>________________________________
>From: David Simcha <[email protected]>
>To: Discuss the phobos library for D <[email protected]>
>Sent: Friday, April 22, 2011 6:58 PM
>Subject: Re: [phobos] Time to get ready for the next release
>
>
> On 4/22/2011 5:05 PM, Steve Schveighoffer wrote:
>> How do you automate 4182? That is, how does the wrapper know that I want
>> the result to be covariant in derived classes?
>>
>>
> Right, I agree with you that this is a bug and should be fixed. I am not
> sure, however, how it relates to the operator overloading issue.
OK, I looked over it again, and it's actually a bug that affects my workaround
for the other bug. Essentially, in order to work around the fact that I can't
have templates in my interfaces, I aliased the virtual functions to opCat,
etc. But that did not pass on the covariance.
However, thinking about it, it would be nice if I could simply alias the
operators to virtual functions. Hm... Would something like this make sense?
class C
{
C opCatAssign(C other) {...}
template opOpAssign(string s) if s == "~="
{
alias opCatAssign opOpAssign;
}
}
This would be awesome, because then we have a way to get all our virtual
functions without having to incur template bloat, or rely on inlining. I can
see how this would be useful for opDispatch.
It would also make for a super-easy mixin implementation. In fact, you
wouldn't need to deal with parameters at all.
But in this case, 4182 would need to be fixed ;)
I need to file another enhancement request...
-Steve
_______________________________________________
phobos mailing list
[email protected]
http://lists.puremagic.com/mailman/listinfo/phobos
