Issue 3143 - Filter elements with :-abp-has()

chrome/content/elemHideEmulation.js

Index: chrome/content/elemHideEmulation.js
===================================================================
--- a/chrome/content/elemHideEmulation.js
+++ b/chrome/content/elemHideEmulation.js
@@ -54,172 +54,372 @@
       }
     }
   }
 
   selectors.push(selector.substring(start));
   return selectors;
 }
 
-function ElemHideEmulation(window, getFiltersFunc, addSelectorsFunc)
+/** Return position of node from parent.
+ * @param {Node} node the node to find the position of.
+ * @return {number} One-based index like for :nth-child(), or 0 on error.
+ */
+function positionInParent(node)
+{
+  let {children} = node.parentNode;
+  for (let i = 0; i < children.length; i++)
+    if (children[i] == node)
+      return i + 1;
+  return 0;
+}
+
+function makeSelector(node, selector)
+{
+  if (!node.parentElement)
+  {
+    let newSelector = ":root";
+    if (selector)
+      newSelector += " > " + selector;
+    return newSelector;
+  }
+  let idx = positionInParent(node);
+  if (idx > 0)
+  {
+    let newSelector = `${node.tagName}:nth-child(${idx})`;
+    if (selector)
+      newSelector += " > " + selector;
+    return makeSelector(node.parentElement, newSelector);
+  }
+
+  return selector;
+}
+
+function parseSelectorContent(content, startIndex)
+{
+  let parens = 1;
+  let quote = null;
+  let i = startIndex;
+  for (; i < content.length; i++)
+  {
+    let c = content[i];
+    if (c == "\\")
+    {
+      // Ignore escaped characters
+      i++;
+    }
+    else if (quote)
+    {
+      if (c == quote)
+        quote = null;
+    }
+    else if (c == "'" || c == '"')
+      quote = c;
+    else if (c == "(")
+      parens++;
+    else if (c == ")")
+    {
+      parens--;
+      if (parens == 0)
+        break;
+    }
+  }
+
+  if (parens > 0)
+    return null;
+  return {text: content.substring(startIndex, i), end: i};
+}
+
+/** Parse the selector
+ * @param {string} selector the selector to parse
+ * @return {Object} selectors is an array of objects,
+ * or null in case of errors. hide is true if we'll hide
+ * elements instead of styles..
+ */
+function parseSelector(selector)
+{
+  if (selector.length == 0)
+    return [];
+
+  let match = abpSelectorRegexp.exec(selector);
+  if (!match)
+    return [new PlainSelector(selector)];
+
+  let selectors = [];
+  if (match.index > 0)
+    selectors.push(new PlainSelector(selector.substr(0, match.index)));
+
+  let startIndex = match.index + match[0].length;
+  let content = parseSelectorContent(selector, startIndex);
+  if (content == null)

[Sebastian Noack, 2017/06/08 08:28:47]

Nit: Given that parseSelectorContent() either returns and object or null,
perhaps `!content` is more appropriate here rather than `content == null`.

[hub, 2017/06/08 15:45:47]

Done.

+  {
+    console.error(new SyntaxError("Failed parsing AdBlock Plus " +

[Sebastian Noack, 2017/06/08 08:28:48]

Nit: It's Adblock Plus (with lower case "B"). But perhaps we should avoid
mentioning the product name in error messages in the first place, in order to
avoid confusion in other products that are based on this code. (Same below.)

[hub, 2017/06/08 15:45:48]

Maybe we could change this to "Failed parsing content blocker selector". It is
neutral.

[Felix Dahlke, 2017/06/09 09:02:33]

Well, we're actually parsing what we call an "ABP selector" (i.e. the :abp-*
pseudo-classes) here, so I think we should call it that in the error message:
Either "Adblock Plus selector" or "ABP selector". I personally prefer the latter
since it seems easier to make the connection to the :abp-* pseudo-classes that
way.

Various products with different names use this code, but it's still an "ABP
selector", similar to how we and others typically refer to the filter list
syntax as something to the effect of "ABP-style filters".

