From: Charles Oliver Nutter <headius@...>
Date: 2010-06-30T18:59:32+09:00
Subject: [ruby-core:30973] Re: [Feature #1478] Ruby archive

On Sat, May 22, 2010 at 1:21 PM, Urabe Shyouhei <shyouhei@ruby-lang.org> wrote:
> That's what I'm in doubt. ��At the time of its appearance jar needed compression
> because java aimed to be network-oriented. ��But definitely our browser don't
> run ruby.

We have users running JRuby in applets in browsers with jar-compressed
scripts. It's pretty easy to do, and very handy to distribute a single
compressed file.

We also have folks starting to use JRuby on Android, where we also
support loading scripts out of compressed jar files (since that is
again the format for all Android application archives).

And JRuby's "jruby-complete.jar" is all of JRuby plus the Ruby
standard library in a single 10-11MB file, where JRuby alone is about
8-9MB of that. It would be closer to 18-19MB if we could not compress
stdlib into the jar file, and the "complete" jar's execution makes
heavy use of loading scripts out of the jar itself, similar to what
the new Ruby archive format could enable. The "complete" jar is used
in many prepackaged JRuby applications that need a single distribution
file.

The use of jars and compressed jars is a very common, very useful tool
for JRuby users, and we've supported it for over four years. If Ruby
is going to adopt a standard archive format, it would seem rather
strange to ignore what we've done (and what users are doing) in JRuby
today.

- Charlie