Current state of duchain/parser

[Hamish Rodda]

On Saturday 12 May 2007 05:43, Jakob Petsovits wrote:
> On Friday, 11. May 2007, Kris Wong wrote:
> > Since my current semester in grad school has come to and end, I expect
> > to have a bit more free time and have started to look at KDevelop 4.  I
> > was wondering what the state of the duchain was and what was planned in
> > this area.
>
> Ok, the state of the duchain is like this:
>
> - It compiles. Other than that, it's pretty much in the same state as when
>   Hamish stopped hacking on it, which is a few months ago.
>
> - Note that there has not been serious work on the duchain since
>   Sublime / the new plugin architecture has been merged.
>   It worked before that, and it doesn't work now.

Well, I just finally managed to get it all compiled, and it does work here. I 
had to comment out the code in CppLanguageSupport::findInclude, because it 
would appear to be incredibly slow.

> - The assert that Andreas mentioned is a direct consequence of
>   locking difficulties. Hamish battled hard with multi-threading, and
>   all of this seems fragile enough to break as soon as something
>   unexpected happens. The other, previously found assert has been
>   overcome by the Kate developers directly commenting it out in Katepart,
>   but that problem has not really been solved either.
>   So if you understand threading and locking well, you're at an advantage.
>
> The plans for duchain are, in order:
>
> - Make it work again for C++, like it did before.
>   That means, let the code be painted in funny colors by which you can tell
>   what variable is defined and referenced where in the code.
>
> - Generalize it, so that it can be used for languages other than C++.
>   Make the C++ one just a specialization of the general duchain.
>
> - Get rid of the codemodel, and replace it by the combination of
>   duchain + parse tree. Adam has a few good ideas there, assuming
>   that the previous two goals are achieved in time.
>
> If Hamish is still listening: we'd appreciate advice on what exactly is
> necessary and how to get there best. If you could spare some time for
> reactivating the duchain, we'd not be disappointed either.

I guess just fixing the findInclude stuff is also important at the moment, but 
you can get by if you comment it out.

> > Since Jacob has said that he doesn't have much time to work
> > on KDevelop 4, and I am familiar with the code model, I figure this
> > would probably be a good place for me to help out.
>
> You could not possibly find any better place to help out!
> Nevertheless, they still spell me with a "k", like KDE ;)

I'm also happy to see someone else interested in hacking on it, I'll try to 
answer any questions you may have about it (I get my ADSL installed next 
Monday, hopefully).

Cheers,
Hamish.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/kdevelop-devel/attachments/20070515/cbd858f8/attachment.sig>