PATCH: Grammar checking basically works :-)

From: Martin Sevior <msevior_at_physics.unimelb.edu.au>
Date: Tue Feb 08 2005 - 04:56:30 CET

> Hi Dom,
> Thanks! But now m_Dict is never created. It is always set to
> NULL. It fails to find a dictionary inside api.c.
> It executes the code path for having failed to find a dictionary path
> starting at line 361
>

Hi Dom,
       The following patch to link-grammar prevents segfaults when
attempting to load a dictionary file when linked to abiword. I guess
there is a symbol conflict with "advance" somewhere.

Now when I place the "data" directory in the same directory I run
AbiWord link-grammar loads and grammar checks the document.

Maybe "link_advance" should be made static?

I'll leave it to you to decide.

It Works!

But long complex sentences seem to bring link grammar to it's knees. I
think we need to implement a cpu-time limit per sentence.

I'm pretty impressed with link-grammar so far. I haven't found a
sentence it has got wrong yet.

On the abiword side we have to come up with some heuristics for when to
re-check a paragraph.

ie Do we want to recheck on every keystroke, when the insertion point
leaves a paragraph, once per second?

I've played with MS Word a bit and I'm not sure what they do.

Cheers

Martin

> Martin
>
>

Received on Tue Feb 8 03:57:36 2005

This archive was generated by hypermail 2.1.8 : Tue Feb 08 2005 - 03:57:37 CET