On Wed, 30 Jul 2003, William A. Rowe, Jr. wrote:
> At 10:09 PM 7/29/2003, Randy Kobes wrote:
>
> >As far getting the right name for the Win32 binary goes with apxs
> >present, I don't think there's an ideal solution ... To
> >summarize, the problem is that apxs -q TARGET ( = httpd) is used
> >in Apache-Test for the name of the apache binary, whereas in
> >other places (in apxs itself, for example) it's used to represent
> >"httpd".conf. What I ended up doing here is following the
> >Apache-Test convention of setting $vars->{target} to Apache.exe
> >for Win32 - this means overwriting the apxs value, which isn't
> >great, but I guess as long as the same convention is followed
> >throughout the package, it should be OK.
>
> I think the right solution here is to provide two seperate
> flags for apxs to determine the name of the "conf file" v.s.
> the name of the "binary".
That does seem the better solution ... It would mean changing
httpd-2.0's apxs.in, where $CFG_TARGET is set to "progname"
(httpd), and this is used, for example, to edit the
$CFG_TARGET.conf file to activate modules.
> Call me crazy but it seems pretty simple :-) Only question, does the
> old option preserve the 'binary name' or the 'conf name'?
Right now, within Apache-Test, apxs -q target is used as the
name of the Apache binary to search for (I haven't seen yet a
a use of it within Apache-Test for the name of the conf file).
And this seems more in tune with the fact that "target" is
derived from "progname" in config_vars.mk.
> I suspect we want to use the original flag as the 'binary
> name'. The new flag to return the '.conf name' should properly
> decorate it with .conf on all platforms. And the original flag
> can be used to retrieve the decorated 'Apache2.exe' program
> name on Win32.
That sounds good ... There already is a "sysconfdir"
(eg., /usr/local/apache2/conf/) - what about "sysconf"
( = httpd.conf ) to represent the conf file? I'll revise
the scripts (apxs_win32, and the configuration script
apxs_win32_PL, under http://theoryx5.uwinnipeg.ca/) to
reflect this.
> I'm really excited about trying your implementation, although I'm a
> bit backlogged after traveling this weekend. Certainly this belongs
> in httpd-2.0/build/ once we dot all the I's and cross the t's.
>
> If you don't object, once I give it the once-over, I'll commit
> it after I can address any minor quibbles to the httpd 2.1 dev
> branch, and we can work from there on improving it.
Sounds good!
> We should definately look to make this usable for our Netware
> users as well, who might not be using a full gcc environment,
> but tools like codewarrior suite, etc. I'm not saying we
> commit to this, but if we can keep some bits pretty abstract,
> it will be easy for another contributor to come along and
> connect the dots. {maybe it's already this way, I promise to
> delve into the code this week :-}
Right now the script has minimal changes, as I wasn't sure
if it was desired to merge the changes into the unix version.
But working with Stas has made me appreciate the value of clean,
extendable code :) - it would be fun to go through this and
make it easier to be more portable.
--
best regards,
randy
