Re: [CZT-Devel] Object-Z parser v1.0

[Timothy M. <tim...@sv...>]

Petra Malik wrote:
> I fixed some bugs in corejava.
> One of them is quite nasty and I am not sure how to convince gnast to 
> generate the correct code.
> I modified the generated code and checked it in to get rid of the bug 
> for now (I have to find
> a long term solution later).
> Unfortunately the validator is still throwing exceptions. But this time 
> I cannot find a meaningful
> error message. All the messages are null as far as I can see. I have got 
> no idea ... perhaps this is a bug in Jaxb.
> Anyway, it will take some time to solve this :-(
>
Sorry for making your life hard! :)

> But first I want to connect my scanner with your parser.
> Hopefully I can check in a first version later this day.
> I am planning to copy the current version of my scanner and Marks 
> SmartScanner into your (oz/parser)
> directory. Then we can modify them till it works without affecting the 
> other tools and later we have to decide
> into which package the scanner should go and how it can be used by 
> different parsers, transformers, etc.
> 
> Hope you are happy with that.
> 
That sounds like a good plan. I already have a SmartScanner version in 
there - I think it's much the same as the original. It's probably best 
to use Mark's original version so we don't have multiple versions of the 
same file floating around.

When I hacked your scanner into the parser, it seemed to go fairly 
smoothly. I ran the bbook.utf8 example through it. The only problem was 
that it seemed to think that "[NAME, DATE]" was a narrative paragraph. 
I'm not sure if there is a hidden ZED token (is there a ZED token in 
unicode?) to start that paragraph, but it didn't seem to recognise it.

By the way, I made a small change to the parser now so it handles ZED 
instead of ZSECTION (or did I mention this already?).

Cheers,
Tim