rib 0.1.0 → 0.9.0

data/2011-02-28.md

@@ -1,203 +0,0 @@
-[original text](http://blogger.godfat.org/2011/02/ripl-rc-1-ripl-irb-replacement.html)
-
-# ripl, an irb replacement
-
-<p></p><a href="http://www.urbandictionary.com/define.php?term=TL%3BDR">TL;DR</a>: see the <a href="#comparison">comparison table/list</a> on
-the bottom of this post. Still TL;DR?
-<pre><code>&gt; gem install ripl
-&gt; ripl rc</code></pre>
-<img src="https://github.com/godfat/ripl-rc/raw/ripl-rc-0.1.3/screenshot.png"/>
-
-Use it in rails console?
-<pre><code>&gt; gem install ripl-rails
-&gt; ripl rc rails</code></pre>
-
-# ripl
-
-Days ago when I was using <a href="https://github.com/banister/pry">pry</a> to debug heroku-scaler,
-(which is a <a href="https://github.com/Ramaze/ramaze">ramaze</a> and <a href="https://github.com/igrigorik/em-http-request">em-http-request</a> application),
-I wanted to make pry use irb history to easy debugging,
-and then I found it was talking about something like
-pry is not an irb replacement, to find an irb replacement,
-see <a href="https://github.com/cldwalker/ripl">ripl</a>. That's where I started playing around with ripl.
-
-At first, I didn't think I need an replacement for irb,
-because it worked for me. But as always, we don't know
-if we need a better tool until we're really using a better
-tool. ripl is this kind of story for me, and so do <a href="https://github.com/lsegal/yard">yard</a> to
-<a href="https://github.com/rdoc/rdoc">rdoc</a> and so on so forth. Interestingly, this is not always
-true for all tools. For example, I like <a href="https://github.com/jimweirich/rake">rake</a> better than
-<a href="https://github.com/wycats/thor">thor</a> (though I didn't really try thor), and didn't really
-need the power of <a href="http://www.kuwata-lab.com/erubis/">erubis</a> over erb.
-
-One can think that, using thor/erubis might force others
-to do the same (though erubis could be used as a drop-in
-replacement for erb), but irb/ripl is just a personal tool
-instead of really a library that once you used it in the
-application, all developers would be forced to use or test
-against it, too. Just like a text editor. Personal taste only..
-
-Back to topic...
-
-So what would we benefit from switching irb to ripl? If
-we're only using the core functionality of ripl, then not
-even talking about gains, but we'll lose some features
-that irb provides but ripl doesn't. What good is, according
-to the <a href="https://github.com/cldwalker/ripl/blob/v0.3.2/README.rdoc">README</a> of ripl, the code size was about ~270
-lines vs 5,000+ lines, so we can expect that ripl is a lot
-easier to extend and customize, and this is also a fact.
-
-ripl is designed to be extended and customized from the
-beginning, and since the code is very modular and
-lightweight, it's very easy to replicate irb's behavior
-and features that are missing in ripl. You can take a look
-at ripl-irb. It might grow to be fat after installing many
-plugins and extensions, but I guess it will never reach
-the size of irb, since irb *has its own ruby parser*...
-<a href="https://github.com/janlelis/ripl-multi_line">ripl-multi_line</a> uses a trick that catches syntax error
-exception to achieve multiline support. This might be
-tricky because it depends on the error message coming
-from syntax error; however irb's own ruby parser might
-not be accurate as well... as long as it's not using ruby's
-own parser.
-
-At first I was trying to add some simple patches with only
-a few of lines to the core of ripl to make it better, but the
-author only accepted my bug fix patches instead of features
-patches, even it's only a few of lines and I don't see why
-others won't want that features. Sometimes this make me
-feel like vim or bash, you'll definitely want some personal
-(might not be so personal though) config files instead of
-using the defaults, which is very different than GUI
-applications, which always try to provide the best defaults.
-I think both have their strength, and I don't mind as long
-as it could be customized to suit my flavor, and that's why
-I've written <a href="https://github.com/godfat/ripl-rc">ripl-rc</a>.
-
-ripl-rc is a ripl plugins collection, each of its require is a
-different plugin. For example:
-<pre><code>require 'ripl/rc/color'
-require 'ripl/rc/anchor'</code></pre>
-The first line would enable colorizing plugin, and the latter
-would enable pry session like plugin. The reason I do this
-instead of releasing ripl-color, ripl-anchor, etc., is I think
-it's a lot easier to maintain and install as a plugin collection,
-since each plugin has only a few of codes. We can take
