* Basic  commands
  Make sure to uninstall code zauker from your gems (gem  uninstall code_zauker)
before start developing
** Environment setup (windows)

Ensure you have Dev kit too 
http://rubyinstaller.org/downloads
https://github.com/oneclick/rubyinstaller/wiki/Development-Kit
It is for hiredis: hiredis is not mandatory, but suggested


#+begin_src sh
gem install bundler
# Dev kit installation...
#ruby /c/rubyinstallkit/dk.rb init
#ruby /c/rubyinstallkit/dk.rb install
bundle install
rake test
# Ensure dev code is reachable
export RUBYLIB=k:/code/code_zauker/lib
#+end_src


** To Run tests
#+begin_src sh
rake test
#+end_src

** To release a new version to rubygem
#+begin_src sh
rake release
#+end_src

** Dependency management
Done with ruby "bundle", you should check periodically dependency
with "bundle update" to be sure to be with latest bug fixes of dependence libs

* Notable facts
** DB Size tradeoff
If trigram size is greather then 3, the database become larger, because of less collisions.
czlist work better with 4-grams then with 3-grams (a lot less false positive) 
but the size can be 50% bigger

2-gram size rocks a lot, because of a very small db but false positive are a nightmare.
czlist give 2188 files with a "for", but grep report only 383 of them (less then 18% of success)


Emacs-lisp files spot a very huge number of trigrams


* Future/Study
To fulfill Google code options:
** Google code input
Offers regula expression search like
 ^java/.*\.java$
and also:

Package		package:linux-2.6
Language		lang:c++
File Path		file:(code|[^or]g)search
Class		class:HashMap
Function		function:toString
License		license:mozilla
Case Sensitive		case:yes

** Reference
   http://www.rubyinside.com/21-ruby-tricks-902.html


* Aws tests
** Micro instance
Without multiplexing you get 4m39.599s
for indexing code_zauker

With
 time find . -type f -print0 | xargs -0 -P 10  -n 20  ./bin/czindexer -v --redis-server awsserver
You get about 0m31.284s