Merge pull request #454 from IDPF/feature/xhtml-extension

Check XHTML Content Docs extensions in EPUB 3
w3c · Sep 12, 2014 · 9cd4e87 · 9cd4e87
2 parents a5fd943 + 6fd290d
commit 9cd4e87
Show file tree

Hide file tree

Showing 33 changed files with 387 additions and 46 deletions.
diff --git a/src/main/java/com/adobe/epubcheck/ctc/EpubHTML5StructureCheck.java b/src/main/java/com/adobe/epubcheck/ctc/EpubHTML5StructureCheck.java
@@ -1,5 +1,15 @@
 package com.adobe.epubcheck.ctc;
 
+import java.io.File;
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.Hashtable;
+import java.util.Scanner;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+import java.util.zip.ZipEntry;
+import java.util.zip.ZipFile;
+
 import com.adobe.epubcheck.api.Report;
 import com.adobe.epubcheck.ctc.epubpackage.EpubPackage;
 import com.adobe.epubcheck.ctc.epubpackage.ManifestItem;
@@ -15,16 +25,6 @@
 import com.adobe.epubcheck.util.SearchDictionary;
 import com.adobe.epubcheck.util.SearchDictionary.DictionaryType;
 
-import java.io.File;
-import java.io.IOException;
-import java.io.InputStream;
-import java.util.Hashtable;
-import java.util.Scanner;
-import java.util.regex.Matcher;
-import java.util.regex.Pattern;
-import java.util.zip.ZipEntry;
-import java.util.zip.ZipFile;
-
 public class EpubHTML5StructureCheck implements DocumentValidator
 {
   static final int hasHtml = 1;
@@ -113,8 +113,12 @@ else if (prop.equals("rendition:layout-reflowable"))
         /***VALIDATE FILE EXTENSION***/
 
         String fileExtension = mi.getHref().substring(mi.getHref().lastIndexOf('.') + 1, mi.getHref().length());
-        if (!(fileExtension.compareToIgnoreCase("html") == 0 || fileExtension.compareToIgnoreCase("htm") == 0 || fileExtension.compareToIgnoreCase("xhtml") == 0))
+        if (epubPackage.getVersion() == EPUBVersion.VERSION_2
+            && !(fileExtension.compareToIgnoreCase("html") == 0
+                || fileExtension.compareToIgnoreCase("htm") == 0 
+                || fileExtension.compareToIgnoreCase("xhtml") == 0))
         {
+          // Note: extension is already checked in OPFChecker30 for EPUB 3 
           report.message(MessageId.HTM_014, new MessageLocation(mi.getHref(), -1, -1));
         }
 

diff --git a/src/main/java/com/adobe/epubcheck/messages/MessageDictionary.java b/src/main/java/com/adobe/epubcheck/messages/MessageDictionary.java
@@ -134,6 +134,7 @@ static Map<MessageId, Severity> getDefaultSeverities()
       map.put(MessageId.HTM_012, Severity.USAGE);
       map.put(MessageId.HTM_013, Severity.USAGE);
       map.put(MessageId.HTM_014, Severity.WARNING);
+      map.put(MessageId.HTM_014a, Severity.WARNING);
       map.put(MessageId.HTM_015, Severity.WARNING);
       map.put(MessageId.HTM_016, Severity.WARNING);
       map.put(MessageId.HTM_017, Severity.ERROR);

diff --git a/src/main/java/com/adobe/epubcheck/messages/MessageId.java b/src/main/java/com/adobe/epubcheck/messages/MessageId.java
@@ -98,6 +98,7 @@ public enum MessageId implements Comparable<MessageId>
   HTM_012("HTM-012"),
   HTM_013("HTM-013"),
   HTM_014("HTM-014"),
+  HTM_014a("HTM-014a"),
   HTM_015("HTM-015"),
   HTM_016("HTM-016"),
   HTM_017("HTM-017"),

diff --git a/src/main/java/com/adobe/epubcheck/opf/OPFChecker.java b/src/main/java/com/adobe/epubcheck/opf/OPFChecker.java
@@ -156,7 +156,6 @@ public void runChecks()
       }
 
