On 25/09/12 14:33, Dave Reisner wrote: > On Tue, Sep 25, 2012 at 02:25:59PM +1000, Allan McRae wrote: >> On 25/09/12 14:24, Dave Reisner wrote: >>> On Tue, Sep 25, 2012 at 02:19:16PM +1000, Allan McRae wrote: >>>> On 23/09/12 02:17, Dave Reisner wrote: >>>>> Avoid adding our own messaging, as autoconf will add this for us with >>>>> the result of the AC_CHECK_FILE test. Reuse the cache variable from >>>>> autoconf to set our local variable. >>>>> >>>>> Signed-off-by: Dave Reisner <[email protected]> >>>>> --- >>>>> configure.ac | 6 +----- >>>>> 1 file changed, 1 insertion(+), 5 deletions(-) >>>>> >>>>> diff --git a/configure.ac b/configure.ac >>>>> index 7fe696a..7c65a75 100644 >>>>> --- a/configure.ac >>>>> +++ b/configure.ac >>>>> @@ -424,13 +424,9 @@ AC_MSG_CHECKING(whether to use git version if >>>>> available) >>>>> if test "x$wantgitver" = "xyes" ; then >>>>> AC_CHECK_PROGS([GIT], [git]) >>>>> AC_CHECK_FILE([.git/], hasgitdir=yes) >>>>> + usegitver=$ac_cv_file__git >>>> >>>> That should be: >>>> >>>> $ac_cv_file__git_ >>>> >>> >>> Interesting that it still appeared to work as intended. Anyways, you're >>> right. Fixed. >>> >> >> Hmm... it did not work for me (which is how I noticed it). >> >> Which brings up the point, do we want to rely on autoconf naming of such >> variables? >> >> > > I don't see why not. The variable naming scheme is documented and > intended to be used. FWIW, we used to use a cache variable for the > purposes of detecting libssl. >
Ah - I did not realise that scheme was documented. All is good then.
