Gabriel Baraldi

I agree with Kristoffer (I can't make that triage time :( ). This is a bug, not a behaviour change we are doing willy nilly, if someone expects codeunits do...

IMO triages change is more breaking and more subtle because it seems it would basically that that codeunits doesn't take an AbstractString as it's argument but something like a `AbstractDenseString`

For reference Java also follows the quota/period variables. So there might be some discussion warranted, but I don't think it's a big priority.

I'm running creduce on it ;)

Creduce didn't do it right, but looking at it. We put some garbage in the GC roots. The function is `+(FixedSizeArrays.FixedSizeArray{FixedSizeArrays.FixedSizeArray{Float64, 1}, 1}, FixedSizeArrays.FixedSizeArray{FixedSizeArrays.FixedSizeArray{Float64, 1}, 1})` ```llvm ; ModuleID =...

I've minimized this a bit further with ```julia @inline function mybroadcast_preserving_zero_d(f, As...) bc = Base.broadcasted(f, As...) r = Base.materialize(bc) GC.gc() return length(axes(bc)) == 0 ? fill!(similar(bc, typeof(r)), r) : r...

I'm a bit confused, because the array you get is empty ```julia julia> a = rand(UInt8, 5) 5-element Vector{UInt8}: 0xda 0x73 0x44 0x3e 0xb5 julia> str = String(a) "\xdasD>\xb5" julia>...

I believe the issue is that ->maxsize shouldn't be reset. The others should be reset.

I think your previous change (only maxsize) is more correct here. Because the buffer hasn't been freed yet, we've just resized the array, the memory is still there.

So we are talking about the case at the bottom where we just make a copy.