problems others have making MLton using old version

Henry Cejtin henry@sourcelight.com
Thu, 11 Oct 2001 11:21:28 -0500


Another  example of how it is bad to have the MLton source not be in standard
ML.  I vote again for the  goal  that  making  MLton  from  sources  (in  the
Makefile)  take  two passes: In the first you use a standard (i.e., any, like
probably an old version of MLton) compiler with no non-standard stuff.   Then
the second pass uses the just-made compiler to re-make MLton, this time using
all the extra facilities provided by MLton.

>From bap@cs.unm.edu  Thu Oct 11 04:01:29 2001
>Date: Thu, 11 Oct 2001 02:58:19 -0600 (MDT)
>From: Barak Pearlmutter <bap@cs.unm.edu>
>To: henry@sourcelight.com
>Subject: Re: new version of MLton
>Reply-to: bap@cs.unm.edu
>
>Can the previous release of mlton be used to compile this one?
>I'm getting an error when I try,
>
>    time mlton @MLton -- -v -o mlton-compile mlton.cm
>    MLton starting
>       Compile SML starting
>	  pre codegen starting
>    ../lib/mlton/basic/exn.sml:6.32-7.0 Error: unbound variable MLton.Exn.history
>	  pre codegen raised in 0.18 + 0.08 (31% GC)
>       Compile SML raised in 0.18 + 0.08 (31% GC)
>    MLton raised in 0.30 + 0.08 (21% GC)
>
>No big deal to work around I guess, just surprised.