       report.info(item.getPath(), FeatureEnum.DECLARED_MIMETYPE, item.getMimeType());
-      checkItem(item, opfHandler);
     }
 
     checkGuide();

diff --git a/src/main/java/com/adobe/epubcheck/opf/OPFChecker30.java b/src/main/java/com/adobe/epubcheck/opf/OPFChecker30.java
@@ -38,6 +38,7 @@
 import com.adobe.epubcheck.util.EPUBVersion;
 import com.adobe.epubcheck.util.GenericResourceProvider;
 import com.adobe.epubcheck.xml.XMLValidator;
+import com.google.common.io.Files;
 
 public class OPFChecker30 extends OPFChecker implements DocumentValidator
 {
@@ -109,7 +110,13 @@ protected void checkItem(OPFItem item, OPFHandler opfHandler)
       // "invalid content for media-type attribute");
       return;
     }
-
+
+    if ("application/xhtml+xml".equals(mimeType) && !"xhtml".equals(Files.getFileExtension(item.getPath())))
+    {
+      report.message(MessageId.HTM_014a,
+          new MessageLocation(path, item.getLineNumber(), item.getColumnNumber()), item.getPath());
+    }
+
     if (fallback != null)
     {
       OPFItem fallbackItem = opfHandler.getItemById(fallback);

diff --git a/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties b/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties
@@ -80,7 +80,8 @@ HTM_011=Entity is undeclared.
 HTM_011_SUG=Define the entity or use the numbered entity instead.
 HTM_012=Found a link to a CFI in an external book.
 HTM_013=Intra-Publication CFIs found in document.
-HTM_014=Invalid file extension for HTML5 file, expecting (html, htm or xhtml).
+HTM_014=Invalid file extension for HTML file, expecting (html, htm or xhtml).
+HTM_014a=XHTML Content Document file name '%1$s' should have the extension '.xhtml'.
 HTM_015=HTML4 DOCTYPE definition within ePub v3.
 HTM_016=HTML5 DOCTYPE definition within ePub v2.
 HTM_017=Content file has different language value in attributes xml:lang and lang.

diff --git a/src/test/java/com/adobe/epubcheck/api/Epub20CheckExpandedTest.java b/src/test/java/com/adobe/epubcheck/api/Epub20CheckExpandedTest.java
@@ -150,4 +150,13 @@ public void testValidateEPUB20_issue267() {
     List<MessageId> expectedWarnings = new ArrayList<MessageId>();
     testValidateDocument("valid/issue267/", expectedErrors, expectedWarnings, "valid/issue267.txt");
 	}
+
+  @Test
+  public void testXHTMLExtension()
+  {
+    List<MessageId> expectedErrors = new ArrayList<MessageId>();
+    List<MessageId> expectedWarnings = new ArrayList<MessageId>();
+    Collections.addAll(expectedWarnings, MessageId.HTM_014);
+    testValidateDocument("invalid/xhtml-extension", expectedErrors, expectedWarnings);
+  }
 }
diff --git a/src/test/java/com/adobe/epubcheck/api/Epub30CheckExpandedTest.java b/src/test/java/com/adobe/epubcheck/api/Epub30CheckExpandedTest.java
@@ -476,7 +476,7 @@ public void testValidateEPUB30_circularFallback()
   public void testValidateEPUB30_nonresolvingFallback()
   {
     List<MessageId> expectedErrors = new ArrayList<MessageId>();
-    Collections.addAll(expectedErrors, MessageId.RSC_005, MessageId.OPF_040, MessageId.OPF_040, MessageId.MED_003);
+    Collections.addAll(expectedErrors, MessageId.RSC_005, MessageId.OPF_040, MessageId.MED_003);
     List<MessageId> expectedWarnings = new ArrayList<MessageId>();
     //dupe messages, tbf
     testValidateDocument("invalid/fallbacks-nonresolving/", expectedErrors, expectedWarnings);
@@ -694,4 +694,14 @@ public void testCollectionPreview() {
 	  testValidateDocument("valid/collections-preview/", expectedErrors, expectedWarnings, expectedFatals,true);
 	}
 
