Software Assurance Reference Dataset

[bearophile via Digitalmars-d]

Tobias Müller:

> Wouldn't it be more useful to have a modified/annotated return 
> statement for that?

I don't know.


> Tail-recursiveness is an implementation detail, for the user of 
> the function it's not really interesting.

Yes, in theory a @tailrec annotation doesn't change the mangling 
of the function. But D has exceptions and stack traces. A stack 
trace for an annotated function is different because there are 
less stack frames...


> Except for the fact that it has bounded stack size which is a 
> useful property by itself and not only for tailrecursive 
> functions.

Yes, that's why I have said that a @continuation is a more 
general solution than @tailrec.

Bye,
bearophile