JRuby as a system programming language

At work, we are starting to refactor some of our system utilities. The scripts interact with these basic components:

• Postgres databases
• FTP servers
• Files on various servers
• Java applications
• Specific system calls

My boss, being from an "older generation" ;-), favors Unix scripting (sed, bash, etc.) . Being a Ruby/JRuby enthusiast myself, I wanted to show my team how easy these tasks are in JRuby. While sed might take fewer characters in some cases, I think JRuby strikes an ideal match between terseness and readability.

Note that the following code will not run on your computer, at least not without setting up a Postgres database and changing the FTP server info, etc. But it will give you a good flavor for system programming in JRuby.

#

# Copyright 2007 Effectiveqa.com. All rights reserved

#



def postgres

# Accessing Postgres from JRuby

# http://ruby.scripting.ca/postgres/rdoc/



  $LOAD_PATH << 'C:/jruby-1.0.1/lib/ruby/gems/1.8/gems/postgres-pr-0.4.0/lib'
   
  require 'postgres'

  con = PGconn.connect('localhost', 5432, nil, nil, 'foo', 'postgres', 'password'

  res = con.query("select * from bar")

  puts res[0][0]

end



def ftp

  # FTP Access

  # http://ruby-doc.org/stdlib/libdoc/net/ftp/rdoc/index.html

  require 'net/ftp'

  Net::FTP.open('effectiveqa.com') do |ftp|

     ftp.login('username', 'password')

     ftp.chdir('public_html')

     ftp.list('*.css').each {|file| puts file}

     ftp.gettextfile('main.css', 'local_main.css')

  end

end



def ftp_password

  # Show how easily JRuby can handle password problem

  require 'net/ftp'

  begin

     ftp = Net::FTP.open('effectiveqa.com')

     ftp.login('effectiv', 'BADPASSWORD')

  rescue

     puts "Here is where I fix the password"

     ftp.login('username', 'password')

     ftp.chdir('public_html')

     ftp.list('*.css').each {|file| puts file}
  end

end



def file_tricks

  # Different file/directory manipulations

  # http://www.ruby-doc.org/stdlib/libdoc/fileutils/rdoc/index.html

  require 'fileutils'

  include FileUtils



  begin cd('dfsdf') rescue puts 'Oops, had a problem...' end

  chmod(0777, 'local_main.css')

  puts "Don't match!" if ! cmp('local_main.css', 'local_main_1.css')

  cp('local_main.css', 'local_main_copy.css')

  puts "Do match!" if cmp('local_main.css', 'local_main_copy.css')

  puts pwd

  rm('local_main_copy.css')

  begin rm('DOESNT_EXIST') rescue puts 'Caught one' end

end



def file_tests

  # Very nice tests

  # http://ruby-doc.org/docs/ProgrammingRuby/html/ref_m_filetest.html



  puts FileTest.directory?('example.zip') # false

  puts FileTest.size('example.zip') # 1723

  puts FileTest.file?('.') # false

  puts FileTest.file?('example.zip') # true

  puts FileTest.readable?('example.zip') # true

  puts FileTest.writable?('example.zip') # true



  include FileTest

  puts directory?('.') # true; needs include

  puts file?('.') # false

end

def use_java

  # Show off how easy it is to use Java from JRuby

  include Java



  zf = java.util.zip.ZipFile.new('example.zip')

  puts "There are #{zf.size} files in the zip."

  zf.entries.each {|entry| puts "  file: #{entry}"}

end



def split_files

  # Easy processing of DB dump files



  lines = IO.readlines('pipeFile.txt')

  lines.each {|line| puts line.split('|')[1]}

end



def system_calls

  # JRuby has full access to OS calls

  lines = `df`.split("\n")

  lines.shift   # Read off header line

  lines.each do |line|

    fields = line.split

    next unless fields[0] =~ /C:$/   # Only care about the C: drive

    percent = fields[4].to_i

    puts "Holy cow, disk almost full" if percent > 95

    puts "Disk > 50% full" if percent > 50

  end

end



# Some interesting Ruby libraries

# http://beaver.net/slides/ruby/10-easy-pieces.html

# Library doc: http://www.ruby-doc.org/stdlib/

# Pathnames: http://www.ruby-doc.org/stdlib/libdoc/pathname/rdoc/index.html

# MD5, sha1: http://www.ruby-doc.org/stdlib/libdoc/digest/rdoc/index.html

# Waking trees: http://ruby-doc.org/stdlib/libdoc/find/rdoc/index.html



puts "\n**** I can access postgres\n"

postgres

puts "\n**** I can access FTP\n"

ftp

puts "\n**** I can tell if an FTP password failed\n"

ftp_password

puts "\n**** I have all kinds of *nix file commands\n"

file_tricks

puts "\n**** I also have all kinds of cool file tests\n"

file_tests

puts "\n**** I can easily use Java code\n"

use_java

puts "\n**** I can split up Dave's db dump files\n"

split_files

puts "\n**** I can make OS calls with ease\n"

system_calls

The only shortcoming I could find with JRuby is that it doesn't support all the command line options that Ruby does. So you can see that JRuby falls behind in the terseness of its one-liner to rip a particular field out of a pipe-delimited dump file, though I would argue it is more readable.

#> ruby -n -a -F"\|" -e "puts $F[1]" pipeFile.txt

#> perl -n -a -F"\|" -e 'print "$F[1]\n"' pipeFile.txt

#> jruby -e "IO.readlines('pipeFile.txt').each {|ln| puts ln.split('|')[1]}"