Issue 6940 - Use underscore prefixes lib/matcher.js

lib/matcher.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
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  * GNU General Public License for more details.
  *
  * You should have received a copy of the GNU General Public License
  * along with Adblock Plus.  If not, see <http://www.gnu.org/licenses/>.
  */
 
 "use strict";
 
 /**
  * @fileOverview Matcher class implementing matching addresses against
  *               a list of filters.
  */
 
-const {WhitelistFilter} = require("./filterClasses");
+const {RegExpFilter, WhitelistFilter} = require("./filterClasses");
 
 /**
  * Regular expression for matching a keyword in a filter.
  * @type {RegExp}
  */
 const keywordRegExp = /[^a-z0-9%*][a-z0-9%]{3,}(?=[^a-z0-9%*])/;
 
 /**
  * Regular expression for matching all keywords in a filter.
  * @type {RegExp}
  */
 const allKeywordsRegExp = new RegExp(keywordRegExp, "g");
+
+/**
+ * Bitmask for "types" that are for exception rules only, like
+ * <code>$document</code>, <code>$elemhide</code>, and so on.
+ * @type {number}
+ */
+const WHITELIST_ONLY_TYPES = RegExpFilter.typeMap.DOCUMENT |
+                             RegExpFilter.typeMap.ELEMHIDE |
+                             RegExpFilter.typeMap.GENERICHIDE |
+                             RegExpFilter.typeMap.GENERICBLOCK;
 
 /**
  * Checks whether a particular filter is slow.
  * @param {RegExpFilter} filter
  * @returns {boolean}
  */
 function isSlowFilter(filter)
 {
   return !filter.pattern || !keywordRegExp.test(filter.pattern);
 }
 
 exports.isSlowFilter = isSlowFilter;
 
 /**
  * Blacklist/whitelist filter matching
  */
 class Matcher
 {
   constructor()
   {
     /**
      * Lookup table for filters by their associated keyword
+     * @type {Map.<string,(Filter|Set.<Filter>)>}
      * @private

[Manish Jethani, 2018/10/21 12:12:19]

I've already started using @private/@protected elsewhere [1], but I put it as
the last thing in the comment. Could you please make @private the last line in
the comment?

[1]: https://hg.adblockplus.org/adblockpluscore/rev/4a0244a174e6#l2.57

[Jon Sonesen, 2018/10/23 01:28:01]

Acknowledged.

-     * @type {Map.<string,(Filter|Set.<Filter>)>}
      */
     this._filterByKeyword = new Map();
   }
 
   /**
    * Removes all known filters
    */
   clear()
   {
     this._filterByKeyword.clear();
@@ skipping 43 matching lines @@
     {
       set.delete(filter);
 
       if (set.size == 1)
         this._filterByKeyword.set(keyword, [...set][0]);
     }
   }
 
   /**
    * Chooses a keyword to be associated with the filter
-   * @protected

[Manish Jethani, 2018/10/21 12:12:19]

Same here.

[Jon Sonesen, 2018/10/23 01:28:04]

Acknowledged.

    * @param {Filter} filter
    * @returns {string} keyword or an empty string if no keyword could be found
+   * @protected
    */
   findKeyword(filter)
   {
     let result = "";
     let {pattern} = filter;
     if (pattern == null)
       return result;
 
     let candidates = pattern.toLowerCase().match(allKeywordsRegExp);
     if (!candidates)
@@ skipping 21 matching lines @@
   /**
    * Checks whether the entries for a particular keyword match a URL
    * @param {string} keyword
    * @param {string} location
    * @param {number} typeMask
    * @param {string} [docDomain]
    * @param {boolean} [thirdParty]
    * @param {string} [sitekey]
    * @param {boolean} [specificOnly]
    * @returns {?Filter}
-   */
-  _checkEntryMatch(keyword, location, typeMask, docDomain, thirdParty, sitekey,

[Manish Jethani, 2018/10/21 12:52:04]

About this function, I think we can make it "protected". Let's remove the
underscore and add the @protected tag.

[Jon Sonesen, 2018/10/23 01:28:02]

Acknowledged.

