> Oh, I agree with that of course. I haven't actually checked out your patch,
> but what I had to do to find the problem (and similar ones) was to do a
> local build and then inspect the output locally. It would of course be nice
> if that wouldn't be necessary (unless the logs get too bloated).

Basically what the patch does is capture and log the stderr output
from the 'curl --version' command that previously was been redirected
to /dev/null and only show that log in case that the command has
failed.

If 'curl --version' does not fail no additional information is given,
so sites with no problem at this point won't see any increase in log
files sizes. Except for a new log file with zero size.

I hope Daniel also takes this patch into consideration.

Regards,
Yang
Received on 2005-12-05