Re: mod_perl-2.0 and Win32

[Randy Kobes <ra...@theoryx5.uwinnipeg.ca>]

On Thu, 13 Jul 2000, Doug MacEachern wrote:

> On Mon, 10 Jul 2000, Randy Kobes wrote:
> 
> >    For one thing, there's various conflicts with
> > some apache and perl symbols, like stat, etc.
> > on Win32, just as with mod_perl-1.2x - these
[ ... ]
> 
> argh.  are those symbol conflicts something you think can be fixed in 
> apache or perl?  i'd really like to throw way apache_inc.h
> don't bother with PERL_IS_5_6, mod_perl-2.0 will require at least 5.6.0.

Do you mean have either perl or apache change their definitions?
I know a couple of the conflicts could be resolved by changing
apache, but might they be hesitant in making such a change?
When I get back next week (I'm away at a workshop) I'll look
at that some more. I do recall a couple of the
conflicts (but not all) could be resolved with a -Dsomething
flag, and so that didn't have to be added to apache_inc.h - would
that be better?
 
> >     Another thing is the use of __FUNCTION__, as in
> > #define MP_FUNC __FUNCTION__
[ ... ] 
> no, __FUNCTION__ is the function name, eg, modperl_interp_init.
> i was planning to do something like this:
> 
> #if  __FUNCTION__NOT_SUPPORTED
> #define __FUNCTION__ __FILE__ ":" __LINE__

Would that be the only change involved related to __FUNCTION__?
I'll try that .....

> >     I'm now left with 3 groups of related warnings/
> > errors. In mod_perl.c,
> >     AP_INIT_ITERATE
> > on line 134 and
> >     AP_INIT_TAKE1
> > on line 141 aren't defined ....
> 
> oh, these are post-alpha4, they are defined in the current cvs tree.

I'll grab that ....

> > line 24 in modperl_callback.c, there's an error
> > about "my_perl" being an undeclared identifier,
> 
> hmm, strange that doesn't break in unix land too.  just comment out that
> line for now, it's never used anyhow.

OK .. If this all works, it'll probably then compile OK; I'll then 
have to get a working Apache, and see how it goes ....

best regards,
randy