On 09/10/02 21:54 +0100, Nicholas Clark wrote: > On Wed, Oct 09, 2002 at 01:06:15PM -0700, Brian Ingerson wrote: > > This seems to nicely solve all our use cases. One thing to remember > > is that symbols aren't loaded (can't be used as barewords) until the > > end. This is the same as with the DATA section today. > > and why not? [you've done so many other crazy impossible things, what's > stopping you here] > > Can't you stub the functions from Inline? ie do perl of the form > > sub function; > > If function is ever called AUTOLOADER still kicks in, but now it's in the > symbol table, so I believe it's a legitimate bareword.
The problem is not stubbing, but knowing what to stub. I definitely don't want to parse C code for the runs that don't need compiling. But you bring up a good point. So for the attribute based subs, I don't need to do anything, because they are already subs. Right? For the multi-part "use Inline C =>" thingies I'd need to do some kind of ugly caching hack. Not sure it's worth it. I think I'll forego it (unless I'm feeling particularly challenged) and see who screams. > > > PS If somebody knows how to do this stuff in 5.6.0 using just > > attributes.pm, please speak up. attributes.pod makes my head hurt. > > as in it works in 5.6.1, but not 5.6.0? > > You could just say "sorry, doesn't work on 5.6.0" > I take it that the basic stuff still works on 5.6.0 My research shows the attributes.pm showed up in 5.6.0. It has a UI that makes me want to use a cuss word. Damian's Attribute::Handlers rocks, but is not core until 5.8.0, although it should work on 5.6.0. I was just wondering if I could pull off the A::H stuff just using, attributes.pm. Cheers, Brian
