<d-bugmail@puremagic.com> wrote:
> http://d.puremagic.com/issues/show_bug.cgi?id=2834
> 
> 
> Max Samukha <samukha@voliacable.com> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|
> |samukha@voliacable.com
> 
> 
> --- Comment #8 from Max Samukha <samukha@voliacable.com> 2010-11-18
> 03:39:17 PST ---
> So what is the verdict? Should we simply specify that struct
> destructors are
> not automatically called except in RAII and remove the struct-in-class
> special
> case?
> 
> BTW, there are other problems (serious IMO):
> 
> auto ss = new S[10];
> ss.length = 5;
> delete ss;
> 
> Destructors are not called on the last 5 elements.
> 
> ----
> auto ss = new S[10];
> ss ~= ss;
> delete ss;
> 
> We have a nasty problem when destructors are called on the appended
> elements
> because postblits was not run for them during append.
> 
> etc
> 
> Essentially, operations on arrays of structs with postblits/dtors
> defined are
> currently unusable.

I think this is unavoidable. Consider:

auto a = new T[5];
auto b = a[4..5];
a.length = 4;

We can't safely destroy a[4] because it's aliased. Also, since there's no concept of an owner reference vs an alias, modifying the length of b could screw up a as well.

For this and other reasons I'm inclined to withdraw this issue, and declare that since structs are value types they won't be automatically destroyed when collected by the GC or when held in arrays.

On 11/21/2010 08:20 PM, Sean Kelly wrote:
> <d-bugmail@puremagic.com>  wrote:
>> http://d.puremagic.com/issues/show_bug.cgi?id=2834
>>
>>
>> Max Samukha<samukha@voliacable.com>  changed:
>>
>>             What    |Removed                     |Added
>> ----------------------------------------------------------------------------
>>                   CC|
>> |samukha@voliacable.com
>>
>>
>> --- Comment #8 from Max Samukha<samukha@voliacable.com>  2010-11-18
>> 03:39:17 PST ---
>> So what is the verdict? Should we simply specify that struct
>> destructors are
>> not automatically called except in RAII and remove the struct-in-class
>> special
>> case?
>>
>> BTW, there are other problems (serious IMO):
>>
>> auto ss = new S[10];
>> ss.length = 5;
>> delete ss;
>>
>> Destructors are not called on the last 5 elements.
>>
>> ----
>> auto ss = new S[10];
>> ss ~= ss;
>> delete ss;
>>
>> We have a nasty problem when destructors are called on the appended
>> elements
>> because postblits was not run for them during append.
>>
>> etc
>>
>> Essentially, operations on arrays of structs with postblits/dtors
>> defined are
>> currently unusable.
>
> I think this is unavoidable. Consider:
>
> auto a = new T[5];
> auto b = a[4..5];
> a.length = 4;
>
> We can't safely destroy a[4] because it's aliased. Also, since there's
> no concept of an owner reference vs an alias, modifying the length of b
> could screw up a as well.
>
> For this and other reasons I'm inclined to withdraw this issue, and
> declare that since structs are value types they won't be automatically
> destroyed when collected by the GC or when held in arrays.

I agree that correct automatic struct destruction is impossible without significant changes to arrays/slices/GC.