Re: new experimental C++ parser (branch)

Roman Yakovenko wrote:
> On 10/24/05, Stefan Seefeld <[email protected]> wrote:
> 
>>Roman Yakovenko wrote:
>>
>>>Why not to take C++ parser from gcc?
>>
>>There are a number of reasons:
>>
>>1) The generated intermediate representation is quite different
>>    from the one synopsis is using.
> 
> 
> May be I wrong, but it seems to me that it should not be too difficult
> to move from
> one intermediate representation to an other

The parser would need to be designed with such a requirement in mind.
As it stands, changing IRs incures a lot of changes. Besides, the parser
closely interacts with the lexer, symbol table, etc.

>>2) It is written in C.
> 
> 
> What wrong with it?

You can't take advantage of some C++ features if half of the code is
written in C (exception handling, notably).

>>3) It is released under the GPL and thus not compatible with
>>    synopsis' LGPL.
> 
> 
> That is really problem, but may be some parts of your code/product
> could be released
> under different license?

Not really. If a central piece such as the parser is released under GPL,
all code linking to it needs to, too.

>>However, there is a big deal of inspiration I get from the gcc C++
>>parser, notably its handling of tentative parsing.
> 
> 
> Are you going to build something similar to boost.wave ? It would be
> nice to have
> standard intermediate representation.

While I'm obviously aiming at a good and generally useful set of
representations, I'm not sure whether it is possible to come up
with a one-size-fits-all representation.

Finally, don't forget that I'm not rewriting synopsis. I'm just replacing
the C++ parser code, making small adjustments to the rest.

Regards,
		Stefan