Charles Oliver Nutter

There are some legitimate failures for that change, so I think I need to modify it to only remove the path if the require failed due to an error. That...

@fxn Well that's a good question. I think it may represent a behavioral difference between how autoload and a normal require react to the in-progress load. I'll try to confirm...

I think the explanation is that autoload checks for a loaded feature *before* invoking require, and require checks for a loaded feature *after* acquiring the lock. Loaded feature checking also...

We're going to call this fixed with the merge of #8871. Only remaining failure is the known issue with canonicalizing loaded features.

I improved this in #708 but eliminating all references is tricky. Basically this is being used as a way to key a cache object per-runtime that holds references to things...

The API in question has been incubating for many years in JDK and is still experimental, but information on the ninth version of the API is here: https://openjdk.org/jeps/489 The API...

I think MRI removed a lot of warnings over the years and we have not done so in many cases.

Sorry for disappearing after dropping this on y'all. Replies below. > One downside is that this moves complexity out of code and into the build As long as the standard...

Oy vey... pretty face-palmy bug. ```diff diff --git a/ext/jruby/org/jruby/ext/strscan/RubyStringScanner.java b/ext/jruby/org/jruby/ext/strscan/RubyStringScanner.java index c0b654d..6e6501f 100644 --- a/ext/jruby/org/jruby/ext/strscan/RubyStringScanner.java +++ b/ext/jruby/org/jruby/ext/strscan/RubyStringScanner.java @@ -651,7 +651,7 @@ public class RubyStringScanner extends RubyObject { setMatched(); adjustRegisters(); -...

Trivial reproduction: ``` require 'strscan' ss = StringScanner.new("asdf\n1") ss.pos = 5 puts ss.scan_integer ``` There must not be tests of scanning for numbers at a non-zero offset.