Re: [DotGNU]Libjit x86 back end enabled

[S??rgio Ferreira]

I think that Mike was speaking about Aubit 4gl where we generate C from 
the 4gl syntax.

The C conversion is allready working and allmost tested.
It is very tied to the bison grammar so the effort to change it will be 
very big.

Mike allready made some work with CINT.

IMW: I have some doubts about the strategy.
Should we concentrate our efforts in separating the syntax validation 
from semanthic validation generating an AST with treecc or use the very 
good working stuff that generates C ?

> On Tuesday 01 June 2004 08:52 pm, Mike Aubury wrote:
>  
>
> > Is there any chance at all of ever getting a very basic/simplified C
> > compiler frontend  to hook onto this rather than the Pascal one ?
> > Only - I have a project that this would be absolutely perfect for...
> >    
>
> I actually used Pascal instead of C on purpose, so as not to get distracted by 
> the language front end implementation.  I wanted to focus my efforts on the 
> libjit core.  C is a pretty quirky language to compile, and the libc 
> necessary to perform even simple tasks is far from trivial.
>
> But anything is possible.  If someone else wants to take on this task, then I 
> won't say no to adding it to the tree.
>
> You might want to check out CINT, an interpreter for C that has been around 
> for a while: "http://root.cern.ch/root/Cint.html";.  It may be easier to plug 
> libjit into that than build a C front end from scratch.
>
> Cheers,
>
> Rhys.
>
> _______________________________________________
> Developers mailing list
> address@hidden
> http://dotgnu.org/mailman/listinfo/developers
>
>