[702] in Coldmud discussion meeting
Re: continuing the debates
daemon@ATHENA.MIT.EDU (Fri Mar 10 23:59:56 1995
)
Date: Fri, 10 Mar 1995 20:55:12 -0800
To: coldstuff@MIT.EDU
From: jeffpk@netcom.com (Jeff Kesselman)
>< I'm just nit picking and stuff. For now add_method will take two args. The
>< name and structure. (add_method('foo, #[['flags, []]...]))
>
>That is probably the best.
>
>< I will have default values for all elements of the dictionary, too. #[] will
>< be a valid program. :)
>
>Actually, it should never assume what is being handed it is a valid program,
>have it check anyway (ugh, this could get scary...)
Would be nice, but I'd settle fo a run-time "exception" (throw) of invalid
op-code. After all, thats what a modern processor does. In general, I want
this level to be fast and efficient ratehr then bullet-proof...
Actually, I guess in the spirit of things I'd rather the checks be in a
"proofread_coldcode()" call that is independant from assignment.
BTW: I love this idea because it also opens up the way ofr
alternate-compilers are seperate yacc/lex programs, or even a Coldmud IDE,
now THERES an idea...
Jeff Kesselman