+  @Test
+  public void testXHTMExtension()
+  {
+    List<MessageId> expectedErrors = new ArrayList<MessageId>();
+    List<MessageId> expectedWarnings = new ArrayList<MessageId>();
+    List<MessageId> expectedFatals = new ArrayList<MessageId>();
+    Collections.addAll(expectedWarnings, MessageId.HTM_014a);
+    testValidateDocument("invalid/xhtml-extension", expectedErrors, expectedWarnings, expectedFatals, true);
+  }
+
 }
diff --git a/src/test/resources/20/expanded/invalid/xhtml-extension/EPUB/lorem.css b/src/test/resources/20/expanded/invalid/xhtml-extension/EPUB/lorem.css
@@ -0,0 +1,4 @@
+body {
+    color:black;
+    font-family: arial, helvetica, sans-serif;
+}
diff --git a/src/test/resources/20/expanded/invalid/xhtml-extension/EPUB/lorem.foo b/src/test/resources/20/expanded/invalid/xhtml-extension/EPUB/lorem.foo
@@ -0,0 +1,100 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
+    "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
+<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="la">
+	<head>
+		<title>Lorem Ipsum</title>
+		<link type="text/css" rel="stylesheet" href="lorem.css" />
+	</head>
+	<body>
+		<h1>Lorem Ipsum</h1>		
+		<div id="ch1">
+			<h2>Chapter 1</h2>
+			<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aliquam vel purus mauris, ut
+				auctor massa. Pellentesque non nunc risus. Fusce a massa augue. Nunc erat ante,
+				auctor id varius ac, vestibulum non purus. Quisque non dui in sem consectetur
+				condimentum non ac quam. Quisque ultricies nulla nec urna fringilla pretium.
+				Pellentesque dictum pulvinar purus in mattis. Aliquam vestibulum orci sed magna
+				vestibulum a sollicitudin lectus pharetra. Suspendisse luctus risus imperdiet nunc
+				condimentum malesuada. Nulla fringilla vulputate vestibulum. Sed diam dui, fringilla
+				quis sagittis nec, viverra et nibh.</p>
+
+			<p>Sed sollicitudin accumsan augue, quis pulvinar sem volutpat at. Vestibulum rutrum
+				bibendum augue sit amet accumsan. Etiam tempus malesuada libero vestibulum
+				fringilla. Maecenas diam nulla, ultricies ac sodales vitae, viverra ut velit.
+				Vivamus posuere, mi sit amet vehicula tempus, nibh purus scelerisque enim, non
+				vestibulum erat arcu in libero. Aliquam vel convallis nibh. Sed in nisi ipsum. Sed
+				sed est justo, in lacinia nulla.</p>
+
+			<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed luctus est vel lacus
+				ullamcorper vestibulum. Mauris est sapien, pharetra id feugiat in, ornare a erat.
+				Nam consectetur vehicula nisi vel faucibus. Morbi blandit augue nec lacus malesuada
+				venenatis. Cum sociis natoque penatibus et magnis dis parturient montes, nascetur
+				ridiculus mus. Maecenas consectetur, odio vitae suscipit ullamcorper, arcu ligula
+				pellentesque sem, quis rhoncus enim eros id lectus. Nam ornare dui est, vel posuere
+				metus. Quisque non nisl metus. Pellentesque id mi nunc, in gravida metus. Nullam
+				neque tellus, ultricies quis laoreet vitae, imperdiet at nunc. Ut laoreet massa quis
+				quam vulputate et ultricies nibh consectetur. Donec convallis, nulla id ultricies
+				ullamcorper, diam tortor interdum dolor, vel tempor lectus urna ut est. Praesent
+				convallis lacus vitae justo lobortis euismod. In at ante elit.</p>
+
+			<p>Aenean quis consectetur justo. Nulla nec enim nisl. Etiam rutrum volutpat tellus, a
+				scelerisque mauris malesuada sit amet. Suspendisse quis urna augue. Proin tempus
+				hendrerit libero non cursus. Praesent non massa at nisl luctus facilisis. Nullam
+				pulvinar, ligula eu porta ornare, mi mi accumsan orci, a iaculis tortor lorem quis
