seaofclouds friends

[C] On A Java-like function split

2008-09-17T18:54:34+00:00

## split.h
#ifndef _SPLIT_H_
#define _SPLIT_H_

char * string_create(char *start, char *end);
char ** split(char *string, char seperator, char escape);

#endif

## split.c
#include <stdlib.h>
#include <string.h>
#include "array.h"

char * string_create(char *start, char *end)
{
    char *result = malloc(sizeof(char) * (end - start + 1));

    strncpy(result, start, end - start);
    
    return result;
}

char ** split(char *string, char seperator, char escape)
{
    char **result;
    int escape_count = 0;
    char *string_offset = string;
    char seperation_character = seperator;
    Array *array = array_init();
    
    while (*string++) {
        if (*string == escape) {
            escape_count++;
            seperation_character = escape;
        }
        
        if (*string == seperation_character || *string == '\0') {
            array_push(array, string_create(string_offset, string));
            string_offset = string + 1;
            
            if (escape_count > 1) {
                escape_count--;
                seperation_character = seperator;
            }
        }
    }
    
    result = array_to_char_array(array);
    array_free(array);    
    
    return result;
}

## array.h
#ifndef _ARRAY_H_
#define _ARRAY_H_

typedef struct _array {
    char **elements;
    int size;
} Array;

Array * array_init();
Array * array_push(Array *array, char *value);
char ** array_to_char_array(Array *array);
void array_free(Array *array);

#endif

## array.c
#include <stdlib.h>
#include <string.h>
#include "array.h"

Array * array_init()
{
    Array *array = malloc(sizeof(Array));

    array->elements = NULL;
    array->size = 0;
    
    return array;
}

Array * array_push(Array *array, char *value)
{
    array->elements = realloc(array->elements, sizeof(char *) * ++array->size);
    array->elements[array->size - 1] = value;
    
    return array;
}

char ** array_to_char_array(Array *array)
{
    char **result = malloc(sizeof(char *) * array->size + 1);
    
    memcpy(result, array->elements, sizeof(char *) * array->size);
    
    return result;
}

void array_free(Array *array)
{
    free(array->elements);
    free(array);
}

[Ruby] On Horrible loop

2008-09-17T16:34:07+00:00

@feedbacks.each do |feedback|
  data.each do |key,value|
    question = feedback.send("question#{key + 1}")
    value[question.to_i - 1] += 1 if question
  end
end

[Ruby] On sort_by_multiple

2008-09-05T18:59:27+00:00

Are you using Ruby 1.9 by any chance? Otherwise you should be getting an OrderedHash. Here's the relevant code from ActiveSupport:

module Enumerable
  # Ruby 1.8.7 introduces group_by, but the result isn't ordered. Override it.
  remove_method(:group_by) if [].respond_to?(:group_by) && RUBY_VERSION < '1.9'
 
  def group_by
    inject ActiveSupport::OrderedHash.new do |grouped, element|
      (grouped[yield(element)] ||= []) << element
      grouped
    end
  end unless [].respond_to?(:group_by)

  ...

[Ruby] On sort_by_multiple

2008-09-05T18:11:13+00:00

That error is normal under that circumstance. OrderedHash works with arrays, not hashes.

>> ActiveSupport::OrderedHash.new([[1,2], [3,4]])
=> [[1, 2], [3, 4]]

However, there shouldn't be a Hash going into the class from the specs you posted.

[Ruby] On sort_by_multiple

2008-09-05T17:29:03+00:00

How strange. It passes all your specs here. Using ActiveSupport 2.1 and Ruby 1.8.5.

[Ruby] On sort_by_multiple

2008-09-05T15:57:49+00:00

This is quite a bit faster on my system.

## Sorter
require "activesupport"

class Sorter < ActiveSupport::OrderedHash
  def initialize(array, &block)
    block ||= lambda { |element| element }
    super(array.map { |element| block.call(element).clone }.group_by { |element| element.delete_at(0) })
  end
  
  def sort(*directions)
    keys.sort(&method(directions.shift)).inject([]) do |result,key|
      next [ result, key ].flatten if self[key].flatten.empty?
      
