Re: namespaces and prefix types in context analysis

Hi Eric,

> I don't know if this will help with your fortran parser or not, but it
> did fix this C++ example.
Thanks for the patch and confirming the missing piece in prefix type 
computation. The commit indeed fixes the fortran - to some extend. If 
the type definition is in the same file, then type analysis succeeds. In 
general, type definitions will be in different files, and your patch 
fails, as it does not get a position for the tag. Even more, for fortran 
scope calculation at this point is like taking a sledgehammer to crack a 
nut. The only options would be members of the type tag reduced by 
protection considerations. I guess C++ is vastly more complex here. (In 
fact scope calculation is much simpler in fortran than in C. I still 
could not bend semantic-calculate-scope to work properly for fortran. 
There is always something which breaks or just plainly does not work if 
I insert an auxiliary type tag or whatever. So I guess I should overload 
this function as well. But I have gone astray...).

Actually, your commit contains a small bug, which I do not quite 
understand, as it should work fine judging by the GNU documentation:
The completion menu occasionally gets misplaced and the completion gets 
inserted at strange places. Replacing save-current-buffer with 
save-excursion in commit 1a68ab5 solves this problem. That confuses me.

Martin


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/