bug-autoconf
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: mmm-mode configure hangs because of autoconf bug


From: Jauhien Piatlicki
Subject: Re: mmm-mode configure hangs because of autoconf bug
Date: Thu, 24 Sep 2015 01:09:38 +0200
User-agent: Mozilla/5.0 (X11; Linux i686; rv:38.0) Gecko/20100101 Thunderbird/38.2.0

Hi,

On 09/24/2015 12:28 AM, Eric Blake wrote:
> On 09/23/2015 03:45 PM, Jauhien Piatlicki wrote:
>> Hi,
>>
>> on my Gentoo system autogenerated configure script for mmm-mode
>> (http://mmm-mode.sourceforge.net/) hangs because configure script
>> generated by autoconf calls emacs without '--no-site-file' flag.
> 
> Thanks for the report.  However, it is not a bug in autoconf.
> 

Yes, sorry, it seems I was wrong and the broken calls are in automake
only (e.g. AM_PATH_LISPDIR). I'm investigating it now, sorry for trouble.

> Here are the only three places that autoconf source code mentions '-batch':
> 
> $ git grep -e -batch
> ChangeLog.3:  (TEXI2DVI): Make it --batch.
> build-aux/gnupload:GPG='gpg --batch --no-tty'
> configure.ac:      [$TEST_EMACS -batch -q -f batch-byte-compile
> conftest.el 1>&2])],
> 
> The first is not code; the second is unrelated to emacs, and the third,
> while probably something that could be patched to use --no-site-file, is
> limited to code run when configuring autoconf itself, and not something
> injected into configure scripts of other packages.
> 
> But even then, I'd rather have an actual bug report of someone failing
> to configure autoconf itself before changing autoconf's configure.ac,
> and not an indirect report of a bug in mmm-mode's configure script.
> 
>>
>> The call that configure makes looks like this:
>>
>>     sudo emacs -batch -q -eval '(while load-path (princ (concat (car
>> load-path) "\n")) (setq load-path (cdr load-path)))'
>>
>> and it hangs on
>>
>>     Opening output file: no such file or directory,
>> /root/.emacs.d/.smex-items
>>
>> The call
>>
>>     sudo emacs -batch -q --no-site-file -eval '(while load-path (princ
>> (concat (car load-path) "\n")) (setq load-path (cdr load-path)))'
>>
>> succeds. It is a bug in the sys-devel/autoconf and can not be fixed by
>> patching configure.in as autoconf itself generates wrong call to emacs.
>>
> 
> I suspect that the bad code is embedded in whoever wrote
> AC_EMACS_VERSION.  That macro is poorly named (it infringes on the
> autoconf namespace, even though it is not part of autoconf).
> 
> $ git grep AC_EMACS_VERSION
> $ echo $?
> 1
> 
> but without seeing the source of that macro, I cannot reproduce your
> problem.
> 
> There may also be problems with the elisp-comp script bundled by
> automake not using --no-site-file, but that would be a bug in the
> automake project, not here.
> 
>> Configure and configure.in are attached.
> 
> By the way, configure.in is an obsolete name; you should consider
> upgrading mmm-mode to name the file configure.ac.
> 

I'm not the mmm-mode upstream )

Attachment: signature.asc
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]