Fix 21484: Infinite loop in core.memory : GC.{get,set,clr}Attr(const scope void*...)

[Geod24]

Those functions never worked, and would cause an infinite recursion.
In a non-optimized build that would exhaust the stack and trigger a SEGV,
however in the optimized build we distribute, it results in an infinite
loop as tail call optimization is performed.

[dlang-bot]

Thanks for your pull request, @Geod24!

Bugzilla references

Auto-close	Bugzilla	Severity	Description
✓	21484	major	Infinite recursion in core.memory : GC.{get,set,clr}Attr(const scope void*...)

Testing this PR locally

If you don't have a local development environment setup, you can use Digger to test this PR:

dub run digger -- build "stable + druntime#3314"

[MoonlightSentinel]

Tests?

[PetarKirov]

We were talking with @Geod24 on Slack PM so that's why I was quite fast with the approval, though on second thoughts I agree with @MoonlightSentinel that it would be good to cover this with a test. Otherwise the test case for the compiler bug that @Geod24 showed me was:

extern (C) uint gc_setAttr( void* p, uint a ) pure nothrow;
​
struct Baguette
{
    static uint setAttr( const scope void* p, uint a ) nothrow
    {
        return setAttr(cast()p, a);
    }
​
    static uint setAttr(void* p, uint a) pure nothrow
    {
        return gc_setAttr( p, a );
    }
}
​
inout(void)* bug (inout(void)* ptr)
{
    Baguette.setAttr(ptr, 42);
    return ptr;
}
​
void main ()
{
    auto ptr = new int*;
    bug(ptr);
}

[Geod24]

Usually I would say "it doesn't make sense to test for infinite loop here", but since the function were not covered at all, I added a small test.

[MoonlightSentinel]

Thanks

[Geod24]

So, anything holding this back ?

[ibuclaw]

Sgtm