[MLton-devel] Re: MLton question

John Reppy jhr@cs.uchicago.edu
Fri, 3 Oct 2003 17:01:27 -0500


On Friday, October 3, 2003, at 02:52  PM, Stephen Weeks wrote:

>
>> The Makefile has the line
>>
>> 	LEX = mllex
>>
>> in it, but I think that it doesn't get passed to the recursive calls.
>
> Right.  LEX is for the generated lexer.
>
>> I'm trying to compile it on a Mac, so using a binary package is not an
>> option.
>
> Ahh.  I suppose the easiest fix is then to copy over the generated lex
> and yacc files from somewhere else.
>
> I'm not sure if the ml-lex that comes with SML/NJ will work.

It did (it appears that your mllex is just a repackaging of the ml-lex
tool).

>
>>> Use CM.make "sources.cm"
>>
>> I tried that too, but there is a conflict with INT_INT being imported
>> from two
>> sources (the SML/NJ basis and also from
>> lib/mlton-stubs-in-smlnj/int-inf.sig.
>
> Ahh.  Which version of MLton are you compiling?  The 20030716 release
> or the CVS head?  20030716 runs with SML/NJ 110.42, while in the CVS
> head we have upgraded to 110.43.
>
> It looks like that error comes from using 110.43 on 20030716, which
> won't work because SML/NJ added INT_INF in 110.43.

I am trying to compile 20030716; switching back to 110.42 seems to
work (except that I hit a MLRISC bug :().

>
>> But you are not doing anything beyond what you get with "gcc -o"?
>
> Right.  Maybe the binutils isn't necessary, but I don't think the
> linker came on the machine that I got from sun.

The gcc toolchain on MacOS X uses its own assembler and linker that
understands Mach-O files, but I don't think that will be a problem.

	- John



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel