[EMAIL PROTECTED] wrote:

Quote:
I found the problem and it made appear that the implementation of this javadoc 
improvement is going to be far more difficult than I thought because there are 
plenty of tests that are going to fail. 
(end of quote)




well, that's the purpose, right ? for every cartridge the tests are going to 
fail because the code generation process is going to add javadoc for parameters 
where possible .. don't forget to add a test for this in each cartridge where a 
template has changed


(end of quote)




Sure I'll do.




[EMAIL PROTECTED] wrote:



Quote:
Actually I was even thinking that theoretically, the best would be to override 
getDocumentation() in the different subclasses of ModelElementFacade, including 
ParameterFacade.
(end of quote)




no, because the reason you're overrriding in the PIM (uml metafacade) layer is 
because you want some specific PSM (javadoc) thing rendered



I think the best way to proceed is to add a method to the StringUtilsHelper 
class, and this method can be completely JavaDoc aware and performing all your 
required logic .. you would then call this method from the templates (not the 
facades) as the templates are in the PSM layer anyway



this is an easy and clean solution to your problem since it only affects the 
templates, I would avoid putting a Javadoc documentation generator in the uml 
metafacades because it's too close to the Java platform (in other words: during 
code generation for .Net it would be entirely useless)



what do you think ?
(end of quote)




I get your point about PIM/PSM separation. So I see another solution. Forgive 
me if it implies another class creation instead of reusing some existing one 
but I think documentation generation deserves a class of its own. Actually, it 
can be interesting to generate documentation for other documentation systems 
like doxygen for example. So why not define a base interface named 
DocumentationGenerator with method like getClassifierDocumentation, 
getMethodDocumentation, getParameterDocumentation and so one, and then a 
JavaDocGenerator that would implement this interface and called from the 
templates. That way, we would not have to modify getDocumentation and all the 
templates that still use this method would not make the tests fail. And we 
would keep documentation generation in one place in order to be able to update 
it (e.g. as javadoc specification evolves) or extend it (e.g. with a doxygen 
generator).



I know it might seem a bit "too much" but I think it would be a really 
interesting feature, to be able to edit documentation in the UML model, without 
caring about the syntax of a particular documentation system.



What do you think ?
--
Sébastien Arbogast
_________________________________________________________
Reply to the post : http://galaxy.andromda.org/forum/viewtopic.php?p=2066#2066
Posting to http://forum.andromda.org/ is preferred over posting to the mailing 
list!


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
_______________________________________________
Andromda-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/andromda-user

Reply via email to