On Thu, 15 Nov 2001, Stas Bekman wrote: > ok, that sounds good. I say we should change it. My reasoning is that > the API never defined the case where fname==NULL, so it wasn't a part of > the public API and users weren't supposed to use this method with > fname==NULL => we can safely add this to the API.
but Apache->server_root_relative() is a documented part of the 1.x api. it might not explicitly say 'returns a trailing /' but anybody who's using it currently is getting a trailing / and might depend on that. > In any case if some code goes broken, the fix is just to add /. then you end up with ugly conditional code to make it work with both 1.x and 2.0 > BTW, we can also expose ap_server_root and then the problem is fixed as > well. maybe what we should do is have Apache->server_root_relative that behaves just like 1.x in Apache::compat and have 2.0's function be named Apache::server_root() which returns ap_server_root if fname == NULL and calls ap_server_root_relative() otherwise. this would also fix the current compat problem that 2.0's Apache::server_root_relative doesn't support Apache->server_root_relative. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
