[PATCH] make apr_sockaddr_info_get() use AI_ADDRCONFIG in APR 0.9.x

[Jeff Trawick <tr...@gmail.com>]

Backporting these changes is expected to work around a bizarro
resolver glitch on z/OS and it ought to save worthless lookups for a
number of people.  (Apache httpd proxy, anyone?)

Is there any particular reason this wasn't already backported to 0.9? 
Any objections, assuming that I ensure that the backport works fine on
a few platforms?

Re: [PATCH] make apr_sockaddr_info_get() use AI_ADDRCONFIG in APR 0.9.x

[Joe Orton <jo...@redhat.com>]

On Thu, Apr 20, 2006 at 08:32:17AM -0400, Jeff Trawick wrote:
> Backporting these changes is expected to work around a bizarro
> resolver glitch on z/OS and it ought to save worthless lookups for a
> number of people.  (Apache httpd proxy, anyone?)
> 
> Is there any particular reason this wasn't already backported to 0.9? 
> Any objections, assuming that I ensure that the backport works fine on
> a few platforms?

Looks fine to me.  The original change (which did not have the autoconf 
test) did cause regressions with a particular vintage of glibc, which 
was why I put off backporting it.  With the autoconf test in there it 
should be safe.

joe

Re: [PATCH] make apr_sockaddr_info_get() use AI_ADDRCONFIG in APR 0.9.x

[Jeff Trawick <tr...@gmail.com>]

On 4/20/06, Jeff Trawick <tr...@gmail.com> wrote:
> Backporting these changes is expected to work around a bizarro
> resolver glitch on z/OS and it ought to save worthless lookups for a
> number of people.  (Apache httpd proxy, anyone?)

BTW... Does anybody know if HAVE_GAI_ADDRCONFIG should be defined in
apr.hw for Windows or apr.hnw for NetWare?