autoconf-patches
[Top][All Lists]
Advanced

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

Re: [PATCH] Support AC_CONFIG_LIBOBJ_DIR


From: Stepan Kasal
Subject: Re: [PATCH] Support AC_CONFIG_LIBOBJ_DIR
Date: Wed, 20 Apr 2005 10:30:19 +0200
User-agent: Mutt/1.4.1i

Hi again,

On Tue, Apr 19, 2005 at 09:49:25AM -0700, Paul Eggert wrote:
> > And why is it necessary to try to handle AC_CONFIG_LIBOBJ_DIR(/)?
> 
> Paranoia, that's all.

The documentation says that AC_CONFIG_LIBOBJ_DIR specifies
        "a name relative to the top level of the source tree"

Thus I believe the paranoia should be directed elsewhere.

Please see the patch attached to this mail.  Can I commit it?

> Come to think of it, perhaps I wasn't paranoid enough.  Is it possible
> for ac_config_libobj_dir to be empty?

I think your previous code handles this case correctly--expr fails and
empty string is returned.

My previous proposal was wrong, but I think we can use this variation of
your code:

ac_libobj_dir=
test "X$ac_config_libobj_dir" = X. ||
  ac_libobj_dir=`expr "X$ac_config_libobj_dir/" : 'X\(.*[^/]\)'`

The most common case is that ac_config_libobj_dir is set to ".", as it
is the default set by Autoconf; it's good that this case is optimized
and doesn't call expr.

_If_ our checks fail, and ac_config_libobj_dir happens to be / (or more
slashes), then expr returns empty string.  I see no problem here.

Have a nice day,
        Stepan

Attachment: autoconf-20050420-libobj_dir.patch
Description: Text document


reply via email to

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