+                                  `selector ${selector}, didn't ` +
+                                  "find closing parenthesis."));
+    return null;
+  }
+  if (match[1] == "properties")
+    selectors.push(new PropsSelector(content.text));
+  else if (match[1] == "has")
+  {
+    let hasSelector = new HasSelector(content.text);
+    if (!hasSelector.valid())
+      return null;
+    selectors.push(hasSelector);
+  }
+  else
+  {
+    // this is an error, can't parse selector.
+    console.error(new SyntaxError("Failed parsing AdBlock Plus " +
+                                  `selector ${selector}, invalid ` +
+                                  `pseudo-class -abp-${match[1]}().`));
+    return null;
+  }
+
+  let suffix = parseSelector(selector.substr(content.end + 1));
+  if (suffix == null)
+    return null;
+
+  selectors.push(...suffix);
+
+  return selectors;
+}
+
+/** Stringified style objects
+ * @typedef {Object} StringifiedStyle
+ * @property {string} style CSS style represented a string.

[Wladimir Palant, 2017/06/08 13:12:40]

This doesn't seem to be quite English to me, also not very helpful. "sorted and
joined list of CSS properties" maybe? Note that this isn't a full sentence, so
there shouldn't be a period at the end.

[hub, 2017/06/08 15:45:47]

Acknowledged.

+ * @property {string[]} subSelectors selectors for the rule.

[Wladimir Palant, 2017/06/08 13:12:39]

"selectors that the CSS properties apply to"?

[hub, 2017/06/08 15:45:46]

Acknowledged.

+ */
+
+/**
+ * Turn a CSSStyleRule into a StringifiedStyle

[Wladimir Palant, 2017/06/08 13:12:39]

You have the types listed below, no need to repeat them. "Produces a string
representation of a stylesheet entry." maybe? Note that this is a sentence, so
it should end with a period.

[hub, 2017/06/08 15:45:47]

Acknowledged.

+ * @param {CSSStyleRule} rule the CSS style rule.
+ * @return {StringifiedStyle} the stringified style.
+ */
+function stringifyStyle(rule)
+{
+  let styles = [];
+  for (let i = 0; i < rule.style.length; i++)
+  {
+    let property = rule.style.item(i);
+    let value = rule.style.getPropertyValue(property);
+    let priority = rule.style.getPropertyPriority(property);
+    styles.push(property + ": " + value + (priority ? " !" + priority : "") +

[Sebastian Noack, 2017/06/08 08:28:47]

Nit: If you'd use a template literal here you woudnl't have to wrap this line:

  `${property}: ${value}${priority ? " !" + priority : ""};`

[hub, 2017/06/08 15:45:46]

This code was simply moved and written pre-ES6. I'll use a template string.

