1.20.2 yanked?

[MrJoy]

I see that 1.20.2 was yanked from Rubygems, and looking at the diff I don't really see an obvious reason as to why it might need to have been yanked. Could you outline what went wrong that resulted in the need to yank this?

[grosser]

change had to be reverted because it broke other peoples stuff #301

[grosser]

@pamgluss-slack @pedromartinez can tell you more / hopefully have and idea if this is good as is or should be moved forward again ... I could not reproduce either of their issues 🤷

[MrJoy]

Ahhhhh, I see what happened, and why the code misbehaved!

https://github.com/grosser/parallel/pull/299/files

Note the trailing comma on the when line. That's going to basically make Ruby read the code like so:

ppc = case RbConfig::CONFIG["target_os"]
        when /darwin[12]/, IO.popen("/usr/sbin/sysctl -n hw.physicalcpu").read.to_i
          # empty block
        when /linux/
           cores = {}  # unique physical ID / core ID combinations

So it will always execute /usr/sbin/sysctl -n hw.physicalcpu, no matter what system you're on. If /usr/sbin/sysctl isn't present, then IO.popen raises an error. If it does exist, then the output is read, and converted to an integer and used as a test value for whether that when block matches RbConfig::CONFIG["target_os"] -- which is probably never gonna happen, obviously.

[grosser]

oh wow 🤦 totally missed that ... 🎉

[grosser]

... rubocop would have found that 😞

[grosser]

welp rubocop does not care either ...

[grosser]

1.21.0

[MrJoy]

Many eyes make all bugs shallow. Happy I could be of assistance, and thank you for your prompt response!