Re: modperl-2 tests on Win32

2001-12-21 Thread Randy Kobes
On Fri, 21 Dec 2001, Doug MacEachern wrote: > ah, that makes sense. i don't really like the name APR::Lib anyhow, > thinking changing it to APR::Util would be the right thing todo. > because even if we fixup the test %INC, Lib.pm would still get installed > in the wrong place. OK, that's great

::test config

2001-12-21 Thread Rick Myers
I guess the better way of asking is to provide the patch. Here's what I meant... diff -ur modperl/lib/Apache/test.pm modperl~/lib/Apache/test.pm --- modperl/lib/Apache/test.pm Sun Aug 26 18:17:26 2001 +++ modperl~/lib/Apache/test.pm Fri Dec 21 22:39:19 2001 @@ -48,8 +48,18 @@ sub write_httpd

fix for not finding .so's in blib

2001-12-21 Thread Rick Myers
adds blib/arch to the httpd.conf While I'm poking about in here, would it be alright to add some optional configuration stuff for write_httpd_conf()? It would sure clean up a lot of after the fact twiddling I'm doing with it. (Like moving httpd, httpd.conf and creating an access_log.) diff -ur m

this fixes a little problem with ``make testdb''

2001-12-21 Thread Rick Myers
I realize work is currently underway for Apache::Test, but this might help out anyone using the old Apache::test stuff. It wasn't setting the PORT envar, or rather, it was setting it to 8529 every time -- regardless of configuration. Severe line wrappage ahead. :) diff -ur modperl/lib/Apache/tes

Re: modperl-2 tests on Win32

2001-12-21 Thread Doug MacEachern
On Fri, 21 Dec 2001, Randy Kobes wrote: > There is a blib/lib/Lib.pm, which is identical to > blib/lib/APR/Lib.pm. I think what's happened is that the > Makefile.PL in WrapXS/APR/ sees a 'Lib' subdirectory, and, > ignoring the case, treats it like a 'lib' subdirectory and > dutifully copies the

Re: [patch] send_fd

2001-12-21 Thread Doug MacEachern
On Fri, 21 Dec 2001, Stas Bekman wrote: > > length required. have you seen $r->sendfile? this is meant to provide an > > interface to the 2.0 send_fd but just for sending files. > > I've seen sendfile, but it cannot replace send_fd, it has the same > requirement of input length. Also it's not