zena 1.0.0.beta3 → 1.0.0.rc1

data/lib/zena/acts/serializable.rb

@@ -75,7 +75,7 @@ module Zena
         end
       end
 
-      module NodeMethods
+      module ModelMethods
         def to_xml(options = {}, &block)
           options = default_serialization_options.merge(options)
           serializer = XmlNodeSerializer.new(self, options)
@@ -83,7 +83,7 @@ module Zena
         end
 
         def default_serialization_options
-          { :only       => %w{node_name created_at updated_at log_at event_at kpath ref_lang fullpath position},
+          { :only       => %w{created_at updated_at log_at event_at kpath ref_lang fullpath position},
             :methods    => %w{v_status klass},
             :properties => export_properties,
             :ids        => export_ids,
@@ -93,10 +93,10 @@ module Zena
         end
 
         def export_properties
-          load_roles!
           res  = {}
           prop = self.prop
-          schema.column_names.each do |key|
+          schema.safe_columns.each do |col|
+            key = col.name
             value = prop[key]
             next if value.blank?
             res[key] = value
@@ -112,16 +112,15 @@ module Zena
           res   = {}
           roles = {}
 
-          sql = [%Q{SELECT nodes.zip, links.relation_id, links.source_id FROM nodes,links WHERE ((nodes.id = links.target_id AND links.source_id = ?) OR (nodes.id = links.source_id AND links.target_id = ?) OR (nodes.id = ? AND links.id = 0)) AND #{secure_scope('nodes')}}, self.id, self.id, self.parent_id]
+          # Yes, UNION is the fastest way to do this (54k data, 600 links). UNION = 0.02s, OR = 124s !!
+          sql = [%Q{SELECT nodes.zip, links.relation_id, links.source_id FROM nodes INNER JOIN links ON nodes.id = links.target_id AND links.source_id = ? WHERE #{secure_scope('nodes')} UNION SELECT nodes.zip, links.relation_id, links.source_id FROM nodes INNER JOIN links ON nodes.id = links.source_id AND links.target_id = ? WHERE #{secure_scope('nodes')}}, self.id, self.id]
 
           Zena::Db.select_all(sql).each do |record|
-            if record['relation_id'].to_i > 0
-              (roles[record['relation_id']] ||= []) << record
-            else
-              res["parent_id"] = record['zip']
-            end
+            (roles[record['relation_id']] ||= []) << record
           end
 
+          res["parent_id"] = parent_zip.to_s
+
           roles.each do |relation_id, records|
             relation = secure(Relation) { Relation.find(relation_id) }
             records.each do |record|

data/lib/zena/app.rb

@@ -1,5 +1,3 @@
-require File.join(File.dirname(__FILE__), 'info')
-
 Zena::Use.module  [
   Zena::Use::Action,
   Zena::Use::Authlogic,

data/lib/zena/code_syntax.rb

@@ -150,7 +150,7 @@ class ZafuTokenizer < Syntax::Tokenizer
         start_group :value, v
       end
       start_group :tag, "#{trailing}>"
-    elsif dotag = scan(/<([^>]+)do\s*=([^>]+)>/)
+    elsif dotag = scan(/<(\w+)([^>]*?)do\s*=('|")([^\3]*?[^\\])\3([^>]*?)(\/?)>/)
         if dotag =~ /\A<(\w+)([^>]*?)do\s*=('|")([^\3]*?[^\\])\3([^>]*?)(\/?)>/
           start_group :tag, "<#{$1}#{$2}"
           start_group :tag, "do="

data/lib/zena/controller/test_case.rb

@@ -48,11 +48,6 @@ module Zena
           "expected not tag, but tag found matching #{match.inspect} in:\n#{target.inspect}"
       end
 
-      def err(obj)
-        obj.errors.each_error do |er,msg|
-          puts "[#{er}] #{msg}"
-        end
-      end
     end
   end
 end

data/lib/zena/core_ext/string.rb

@@ -1,5 +1,5 @@
 # encoding: utf-8
-# FIXME: do we really need to monkey patch String ?
+require 'cgi'
 