-advantage of fixing plugins at the same time, too, since
-some plugins would work with other plugins. (e.g.
-ripl/rc/anchor and ripl/rc/color) It would be tedious for
-users and developers to release or update a bunch of new
-releases... I really hate rails doing that, though the code
-base is not the same level. It might be a must for rails
-somehow...
-
-Enough of rubbish, let's see the comparison.<a id="comparison"></a>
-
-# Comparison
-
-ripl has:
-
-* a lot fewer codes and is a lot easier to customize
-* better auto-complete (from <a href="https://github.com/cldwalker/bond">bond</a>)
-
-irb has:
-
-* multiline support (ripl uses ripl-multi_line)
-* subsessions and workspaces (actually I don't know
-  what are they, never used. ripl uses <a href="https://github.com/cldwalker/ripl-commands">ripl-commands</a>)
-
-ripl-rc has, upon session ends:
-
-* <span style="color:gold">require 'ripl/rc/squeeze_history'</span>
-
-  which squeezes the same input in history, both in memory
-  and history file.
-
-* <span style="color:gold">require 'ripl/rc/mkdir_history'</span>
-
-  which calls `mkdir -p` on directory which contains history
-  file. For example, I put my irb_history in an directory
-  might not exist before use: `~/.config/irb/irb_history`
-
-* <span style="color:gold">require 'ripl/rc/ctrld_newline'</span>
-
-  ruby 1.9.2 has no this problem in irb, but 1.8 and ripl do.
-  When hitting ctrl+d to exit ripl, it would print a newline
-  instead of messing up with shell prompt.
-
-upon formatting output:
-
-* <span style="color:gold">require 'ripl/rc/strip_backtrace'</span>
-
-  ripl prints the full backtrace upon exceptions, even the
-  exceptions come from interactive environment, making it
-  very verbose. This ripl plugin strips those backtrace.
-
-* <span style="color:gold">require 'ripl/rc/color'</span>
-
-  There's ripl-color_result that make use of <a href="https://github.com/michaeldv/awesome_print">awesome_print</a>,
-  <a href="http://coderay.rubychan.de/">coderay</a>, or <a href="https://github.com/janlelis/wirb">wirb</a>. The problem of awesome_print is it's too
-  awesome and too verbose, and the problem of coderay and
-  wirb is that they are both parser based. In ripl, this should
-  be as simple as just print different colors upon different
-  objects, instead of inspecting it and parsing it.
-
-  ripl/rc/color just uses a hash with Class to color mapping
-  to pick up which color should be used upon a ruby object.
-
-  To customize the color schema, inspect `Ripl.config[:rc_color]`
-
-upon input:
-
-* <span style="color:gold">require 'ripl/rc/multiline'</span>
-
-  I need some modification on ripl-multi_line to make prompt
-  work better, but not sure if I can come up a good fix and
-  try to convince the author to accept those patches. So I
-  just bundle and maintain it on my own. If you're using
-  ripl-rc, you could use this plugin, otherwise, keep using
-  ripl-multi_line.
-
-* <span style="color:gold">require 'ripl/rc/eat_whites'</span>
-
-  irb will just give you another prompt upon an empty input,
-  while ripl would show you that your input is nil. I don't like
-  this, because sometimes I'll keep hitting enter to separate
-  between inspects. This plugin would skip inspect if the input
-  is empty just like irb.
-
-special tool:
-
-* <span style="color:gold">require 'ripl/rc/anchor'</span>
-
-  So this is my attempt to emulate pry in ripl. Instead
-  trying to make pry support irb_history, colorizing, etc.,
-  I think implement pry like feature in ripl is a lot easier.
-  No need to be fancy, I just need the basic functionality.
-
-  To use it, use:
-  <pre><code>Ripl.anchor your_object_want_to_be_viewed_as_self</code></pre>
-  or
-  <pre><code>Ripl.anchor binding</code></pre>
-  in your code. Other than pry ripl support, you might be
-  interested in <a href="https://github.com/cldwalker/ripl-rails">ripl-rails</a> and <a href="https://github.com/cldwalker/ripl-hijack">ripl-hijack</a>, too.
-
-about config:
-
-* <span style="color:gold">require 'ripl/rc/noirbrc'</span>
-
-  By default ripl is reading `~/.irbrc`. I don't think this
-  is what people still using irb would want, because the
-  configuration is totally different. This suppress that,
-  make it only read `~/.riplrc`
-
-for lazies:
-
-* <span style="color:gold">require 'ripl/rc'</span>
-
-  This requires anything above for you, and is what `ripl rc`
-  and `ripl rc rails` shell commands did.
-
-So that's all at the moment for <a href="https://github.com/godfat/ripl-rc/tree/ripl-rc-0.1.3">ripl-rc 0.1.3</a>. Enjoy,
-
-2010-02-28 (16:36~21:49)

