stability

Robert Fraser fraserofthenight at gmail.com
Sun Feb 24 08:14:04 PST 2008


Jarrett Billingsley wrote:
> Exactly.  The "struct tupleof not able to access private members" that was 
> fixed in the most recent D2 update is a perfect example of this.  tupleof is 
> very ill-defined in the spec, so it's not entirely clear _what_ it should 
> do, so it seems kind of arbitrary to classify that ability as an enhancement 
> rather than a bug.
> 
> This is the kind of stuff that would not happen with an exhaustively 
> detailed spec. 

True about the spec. However, this is a _breaking_ change (that is, code 
that compiled frine with older versions has a slight chance of acting 
differently or breaking when this change is introduced). The same was 
also true of .init, and the same was also true when:

version(Windows) { extern(Windows): }
else { extern(C): }

was fixed, and a few other bugs. So D1 isn't non-breaking. But I sort of 
would prefer it was.



More information about the Digitalmars-d mailing list