Re: Framework support for darwin

[Mike Stump <mrs at apple dot com>]

On Tuesday, August 5, 2003, at 10:51 PM, Neil Booth wrote:
> However, my previous points still apply.  We need to decide what
> we want cpplib to do, in general, when a header isn't found.  Do
> we hand it to the client and give the client a stab?  Or do
> we do what you (appear to) do in this patch, and go through the include
> stack giving the client a stab at each?

I was worried that the data structures (_cpp_file) are private to 
cpplib (just to cppfiles.c), and I didn't want to export them.  I think 
we can export the notion as an opaque type and provide an accessor to 
the data we want.  That way, I can improve the usability of the 
callback to not do the loop and move the entire loop into the port 
files.  These updates are reflected in the patch below.

> Whatever we decide I think most of the changes that are in c-incpath.c
> should be in darwin.c.

Ok, gosh, no one wants it for Linux, sniff, I'll move em...

Ok, here is a updated patch...  Suggestions on improvements welcome.  
Do don't really like the shape of the F stuff, the only problem, I 
don't know what I like that is better.

Attachment: frameworkc.diffs.txt
Description: Text document