That was a small surprise, yes. Some checks in trunk did not like r->hostname to still have the port.
> Am 10.07.2015 um 16:59 schrieb Eric Covener <cove...@gmail.com>: > >> On Fri, Jul 10, 2015 at 7:45 AM, <ic...@apache.org> wrote: >> Copied: httpd/httpd/trunk/modules/http2/h2_task.c (from r1690246, >> httpd/httpd/trunk/modules/http2/mod_h2/h2_task.c) >> URL: >> http://svn.apache.org/viewvc/httpd/httpd/trunk/modules/http2/h2_task.c?p2=httpd/httpd/trunk/modules/http2/h2_task.c&p1=httpd/httpd/trunk/modules/http2/mod_h2/h2_task.c&r1=1690246&r2=1690248&rev=1690248&view=diff >> ============================================================================== >> --- httpd/httpd/trunk/modules/http2/mod_h2/h2_task.c (original) >> +++ httpd/httpd/trunk/modules/http2/h2_task.c Fri Jul 10 11:45:46 2015 >> @@ -390,11 +390,12 @@ static request_rec *h2_task_create_reque >> r->protocol = (char*)"HTTP/1.1"; >> r->proto_num = HTTP_VERSION(1, 1); >> >> - r->hostname = env->authority; >> - >> /* update what we think the virtual host is based on the headers we've >> * now read. may update status. >> + * Leave r->hostname empty, vhost will parse if form our Host: header, >> + * otherwise we get complains about port numbers. >> */ >> + r->hostname = NULL; >> ap_update_vhost_from_headers(r); > > Looks like your sandbox may have had some unexpected changes?