+                ";");
+  }
+  styles.sort();
+  return {
+    style: styles.join(" "),
+    subSelectors: splitSelector(rule.selectorText)
+  };
+}
+
+function* evaluate(chain, index, prefix, subtree, styles)
+{
+  if (index >= chain.length)
+  {
+    yield prefix;
+    return;
+  }
+  for (let [selector, element] of
+       chain[index].getSelectors(prefix, subtree, styles))
+    yield* evaluate(chain, index + 1, selector, element, styles);
+}
+
+function PlainSelector(selector)
+{
+  this._selector = selector;
+}
+
+PlainSelector.prototype = {
+  /**
+   * Generator function returning a pair of selector
+   * string and subtree.
+   * @param {string} prefix the prefix for the selector.
+   * @param {Node} subtree the subtree we work on.
+   * @param {StringifiedStyle[]} styles the stringified style objects.
+   */
+  *getSelectors(prefix, subtree, styles)
+  {
+    yield [prefix + this._selector, subtree];
+  }
+};
+
+const incompletePrefixRegexp = /[\s>+~]$/;
+
+function HasSelector(selector)
+{
+  this._innerSelectors = parseSelector(selector);
+}
+
+HasSelector.prototype = {
+  requiresHiding: true,
+
+  valid()
+  {
+    return this._innerSelectors != null;
+  },
+
+  *getSelectors(prefix, subtree, styles)
+  {
+    for (let element of this.getElements(prefix, subtree, styles))
+      yield [makeSelector(element, ""), element];
+  },
+
+  /**
+   * Generator function returning selected elements.
+   * @param {string} prefix the prefix for the selector.
+   * @param {Node} subtree the subtree we work on.
+   * @param {StringifiedStyle[]} styles the stringified style objects.
+   */
+  *getElements(prefix, subtree, styles)
+  {
+    let actualPrefix = (!prefix || incompletePrefixRegexp.test(prefix)) ?
+        prefix + "*" : prefix;
+    let elements = subtree.querySelectorAll(actualPrefix);
+    for (let element of elements)
+    {
+      let newPrefix = makeSelector(element, "");
+      let iter = evaluate(this._innerSelectors, 0, newPrefix + " ",
+                          element, styles);
+      for (let selector of iter)
+        // we insert a space between the two. It becomes a no-op if selector
+        // doesn't have a combinator
+        if (subtree.querySelector(selector))
+          yield element;
+    }
+  }
+};
+
+function PropsSelector(propertyExpression)
+{
+  let regexpString;
+  if (propertyExpression.length >= 2 && propertyExpression[0] == "/" &&
+      propertyExpression[propertyExpression.length - 1] == "/")
+  {
+    regexpString = propertyExpression.slice(1, -1)
+      .replace("\\x7B ", "{").replace("\\x7D ", "}");
+  }
+  else
+    regexpString = filterToRegExp(propertyExpression);
+
+  this._regexp = new RegExp(regexpString, "i");
+}
+
+PropsSelector.prototype = {
+  *findPropsSelectors(styles, prefix, regexp)
+  {
+    for (let style of styles)
+      if (regexp.test(style.style))
+        for (let subSelector of style.subSelectors)
+          yield prefix + subSelector;
+  },
+
+  *getSelectors(prefix, subtree, styles)
+  {
+    for (let selector of this.findPropsSelectors(styles, prefix, this._regexp))
+      yield [selector, subtree];
+  }
+};
+
+function ElemHideEmulation(window, getFiltersFunc, addSelectorsFunc,
+                           hideElemsFunc)
 {
   this.window = window;
   this.getFiltersFunc = getFiltersFunc;
   this.addSelectorsFunc = addSelectorsFunc;
+  this.hideElemsFunc = hideElemsFunc;
 }
 
 ElemHideEmulation.prototype = {
-  stringifyStyle(style)
-  {
-    let styles = [];
-    for (let i = 0; i < style.length; i++)
-    {
-      let property = style.item(i);
-      let value = style.getPropertyValue(property);
-      let priority = style.getPropertyPriority(property);
-      styles.push(property + ": " + value + (priority ? " !" + priority : "") +
-                  ";");
-    }
-    styles.sort();
-    return styles.join(" ");
-  },
-
   isSameOrigin(stylesheet)
   {
     try
     {
       return new URL(stylesheet.href).origin == this.window.location.origin;
     }
     catch (e)
     {
       // Invalid URL, assume that it is first-party.
       return true;
     }
   },
 
-  findSelectors(stylesheet, selectors, filters)
+  addSelectors(stylesheets)
   {
-    // Explicitly ignore third-party stylesheets to ensure consistent behavior
-    // between Firefox and Chrome.
-    if (!this.isSameOrigin(stylesheet))
-      return;
+    let selectors = [];
+    let selectorFilters = [];
+
+    let elements = [];
+    let elementFilters = [];
+
+    let cssStyles = [];
 
-    let rules = stylesheet.cssRules;
-    if (!rules)
-      return;
+    for (let stylesheet of stylesheets)
+    {
+      // Explicitly ignore third-party stylesheets to ensure consistent behavior
+      // between Firefox and Chrome.
+      if (!this.isSameOrigin(stylesheet))
+        continue;
 
-    for (let rule of rules)
-    {
-      if (rule.type != rule.STYLE_RULE)
+      let rules = stylesheet.cssRules;
+      if (!rules)
         continue;
 