+				dolor. Phasellus ante nibh, pulvinar ac pulvinar eu, pulvinar ac enim.</p>
+
+			<p>Donec vel velit id elit volutpat vestibulum vitae a erat. Duis id est id magna
+				aliquam pretium nec sit amet nibh. Nullam condimentum suscipit felis, sed interdum
+				felis dictum ac. Phasellus non nisi quis magna pellentesque auctor. Cras risus
+				lectus, viverra eu fringilla malesuada, rhoncus et est. Etiam rhoncus pharetra
+				accumsan. Nullam suscipit tellus felis.</p>
+		</div>
+		<div id="ch2">
+			<h2>Chapter 2</h2>
+			<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla laoreet nibh felis.
+				Vestibulum ante ipsum primis in faucibus orci luctus et ultrices posuere cubilia
+				Curae; Etiam est sapien, dapibus eget gravida nec, accumsan a turpis. Nunc in nisi
+				ut dolor elementum porttitor. Mauris hendrerit pulvinar tincidunt. Etiam metus
+				metus, ullamcorper ut varius lacinia, luctus et nibh. Donec ut metus enim, id
+				faucibus nunc. Quisque ut iaculis mauris. Duis pellentesque nulla ut eros ultricies
+				quis condimentum eros adipiscing. Sed porta ultrices diam, ut sagittis lectus mattis
+				a. Phasellus gravida, sapien vitae mollis interdum, dui neque tempor arcu, ac ornare
+				leo ipsum ut nisl.</p>
+
+			<p>Donec porta, odio et aliquet molestie, felis tellus fermentum leo, id interdum magna
+				massa quis ligula. Integer elementum mauris eget nisl eleifend facilisis nec sit
+				amet tellus. Morbi consectetur dignissim egestas. Donec pulvinar, enim eu auctor
+				cursus, turpis arcu venenatis turpis, eu cursus magna nisl sit amet ante. Curabitur
+				eleifend arcu eget nibh facilisis mattis. Etiam nisl nunc, semper vitae condimentum
+				sed, viverra sit amet lacus. Curabitur et orci augue. Suspendisse sollicitudin
+				vulputate risus, sit amet consequat erat mollis eu. Nunc sodales tincidunt
+				tincidunt.</p>
+
+			<p>Aliquam erat volutpat. Aliquam ornare augue et nulla consequat commodo. Quisque
+				dictum rhoncus orci vel euismod. Proin leo turpis, adipiscing quis facilisis id,
+				condimentum sed metus. Nullam pellentesque scelerisque est nec tristique. Nunc augue
+				turpis, consequat non varius quis, aliquam auctor dolor. Cras luctus dignissim justo
+				sit amet laoreet. Quisque vel ipsum quis massa suscipit vehicula.</p>
+
+			<p>Pellentesque habitant morbi tristique senectus et netus et malesuada fames ac turpis
+				egestas. Vivamus fringilla eleifend magna, vel commodo turpis egestas at.
+				Pellentesque habitant morbi tristique senectus et netus et malesuada fames ac turpis
+				egestas. Sed eu lorem quam, et sagittis libero. Maecenas vel ante id sem bibendum
+				laoreet nec dignissim justo. Class aptent taciti sociosqu ad litora torquent per
+				conubia nostra, per inceptos himenaeos. Fusce eu lorem orci, eu viverra nisi. Lorem
+				ipsum dolor sit amet, consectetur adipiscing elit. Vestibulum dapibus commodo
+				pellentesque. Maecenas quis est accumsan est interdum pharetra egestas nec lorem.
+				Nam a lectus sit amet justo facilisis suscipit.</p>
+
+			<p>Integer dolor dolor, volutpat id commodo id, gravida id risus. Donec consectetur
+				sollicitudin sem, non auctor urna pulvinar non. Vivamus ipsum nisi, commodo sed
+				scelerisque id, porta nec massa. Vestibulum ac risus et augue faucibus fermentum ut
+				et nisi. Integer tincidunt suscipit ipsum, sed interdum felis mollis sed.
+				Suspendisse potenti. Praesent et mauris et quam consequat tristique. Morbi mi dolor,
+				pharetra quis rutrum quis, fringilla in tortor. Sed a nulla vitae leo dapibus
