Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [cdt-dev] CDT API hover help and completion

I'd heard from a fellow Hatter earlier today that IBM was doing some work in completion. It sounds like we lucked into a workable division of labor: you folks concentrate on completion, we'll do hover.

We should, however, I suspect, work from a common database, or at least a common db format and mechanism for handling a library of dbs for different APIs. Do you have anything in mind? What I have right now is a smallish regexec-based hunk of C code that parses .texi files -- fairly easy since texi files use text tags -- but doesn't yet produce a real db. When I have a better idea of how to impliment hover/completion I'll tinker the code to produce whatever is necessary.

I just looked at doxygen. While it looks slick, it also looks like it won't be very useful for code --like glibc source -- that's not marked up. My guess is that we'll need several ways to create dbs, possibly basing future work on doxygen, but using custom code and other stuff for legacy APIs.


Douglas Schaefer wrote:
Hi Chris,

While the hover help with content assist will be very useful, we are already working on a more generic mechanism for code assist (the completion side of things) using the new parser. Please make sure you sync up with Hoda Amer (hamer@xxxxxxxxxx) and the gang over here as you proceed.

Another thought is how about integrating something like doxygen to generate the help from the files in the workspace. My guess is that we will need a framework to enable us to plug in API help from a number of different places, including the texi files.

Doug Schaefer, Senior Software Developer
IBM Rational Software, Ottawa, Ontario, Canada

Chris Moller <cmoller@xxxxxxxxxx> Sent by: cdt-dev-admin@xxxxxxxxxxx
07/03/2003 02:19 PM
Please respond to


[cdt-dev] CDT API hover help and completion

To the world at large:

We here at Le Chapeau Rouge have started a project to supply API "hover help" and completion mechanisms for CDT.

The basis for both of these is the construction of a collection of indexable databases containing the necessary information. What's at least a first pass at creating one such db has already been done by writing a custom parser that extracts info from the GNU glibc manual .texi sources. As time goes on, we plan to add dbs for libs like X, GTK, and whatever else looks useful.

Our next step will be to tinker the CDT plugin to make use of the dbs, using the mechanisms already in place for JDT.

Any input -- suggestions, wish-lists, whatever -- is welcome.

Chris Moller
Tinker at Large
Red Hat, Inc.

cdt-dev mailing list

cdt-dev mailing list

Back to the top