c++ code completion status report
ian reinhart geiser
geiseri at yahoo.com
Fri Jan 4 14:32:04 UTC 2002
-----BEGIN PGP SIGNED MESSAGE-----
On Friday 04 January 2002 06:09, you wrote about Re: c++ code completion
> This is a nice idea. Does the doxygen parser parse also the implementation
> stuff or only *toplevel*?
> Our idea was that we could call *our* parser in a flexible way so that only
> those statements are examined which are from interest. Maybe this advantage
> of an own parser get's lost when we use doxygen - but I think I will have
> look at it when "I'm back and ready to go".
> Btw.: The class store is persistant but I have found some - hmmm -
> mysterious behave(?).
> I will have a look, too. And we also think that the class store has to
> become a *real* database one day.
well i was calling doxygen and then parsing the xml file. the
entire process takes about 1.5 seconds for kdevelop's source. really i see
no use for a persistant class store information, the reason why i like XML is
that it is fast to access the information from a DOM. The hardest issue we
have at this point is that a) Doxygen's XML has some "qwirks" and b) it
seems no-one here knows anything about what a DOM really does.
The other approach is to look at doxygen's souce and just use the data
structure he is using there internally. I must warn you though it is a two
drink minimum to look at that code :P
I give my sympathies to anyone here who wishes to write there own parser.
-ian reinhart geiser
:-- Ian Reinhart Geiser --:
GPG Key: D6A6 7E16 13A9 B5A7 9E18 D1A7 3F2E B64D 19BC 76F8
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
-----END PGP SIGNATURE-----
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
More information about the KDevelop-devel