+				cursus. Aliquam erat volutpat. Integer purus purus, dictum id bibendum at, lobortis
+				quis metus.</p>
+		</div>
+	</body>
+</html>
diff --git a/src/test/resources/20/expanded/invalid/xhtml-extension/EPUB/lorem.ncx b/src/test/resources/20/expanded/invalid/xhtml-extension/EPUB/lorem.ncx
@@ -0,0 +1,29 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<ncx xmlns:ncx="http://www.daisy.org/z3986/2005/ncx/"
+    xmlns="http://www.daisy.org/z3986/2005/ncx/"
+    version="2005-1"
+    xml:lang="en">
+    <head>
+        <meta name="dtb:uid" content="urn:uuid:550e8400-e29b-41d4-a716-4466674412314"/>
+        <meta name="dtb:depth" content="1"/>
+        <meta name="dtb:totalPageCount" content="0"/>
+        <meta name="dtb:maxPageNumber" content="0"/>
+    </head>
+    <docTitle>
+        <text>Lorem Ipsum</text>
+    </docTitle>
+    <navMap>
+        <navPoint id="ch1" playOrder="1">
+            <navLabel>
+                <text>Chapter 1</text>
+            </navLabel>
+            <content src="lorem.foo#ch1"/>
+        </navPoint>
+        <navPoint id="ch2" playOrder="2">
+            <navLabel>
+                <text>Chapter 2</text>
+            </navLabel>
+            <content src="lorem.foo#ch2"/>
+        </navPoint>
+    </navMap>
+</ncx>
diff --git a/src/test/resources/20/expanded/invalid/xhtml-extension/EPUB/lorem.opf b/src/test/resources/20/expanded/invalid/xhtml-extension/EPUB/lorem.opf
@@ -0,0 +1,17 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<package xmlns="http://www.idpf.org/2007/opf" version="2.0" unique-identifier="uid">
+    <metadata xmlns:dc="http://purl.org/dc/elements/1.1/">
+        <dc:identifier id="uid">urn:uuid:550e8400-e29b-41d4-a716-4466674412314</dc:identifier>
+        <dc:title>Lorem Ipsum</dc:title>        
+        <dc:language>la</dc:language>
+        <dc:date>2011-09-01</dc:date>        
+    </metadata> 
+    <manifest>
+        <item id="t1" href="lorem.foo" media-type="application/xhtml+xml" /> 
+        <item id="ncx" href="lorem.ncx" media-type="application/x-dtbncx+xml" /> 
+        <item id="css" href="lorem.css" media-type="text/css" />
+    </manifest>
+    <spine toc="ncx">
+        <itemref idref="t1" />        
+    </spine>    
+</package>
diff --git a/src/test/resources/20/expanded/invalid/xhtml-extension/META-INF/container.xml b/src/test/resources/20/expanded/invalid/xhtml-extension/META-INF/container.xml
@@ -0,0 +1,7 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<container xmlns="urn:oasis:names:tc:opendocument:xmlns:container" version="1.0">
+	<rootfiles>
+		<rootfile full-path="EPUB/lorem.opf" 	
+			media-type="application/oebps-package+xml"/>
+	</rootfiles>
+</container>
diff --git a/src/test/resources/20/expanded/invalid/xhtml-extension/mimetype b/src/test/resources/20/expanded/invalid/xhtml-extension/mimetype
@@ -0,0 +1 @@
+application/epub+zip
diff --git a/src/test/resources/30/epub/invalid/issue265c.epub b/src/test/resources/30/epub/invalid/issue265c.epub
diff --git a/src/test/resources/30/epub/valid/issue158.epub b/src/test/resources/30/epub/valid/issue158.epub
diff --git a/src/test/resources/30/expanded/invalid/xhtml-extension/EPUB/lorem.css b/src/test/resources/30/expanded/invalid/xhtml-extension/EPUB/lorem.css
@@ -0,0 +1,4 @@
+body {
+    color:black;
+    font-family: arial, helvetica, sans-serif;
+}
-Original file line number
+Diff line change
@@ Expand Up / @@ -156,7 +156,6 @@ public void runChecks() @@
           }
           report.info(item.getPath(), FeatureEnum.DECLARED_MIMETYPE, item.getMimeType());
-          checkItem(item, opfHandler);
         }
         checkGuide();
@@ Expand Down @@