[437] in Coldmud discussion meeting

Re: code formatting in-db

daemon@ATHENA.MIT.EDU (Thu Sep 22 19:20:52 1994 )

From: stewarta@netcom.com (Alex Stewart)
To: deforest@netcom.com (Robert de Forest)
Date: Thu, 22 Sep 1994 15:44:08 -0700 (PDT)
Cc: coldstuff@netcom.com
In-Reply-To: <199409221807.LAA25521@netcom14.netcom.com> from "Robert de Forest" at Sep 22, 94 11:05:11 am

> @program me.example
>  arg x, y, z;
>  
>  // FORMATTED: info starts at 5
>  .tell(x, y, z);
>  // FMT_INFO: line_break, line 4, col 10
>  // FMT_INFO: indent, line 5, spaces 6
>  ...
>  // END_FMT
> .

Ok, I stand corrected, but ICK! :)

Actually, what I suggested back when this came up was two things.  a) for
parsing, a newline should be treated like any other whitespace.  I still
(strongly) think think this should be true in any case (except for comment
termination, of course).  b) when compiling, a
newline is stored as a "newline" opcode in the compiled code.  This opcode
would have no function during execution but to increment the line counter (for
traceback reports), and would decompile into a newline in output code.

The only drawback to this would be that each line would take one more tick when
executing, but given that it's almost a NOOP, and ticks aren't as precious in
ColdMUD as they are in MOO, this isn't as big a deal as it might be, I think.
This would also avoid a problem which the in-DB method still has, which is
reporting line numbers correctly in tracebacks.  With a "linecounter-increment"
opcode, this would be guaranteed to be accurate.

(the other thing, of course, is that an opcode takes up a helluva lot less
space than lots of long comment strings would, reducing DB size considerably if
this type of thing was used for lots of methods)

I'm actually somewaht undecided about whether indenting should be controllable,
however..  If we go too far in this direction we'll have everybody programming
with a different style and no two people's code will look at all the same,
which I've always hated about C in particular (it's really nice to be able to
look at anybody's code and know it's going to be in a format you can understand
immediately and work with).

-R (waiting for another "defence against fascism" post from Quinn about all
this now :)
-------------------------------------------------------------------------------
   Alex Stewart - stewarta@netcom.com - Richelieu @ Diversity University MOO
             ftp://ftp.netcom.com/pub/stewarta/html/stewarta.html