 # Avoid incompatibility with rails 'chars' version in Ruby 1.8.7
 unless '1.9'.respond_to?(:force_encoding)
@@ -13,29 +13,45 @@ unless '1.9'.respond_to?(:force_encoding)
 end
 
 class String
-  # could not name this extension 'camelize'. 'camelize' is a Rails core extension to String.
+  ALLOWED_CHARS_IN_URL      = " a-zA-Z0-9_\\."
+  # in filename, allow '-' because it does not represent a space
+  ALLOWED_CHARS_IN_FILENAME = "#{ALLOWED_CHARS_IN_URL}\\-"
+  # Everything apart from a-zA-Z0-9_.-/$ are not allowed in template paths
+  ALLOWED_CHARS_IN_FILEPATH = "#{ALLOWED_CHARS_IN_FILENAME}+/$"
+  TO_FILENAME_REGEXP = %r{([^ #{ALLOWED_CHARS_IN_FILENAME}]+)}n
+  TO_URL_NAME_REGEXP = %r{([^ #{ALLOWED_CHARS_IN_URL}])}
+
+  # Change a title into a valid url name
   def url_name
-    dup.url_name!
+    gsub(TO_URL_NAME_REGEXP) do
+      '%' + $1.unpack('H2' * $1.size).join('%').upcase
+    end.tr(' ', '-')
   end
 
-  def url_name!
-    accents = {
-      ['á',    'à','À','â','Â','ä','Ä','ã','Ã'] => 'a',
-      ['é','É','è','È','ê','Ê','ë','Ë',       ] => 'e',
-      ['í',    'ì','Ì','î','Î','ï','Ï'        ] => 'i',
-      ['ó',    'ò','Ò','ô','Ô','ö','Ö','õ','Õ'] => 'o',
-      ['ú',    'ù','Ù','û','Û','ü','Ü'        ] => 'u',
-      ['œ'] => 'oe',
-      ['ß'] => 'ss',
-    }
-    accents.each do |ac,rep|
-      ac.each do |s|
-        gsub!(s, rep)
-      end
+  # Retrieve original title from an url_name
+  def self.from_url_name(str)
+    CGI.unescape(str.tr('-', ' '))
+  end
+
+  # Change a title into a valid filename
+  def to_filename
+    gsub(TO_FILENAME_REGEXP) do
+      '%' + $1.unpack('H2' * $1.size).join('%').upcase
     end
-    gsub!(/[^a-zA-Z0-9\.\-\+ ]/,' ')
-    replace(split.join(' '))
-    gsub!(/ (.)/) { $1.upcase }
+  end
+
+  # Retrieve original title from filename
+  def self.from_filename(str)
+    CGI.unescape(str.gsub('+', '%2B'))
+  end
+
+  def url_name!
+    replace(url_name)
+    self
+  end
+
+  def to_filename!
+    replace(to_filename)
     self
   end
 
@@ -72,4 +88,16 @@ class String
     (root + path).join('/')
   end
 
+  # Limit the number of characters and append the 'readmore' argument
+  # or "…" by default if the string is longer then limit.
+  # If you limit to '20' characters, the final size will max 20 + the size
+  # of the readmore argument.
+  def limit(size, readmore = '…')
+    if self.size > size
+      self[0..(size-1)] + readmore
+    else
+      self
+    end
+  end
+
 end

