Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [cdt-dev] Code Completion Proposal

Hey this is great!  (One thing, Kleo, if you are listening, can you take all
of these wicked
proposals and have them slowly percolate up to the CDT main site).

A couple of questions/comments:
- Examples where code completion is requested:  Would it be worthwhile to
just straight out
   list the expected completions rather than just highlight three (global
variable, scope variable,
   macros, functions, class/structure variables, class methods).  The fact
that you may have
   to do a lot of "work" (ie in the case of the Complex Qualified Name
Reference) doesn't
   really matter since it is the "completion" part that matters (IMHO).

- How does the Doxygen/JavaDoc additional description information fit into
this?  Looking at
the Java Source they seem to be part of the "or more completion processors",
but since the
description (for the context help) is bound in at the time you present the
proposal, how would
this work?  Does our parser/model need to be aware and capable of dealing
with these comment
blocks so that they can be referenced once a completion option is
considered?

- What about files outside of the workspace?  Here I'm thinking of system
header files and
the like.  Currently vendors can work around this by using the custom
completion/proposal
extention point, but really we would like the parser/indexer to be able to
run over these
items based on the "source/include" paths.   Is this going to be a problem
(I guess I'm really
just advocating that we can't be totally reliant on an IFile/IResource
world).

Nice work Hoda ... I especially liked the JDT overview since there were some
things that
I missed there when I was jogging through the code.

Thomas

----- Original Message -----
From: "Hoda Amer" <hamer@xxxxxxxxxx>
To: <cdt-dev@xxxxxxxxxxx>
Sent: Wednesday, July 23, 2003 2:30 PM
Subject: [cdt-dev] Code Completion Proposal


>
> Hello All
>
> Attached is a proposal for implementing Code Completion in CDT. As Doug
> indicated before, we have two phases in mind to enhance code completion.
> This document is mostly concerned with the second phase only. All
> comments are appreciated.
>
> For the first phase, I still have to work with Bogdan and see what are
> the capabilities of the new indexer and how we could make the most of it
> to enhance Code Completion for this release. Thanks Thomas for your
> comments below.  We will put it into consideration. What we are sure of
> is that Code Completion will not be broken due to the retirement of the
> old indexer. We will certainly try our best to add to its existing
> functionality.
>
>
>
> Thanks,
> Hoda Amer
> Staff Software Engineer
> Rational Software - IBM Software Group
>
>
>
> "Thomas Fletcher" <thomasf@xxxxxxx>
> Sent by: cdt-dev-admin@xxxxxxxxxxx
>
>
> 07/22/2003 08:21 PM
> Please respond to cdt-dev
>         To:        <cdt-dev@xxxxxxxxxxx>
>         cc:
>         Subject:        Re: [cdt-dev] Re: [cdt-patch] Indexer Patch
>
>
>
>
> So having worked on some very small pieces of the code completion which
> is
> in there
> today I have these comments:
>
> - We need to make sure that we continue to support the contributed
> function
> information
>   extention point (in some manner, not necessarily as is).
>   - The current contributed extention point needs to have a mechanism by
> which
>      users can "toggle" which sets of information they want contributed
> for
> a given
>      project (currently missing, but will be needed when RH finishes
> their
> contribution).
>
> - It would be great if the clean up allowed multiple completion models
> to be
> inserted
>  (ie not just functions, but macros, preprocessors, types etc) easily.
>  Currently however, there is no "clean" way to implement
> #[define|include|if|else|etc]
>  completion even though it should be dead easy (and something I want to
> add
> for
>  1.2).  This might involve a modification of the current extention point
> where if the
>  "context" can be determined (out of scope, in a class/structure, in a
> method/function etc)
>  this information gets passed along.
>
> - Being able to build up the associated "live help" based on source code
> comments is
>   really valuable.  I haven't looked clearly at how the JavaDoc/Java
> world
> works this
>   one, but I think that they actually have two totally different context
> completion
>   contributors.
>
> If you guys are having discussions about this, I'm keen to listen in and
> get
> educated.
>
> Thanks,
> Thomas
>
> ----- Original Message -----
> From: "Douglas Schaefer" <dschaefe@xxxxxxxxxx>
> To: <cdt-dev@xxxxxxxxxxx>
> Sent: Tuesday, July 22, 2003 5:15 PM
> Subject: [cdt-dev] Re: [cdt-patch] Indexer Patch
>
>
> > Hey Thomas,
> >
> > Sorry for the delay but we had to finish a couple of investigation
> items
> >
> > before I could send this.
> >
> > Yes, we've taken a look at the work needed to do code complete the way
> > we
> > want to, but unfortunately the parser just isn't ready to handle this
> > new
> > requirement yet.  At a minimum for 1.2, we are now looking at just
> > porting
> > the current code completion over to the new indexer as you suggest.
> > Hoda
> > will start looking at this soon and work with Bogdan and Andrew to
> > ensure
> > she can get enough information out of the new indexer for the job.  If
> > you
> > or Alain have any help to offer, that would be great.  The super-duper
> > code complete will have to wait until the next release :-(.
> >
> > Cheers,
> > Doug
> >
> >
> >
> >
> > "Thomas Fletcher" <thomasf@xxxxxxx>
> > Sent by: cdt-patch-admin@xxxxxxxxxxx
> > 07/21/2003 04:41 PM
> > Please respond to
> > cdt-patch@xxxxxxxxxxx
> >
> >
> > To
> > <cdt-patch@xxxxxxxxxxx>
> > cc
> >
> > Subject
> > Re: [cdt-patch] Indexer Patch
> >
> >
> >
> >
> >
> >
> > Bogdan, Hoda,
> >
> >   Does this mean that with the indexing of the functions now, we could
> > start
> > using
> > this for code completion in the editor?  If I understood correctly,
> this
> >
> > was
> > something
> > that Hoda was looking at, I'm interested in finding out where we are
> at
> > with
> > this.
> >
> > Thanks,
> >   Thomas
> >
> > ----- Original Message -----
> > From: "Bogdan Gheorghe" <gheorghe@xxxxxxxxxx>
> > To: <cdt-patch@xxxxxxxxxxx>
> > Sent: Monday, July 21, 2003 3:02 PM
> > Subject: [cdt-patch] Indexer Patch
> >
> >
> > >
> > > This patch adds namespaces, enums, typedefs, functions, methods,
> > fields
> > > and vars to the index. It also fixes a problem with the Search label
> > > provider which caused it to not display properly under some
> > conditions.
> > > (Note this patch does not affect any other components).
> > >
> > > - Bogdan
> > >
> > >
> > >
> >
> > _______________________________________________
> > cdt-patch mailing list
> > cdt-patch@xxxxxxxxxxx
> > http://dev.eclipse.org/mailman/listinfo/cdt-patch
> >
> >
> > _______________________________________________
> > cdt-dev mailing list
> > cdt-dev@xxxxxxxxxxx
> > http://dev.eclipse.org/mailman/listinfo/cdt-dev
> >
>
> _______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/cdt-dev
>
>
>



Back to the top