      Sorter.new(self[key]).sort(*directions).each do |array|
        result << [ key, *array ].flatten
      end
      
      result
    end
  end
  
  private
    def ascending(a, b)
      a <=> b
    end
    
    def descending(a, b)
      b <=> a
    end
end

## Enumerable
module Enumerable
  def sort_by_multiple(*directions, &block)
    Sorter.new(self, &block).sort(*directions)
  end
end

[Ruby] On Rails Active Record Time Summary

2008-09-03T16:21:26+00:00

This is completely untested since I don't have a good environment to play in, but this should hopefully give you a pretty good idea of where you can go with this. Use mixins if you are performing the same operation on more than one model.

class Summary < Struct.new(:klass, :options)
  def range
    options[:range] || options[:from]..options[:to]
  end
  
  def step
    options[:step]
  end

  def calculate(method, *args)
    klass.between(range).by_timestamp(step).send(method, *args)
  end
  
  def method_missing(method, *args, &block)
    records = Hash[calculate(method, *args).map(&:reverse).flatten]
    range.map { |time| records[time.to_i].to_f }
  end
end
    
class Order < ActiveRecord::Base
  named_scope :between, lambda { |range| { :conditions => { :created_at => range } } }
  named_scope :by_timestamp, lambda { |step| { :group => "round(UNIX_TIMESTAMP(created_at)/#{step})" } }
  
  def self.summary(options)
    Summary.new(self, options)
  end
end    

# Example usage:
#
# Order.summary(:from => Time.now, :to => 6.hours.from_now, :step => 10).sum(:amount)
# Order.summary(:from => Time.now, :to => 6.hours.from_now, :step => 10).count(:amount)

[Ruby] On A Rails rake task to lookup status codes.

2008-09-03T05:14:34+00:00

task :status_codes do
  require 'action_controller/status_codes'
  include ActionController::StatusCodes  
  
  SYMBOL_TO_STATUS_CODE.sort_by(&:last).map do |symbol,status|    
    "%-4s :%-32s %s\n" % [ status, symbol, STATUS_CODES[status] ]
  end.grep(/(^|:)#{ENV['CODE'] || ENV['code']}/i).display
end

[Ruby] On Too many assignments and conditional statements

2008-08-29T21:11:06+00:00

Edit: Double post.

[Ruby] On Too many assignments and conditional statements

2008-08-29T21:03:20+00:00

I couldn't quite discern what exactly you are trying to do. I'm guessing there is a much better way to do it. But based on what I could gather:

class Commentor
  def comment_type
    self.class.to_s.downcase
  end
  
  def comment
    comments.find_by_comment_type(comment_type)
  end
  
  def build_comment(*args)
    comments.build(*args)
  end
  
  def add_comment(text)
    if comment
      comment.text += text
    else
      build_comment(:text => text, :comment_type => comment_type)
    end
  end
end

class Header < ActiveRecord::Base
  include Commentor
  has_many :comments
end

class Project < ActiveRecord::Base
  include Commentor
  has_one :comment
end

class Comment < ActiveRecord::Base
  belongs_to :header
  belongs_to :traslation
end

# Usage:
# translation.add_comment("This is a comment")

[Ruby] On Code to detect the web browser

2008-08-28T22:02:18+00:00

@Nicolás - I don't think the detect method does what you think it does. You can do the following, but I decided I liked the other form better.

def test_web_browser
  %w(Firefox/3 Firefox/2 MSIE\ 6 MSIE\ 7 Opera).detect do |agent|
    request.user_agent =~ Regexp.new(agent)
  end.gsub(/\W+/, "")
end

[Ruby] On simplify multiple gsub

2008-08-28T18:50:37+00:00

class Normalizer < Struct.new(:string)
  REPLACEMENTS = {
    0x92 => "&#39;",
    0x91 => "&#39;",
    0x96 => "-",
    0xAE => "(R)",
    0x99 => "(TM)",
    0xBD => "1/2",
    0xBE => "3/4",
    0xBC => "1/4",
    0xB0 => " degree ",
    0xBA => " degrees "
  }
  
  def normalize
    string.enum_for(:each_byte).map { |char| REPLACEMENTS[char] || char.chr }.join
  end
  
  def normalize!
    string.replace(normalize)
  end
end

# Normalizer.new(data).normalize!

[Ruby] On Code to detect the web browser

2008-08-28T18:32:22+00:00

def test_web_browser
  catch(:match) do
    ["Firefox/3", "Firefox/2", "MSIE 6", "MSIE 7", "Opera"].each do |ua|
      throw(:match, ua.gsub(/[^a-z0-9]/i, "")) if request.user_agent =~ Regexp.new(ua)
    end
    nil
  end
end

[Ruby] On simplify multiple gsub

2008-08-27T13:42:51+00:00

Perhaps I wasn't clear with my previous post. The normalize method does the same thing you are doing with gsub.

