Benoit Daloze

@bjfish Could you take a look?

@bjfish found that this is caused by prepending twice the same module in the same ancestor chain, in different classes. TruffleRuby currently has the traditional Ruby behavior to only allow...

Right, that explains why it only happens on macOS then, it's due to the subprocesses of https://github.com/oracle/truffleruby/blob/634c0ae2517d04261b05e357b79da009eeac62ff/lib/truffle/truffle/openssl-prefix.rb. I'll look into it (running an extra subcommand when loading RubyGems to find...

https://github.com/oracle/truffleruby/commit/8ca9852433bb4974bdf4e7306feee7e06235faec avoids creating autorelease pointers for spawning subprocesses. It would be interesting to know how long `brew --prefix` takes to see if it's worth optimizing for the common case of...

Right, so that's more like 20-30ms, seems worth optimizing the common case.

https://github.com/oracle/truffleruby/commit/3b05154d2bc1344eb161514388d8d8917b847c3e avoids that extra subprocess if Homebrew is in `/usr/local`.

We also noticed this one when running https://github.com/ruby/ruby/blob/master/benchmark/app_aobench.rb (GR-13417). PR welcome, it seems a bug in format nodes. My notes from then: This turns out to be a bug in...

Mmh, so the image is represented as a long String and heavily mutated by String#[]= (used like a buffer)? https://github.com/tario/imageruby/blob/ba38c3e4a5c8c068566501f0e73357fd219535e4/lib/imageruby/bitmap/rbbitmap.rb#L65 Also that code could probably use `setbyte` for clarity and...

So that's 1.15x (17894ms -> 15496ms) faster on CRuby when using an intuitive representation (Array) vs a String used like a buffer. So even on CRuby it seems problematic to...

We actually have a similar issue with `setbyte`: https://github.com/oracle/truffleruby/issues/2044#issuecomment-668524452 And some discussion there on possible fixes.