data/lib/zena/db.rb

@@ -1,443 +1,11 @@
-# FIXME: we should patch the connection adapters instead of having 'case, when' evaluated each time
-# For example:
-# module ActiveRecord
-#   module ConnectionAdapters
-#     class MysqlAdapter
-#       include Zena::Db::MysqlAdditions
-#     end
-#   end
-# end
-
-
 module Zena
-  module Db
-    extend self
-
-    # constants
-    case ActiveRecord::Base.configurations[RAILS_ENV]['adapter']
-    when 'mysql'
-      NOW         = 'now()'
-      TRUE        = '1'
-      TRUE_RESULT = '1'
-      FALSE       = '0'
-
-    when 'postgresql'
-      NOW         = 'now()'
-      TRUE        = 'true'
-      TRUE_RESULT = 't'
-      FALSE       = 'false'
-
-    when 'sqlite3'
-      NOW         = "datetime('now')"
-      TRUE        = '1'
-      TRUE_RESULT = 't'
-      FALSE       = '0'
-
-    end
-
-    def set_attribute(obj, key, value)
-      obj.send("#{key}=", value)
-      execute "UPDATE #{obj.class.table_name} SET #{key}=#{quote(value)} WHERE id=#{obj[:id]}"
-      obj.send(:changed_attributes).delete(key.to_s)
-    end
-
-    def quote(value)
-      connection.quote(value)
-    end
-
-    def insensitive_find(klass, count, attributes)
-      case adapter
-      when 'postgresql'
-        cond = [[]]
-        attributes.each do |attribute, value|
-          cond[0] << (value.kind_of?(String) ? "#{attribute} ILIKE ?" : "#{attribute} = ?")
-          cond << value
-        end
-        cond[0] = cond[0].join(' AND ')
-        klass.find(count, :conditions => cond)
-      when 'sqlite3'
-        cond = [[]]
-        attributes.each do |attribute, value|
-          if value.kind_of?(String)
-            cond[0] << "lower(#{attribute}) = ?"
-            cond << value.downcase
-          else
-            cond[0] << "#{attribute} = ?"
-            cond << value
-          end
-        end
-        cond[0] = cond[0].join(' AND ')
-        klass.find(count, :conditions => cond)
-      else
-        klass.find(count, :conditions => attributes)
-      end
-    end
-
-    def quote_date(date)
-      if date.kind_of?(Time)
-        case adapter
-        when 'mysql'
-          date.strftime('%Y%m%d%H%M%S')
-        when 'postgresql', 'sqlite3'
-          "'#{date.strftime('%Y-%m-%d %H:%M:%S')}'"
-        end
-      else
-        "''"
-      end
-    end
-
-    def adapter
-      # Loads the wrong adaper when running rake tasks: ActiveRecord::Base.configurations[RAILS_ENV]['adapter']
-      @adapter ||= ActiveRecord::Base.connection.class.name.split('::').last[/(.+)Adapter/,1].downcase
-    end
-
-    def execute(*args)
-      ActiveRecord::Base.connection.execute(*args)
-    end
-
-    def update(*args)
-      ActiveRecord::Base.connection.update(*args)
-    end
-
-    def connection
-      ActiveRecord::Base.connection
-    end
-
-    def quote_column_name(column_name)
-      connection.quote_column_name(column_name)
-    end
-
-    def table_options
-      case adapter
-      when 'mysql'
-        'type=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci'
-      else
-        ''
-      end
-    end
-
-    def update_value(name, opts)
-      tbl1, fld1 = name.split('.')
-      tbl2, fld2 = opts[:from].split('.')
-      case adapter
-      when 'mysql'
-        execute "UPDATE #{tbl1},#{tbl2} SET #{tbl1}.#{fld1}=#{tbl2}.#{fld2} WHERE #{opts[:where]}"
-      when 'postgresql'
-        execute "UPDATE #{tbl1} SET #{fld1}=#{tbl2}.#{fld2} FROM #{tbl2} WHERE #{opts[:where]}"
-      when 'sqlite3'
-        execute "UPDATE #{tbl1} SET #{fld1} = (SELECT #{fld2} FROM #{tbl2} WHERE #{opts[:where]})"
-      else
-        raise Exception.new("Database Adapter #{adapter.inspect} not supported yet (you can probably fix this).")
-      end
-    end
-
-    def change_engine(table, engine)
-      case adapter
-      when 'mysql'
-        execute "ALTER TABLE #{table} ENGINE = #{engine}"
-      else
-        # do nothing