data/CHANGES

@@ -1,86 +0,0 @@
-= ripl-rc changes history
-
-== ripl-rc 0.2.4 -- 2011-08-03
-
-* [ensure_after_loop] A plugin which would make sure after_loop is called.
-                      (Some plugins need to do cleaning up in after_loop)
-
-* [           anchor] Fixed a bug for buggy EditLine, which would raise
-                      weird exception upon calling `HISTORY == nil`.
-                      Now we use `.nil?` to lookup if it's nil or not.
-
-* [multiline_history] Fixed a bug when multiline evaluation raised an
-                      exception, it wouldn't handle the history correctly.
-                      Moving the handling to an ensure block fixed this.
-
-* [  squeeze_history] (Internal) Take the advantage of the new history API
-
-== ripl-rc 0.2.3 -- 2011-06-16
-
-Please read this for detail:
-http://blogger.godfat.org/2011/06/new-feature-mainly-for-anchor-in-ripl.html
-
-* [ripl-rc] Ripl.enable/disable_??? now accepts a block to ease switching.
-* [  debug] This plugin is simply calling Ripl.disable_anchor at the
-            beginning, make it easier to use [anchor] to inspect a certain
-            place under a certain condition. See the blog post for detail.
-
-== ripl-rc 0.2.2 -- 2011-06-01
-
-* [multiline_history     ] fixed multiline_history
-* [multiline_history_file] now we have persistent multiline_history
-
-== ripl-rc 0.2.1 -- 2011-04-11
-
-* [multiline] fixed history with editline (e.g. mac's build-in ruby)
-
-== ripl-rc 0.2.0 -- 2011-04-10
-
-* [  ripl-rc] tested with MRI 1.8.7, 1.9.2 and Rubinius 1.2.3, JRuby 1.6.0
-* [  ripl-rc] fixed some conflicts with bundler
-* [  ripl-rc] internal structure rearrangement
-* [  ripl-rc] all plugins now have runtime enable/disable ability
-              e.g. run `Ripl.disable_color` to disable coloring,
-                       `Ripl.enable_color` to enable again.
-* [multiline] some tweak to make it more accurate
-* [last_exception] save last exception in Ripl.last_exception
-
-== ripl-rc 0.1.5 -- 2011-03-26
-
-* [      bin] shows a better error message when ripl-rails can't be found
-* [  ripl-rc] made require 'ripl-rc' works like require 'ripl/rc'
-* [multiline] fixed regexp warning on ruby 1.8.7
-* [multiline] fixed history when using multiline
-
-== ripl-rc 0.1.4 -- 2011-03-01
-
-* [  color] fix syntax error; sorry, i definitely need tests.. :s
-
-== ripl-rc 0.1.3 -- 2011-02-28
-
-* [    bin] added `ripl rc` and `ripl rc rails` commands
-* [noirbrc] added that don't read ~/.irbrc by default
-* [ anchor] use short_inspect
-* [ anchor] fixed a bug that cannot display nil
-* [  color] fixed a bug that displaying some values as Object. see [9a64495]
-* [  color] colorize error output as well
-* [squeeze_history] max 500 items
-* [strip_backtrace] fixed cwd, don't trace if it's a syntax error
-
-== ripl-rc 0.1.2 -- 2011-02-25
-
-* [     rc] rearranged require order
-* [ plugin] added anchor, like pry. usage: Ripl.anchor(binding) # or obj
-* [ plugin] added multiline, which works better with anchor
-* [ plugin] added mkdir_history, which tries to mkdir -p on history directory
-* [strip_backtrace] fix functionality for custom name (anchor)
-
-== ripl-rc 0.1.1 -- 2011-02-24
-
-* [ plugin] added strip_backtrace
-* [general] now use Ripl::Rc::U.include(YourExtension) to customize details
-* [  color] use Ripl.config[:rc_color] to configure color schema
-
-== ripl-rc 0.1.0 -- 2011-02-23
-
-* release early, release often

