Thoughts on Backward Compatibility

[Dukc]

On Tuesday, 20 February 2024 at 09:03:15 UTC, Atila Neves wrote:
>
> In the case of DIP1000 specifically I think maybe Robert's idea 
> of moving its checks to `@trusted` may be that way forward, and 
> making `@safe` regular GC D. Once I'm done with editions I'm 
> going to write a DIP for this.

Please remember that it won't be any better than DIP1000 for 
backwards compatibility - in fact it's even worse, since:

  - With DIP1000 your slice of static array or pointer to a struct 
field will still compile if you didn't happen to escape it. With 
Roberts proposal it will always need fixing.

  - With Roberts proposal your choices for breaking code are 
either removing `@safe` or moving your variables to the heap. 
With DIP1000 you can also do either, but you also have a third 
choice: adding `scope` annotations.