Environment

JRuby version:
jruby 9.1.2.0 (2.3.0) 2016-05-26 7357c8f OpenJDK 64-Bit Server VM 25.91-b14 on 1.8.0_91-8u91-b14-1~bpo8+1-b14 +jit [linux-x86_64]
No special command line arguments.

OS:
Linux 29ea782aaebe 4.4.15-boot2docker #1 SMP Wed Jul 13 14:05:46 UTC 2016 x86_64 GNU/Linux

Expected Behavior

MRI (ruby 2.2.2p95 (2015-04-13 revision 50295) [x86_64-darwin14]):

f = "eyJzY2hlbWEiOiJpZ2x1OmNvbS5zbm93cGxvd2FuYWx5dGljcy5zbm93cGxvdy9jb250ZXh0cy9qc29uc2NoZW1hLzEtMC0wIiwiZGF0YSI6W3t9XX0"
f.size
 => 115
require 'base64'
Base64.decode64(f)
 => "{\"schema\":\"iglu:com.snowplowanalytics.snowplow/contexts/jsonschema/1-0-0\",\"data\":[{}]}"

As you can see, MRI correctly determines if a string is missing Base64 paddings and correctly decodes the string.

Actual Behavior

JRuby:

f = "eyJzY2hlbWEiOiJpZ2x1OmNvbS5zbm93cGxvd2FuYWx5dGljcy5zbm93cGxvdy9jb250ZXh0cy9qc29uc2NoZW1hLzEtMC0wIiwiZGF0YSI6W3t9XX0"
f.size
 => 115
require 'base64'
Base64.decode64(f)
=> "{\"schema\":\"iglu:com.snowplowanalytics.snowplow/contexts/jsonschema/1-0-0\",\"data\":[{}"

As you can see, JRuby doesn't detect missing paddings and incorrectly decodes the string (ignores the last three-octet group). Here's a workaround I use:

...
Base64.decode64(f + '=' * (f.size % 4))
 => "{\"schema\":\"iglu:com.snowplowanalytics.snowplow/contexts/jsonschema/1-0-0\",\"data\":[{}]}"

Base64.decode64 uses String#unpack under the hood. Here's the piece in MRI source codes responsible for the correct implementation: https://github.com/ruby/ruby/blob/trunk/pack.c#L1618