[MLton] Patch ready

[Matthew Fluet]

Wesley W. Terpstra wrote:
> On Feb 11, 2007, at 6:18 AM, Matthew Fluet wrote:
>> Wesley W. Terpstra wrote:
>>>> The other bug is that MLton is leaking polymorphism of string types.
>>>> This bug is not specific to my changes; svn MLton does this too.
>>> I'm using svn/HEAD now, and this still applies.
>>
>> I don't know what you mean by "leaking polymorphism".
> 
> In user code, a 'string' is the same as a 'char vector'. I don't believe 
> this equivalence is supposed to be exposed..? eg:
> val y : string = "fdsdfsdg"
> val x : char vector = y

The Basis specifies that String.string is equivalent to 
CharVector.vector.  I think that the Basis would technically have 
CharVector.vector as an abstract type (since in most implementations it 
won't be equivalent to char vector).  I seem to recall some reason that 
MLton has for exposing the equivalence, but I can't recall it at the 
present time.