[MLton] Few problems with porting...

[Anoq of the Sun]

Matthew Fluet wrote:
> Not a bug report, per se, but some forward looking advice: I would suggest
> pointing a prominent pointer in the changelog/release notes of the next
> public release to the Deviations of SML/NJ from Standard ML section of the
> user's guide.  In playing around with the new front-end on some old ML
> code I had, I ran into a (documented) deviation in SML/NJ, which I at
> first thought might be a MLton bug.

I just read this section of the user's guide. I also found something else
in SML/NJ 110.0.7 - not sure if it's a bug which has been fixed by now:

datatype List = list

val a = nil : 'a List

fails to type check.

So datatype replication does not seem to work properly in SML/NJ.

I just tried it quickly at some point and never looked into it
any further. I mean, I didn't look into the SML/NJ issues any
further - but I believe I know what the definition says here,
rule 18 p. 23 - no new t is introduced in \theta as in rule
28 p. 25 :)


Cheers
-- 
http://www.HardcoreProcessing.com