Linus with some good observations on garbage collection

[Piotr Szturmaj]

Timon Gehr wrote:
> Andrei Alexandrescu wrote:
>> Allocation and deallocation are not symmetric and as such handling them
>> in a uniform way would be a mistake that perpetuates a poor
>> understanding of the underlying realities of memory allocation and
>> object creation. I suggest you reconsider.
>
> 'char' is completely orthogonal to memory allocation and it still gets handled
> uniformly to it syntactically. So I do not really get the point here.
>
>>
>>> And it discourages people to get informed about
>>> custom allocators etc.
>>
>> I don't see the relationship here.
>
> I cannot imagine that the syntax for them will be particularly nice after the
> removal of 'delete'. Who likes features with unnecessarily clumsy syntax?
>
>> The cost of keeping "delete" in the language is making a rarely-used,
>> dangerous facility with loosely-defined semantics straight inside the
>> language. It reflects poor language design for many reasons.
>>
>> Andrei
>
> I think we are talking about different things. I do _not_ want to use the keyword
> to somehow try to deallocate GC allocated memory. (who would actually want to do
> this anyways?) I want it for custom allocators as in
> http://www.digitalmars.com/d/2.0/memory.html. In that case, the semantics would be
> completely specified by me. Removing the keyword means ruining that use case. You
> can also find a nice, clean and strict semantic specification for delete when
> called on GC memory that works on all platforms and with all GC implementations:
> "do nothing, GC memory is managed automatically." As it is done now, I agree it is
> not particularly nice.
>
> Timon

Why not allow "delete" only in classes with custom allocators? AFAIK 
they could be determined at compile time. Please let me know if I'm wrong.