Re: [MIT-Scheme-devel] two changes to push

[Taylor R Campbell]

   Date: Sat, 19 Sep 2009 15:37:33 -0700
   From: Joe Marshall <address@hidden>

   That's pretty hard to prove unless all the calls between the
   argument binding and the argument use are primitives or known to be
   effect-free.  For the most part, you can get away with it because
   there isn't much Scheme code that uses side effects.

It would be nice if the compiler knew a good deal more of this
information.  Currently SF knows nothing about it, and LIAR knows only
about the effects of a few primitives (and still does hardly anything
with even that, as far as I can tell).  Maybe some day we'll have a
principled way to transmit compile-time information between components
of programs...

   For SYMBOL?, I thought it was worth integrating because
   GUARANTEE-SYMBOL is called on nearly every I/O operation.

It is?  I don't see why it would, unless you are changing a generic
port's line-ending or using port properties at every I/O operation.