Support for JRuby in executable jar

[philwebb]

I noticed that Sagan often throws a no "!/" in URL exception when browsing the guides.

[philwebb]

I've made a little bit of headway with the root cause of this, but fixing the underlying problem actually makes things worse. One of the main problems is due to the fact that JarURLConnection can always look for the entry in the JAR that it is given, even if the user actually wants something in a nested JAR.

so given jar:file://myfile.jar!/n.jar!/b creating a JarFile(myfile.jar) then calling new URL(jar.getUrl(), "jar:file://myfile.jar!/n.jar!/b") will create a URL using myfile.jar as context and looking directly inside it for b. The nested n.jar is ignored.

Fixing this is relatively easy:
https://github.com/philwebb/spring-boot/tree/gh-1070

However, once this is in place sagan fails with a different error. Hitting the guides causes:

org.jruby.exceptions.RaiseException: (LoadError) load error: openssl -- java.lang.NoClassDefFoundError: org/bouncycastle/asn1/DERBoolean
    at org.jruby.RubyKernel.require(org/jruby/RubyKernel.java:1054)
    at RUBY.require(file:/Users/pwebb/projects/spring-io-site/code/sagan-site/build/libs/sagan-site.jar!/lib/jruby-complete-1.7.4.jar!/META-INF/jruby.home/lib/ruby/shared/rubygems/custom_require.rb:36)
    at RUBY.(root)(file:/Users/pwebb/projects/spring-io-site/code/sagan-site/build/libs/sagan-site.jar!/lib/jruby-complete-1.7.4.jar!/META-INF/jruby.home/lib/ruby/1.9/net/https.rb:22)
    at org.jruby.RubyKernel.require(org/jruby/RubyKernel.java:1054)

The bouncy castle ruby file shipped with JRuby using the following:

Dir[File.expand_path('bc*.jar', File.dirname(__FILE__))].each do |file|
  require File.basename(file)
end

This calls RubyDir.aref which opens the root JarFile and as a result cannot find any bc*.jar entries.

[philwebb]

There seems to be no good way to fix this in the short terms so we are probably best leaving the broken implementation in place since that at least seems to work with Sagan.

Long term we have a couple of options:

1. We could enhance the nested JAR support to unpack specific URL entries to the temp folder and work directly from the unpacked file. This would allow JRuby to operate as is.

2. We could submit a patch for JRuby so that it never creates new JarFile instances and instead always goes via a JAR URLs. We still might have problems with nested JARs (ie those inside jruby-complete) since they are probably stored as compressed entries.