[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
RE: [cdt-dev] Code Assist in 3.0M6
|
Thanks Douglas,
i filed the bug:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=94173
thank you for looking at the problem, its very appreciated
cheers
Emanuel
-----Original Message-----
From: cdt-dev-bounces@xxxxxxxxxxx [mailto:cdt-dev-bounces@xxxxxxxxxxx] On
Behalf Of Douglas Schaefer
Sent: Montag, 9. Mai 2005 14:07
To: CDT General developers list.
Subject: Re: [cdt-dev] Code Assist in 3.0M6
Hi Emanuel,
I have heard of Qt giving us trouble in the past. It does have some pretty
weird macros to make it look like they are adding concepts to the
language. I'm surprised it is still a problem since we've made the parser
more completed in 3.0, especially with it's handling of macros. This is
something we'll look into. Raising a bug in bugzilla with an example of
code that is returning bad results in a content assist would be most
helpful.
Our strategy has a number of prongs to it. For things you are seeing with
Qt, which is heavily C++, it is really based on the parser that we've
upgraded in 3.0 to build a more complete AST. The parser also detects the
context for the content assist and we use the semantic look-up built into
the tree to find the possible completions. There are other strategies as
well, such as using search for C functions, code templates, and pluggable
API help.
The best way to help at this stage is to test and report bugs. If you feel
brave enough to look at the code, check out the CCompletionProcessor2
class which drives the show.
Cheers,
Doug
cdt-dev-bounces@xxxxxxxxxxx wrote on 05/09/2005 03:13:12 AM:
> Hi all,
>
> i have tried to develop a QT ( http://www.trolltech.com )
> application using CDT 3.0M6 / Windows / QT 3.3.4 ? and I noticed
> that the source completion didn?t really do much for me (It
> recognized the classes, but no parameter hints, no constructor hints
> etc). While I know that the Code Assist is a work-in-progress, it
> would be interesting to see why Code Assist worked quite solid for
> me so far, but fails (tried the new and old Code Assist) so bad for
> the QT headers. It may be related to their special header macros or
something.
>
> What is the CDT code assist strategy in a nutshell? If it cannot
> parse a line in a header file, will it give up on the files, or are
> there recovery points? If there is a design document for this
> component, could someone kindly point me to it? I may be interested
> in helping out, if the thing is not too obscure J
>
> thanks for any help/clues on this,
> Emanuel
> http://www.ngine.de_______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/cdt-dev
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev