[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Tinycc-devel] bravo and wishes
From: |
Fabrice Bellard |
Subject: |
Re: [Tinycc-devel] bravo and wishes |
Date: |
Thu, 24 Oct 2002 00:30:11 +0200 |
User-agent: |
Mutt/1.2.5i |
On Wed, Oct 23, 2002 at 10:48:21PM +0200, Basile STARYNKEVITCH wrote:
> First of all, bravo to Fabrice Bellard for TCC.
>
> I am impressed!
>
> The C-- project might interest people here. It does not target a
> subset of C per se (but a language which is inspired by C), but it
> targets a language specifically designed to be an intermediate
> language (with a C like syntax). See www.cminusminus.org for details
> on C--.
>
> There is one feature which I don't know if it is inside TCC or not is
> the GNU gcc computed goto extension.
>
> This means the ability to code (I am pasting from gcc info file)
>
> void *ptr;
> // ...
> ptr = &&foo;
>
> goto *ptr;
It is not handled because it requires an architectural change in the
way the asm 'jump's are generated. I hope to add it someday.
> By the way, how is the volatile keyword handled in TCC? I could be
> interested to use TCC in conjonction with my Qish project [Qish is
> opensource, in alpha stage; see my home page] which contains a copying
> generational garbage collector for C, and expect that the compile
> won't cache in registers any volatile argument, or any field in a
> volatile structure (this is needed because such volatile stuff are
> GC-ed pointers which can be changed by the GC).
Currently tcc ignores the volatile keyword because all local variables
are saved on the stack.
Fabrice.