On Sat, 17 Feb 2001, Jeffrey W. Baker wrote:
> On Sat, 17 Feb 2001, Martin Kraemer wrote:
>
> > On Wed, Feb 14, 2001 at 07:56:05AM -0800, Jeffrey W. Baker wrote:
> > > > It looks good, although in the line
> > > > 29 uptr->scheme = DEFAULT_URI_SCHEME;
> > > > you assign to a "const" structure, and some compilers dislike that.
> > >
> > > Ah, you are right I meant to remove that but my versions got out of sync.
> >
> > Does that mean: your patch, minus exactly this one line?
> > But that breaks the check for the default port further down.
> > (uptr->port == ap_default_port_for_scheme(uptr->scheme)))
> > although if no scheme is set, the port and per_str are probably
> > unset as well?!
>
> I don't see how it breaks the test, since the call to
> ap_default_port_for_scheme() is preceded by a test of the existence of
> uptr->scheme:
>
> if (uptr->port_str &&
> !(uptr->port &&
> uptr->scheme &&
> uptr->port == ap_default_port_for_scheme(uptr->scheme))) {
>
> Thus we never get to the function call unless there is something besides
> NULL in uptr->scheme.
Wait, now I hear you. Five more minutes on the patch. -jwb