-      let style = this.stringifyStyle(rule.style);
-      for (let pattern of this.patterns)
+      for (let rule of rules)
       {
-        if (pattern.regexp.test(style))
+        if (rule.type != rule.STYLE_RULE)
+          continue;
+
+        cssStyles.push(stringifyStyle(rule));
+      }
+    }
+
+    for (let pattern of this.patterns)
+    {
+      for (let selector of evaluate(pattern.selectors,
+                                    0, "", document, cssStyles))
+      {
+        if (!pattern.selectors.some(s => s.requiresHiding))
         {
-          let subSelectors = splitSelector(rule.selectorText);
-          for (let subSelector of subSelectors)
+          selectors.push(selector);
+          selectorFilters.push(pattern.text);
+        }
+        else
+        {
+          for (let element of document.querySelectorAll(selector))
           {
-            selectors.push(pattern.prefix + subSelector + pattern.suffix);
-            filters.push(pattern.text);
+            elements.push(element);
+            elementFilters.push(pattern.text);
           }
         }
       }
     }
-  },
 
-  addSelectors(stylesheets)
-  {
-    let selectors = [];
-    let filters = [];
-    for (let stylesheet of stylesheets)
-      this.findSelectors(stylesheet, selectors, filters);
-    this.addSelectorsFunc(selectors, filters);
+    this.addSelectorsFunc(selectors, selectorFilters);
+    this.hideElemsFunc(elements, elementFilters);
   },
 
   onLoad(event)
   {
     let stylesheet = event.target.sheet;
     if (stylesheet)
       this.addSelectors([stylesheet]);
   },
 
   apply()
   {
     this.getFiltersFunc(patterns =>
     {
       this.patterns = [];
       for (let pattern of patterns)
       {
-        let match = abpSelectorRegexp.exec(pattern.selector);
-        if (!match || match[1] != "properties")
-        {
-          console.error(new SyntaxError(
-            `Failed to parse Adblock Plus selector ${pattern.selector}, ` +
-            `invalid pseudo-class :-abp-${match[1]}().`
-          ));
-          continue;
-        }
-
-        let expressionStart = match.index + match[0].length;
-        let parens = 1;
-        let quote = null;
-        let i;
-        for (i = expressionStart; i < pattern.selector.length; i++)
-        {
-          let c = pattern.selector[i];
-          if (c == "\\")
-          {
-            // Ignore escaped characters
-            i++;
-          }
-          else if (quote)
-          {
-            if (c == quote)
-              quote = null;
-          }
-          else if (c == "'" || c == '"')
-            quote = c;
-          else if (c == "(")
-            parens++;
-          else if (c == ")")
-          {
-            parens--;
-            if (parens == 0)
-              break;
-          }
-        }
-
-        if (parens > 0)
-        {
-          console.error(new SyntaxError(
-            `Failed to parse Adblock Plus selector ${pattern.selector} ` +
-            "due to unmatched parentheses."
-          ));
-          continue;
-        }
-
-        let propertyExpression = pattern.selector.substring(expressionStart, i);
-        let regexpString;
-        if (propertyExpression.length >= 2 && propertyExpression[0] == "/" &&
-            propertyExpression[propertyExpression.length - 1] == "/")
-        {
-          regexpString = propertyExpression.slice(1, -1)
-              .replace("\\x7B ", "{").replace("\\x7D ", "}");
-        }
-        else
-          regexpString = filterToRegExp(propertyExpression);
-
-        this.patterns.push({
-          text: pattern.text,
-          regexp: new RegExp(regexpString, "i"),
-          prefix: pattern.selector.substr(0, match.index),
-          suffix: pattern.selector.substr(i + 1)
-        });
+        let selectors = parseSelector(pattern.selector);
+        if (selectors != null && selectors.length > 0)
+          this.patterns.push({selectors, text: pattern.text});
       }
 
       if (this.patterns.length > 0)
       {
         let {document} = this.window;
         this.addSelectors(document.styleSheets);
         document.addEventListener("load", this.onLoad.bind(this), true);
       }