-      end
-    end
-
-    def add_unique_key(table, keys)
-      case adapter
-      when 'mysql'
-        execute "ALTER IGNORE TABLE #{table} ADD UNIQUE KEY(#{keys.join(', ')})"
-      when 'postgresql'
-        execute "ALTER TABLE #{table} ADD CONSTRAINT #{keys.join('_')} UNIQUE (#{keys.join(', ')})"
-      when 'sqlite3'
-        execute "CREATE UNIQUE INDEX IF NOT EXISTS #{([table] + keys).join('_').gsub(/[^\w]/,'')} ON #{table} (#{keys.join(', ')})"
-      else
-        raise Exception.new("Database Adapter #{adapter.inspect} not supported yet (you can probably fix this).")
-      end
-    end
-
-    # 'DELETE' depending on a two table query.
-    def delete(table, opts)
-      tbl1, tbl2 = opts[:from]
-      fld1, fld2 = opts[:fields]
-      case adapter
-      when 'mysql'
-        execute "DELETE #{table} FROM #{opts[:from].join(',')} WHERE #{tbl1}.#{fld1} = #{tbl2}.#{fld2} AND #{opts[:where]}"
-      when 'postgresql'
-        using = opts[:from].reject {|t| t == table}
-        execute "DELETE FROM #{table} USING #{using.join(', ')} WHERE #{tbl1}.#{fld1} = #{tbl2}.#{fld2} AND #{opts[:where]}"
-      when 'sqlite3'
-        execute "DELETE FROM #{table} WHERE #{fld1} = (SELECT #{fld2} FROM #{tbl2} WHERE #{opts[:where]})"
-      else
-        raise Exception.new("Database Adapter #{adapter.inspect} not supported yet (you can probably fix this).")
-      end
-    end
-
-    # Insert a list of values (multicolumn insert). The values should be properly escaped before
-    # being passed to this method.
-    def insert_many(table, columns, values)
-      values = values.compact.uniq.map do |list|
-        list.map {|e| quote(e)}
-      end
-
-      columns = columns.map{|c| quote_column_name(c)}.join(',')
-
-      case adapter
-      when 'sqlite3'
-        pre_query = "INSERT INTO #{table} (#{columns}) VALUES "
-        values.each do |value|
-          execute pre_query + "(#{value.join(',')})"
-        end
-      else
-        values = values.map {|value| "(#{value.join(',')})"}.join(', ')
-        execute "INSERT INTO #{table} (#{columns}) VALUES #{values}"
-      end
-    end
-
-    # Fetch a single row of raw data from db
-    def fetch_attribute(sql)
-      case adapter
-      when 'sqlite3'
-        res = execute(sql)
-        res.empty? ? nil : res.first[0]
-      when 'mysql'
-        res = execute(sql).fetch_row
-        res ? res.first : nil
-      when 'postgresql'
-        res = connection.select_rows(sql)
-        res.empty? ? nil : res.first.first
-      else
-        raise Exception.new("Database Adapter #{adapter.inspect} not supported yet (you can probably fix this).")
-      end
-    end
-
-    def fetch_ids(sql, attr_name='id')
-      connection.select_all(sql, "#{name} Load").map! do |record|
-        record[attr_name].to_i
-      end
-    end
-
-    def fetch_attributes(attributes, table_name, sql)
-      sql = "SELECT #{attributes.map{|a| connection.quote_column_name(a)}.join(',')} FROM #{table_name} WHERE #{sql}"
-      connection.select_all(sql)
-    end
-
-    def select_all(sql_or_array)
-      if sql_or_array.kind_of?(String)
-        connection.select_all(sql_or_array)
-      else
-        connection.select_all(Node.send(:sanitize_sql, sql_or_array))
-      end
-    end
-
-    def next_zip(site_id)
-      case adapter
-      when 'mysql'
-        res = update "UPDATE zips SET zip=@zip:=zip+1 WHERE site_id = '#{site_id}'"
-        if res == 0
-          # error
-          raise Zena::BadConfiguration, "no zip entry for (#{site_id})"
-        end
-        rows = execute "SELECT @zip"
-        rows.fetch_row[0].to_i
-      when 'postgresql'
-        res = execute("UPDATE zips SET zip=zip+1 WHERE site_id = #{site_id} RETURNING zip").first
-        if res.nil?
-          # error
-          raise Zena::BadConfiguration, "no zip entry for (#{site_id})"
-        end
-        res['zip'].to_i
