On 14/07/05, Martin Bähr <[EMAIL PROTECTED]> wrote: > On Wed, Jul 13, 2005 at 11:49:32AM +1200, Carl Cerecke wrote: > > You are, of course, very right there Volker. But tcc can be useful for > > scripting (in a round-about way). tcc can be used as a back-end for a > > scripting language - in other words, C is the intermediate language. > > It's dynamic abilities would even allow eval/exec statements in your > > scripts. > > it doesn't need c as intermediary. > all it needs is the capability to produce machinecode.
Well, yes. That's true. But it is easier to produce the higher-level c code and let somebody else to the translating to machine-code (ia32/mips/sparc/ppc/etc.) > psyco does that for python on ia32. Yes. psyco is an interesting beast. Not only does it translate to ia32, but it also dynamically optimises the code while the program is running. > pike does it nativly for a few more architectures. Don't know about pike. The point is, generating C has a number of benefits over directly generating assembly. In particular it is (1) higher-level, and (2) widely portable. > > I imagine python implemented this way would be significantly faster, > > with a small loss in flexibility. > > like pyrex? > http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/ > (just why does this url look so familiar?) Yes, I know Greg and pyrex. Very nifty it is too. And fast. Cheers, Carl.