"\xBD + \xBC".chars.normalize.to_s
=> "1/2 + 1/4"

[Ruby] On simplify multiple gsub

2008-08-27T04:07:04+00:00

Why not use ActiveSupport::Multibyte::Chars instead?

data.chars.normalize!

[Ruby] On Code transactions

2008-08-20T14:06:04+00:00

require 'transaction/simple/group'

def transact(objects, action)
  group = Transaction::Simple::Group.new(*objects)
  group.start_transaction
  
  if objects.all? { |object| object.send(action) }
    group.commit_transaction
  else
    group.abort_transaction
  end
end

[Ruby] On Link to if else simple condition.

2008-08-20T13:51:49+00:00

You should be using named routes, not hashes, for URLs. After you change that it won't be so bad to have just a plain old if statement in your view.

## View [html_rails]
<span class="bluelinks-allsmall">
	<% if user.normal? %>
		<%= link_to "Normal", formatted_howto_path(howto, :normal) %>
	<% else %>
		<%= link_to "Executive", formatted_howto_path(howto, :executive) %>
	<% end %>
</span>

[Ruby] On Method for sorting

2008-08-20T05:21:15+00:00

Here is your sort algorithm refactored in a more Ruby-like fashion. Hopefully it makes it much clearer as to what is going on.

def sort(array)
  array.inject([]) do |sorted,element|
    small, large = sorted.partition { |item| item < element }
    (small << element) + large
  end
end

[Ruby] On Default value for user column

2008-08-20T04:22:14+00:00

I disagree with the others that it belongs in the model. This is most certainly display logic and would traditionally belong in a helper method. Except helper methods are rather ugly for this kind of thing. Thus, I give you a (basic) presenter:

## Helper
module ApplicationHelper
  def present(object)
    object.class.const_get("Presenter").new(object)
  end
end

## Presenter
class User::Presenter < Struct.new(:user)
  def city
    user.city.blank? ? "Not specified" : user.city.capitalize
  end
  
  def method_missing(method, *args, &block)
    user.send(method, *args, &block)
  end
end

## View [html_rails]
<%= present(@user).city %>

[Ruby] On Method for sorting

2008-08-19T17:43:24+00:00

Although I understand that you are probably trying to learn the logic behind sorting, Ruby has it's own built-in sort method. So, from a refactoring point of view, all you need is something like this:

puts "Enter words, blank row quits."        

while !(line = gets.chomp).empty?
  (lines ||= []) << line
end

puts "sorted: #{lines.sort}"

[Ruby] On How to benchmark

2008-08-19T15:43:59+00:00

Use script/performance/benchmarker to call each version. However, I would suggest that you only worry about the quality of the code here. The performance differences will be negligible.

Also, I present to you an additional refactoring:

def self.favorite?(object)
  if reflection = object.class.reflect_on_association(:favorites)
    send("find_by_#{reflection.primary_key_name}", object)
  end
end

# Example usage:
# script/performance/benchmarker 50 "User.first.favorites.favorite?(Yellowpage.first)"

[Ruby] On simple pluralize

2008-08-19T14:16:03+00:00

<%= pluralize(@resource.favorites.length, "Favorite") %>

[Ruby] On CSV turn in to an array of hashes

2008-08-15T13:23:37+00:00

require "csv"
require "enumerator"

def fix_send_to_array(recipients)
  CSV.parse_line(recipients).enum_slice(2).map do |name,email|
    { "name" => name, "email" => email }
  end
end

[Ruby] On A shorter way to declare class constants?

2008-08-14T20:13:05+00:00

class Report < ActiveRecord::Base
  %w[hits uniques].each { |const| const_set(const.upcase, const) }
end

[Ruby] On Split an array into half

2008-08-14T14:03:15+00:00

require "enumerator"

artists = [ 1, 2, 3 ]
column_1, column_2 = artists.enum_slice((artists.size / 2.0).round).to_a