data/TODO

@@ -1,6 +0,0 @@
-= ripl-rc todo list
-
-* tests for each plugin
-* doc
-
-* read multiline histroy in irb_history

data/lib/rib/core/history_file.rb

@@ -1,38 +0,0 @@
-
-require 'rib'
-require 'fileutils'
-
-module Rib::HistoryFile
-  include Rib::Plugin
-  Shell.use(self)
-
-  def before_loop
-    return super if HistoryFile.disabled?
-    config[:history_file] ||= '~/.config/rib/history.rb'
-    config[:history_size] ||= 500
-    FileUtils.mkdir_p(File.dirname(history_file))
-    super
-  end
-
-  def get_input
-    return super if HistoryFile.disabled?
-    (history << super).last
-  end
-
-  def read_history
-    return super if HistoryFile.disabled?
-    File.exist?(history_file) && history.empty? &&
-      File.readlines(history_file).each{ |e| history << e.chomp }
-  end
-
-  def write_history
-    return super if HistoryFile.disabled?
-    File.open(history_file, 'w'){ |f|
-      f.puts(history.to_a.last(config[:history_size]).join("\n")) }
-  end
-
-  private
-  def history_file
-    @history_file ||= File.expand_path(config[:history_file])
-  end
-end

data/lib/rib/more/multiline.rb

@@ -1,77 +0,0 @@
-
-require 'rib'
-
-# from https://github.com/janlelis/ripl-multi_line
-module Rib::Multiline
-  include Rib::Plugin
-  Shell.use(self)
-
-  # test those:
-  # ruby -e '"'
-  # ruby -e '{'
-  # ruby -e '['
-  # ruby -e '('
-  # ruby -e '/'
-  # ruby -e 'class C'
-  # ruby -e 'def f'
-  # ruby -e 'begin'
-  ERROR_REGEXP = Regexp.new(
-    [ # string or regexp
-      "unterminated \\w+ meets end of file",
-      # mri and rubinius
-      "syntax error, unexpected \\$end",
-      # rubinius
-      "expecting '.+'( or '.+')*",
-      # jruby
-      "syntax error, unexpected end-of-file",
-    ].join('|'))
-
-  def prompt
-    return super if Multiline.disabled?
-    if multiline_buffer.empty?
-      super
-    else
-      "#{' '*(config[:prompt].size-2)}| "
-    end
-  end
-
-  def loop_once
-    return super if Multiline.disabled?
-    catch(:multiline_cont) do
-      super
-      multiline_buffer.clear
-    end
-  end
-
-  def loop_eval(input)
-    return super if Multiline.disabled?
-    multiline_buffer << input
-    super(multiline_buffer.join("\n"))
-  end
-
-  def print_eval_error(e)
-    return super if Multiline.disabled?
-    if e.is_a?(SyntaxError) && e.message =~ ERROR_REGEXP
-      throw :multiline_cont
-    else
-      super
-    end
-  end
-
-  def handle_interrupt
-    return super if Multiline.disabled?
-    if multiline_buffer.empty?
-      super
-    else
-      line = multiline_buffer.pop
-      print "[removed this line: #{line}]"
-      super
-      throw :multiline_cont
-    end
-  end
-
-  private
-  def multiline_buffer
-    @multiline_buffer ||= []
-  end
-end

data/lib/rib/more/squeeze_history.rb

@@ -1,37 +0,0 @@
-
-require 'rib/core/history_file'
-
-module Rib::SqueezeHistory
-  include Rib::Plugin
-  Shell.use(self)
-
-  # squeeze history on memory too
-  def eval_input input
-    return super if SqueezeHistory.disabled?
-    history.pop if input.strip == '' ||
-                  (history.size > 1 && input == history.to_a[-2])
-                  # EditLine is really broken, to_a is needed for it
-    super
-  end
-
-  # write squeezed history
-  def write_history
-    return super if SqueezeHistory.disabled?
-    @history = P.squeeze_history(history)
-    super
-  end
-
-  module Imp
-    def squeeze_history history
-      history.to_a.inject([]){ |result, item|
-        if result.last == item || item.strip == ''
-          result
-        else
-          result << item
-        end
-      }
-    end
-  end
-
-  Plugin.extend(Imp)
-end