-      when 'sqlite3'
-        # FIXME: is there a way to make this thread safe and atomic (like it is with mysql) ?
-        res = update "UPDATE zips SET zip=zip+1 WHERE site_id = '#{site_id}'"
-        if res == 0
-          # error
-          raise Zena::BadConfiguration, "no zip entry for (#{site_id})"
-        end
-        fetch_attribute("SELECT zip FROM zips WHERE site_id = '#{site_id}'").to_i
-      else
-        raise Exception.new("Database Adapter #{adapter.inspect} not supported yet (you can probably fix this).")
-      end
-    end
-
-    # Return a string matching the pseudo sql function.
-    def sql_function(function, key)
-      return key unless function
-      res = case adapter
-      when 'mysql'
-        case function
-        when 'year'
-          "year(#{key})"
-        when 'month'
-          "date_format(#{key},'%Y-%m')"
-        when 'week'
-          "date_format(#{key},'%Y-%v')"
-        when 'day'
-          "DATE(#{key})"
-        when 'random'
-          'RAND()'
-        end
-      when 'sqlite3'
-        case function
-        when 'year'
-          # we multiply by '1' to force a cast to INTEGER so that comparaison against
-          # numbers works.
-          "strftime('%Y', #{key})*1"
-        when 'month'
-          "strftime('%Y-%m', #{key})"
-        when 'week'
-          "strftime('%Y-%W', #{key})"
-        when 'day'
-          "DATE(#{key})"
-        when 'random'
-          'random()'
-        end
-      end
-      raise Exception.new("Database Adapter #{adapter.inspect} does not support function #{function.inspect}.") unless res
-      res
-    end
-
-    # This is used by zafu and it's a mess.
-    # ref_date can be a string ('2005-05-03') or ruby ('Time.now'). It should not come uncleaned from evil web.
-    def date_condition(date_cond, field, ref_date)
-      case adapter
-      when 'mysql'
-        case date_cond
-        when 'today', 'current', 'same'
-          "DATE(#{field}) = DATE(#{ref_date})"
-        when 'week'
-          "date_format(#{ref_date},'%Y-%v') = date_format(#{field}, '%Y-%v')"
-        when 'month'
-          "date_format(#{ref_date},'%Y-%m') = date_format(#{field}, '%Y-%m')"
-        when 'year'
-          "date_format(#{ref_date},'%Y') = date_format(#{field}, '%Y')"
-        when 'upcoming'
-          "#{field} >= #{ref_date}"
-        else
-          # '2008-01-31 23:50' + INTERVAL 1 hour
-          if date_cond =~ /^(\+|-|)\s*(\d+)\s*(second|minute|hour|day|week|month|year)/
-            count = $2.to_i
-            if $1 == ''
-              # +/-
-              "#{field} > #{ref_date} - INTERVAL #{count} #{$3.upcase} AND #{field} < #{ref_date} + INTERVAL #{count} #{$3.upcase}"
-            elsif $1 == '+'
-              # x upcoming days
-              "#{field} > #{ref_date} AND #{field} < #{ref_date} + INTERVAL #{count} #{$3.upcase}"
-            else
-              # x days in the past
-              "#{field} < #{ref_date} AND #{field} > #{ref_date} - INTERVAL #{count} #{$3.upcase}"
-            end
-          end
-        end
-      when 'sqlite3'
-        case date_cond
-        when 'today', 'current', 'same'
-          "DATE(#{field}) = DATE(#{ref_date})"
-        when 'week'
-          "strftime('%Y-%W', #{ref_date}) = strftime('%Y-%W', #{field})"
-        when 'month'
-          "strftime('%Y-%m', #{ref_date}) = strftime('%Y-%m', #{field})"
-        when 'year'
-          # we multiply by '1' to force a cast to INTEGER so that comparaison against
-          # numbers works.
-          "strftime('%Y', #{ref_date}) = strftime('%Y', #{field})"
-        when 'upcoming'
-          "#{field} >= #{ref_date}"
-        else
-          # date('2008-01-31 23:50','+1 hour')
-          if date_cond =~ /^(\+|-|)\s*(\d+)\s*(second|minute|hour|day|week|month|year)/
-            count = $2.to_i
-            if $1 == ''
-              # +/-
