Worst ideas/features in programming languages?

russhy russhy at gmail.com
Mon Oct 18 18:12:49 UTC 2021


On Monday, 18 October 2021 at 15:22:30 UTC, deadalnix wrote:
> On Monday, 18 October 2021 at 14:39:35 UTC, russhy wrote:
>> and then you freeze all threads and must scan all allocated 
>> pointers whenever your "gc'd malloc"'s buffer needs to grow
>>
>
> An unsolvable problem, clearly, as it is inconceivable that the 
> user could set a  heap size.
>
> Don't believe anyone that would tell you this could boils down 
> to a simple call to a function, such as GC.setHeapSize(XXXX), 
> they are clearly manipulating you.
>
>> you can't compete with highlevel languages and their sub 1ms 
>> incremental GC
>>
>
> Let me tell you a secret. most modern languages use an hybrid 
> approach as the one I described, and even old one have been 
> repurposed to use that when possible (for instance python).
>
>> AGAIN
>>
>> ASP.net team is working hard on reducing the impact of the GC 
>> in their library!!!! even thought they have a competitive GC, 
>> they are making sure they don't do useless GC allocations and 
>> are using Span/stackalloc/ValueTysk everywhere!!
>>
>
> Ho, damn, a hybrid approach that allocates on the stack or the 
> GC and frees instead of leaking. Who could have though of that?
>
>> The smart and pragmatic approach is to give tools for people 
>> to write efficient software
>>
>> Allocators is one of them
>>
>
> How the strawman screaming in between you hear doing?
>
>> Have we lost our mind focusing on the GC? i think yes
>
> With all due respect, you clearly are not able to address the 
> points made in the discussions you are participating in. It's 
> harsh, but it's true. You shouldn't be making statements about 
> others losing their mind.

You miss the point..

What problem are you trying to solve?

I'm trying to solve the problem that we don't encourage people to 
write their latency sensitive programs in D and instead fallback 
to C++ (game engines for example), because std isn't built in the 
idea that they can provide their own allocation schemes

Be pragmatic, please

Can we at least agree that the work on 
https://dlang.org/phobos/std_experimental_allocator.html must be 
resumed and moved out of experimental?




More information about the Digitalmars-d mailing list