ExtractorHTML: Fix srcset by normalizing elementContext() to lowercase

This ensures that when we later compare the context in processEmbed()
we don't need to deal with variants like srcSet or SRCSET. Note that
we're already sometimes lowercasing it later in HTMLLinkContext.get().

Fixes internetarchive#477.

Author: ato

modules/src/main/java/org/archive/modules/extractor/ExtractorHTML.java

@@ -1090,7 +1090,7 @@ protected void processStyle(CrawlURI curi, CharSequence sequence,
      * @return CharSequence context
      */
     public static CharSequence elementContext(CharSequence element, CharSequence attribute) {
-        return attribute == null? "": element + "/@" + attribute;
+        return attribute == null? "": (element + "/@" + attribute).toLowerCase(Locale.ROOT);
     }
 }
 

modules/src/test/java/org/archive/modules/extractor/ExtractorHTMLTest.java

@@ -545,8 +545,8 @@ public void testSourceSrcSetAttribute() throws URIException {
 
         CharSequence cs = "<picture>"
                 + "<source media=\"(min-width: 992px)\" srcset=\"images/foo1.jpg\"> "
-                + "<source media=\"(min-width: 500px)\" srcset=\"images/foo2.jpg\"> "
-                + "<source media=\"(min-width: 0px)\" srcset=\"images/foo3.jpg\"> "
+                + "<source media=\"(min-width: 500px)\" SRCSET=\"images/foo2.jpg\"> "
+                + "<source media=\"(min-width: 0px)\" srcSet=\"images/foo3-1x.jpg 1x, images/foo3-2x.jpg 2x\"> "
                 + "<img src=\"images/foo.jpg\" alt=\"\"> "
                 + "</picture>";
 
@@ -559,7 +559,9 @@ public void testSourceSrcSetAttribute() throws URIException {
                 "http://www.example.com/images/foo.jpg",
                 "http://www.example.com/images/foo1.jpg",
                 "http://www.example.com/images/foo2.jpg",
-                "http://www.example.com/images/foo3.jpg" };
+                "http://www.example.com/images/foo3-1x.jpg",
+                "http://www.example.com/images/foo3-2x.jpg",
+        };
 
         for (int i = 0; i < links.length; i++) {
             assertEquals("outlink from picture", dest[i], links[i].getURI());