GNU bug report logs - #21789
25.0.50; Can' build Emacs (my first build since GIT migration)

Previous Next

Package: emacs;

Reported by: Vincent Belaïche <vincent.b.1 <at> hotmail.fr>

Date: Fri, 30 Oct 2015 11:55:02 UTC

Severity: minor

Found in version 25.0.50

Done: npostavs <at> users.sourceforge.net

Bug is archived. No further changes may be made.

Full log


Message #23 received at 21789 <at> debbugs.gnu.org (full text, mbox):

From: Vincent Belaïche <vincent.b.1 <at> hotmail.fr>
To: Glenn Morris <rgm <at> gnu.org>
Cc: Eli Zaretskii <eliz <at> gnu.org>,
 "21789 <at> debbugs.gnu.org" <21789 <at> debbugs.gnu.org>
Subject: RE: bug#21789: 25.0.50; Can' build Emacs (my first build since GIT
 migration)
Date: Fri, 30 Oct 2015 21:41:09 +0100
Dear Glenn,

I don't think that my problem is the toolchain.

But git was not configured for setting attributes to configure.ac to have LF
EOL's. So when configure.ac was pulled from repo it got native EOL's, ie
CRLF, and that was the root cause IMHO.

As per your suggestion, I think that what could be tested is that :

   perl -e 'print "\n"'

will output an LF, and not a CRLF.

For instance on my MSYS console I did this :

 $ if perl -e 'print "x\n"' | grep -U '^x$'> /dev/null; then echo "ok"; else echo "not ok"; fi
-| ok
$ if /c/Programmes/Perl64/bin/perl.exe -e 'print "x\n"' | grep -U '^x$'> /dev/null; then echo "ok"; else echo "not ok"; fi
-| not ok

In the former case the MSYS perl is used, so the verdict is OK, and in
the latter, that is the native MSW perl that is used, so the verdict is
NOT OK.

   Vincent.

----------------------------------------
> From: rgm <at> gnu.org
> To: eliz <at> gnu.org
> CC: vincent.b.1 <at> hotmail.fr; 21789 <at> debbugs.gnu.org
> Subject: Re: bug#21789: 25.0.50; Can' build Emacs (my first build since GIT migration)
> Date: Fri, 30 Oct 2015 14:42:26 -0400
>
> Eli Zaretskii wrote:
>
>> Your Automake/Autoconf installation is not an MSYS installation, but a
>> native Windows one, so it produces Windows CRLF EOLs, and that
>> confuses all kinds of scripts and/or utilities that are involved in
>> the configure process.
>
> If it is possible to test for that, sounds like autogen.sh should do so
> and reject those versions of autotools.
>
 		 	   		  



This bug report was last modified 8 years and 58 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.