Re: [PATCH 1/4] AC_CACHE_SAVE: treat x: as an absolute path as well

Eric Blake <[email protected]> Mon, 22 Sep 2014 09:34:05 -0600

Newsgroups	gmane.comp.sysutils.autoconf.patches
Organization	Red Hat, Inc.
Message-ID	<[email protected]>

On 09/22/2014 12:59 AM, KO Myung-Hun wrote:
> * lib/autoconf/general.m4 (AC_CACHE_SAVE): Subject.

That's a poor changelog entry.  Better might be:

AC_CACHE_SAVE: treat x: as an absolute path as well

* lib/autoconf/general.m4 (AC_CACHE_SAVE): Treat raw drive as
absolute for OS/2 and mingw.

> ---
>  lib/autoconf/general.m4 |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/lib/autoconf/general.m4 b/lib/autoconf/general.m4
> index 59d204f..77f71d2 100644
> --- a/lib/autoconf/general.m4
> +++ b/lib/autoconf/general.m4
> @@ -942,7 +942,7 @@ do
>    esac
>    # Be sure to have absolute directory names.
>    case $ac_val in
> -    [[\\/$]]* | ?:[[\\/]]* )  continue;;
> +    [[\\/$]]* | ?:[[\\/]]* | ?: )  continue;;

On mingw, 'c:' is not necessarily absolute; Windows has up to 27 current
working directories (the current working directory when no drive letter
is given, and then one current directory for each drive), so 'c:foo'
could mean 'c:/path/to/foo' instead of 'c:/foo'.  I'm not sure how it
works on OS/2.  I also don't know the rationale for your change - how
often do you actually encounter someone using AC_CACHE_SAVE with a raw
drive letter and no path as the destination?  That is, you don't try to
ever save the cache as 'c:', but as 'c:/file', in which case the pattern
already matches.  Without an example of something that is broken without
this patch, I'm less inclined to apply it.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

signature.asc (application/pgp-signature, 539 B)