[Uml-devel] Re: Umbrello design (was UML XML tree viewer)

[Andrew Sutton]

On Wednesday 06 November 2002 10:35 am, Jean Vittor wrote:
>     * first one (harder to implement) consists in reversing all
>       generated files before next code generation, getting annotation
>       contents and comparing with previous version,
>     * second one (chosen for Objecteering) consists in including
>       annotations between start/end tags; then getting them back is
>       easier (and depends less of a good parser which is hard to do for
>       C++ - I think).

i thought i did... hmm... interesting options. i don't know if we want to 
tackle the problems of really complex code generation at this point. i think 
it should just be one-shot/overwrite. that's really what generation is.

when we really start thinking about roundtrip engineering (which we will want 
to eventually), then we go to kdevelop and "duct tape" their code to create 
UML model elements (as best as possible). we would also connect kdevelops 
code browser with the signals emitted by the UML repository. another solution 
would be to replace their code browser with the UML object tree.

andy