I don’t strictly think this is a bug in jRuby, exactly, since it breaks across environments. However I’m having a hard time reasoning about what could be going on. And it’s happening on a server that’s hard to debug.

Environments

The following behavior is different across two environments running jRuby 9.1.8.0:

It works on:

jruby 9.1.8.0 (2.3.1) 2017-03-06 90fc7ab Java HotSpot(TM) 64-Bit Server VM 25.60-b23 on 1.8.0_60-b27 +jit [darwin-x86_64]

OSX: (Darwin Ilarnek.local 15.6.0 Darwin Kernel Version 15.6.0: Fri Feb 17 10:21:18 PST 2017; root:xnu-3248.60.11.4.1~1/RELEASE_X86_64 x86_64)

and fails on:

jruby 9.1.8.0 (2.3.1) 2017-03-06 90fc7ab Java HotSpot(TM) 64-Bit Server VM 25.51-b03 on 1.8.0_51-b16 +jit [linux-x86_64]

Amazon Linux: (Linux export-service-worker-0bc3106f0a0098066 4.1.10-17.31.amzn1.x86_64 #1 SMP Sat Oct 24 01:31:37 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux)

In both cases I’m running a simple irb with JRUBY_OPTS=''

Expected Behavior

On the Darwin machine, it behaves the same way MRI does:

irb(main):003:0> system('pwd', chdir: '/etc', out: '/tmp/java_issue') && File.read('/tmp/java_issue')
=> "/etc\n"

Actual Behavior

In the Linux environment, it fails to write stdout to the file (though it still returns true, it is the File.read that is throwing the exception.)

irb(main):001:0> system('pwd', chdir: '/etc', out: '/tmp/java_issue') && File.read('/tmp/java_issue')
Errno::ENOENT: No such file or directory - /tmp/java_issue
        from org/jruby/RubyIO.java:3583:in `read'
        from (irb):1:in `<eval>'
        from org/jruby/RubyKernel.java:1000:in `eval'
        from org/jruby/RubyKernel.java:1298:in `loop'
        from org/jruby/RubyKernel.java:1120:in `catch'
        from org/jruby/RubyKernel.java:1120:in `catch'
        from /usr/local/rbenv/versions/jruby-9.1.8.0/bin/irb:13:in `<main>'

I may try to match Java versions, just to see if it is a bug that has been patched, but it is puzzling.