www

commit 60e5ff103c6a05103b4700c184dd47a1245e3a9c
parent 917910d1d0c4f4a53e4061df4205bd578b08ee9a
Author: Dan Stillman <dstillman@zotero.org>
Date:   Wed, 15 Sep 2010 18:43:52 +0000

Fix COinS and unAPI on pages served as application/xhtml+xml

This requires testing, since it could easily break COinS/unAPI on some pages.


Diffstat:
M
translators/COinS.js
 | 
20
+++++++++++++++++---
M
translators/unAPI.js
 | 
12
+++++++++---

2 files changed, 26 insertions(+), 6 deletions(-)
diff --git a/translators/COinS.js b/translators/COinS.js
@@ -8,7 +8,7 @@
 	"maxVersion":"",
 	"priority":300,
 	"inRepository":true,
-	"lastUpdated":"2010-06-16 18:15:00"
+	"lastUpdated":"2010-09-15 18:42:44"
 }
 
 function detectWeb(doc, url) {
@@ -16,7 +16,15 @@ function detectWeb(doc, url) {
 	
 	var encounteredType = false;
 	
-	var spans = doc.evaluate('//span[contains(@class, " Z3988") or contains(@class, "Z3988 ") or @class="Z3988"][@title]', doc, null, XPathResult.ANY_TYPE, null);
+	// This and the x: prefix in the XPath are to work around an issue with pages
+	// served as application/xhtml+xml
+	//
+	// https://developer.mozilla.org/en/Introduction_to_using_XPath_in_JavaScript#Implementing_a_default_namespace_for_XML_documents
+	function nsResolver() {
+		return 'http://www.w3.org/1999/xhtml';
+	}
+	
+	var spans = doc.evaluate('//x:span[contains(@class, " Z3988") or contains(@class, "Z3988 ") or @class="Z3988"][@title]', doc, nsResolver, XPathResult.ANY_TYPE, null);
 	var span;
 	while(span = spans.iterateNext()) {
 		// determine if it's a valid type
@@ -141,7 +149,13 @@ function doWeb(doc, url) {
 	var needFullItems = new Array();
 	var couldUseFullItems = new Array();
 	
-	var spans = doc.evaluate('//span[contains(@class, " Z3988") or contains(@class, "Z3988 ") or @class="Z3988"][@title]', doc, null, XPathResult.ANY_TYPE, null);
+	
+	// See note in detectWeb()
+	function nsResolver() {
+		return 'http://www.w3.org/1999/xhtml';
+	}
+	
+	var spans = doc.evaluate('//x:span[contains(@class, " Z3988") or contains(@class, "Z3988 ") or @class="Z3988"][@title]', doc, nsResolver, XPathResult.ANY_TYPE, null);
 	var span;
 	while(span = spans.iterateNext()) {
 		var spanTitle = span.title;
diff --git a/translators/unAPI.js b/translators/unAPI.js
@@ -8,7 +8,7 @@
 	"maxVersion":"",
 	"priority":200,
 	"inRepository":true,
-	"lastUpdated":"2010-09-04 20:28:04"
+	"lastUpdated":"2010-09-15 18:42:44"
 }
 
 var RECOGNIZABLE_FORMATS = ["mods", "marc", "endnote", "ris", "bibtex", "rdf"];
@@ -34,10 +34,16 @@ function detectWeb(doc, url) {
 	// Set the domain we're scraping
 	domain = doc.location.href.match(/https?:\/\/([^/]+)/);
 	
-	var nsResolver = doc.createNSResolver(doc.documentElement);
+	// This and the x: prefix in the XPath are to work around an issue with pages
+	// served as application/xhtml+xml
+	//
+	// https://developer.mozilla.org/en/Introduction_to_using_XPath_in_JavaScript#Implementing_a_default_namespace_for_XML_documents
+	function nsResolver() {
+		return 'http://www.w3.org/1999/xhtml';
+	}
 	
 	// look for a resolver
-	unAPIResolver = doc.evaluate('//link[@rel="unapi-server"]', doc, nsResolver, XPathResult.ANY_TYPE, null).iterateNext();
+	unAPIResolver = doc.evaluate('//x:link[@rel="unapi-server"]', doc, nsResolver, XPathResult.ANY_TYPE, null).iterateNext();
 	if(!unAPIResolver) return false;
 	unAPIResolver = unAPIResolver.getAttribute("href");