-                   specificOnly)
+   * @protected
+   */
+  checkEntryMatch(keyword, location, typeMask, docDomain, thirdParty, sitekey,
+                  specificOnly)
   {
     let set = this._filterByKeyword.get(keyword);
     if (typeof set == "undefined")
       return null;
 
     for (let filter of set)
     {
       if (specificOnly && filter.isGeneric() &&
           !(filter instanceof WhitelistFilter))
         continue;
@@ skipping 22 matching lines @@
    *   matching filter or <code>null</code>
    */
   matchesAny(location, typeMask, docDomain, thirdParty, sitekey, specificOnly)
   {
     let candidates = location.toLowerCase().match(/[a-z0-9%]{3,}/g);
     if (candidates === null)
       candidates = [];
     candidates.push("");
     for (let i = 0, l = candidates.length; i < l; i++)
     {
-      let result = this._checkEntryMatch(candidates[i], location, typeMask,
-                                         docDomain, thirdParty, sitekey,
-                                         specificOnly);
+      let result = this.checkEntryMatch(candidates[i], location, typeMask,
+                                        docDomain, thirdParty, sitekey,
+                                        specificOnly);
       if (result)
         return result;
     }
 
     return null;
   }
 }
 
 exports.Matcher = Matcher;
 
 /**
  * Combines a matcher for blocking and exception rules, automatically sorts
  * rules into two {@link Matcher} instances.
  */
 class CombinedMatcher
 {
   constructor()
   {
     /**
      * Maximal number of matching cache entries to be kept
      * @type {number}
      */
     this.maxCacheEntries = 1000;
 
     /**
      * Matcher for blocking rules.
      * @type {Matcher}
+     * @private
      */
-    this.blacklist = new Matcher();
+    this._blacklist = new Matcher();
 
     /**
      * Matcher for exception rules.
      * @type {Matcher}
+     * @private
      */
-    this.whitelist = new Matcher();
+    this._whitelist = new Matcher();
 
     /**
      * Lookup table of previous {@link Matcher#matchesAny} results
      * @type {Map.<string,Filter>}
+     * @private
      */
-    this.resultCache = new Map();

[Manish Jethani, 2018/10/21 12:52:04]

Let's do this entire file while we're at it. I don't think resultCache needs to
be public. It can private (underscore and @private).

[Manish Jethani, 2018/10/22 22:57:50]

FYI we're having a discussion [1] about making blacklist and whitelist private
as well. I think it makes sense to make them private.

[1]: https://codereview.adblockplus.org/29896562/

[Jon Sonesen, 2018/10/23 01:28:02]

Acknowledged. In the new patch set, I made them private and realized they are in
fact not private. in lib/subscriptionClasses.js the "defaultsMap" literal is
used to fetch the filter groups (not a very good design choice as it seems
everyone has thought they are private.) The same applies in lib/filterClassesjs
and test/filterListener.js so my first reaction is to just make them protected.

[Manish Jethani, 2018/10/23 03:10:14]

I don't think that has anything to do with the change here. We can make these
properties private.

[Jon Sonesen, 2018/10/23 03:16:42]

It does, have you tested making them private? I can upload the patch set that
only changes properties but it breaks *alot* of tests

[Manish Jethani, 2018/10/23 03:23:17]

You'll have to modify the tests of course, so they use the underscored names of
the properties. I think you have to change exactly one line for this.

[Manish Jethani, 2018/10/23 03:28:42]

Please try just changing this line:

https://hg.adblockplus.org/adblockpluscore/file/e8254db792d0/test/filterListe...

To this:

  let matcher = defaultMatcher["_" + type];

[Jon Sonesen, 2018/10/23 04:08:03]

Yeah I am the worst, thanks for your help and patience

[Manish Jethani, 2018/10/23 10:38:26]

Not at all, it happens to me too :)

+    this._resultCache = new Map();
   }
 
   /**
    * @see Matcher#clear
    */
   clear()
   {
-    this.blacklist.clear();
-    this.whitelist.clear();
-    this.resultCache.clear();
+    this._blacklist.clear();
+    this._whitelist.clear();
+    this._resultCache.clear();
   }
 
   /**
    * @see Matcher#add
    * @param {Filter} filter
    */
   add(filter)
   {
     if (filter instanceof WhitelistFilter)
-      this.whitelist.add(filter);
+      this._whitelist.add(filter);
     else
-      this.blacklist.add(filter);
-
-    this.resultCache.clear();
+      this._blacklist.add(filter);
+
+    this._resultCache.clear();
   }
 
   /**
    * @see Matcher#remove
    * @param {Filter} filter
    */
   remove(filter)
   {
     if (filter instanceof WhitelistFilter)
-      this.whitelist.remove(filter);
+      this._whitelist.remove(filter);
     else
-      this.blacklist.remove(filter);
-
-    this.resultCache.clear();
+      this._blacklist.remove(filter);
+
+    this._resultCache.clear();
   }
 
   /**
    * @see Matcher#findKeyword
    * @param {Filter} filter
    * @returns {string} keyword
+   * @protected
    */
   findKeyword(filter)

[Manish Jethani, 2018/10/21 12:52:03]

If we're marking fineKeyword @protected in Matcher, let's do it here in
CombinedMatcher as well.

