[MLton-devel] Successfully ported MLton to FreeBSD

Matthew Fluet Matthew Fluet <fluet@CS.Cornell.EDU>
Fri, 28 Jun 2002 08:56:46 -0400 (EDT)


> I tried to compile a native FreeBSD version, but I've yet too less
> resources (~700 MB). Do you know how much memory I will need to compile
> it? 

I've been able to bootstrap the compiler with 500MB; you shouldn't have
any problems with 700MB (assuming a significant portion of memory is free 
for the compilation). It takes about 10 mins on an 800Mhz machine.

With regards to the regression tests:

testing bytechar
---
> Unable to allocate string constant "".

This is dying right at the program start when the static strings are
copied into the heap.  The minimum heap size to hold static data is
computed during compilation and the initialization of the heap should
always result in a heap large enough to hold the static strings.  Nothing
obvious comes to mind for why this is dying under FreeBSD.

testing harmonic
---
> basis/IntInf.c 577: assert(&dp[dsize] <= &spbp->limbs[spbp->card - 1]) 

That's an assert in the implmentation of IntInf.{quot,rem}.  As best I can
make out, it's asserting that the denominator and the space in which to
allocate the result aren't overlapping.  Don't know why that's asserting.

testing math
71c71
< test12n       OK
---
> test12n       WRONG

That's checking that Real.==(Math.pow(3.0, 1000000.0),Real.posInf) is
true.  It's implemented by a C-call to math.h's pow function.  I wouldn't
imagine that would be different under FreeBSD.

testing real.maxFinite
compilation of real.maxFinite failed

This one looks simple: 
/src/runtime/basis/Real_const.S has a
#elif (defined (__linux__))
that wasn't changed to 
#elif (defined (__linux__) || defined (__FreeBSD__))
so the constants defined in that file never get compiled.


Sorry, not much help here.  And you didn't even get to signals and
threads, which are ususally the worst.

-Matthew



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Caffeinated soap. No kidding.
http://thinkgeek.com/sf
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel