Nick Logan

> META6.json may not even exists as such. There are other options, such as hashing the contents of e.g. `$dist.meta.hash` inside of the `id` method I showed earlier.

It also wouldn't surprise me if the solution can be implemented inside of https://github.com/rakudo/rakudo/blob/71f5c50bed11dc6eb8aacbdc1fe353bd6aa0ba0e/src/core.c/CompUnit/Repository/Distribution.pm6#L33-L37

Another thing that might be wise is to make the method an implementation detail, but also the problem of needing a way to tell if something changed doesn't require an...

I suspect they were trying to recreate these steps with `panda`

Strange, maybe there are leftover precompilation files that are not being invalidated. @zackvolz If you run the follow commands and show the output (where requested) we could rule out OO::Monitors...

Thanks @zackvolz You've clearly shown there is a problem, but I'm not sure where to investigate further. The only other thing I would ask is if you can run `perl6...

Here is the ticket from when the same thing used to be broken (although possibly in a different way) on non-windows systems: https://rt.perl.org/Public/Bug/Display.html?id=125813 > Code: > say ('a' x 970).IO.open;...

Actually that still **seems** busted on non-windows systems: > $ perl6 -e '("a" x 970).IO.open' > Failed to open file /Users/ugexe/repos/rakudo/aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa: File in block at -e line 1

So on windows `vsnprintf` returns -1 on overflow ( if it cannot write the the final NULL byte ). So it seems the error message becomes significantly large enough to...

I notice the same broken behavior on OSX (Catalina) ``` $ raku -v Welcome to Rakudo(tm) v2020.10-63-g08f5d028e. Implementing the Raku(tm) programming language v6.d. Built on MoarVM version 2020.10-31-gb13f6a24c. ```