[Jon Sonesen, 2018/10/23 01:28:04]

Done.

   {
     if (filter instanceof WhitelistFilter)
-      return this.whitelist.findKeyword(filter);
-    return this.blacklist.findKeyword(filter);
+      return this._whitelist.findKeyword(filter);
+    return this._blacklist.findKeyword(filter);
   }
 
   /**
    * Optimized filter matching testing both whitelist and blacklist matchers
    * simultaneously. For parameters see
      {@link Matcher#matchesAny Matcher.matchesAny()}.
    * @see Matcher#matchesAny
    * @inheritdoc
-   */
-  matchesAnyInternal(location, typeMask, docDomain, thirdParty, sitekey,

[Manish Jethani, 2018/10/21 12:52:03]

Let's make this private (underscore and @private).

[Jon Sonesen, 2018/10/23 01:28:02]

Acknowledged.

-                     specificOnly)
+   * @private
+   */
+  _matchesAnyInternal(location, typeMask, docDomain, thirdParty, sitekey,
+                      specificOnly)
   {
     let candidates = location.toLowerCase().match(/[a-z0-9%]{3,}/g);
     if (candidates === null)
       candidates = [];
     candidates.push("");
 
+    let whitelistHit = null;
     let blacklistHit = null;
-    for (let i = 0, l = candidates.length; i < l; i++)
-    {
-      let substr = candidates[i];
-      let result = this.whitelist._checkEntryMatch(
-        substr, location, typeMask, docDomain, thirdParty, sitekey
-      );
-      if (result)
-        return result;
-      if (blacklistHit === null)
+
+    // If the type mask includes no types other than whitelist-only types, we
+    // can skip the blacklist.
+    if ((typeMask & ~WHITELIST_ONLY_TYPES) != 0)
+    {
+      for (let i = 0, l = candidates.length; !blacklistHit && i < l; i++)
       {
-        blacklistHit = this.blacklist._checkEntryMatch(
-          substr, location, typeMask, docDomain, thirdParty, sitekey,
-          specificOnly
-        );
+        blacklistHit = this._blacklist.checkEntryMatch(candidates[i], location,
+                                                       typeMask, docDomain,
+                                                       thirdParty, sitekey,
+                                                       specificOnly);
       }
     }
-    return blacklistHit;
+
+    // If the type mask includes any whitelist-only types, we need to check the
+    // whitelist.
+    if (blacklistHit || (typeMask & WHITELIST_ONLY_TYPES) != 0)
+    {
+      for (let i = 0, l = candidates.length; !whitelistHit && i < l; i++)
+      {
+        whitelistHit = this._whitelist.checkEntryMatch(candidates[i], location,
+                                                       typeMask, docDomain,
+                                                       thirdParty, sitekey);
+      }
+    }
+
+    return whitelistHit || blacklistHit;
   }
 
   /**
    * @see Matcher#matchesAny
    * @inheritdoc

[Manish Jethani, 2018/10/21 12:52:04]

There are two instances of @inheritdoc in this file, neither of them makes sense
because there is no inheritance. Let's remove them.

[Jon Sonesen, 2018/10/23 01:28:01]

Doing this makes eslint throw errors for the valid-jsdoc plugin

[Manish Jethani, 2018/10/23 03:10:14]

Oh yeah, I think you had mentioned this earlier as well. I forgot. Welp, let's
leave it then.

    */
   matchesAny(location, typeMask, docDomain, thirdParty, sitekey, specificOnly)
   {
     let key = location + " " + typeMask + " " + docDomain + " " + thirdParty +
       " " + sitekey + " " + specificOnly;
 
-    let result = this.resultCache.get(key);
+    let result = this._resultCache.get(key);
     if (typeof result != "undefined")
       return result;
 
-    result = this.matchesAnyInternal(location, typeMask, docDomain,
-                                     thirdParty, sitekey, specificOnly);
-
-    if (this.resultCache.size >= this.maxCacheEntries)
-      this.resultCache.clear();
-
-    this.resultCache.set(key, result);
+    result = this._matchesAnyInternal(location, typeMask, docDomain,
+                                      thirdParty, sitekey, specificOnly);
+
+    if (this._resultCache.size >= this.maxCacheEntries)
+      this._resultCache.clear();
+
+    this._resultCache.set(key, result);
 
     return result;
   }
 }
 
 exports.CombinedMatcher = CombinedMatcher;
 
 /**
  * Shared {@link CombinedMatcher} instance that should usually be used.
  * @type {CombinedMatcher}
  */
 let defaultMatcher = new CombinedMatcher();
 
 exports.defaultMatcher = defaultMatcher;