> FrontPage is therefore dependant on that second call to init_modules()
> that the SSL patch has removed (because the SSL module does not like to
> be initialized twice.)
A similar problem affects FastCGI initialization in mod_fastcgi. But it
is possible to patch mod_fastcgi in an accommodating fashion, because it
is available as source :-)
Ben, would it be safe to invert the sense of this operation in Apache-SSL?
-- mod_fastcgi, for example, does the following detect:
/*
* This hack will prevent the starting of the process manager
* the first time Apache reads its configuration files.
*/
if((restarts==0)&&(standalone==1)) {
restarts++;
readingConfig = FALSE;
return;
}
which, I believe, causes it to run only on the second round of
init_modules. Could Apache-SSL do something similar?
As more modules appear with picky behavior concerning this startup
activity of Apache, it does raise a number of interoperability problems
with the Apache-SSL patch set.
- Rob