On Mon, May 25, 2020 at 12:31 PM Roberto Ierusalimschy
<roberto@inf.puc-rio.br> wrote:
>
> It was an implementation change. Nothing changes to the user, as
> far as we know. (At least, no one pointed to any real change to the
> user.) Things seem to be exactly as they were in 5.3, despite all this
> noise.
I don't think that's true. In Lua 5.3, an asynchronously set hook
would always eventually interrupt an infinite loop, but in Lua 5.4,
there are some circumstances in which it never will.
Could you write a short example (lua code) that proves it?
It would be interesting to test it while 5.4 it’s still in the RC phase.
regards,
Ranier Vilela
