apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Trawick <trawi...@bellsouth.net>
Subject Re: cvs commit: apr/network_io/unix sa_common.c
Date Thu, 14 Dec 2000 14:42:00 GMT
Greg Stein <gstein@lyra.org> writes:

> While a bit more complicated, it should be completely possible to assign
> non-intersecting ranges at configure time, based on discovery of the actual
> ranges used on the box.
> e.g. have a config test to see whether EAI_* is positive or negative, and
> adjust the APR error space accordingly.

This is not hard, but it introduces even more magic to get from an APR
error code to the specific error it represents.  It is already hard
for me to keep straight how the four basic platforms use the error
space.  Recently Brian Havard had to straighten out both OtherBill and
myself on the meaning of a particular error code I was getting on
Win32.  Introducing configure-time tests to find out how the error
space will be used is going to make this even more error prone.

Jeff Trawick | trawickj@bellsouth.net | PGP public key at web site:
             Born in Roswell... married an alien...

View raw message