-              "#{field} > DATE(#{ref_date}, '-#{count} #{$3.upcase}') AND #{field} < DATE(#{ref_date}, '+#{count} #{$3.upcase}')"
-            elsif $1 == '+'
-              # x upcoming days
-              "#{field} > #{ref_date} AND #{field} < DATE(#{ref_date}, '+#{count} #{$3.upcase}')"
-            else
-              # x days in the past
-              "#{field} < #{ref_date} AND #{field} > DATE(#{ref_date}, '-#{count} #{$3.upcase}')"
-            end
-          end
-        end
-      else
-        raise Exception.new("Database Adapter #{adapter.inspect} not supported yet (you can probably fix this).")
-      end
-
-    end # date_condition
-
-    # Insert a dummy (empty) link to use when mixing queries (QueryNode) with links and without.
-    def insert_dummy_ids
-      tables = []
-      connection.tables.each do |table|
-        if table =~ /^idx_/ || table == 'links'
-          if table =~ /^idx_nodes/
-            next if fetch_attribute("SELECT node_id FROM #{table} WHERE node_id = 0")
-          else
-            next if fetch_attribute("SELECT id FROM #{table} WHERE id = 0")
-          end
-
-          # index table
-          klass = Class.new(ActiveRecord::Base) do
-            set_table_name table
-          end
-
-          dummy_hash = {}
-          klass.columns.each do |col|
-            if !col.null
-              if col.name =~ /_id$/
-                dummy_hash[col.name] = 0
-              elsif col.name != 'id'
-                dummy_hash[col.name] = ''
-              end
-            end
-          end
-
-          if dummy = klass.create(dummy_hash)
-            tables << table
-            if klass.column_names.include?('id')
-              connection.execute "UPDATE #{table} SET id = 0 WHERE id = #{dummy.id}"
-            end
-          else
-            raise "Could not create dummy record for table #{table}"
-          end
-        else
-          next
-        end
-      end
-
-      tables
-    end
-    
-    def prepare_connection_for_timezone
-      # Fixes #98
-      case ActiveRecord::Base.configurations[RAILS_ENV]['adapter']
-      when 'mysql'
-        # Fixes timezone to "+0:0"
-        raise "MySQL timezone UTC required too late, connection already active." if Class.new(ActiveRecord::Base).connected?
-
-        ActiveRecord::ConnectionAdapters::MysqlAdapter.class_eval do
-          def configure_connection_with_timezone
-            configure_connection_without_timezone
-            tz = ActiveRecord::Base.default_timezone == :utc ? "+0:0" : "SYSTEM"
-            execute("SET time_zone = '#{tz}'")
-          end
-          alias_method_chain :configure_connection, :timezone
-        end
-      when 'postgresql'
-        # Fixes timezone to "+0:0"
-        raise "PostgreSQL timezone UTC required too late, connection already active." if Class.new(ActiveRecord::Base).connected?
-
-        ActiveRecord::ConnectionAdapters::PostgreSQLAdapter.class_eval do
-          def configure_connection_with_timezone
-            configure_connection_without_timezone
-            tz = ActiveRecord::Base.default_timezone == :utc ? "UTC" : "SYSTEM"
-            execute("SET TIMEZONE = '#{tz}'")
-          end
-          alias_method_chain :configure_connection, :timezone
-        end
-      end
-    end
-
-    # Return true if we can load models because the database has the basic tables.
-    def migrated_once?
-      connection.tables.include?('nodes')
-    end
-  end # Db
-end # Zena
+  klass = ActiveRecord::Base.configurations[RAILS_ENV]['adapter'].capitalize
+  # Loads the wrong adaper when running rake tasks (RAILS_ENV not correct ?)
+  # Is this fixed ? If not, use
+  # ActiveRecord::Base.connection.class.name.split('::').last[/(.+)Adapter/,1].downcase
+  begin
+    Db = Zena.resolve_const("Zena::DbHelper::#{klass}")
+  rescue NameError
+    raise NameError.new("Could not find db helper 'Zena::DbHelper::#{klass}'.")
+  end
+end # Zena