback that it more clearly suggests 2.50 is configure's
> version number.

Unfortunately this diverges from the GNU coding standard (
http://www.gnu.org/prep/standards_17.html#SEC17 ).  Succintly put, it says
the output should be like:

    PROGRAMRUNNING (PACKAGE PACKAGE_VERSION) PROGRAMRUNNING_VERSION

where PACKAGE_VERSION is included only when != PROGRAMRUNNING_VERSION and
(PACKAGE .*) is only included when != PROGRAMRUNNING.

I think we're debating this only because RMS doesn't seem to have left much
room for generated programs and we seem to agree that it would be useful to
know the version of Autoconf which generated a configure script as well as
the package it came from without having to track down and ask a user
reporting a bug what version they were running when they already sent the
output of 'configure --version'.

Derek

--
Derek Price                      CVS Solutions Architect ( http://CVSHome.org )
mailto:address@hidden         CollabNet ( http://collab.net )
--
All those who believe in psychokinesis raise my hand.