Re: [Pasdoc-main] public type problem

[Hans-Peter D. <DrD...@ao...>]

Michalis Kamburelis schrieb:
> Hans-Peter Diettrich wrote:
>> What's wrong with the "patches" in PasDoc2?
>>
> 
> We already discussed this long time ago, way too many times. To repeat:
> 1. There are bugs in PasDoc2 branch. And it seems noone had time to fix
> them.

Depends on what is considered a bug. I've stripped down the accepted 
syntax to D7, because all later extensions require changes to the 
internal PasDoc structure.

> 2. They break the design (like separation between parser and
> generators) too much.

I've touched the generators only for debugging purposes, i.e. to obtain 
some output from given input. (see below)

> And they are using large commits, that change many
> things at once --- so they cannot be separated and tested easily.

Right, many commits cannot be broken down, because intermediate steps 
won't compile (e.g. lots of units affected).

> And by the way, some useful changes *were* applied to trunk from PasDoc2
> (like PasDoc_Languages, or SimpleXML changes). The rest is not usable,
> as far as I'm concerned. Of course, anyone is always welcome to extract
> useful changes from it and submit.

I only see one chance for further cooperation:

1. Provide a renewed internal structure, that can handle all the new 
syntax elements, and optionally the parsing of the implementation 
section. If you don't like the PasDoc2 model, provide your own or accept 
some other model.

I can provide some information about the requirements of these 
extensions, the complete syntax still has to be explored.

2. Make the generators work with that new structure, e.g. add 
capabilities for handling local type declarations. The according 
implementation may require further changes to the internal structures.

3. Make sure that the new structure is usable by the new parser.

Most requirements herefore should be known from step 1, I'm willing to 
check the new parser interface for completeness.


When all this has been done, the existing parser can be extended 
according to the new syntax, or an alternative parser can be implemented 
independently. Only then I can resume my work, without being made 
responsible for anything outside the parser.


> See the thread "[Pasdoc-main] PasDoc 0.12.0 release next week, and
> decisions for PasDoc2 branch" in mailing list archives. And see threads
> linked from there.

Sorry, the supplied link (about PasDoc2) is broken, leads to "VR 
Juggler" :-(

> Please, let's not repeat the same discussion over and
> over, and please don't repeat something you already said many times
> (like "I'm waiting for someone to design a better model" or "the bugs
> are unavoidable").

I've tried to summarize my expectations above. If these don't please 
you, suggest an alternative procedure.

DoDi