httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Brian Havard" <>
Subject Re: Building APR.
Date Wed, 15 Sep 1999 03:09:33 GMT
On Tue, 14 Sep 1999 19:26:31 +0200, Ralf S. Engelschall wrote:

>In article <> you wrote:
>> It doesn't belong in CVS. If you can install CVS you can install
>> autoconf too. If you want to make it easy for new developers just add a test
>> to Configure that prints a 'please run autoconf' type message if configure
>> doesn't exist.
>That's the reason why I personally always keep configure scripts in my CVS
>trees, although they are redundant. This way it's easier to work with the
>source tree if you check it out as often as I do it. But again: That's
>all just a CVS issue and has nothing to do with the configuration, of course.

Well, as Manjo suggested, it wouldn't be hard to make Configure run autoconf
automatically if there's no configure (and perhaps also if is
newer than configure).

>Or in short: If you want to get rid of configure and friends in CVS, then you
>should provide a README.DEV explaining the bootstrap step or even better
>provide a Makefile.boot at the top-level. Or alternatively (what I appreciate)
>keep the generated files in CVS and all are happy. It doesn't harm that they
>are kept in CVS. If you dislike to see them growing in CVS you can even remove
>the configure,v from time to time ;)

Another problem I have with configure being in CVS is that it's not the same
on my platform. Autoconf on OS/2 generates a configure that can handle extra
stuff like different path separators, drive letters etc, so others committing
changes to configure will mess up mine and my diffs will keeps spewing out
differences in configure. I don't know if this applies to any other

 |  Brian Havard                 |  "He is not the messiah!                   |
 |  |  He's a very naughty boy!" - Life of Brian |

View raw message