On 06/02/2010 03:13 PM, Eric Blake wrote:
> On 06/02/2010 02:58 PM, Vincent Torri wrote:
>>
>> Hey,
>>
>> Our configure.ac contains, at the top:
>>
>> ##--##--##--##--##--##--##--##--##--##--##--##--##--##--##--##--##
>> ##--##--##--##--##--##--##--##--##--##--##--##--##--##--##--##--##
>> m4_define([v_maj], [0])
>> m4_define([v_min], [9])
>> m4_define([v_mic], [9])
>> m4_define([v_rev], m4_esyscmd([(svnversion . | grep -v exported || echo
>> 0) | awk -F : '{printf("%s", $1);}' | tr 'A-z' ' ' | sed 's/ //g']))
>
> $1 of the m4 macro v_rev, or $1 of the awk script?
>
>>
>> Does someone see the problem ?
>
> My guess is that $1 is being expanded by m4. Awk understands '$ 1'
> instead of '$1', try that to see if it helps.
Nope, that's not it. The m4_esyscmd is being evaluated prior to the
definition of v_rev, so the m4 macro v_rev does not contain the sequence $1.
However, I do see another portability pitfall. tr 'A-z' ' ' is nowhere
near portable. What are you really trying to do with that sequence?
And why not use sed's y operator instead of using tr | sed, to shave a
process?
--
Eric Blake eblake@xxxxxxxxxx +1-801-349-2682
Libvirt virtualization library http://libvirt.org
Attachment:
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Autoconf mailing list Autoconf@xxxxxxx http://lists.gnu.org/mailman/listinfo/autoconf
