Re: GNUstep ROADMAP

[Stefan Urbanek]

Citát Richard Frith-Macdonald <address@hidden>:

> 
> On 28 Nov 2005, at 09:17, Stefan Urbanek wrote:
> 
> > How difficult it would be to hack autogsdoc objective-c parser to  
> > parse GNUstep
> > sources and generate a list of unimplemented methods (either marked  
> > as not
> > implemented or being only in @interface)? Simple html table (with  
> > css):
> >
> > | Class | Method | Description |
> 
> Easy to get the parser to recognise empty implementations ... but  
> what to do about it is not clear.

Put on the webpage: "development wanted".

> I like the idea though.
> 
> How about ... if the parser could warn about empty implementations,  
> so we know when something needs doing, and if no documentation  
> comment is provided for the method, it could generate standard stuff  
> about the method not being implemented yet (in the place where is  
> currently generates the 'documentation forthcoming' message.
> 

It can be good for documentation where target would be GNUstep-core developers.
However, for GNUstep users (developers of gnustep frameworks or applications)
it would be a noise in the documentation. On the other hand, it can serve as a
signal for developers that are willing to contribute...

> Generally, if a class is abstract/semi-abstract then subclasses are  
> supposed to override methods ... so an empty implementation here  
> could be quite OK ... but we could get the parser to check to see if  
> it has seen the <override-subclass /> markup in the comment for the  
> method, and accept an empty implementation as OK in that case.
> 

Sounds good.

Stefan
--
http://stefan.agentfarms.net

First they ignore you, then they laugh at you, then they fight you, then
you win.
- Mahatma Gandhi