FEATURE: permalink normalization

Optionally allow admins to apply regex based normalization
to permalinks prior to matching.

This allows us to drop query string, or cleanly ignore slugs, etc.

app/models/permalink.rb

@@ -5,11 +5,72 @@ class Permalink < ActiveRecord::Base
 
   before_validation :normalize_url
 
-  def normalize_url
-    if self.url
-      self.url = self.url.strip
-      self.url = self.url[1..-1] if url[0,1] == '/'
+  class Normalizer
+    attr_reader :source
+
+    def initialize(source)
+      @source = source
+      if source.present?
+        @rules = source.split("|").map do |rule|
+          parse_rule(rule)
+        end.compact
+      end
     end
+
+    def parse_rule(rule)
+      return unless rule =~ /\/.*\//
+
+      escaping = false
+      regex = ""
+      sub = ""
+      c = 0
+
+      rule.chars.each do |l|
+        c += 1 if !escaping && l == "/"
+        escaping = l == "\\"
+
+        if c > 1
+          sub << l
+        else
+          regex << l
+        end
+      end
+
+      if regex.length > 1
+        [Regexp.new(regex[1..-1]), sub[1..-1] || ""] rescue nil
+      end
+
+    end
+
+    def normalize(url)
+      return url unless @rules
+      @rules.each do |(regex,sub)|
+        url = url.sub(regex,sub)
+      end
+
+      url
+    end
+
+  end
+
+  def self.normalize_url(url)
+    if url
+      url = url.strip
+      url = url[1..-1] if url[0,1] == '/'
+    end
+
+    normalizations =  SiteSetting.permalink_normalizations
+
+    @normalizer = Normalizer.new(normalizations) unless @normalizer && @normalizer.source == normalizations
+    @normalizer.normalize(url)
+  end
+
+  def self.find_by_url(url)
+    find_by(url: normalize_url(url))
+  end
+
+  def normalize_url
+    self.url = Permalink.normalize_url(url) if url
   end
 
   def target_url