Code completion in KDEStudio

[Ralf Nolden]

John Zedlewski wrote:
> 
> Hey guys,
>   I haven't looked into the studio source for a long time (like a
> year+), but here's how it used to work (hope you haven't heard this one
> already).
>   If the user has typed like "someVar->", Studio (secretly) inserts some
> text like FAKESTUDIOMETH after the "->", so it becomes
> "someVar->FAKESTUDIOMETH" (but the user doesn't see that).  It then
> dumps the modified code to a temp file and tries to compile it with
> g++.   Of course, it won't compile, because it has this fake method
> name, so Studio then greps through the error messages for an error of
> the form "no matching function for call to Foo::FAKESTUDIOMETH".  Then
> Studio knows that Foo is the class type of "someVar", and it can just
> get the method list from the class tree.
> 
>   Now THAT is like the definition of an ugly hack.  If I have time this
> summer, I'll definitely look into integrating the g++ parser, but it
> would be a tough job...
>  --JRZ

Yes, we know that. Everyone is waiting for gcc 3.0 where things will be
easier. I think if we could manage to use the kdestudio stuff for 2.x
hacked in and a clean gcc 3.0 solution for kdevelop 3 that would be the
best and is the way to go. It actually doesn't matter that much if it
isn't in the initial 3.0 release as people can always fall back to 2.x
if they find it that necessary - or help to write the gcc one :-)

Another issue is that I just returned from GUADEC 2 in Kopenhagen, the
gnome conference for planning/scheduling gnome 2.0. Besides that, I met
matthias ettrich there and talked with him and other guys. The KDE
schedule will be for kde 2.2 and then target kde 3.0 for autumn. Maybe
October. So that is our schedule to have kdevelop 3.0 ready to ship in a
usable, almost feature complete state together with it and ported to qt
3.0 (which will be minor compared to the qt 1.44 -> 2.1 switch). Just as
a side-note so everyone knows a certain timeframe.

I think the most important thing to watch out for is doing a kde-emacs/
qt-emacs. RMS would *love* us if we did that. He was bugging the gnome
guys to do it in gtk+ - and we would gain a lot of developers/users if
we do so. Sure other kde developers will help out with that for a
standalone version of emacs as well. The other top-priority is the
gdb/jdb integration which we should handle within the next 2 month to
the max. of 4 month if possible.

Happy hacking, (I'll have a shower now, 13 hours in the train with no
facilities makes you smell like a pig actually :-) and *then* I will dig
myself into gideon to start with kde/qt templates stuff; maybe writing
some docs how to do these and a little frontend-plugin. Also I'll try to
get Lauri over the week to get her finish up the kdevelop docbook docs
for integrating them although she's busy with the sgml-xml port.

Ralf
-- 
Finally, even I have to admit that being myself was the best thing
that ever could have happened to me. - Le Grand Charmeur

**********************************
Ralf Nolden

The KDevelop Project
http://www.kdevelop.org

nolden at kde.org
rnolden at kdevelop.org
**********************************

-
to unsubscribe from this list send an email to kdevelop-devel-request at kdevelop.org with the following body:
unsubscribe »your-email-address«