Issue 6937 - Lowercase RegExpFilter domains on demand

lib/filterClasses.js

 /*
  * This file is part of Adblock Plus <https://adblockplus.org/>,
  * Copyright (C) 2006-present eyeo GmbH
  *
  * Adblock Plus is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 3 as
  * published by the Free Software Foundation.
  *
  * Adblock Plus is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ skipping 429 matching lines @@
   domainSource: null,
 
   /**
    * Separator character used in domainSource property, must be
    * overridden by subclasses
    * @type {string}
    */
   domainSeparator: null,
 
   /**
-   * Determines whether domainSource is already lower-case,
-   * can be overridden by subclasses.
-   * @type {boolean}
-   */
-  domainSourceIsLowerCase: false,
-
-  /**
    * Map containing domains that this filter should match on/not match
    * on or null if the filter should match on all domains
    * @type {?Map.<string,boolean>}
    */
   get domains()
   {
     let domains = null;
 
     if (this.domainSource)
     {
-      let source = this.domainSource;
-      if (!this.domainSourceIsLowerCase)
-      {
-        // RegExpFilter already have lowercase domains
-        source = source.toLowerCase();
-      }
+      let source = this.domainSource.toLowerCase();

[Jon Sonesen, 2018/09/07 03:09:33]

Im super happy to see things like this go away hehe. Why did subclasses ever
change case?

[Manish Jethani, 2018/09/07 03:42:44]

I'm guessing "historical reasons." It may have been the case that the domains
were already being lowercased as part of the parsing in `RegExpFilter.fomText`,
so it made sense to avoid lowercasing them again. This is definitely not the
case now; as in, we can just choose not to lowercase the domains for any filter
and then lowercase them later when needed.

[Jon Sonesen, 2018/09/17 18:11:33]

Yeah makes sense, thanks for explaining. I tested this and seems good so unless
hub or snoack have an issue I say LGTM

[Sebastian Noack, 2018/09/17 19:20:59]

This was an optimization at some point, but indeed, with the current design not
so much anymore. LGTM.

 
       let knownMap = knownDomainMaps.get(source);
       if (knownMap)
       {
         domains = knownMap;
       }
       else
       {
         let list = source.split(this.domainSeparator);
         if (list.length == 1 && list[0][0] != "~")
@@ skipping 200 matching lines @@
   else
   {
     // No need to convert this filter to regular expression yet, do it on demand
     this.pattern = regexpSource;
   }
 }
 exports.RegExpFilter = RegExpFilter;
 
 RegExpFilter.prototype = extend(ActiveFilter, {
   /**
-   * @see ActiveFilter.domainSourceIsLowerCase
-   */
-  domainSourceIsLowerCase: true,
-
-  /**
    * Number of filters contained, will always be 1 (required to
    * optimize {@link Matcher}).
    * @type {number}
    */
   length: 1,
 
   /**
    * The filter itself (required to optimize {@link Matcher}).
    * @type {RegExpFilter}
    */
@@ skipping 154 matching lines @@
       else
       {
         switch (option.toLowerCase())
         {
           case "match-case":
             matchCase = !inverse;
             break;
           case "domain":
             if (!value)
               return new InvalidFilter(origText, "filter_unknown_option");
-            domains = value.toLowerCase();
+            domains = value;
             break;
           case "third-party":
             thirdParty = !inverse;
             break;
           case "collapse":
             collapse = !inverse;
             break;
           case "sitekey":
             if (!value)
               return new InvalidFilter(origText, "filter_unknown_option");
@@ skipping 364 matching lines @@
 
   /**
    * Script that should be executed
    * @type {string}
    */
   get script()
   {
     return this.body;
   }
 });