New parser branch (Was: Dumping the source DOM?)

[Vladimir Prus]

On Wednesday 13 July 2005 21:22, Steven T. Hatton wrote:

> > > a parser for a compiler expects complete code, while an IDE works on
> > > incomplete code.
> >
> > One possible approach I had in mind was to make parser restartable. First
> > you run g++ parser on the code till the first token it cannot parse. As
> > you type more tokens you feed them to the parser. If you go to the
> > beginning of the file and start typing there, you rewind parser state and
> > start parsing again.
>
> Do you guys even agree as to /why/ the boost is failing?  Vladimir, have
> you run the parser under gdb?  

No. My original bug report had "What should I do" at the end. I expected reply 
like "set breakpoint on 'complain_loudly_about_wrong_token'", but received 
somewhat sarcastic reply.

> It may take a bit of doing to figure out who 
> to feed it by hand, but I suspect that might turn up something useful.

It appears that r++ messes up compiler identification macros. See r++.macros:

   #define __GNUC__                 3
   //#define __GNUC_MINOR__           4

This defines __GNUC__ but does not defines __GNUC_MINOR__, which blows up 
Boost's compiler workarounds.

- Volodya