Unescape URLs from EPUB 2 NCX and EPUB 3 nav docs (#98)

vers-one · May 27, 2023 · 1d3ae21 · 1d3ae21
1 parent 88d0cb0
commit 1d3ae21
Show file tree

Hide file tree

Showing 6 changed files with 139 additions and 4 deletions.
diff --git a/Source/VersOne.Epub.Test/Unit/Readers/Epub2NcxReaderTests.cs b/Source/VersOne.Epub.Test/Unit/Readers/Epub2NcxReaderTests.cs
@@ -256,6 +256,22 @@ public class Epub2NcxReaderTests
             </ncx>
             """;
 
+        private const string NCX_FILE_WITH_ESCAPED_CONTENT_SRC_ATTRIBUTE = """
+            <?xml version='1.0' encoding='utf-8'?>
+            <ncx xmlns="http://www.daisy.org/z3986/2005/ncx/">
+              <head />
+              <docTitle />
+              <navMap>
+                <navPoint id="navpoint-1">
+                  <navLabel>
+                    <text>Chapter 1</text>
+                  </navLabel>
+                  <content src="chapter%31.html" />
+                </navPoint>
+              </navMap>
+            </ncx>
+            """;
+
         private const string NCX_FILE_WITHOUT_CONTENT_SRC_ATTRIBUTE = """
             <?xml version='1.0' encoding='utf-8'?>
             <ncx xmlns="http://www.daisy.org/z3986/2005/ncx/">
@@ -877,6 +893,42 @@ public async void ReadEpub2NcxAsyncWithoutNavLabelTextTest()
             await TestFailingReadOperation(NCX_FILE_WITHOUT_NAVLABEL_TEXT_ELEMENT);
         }
 
+        [Fact(DisplayName = "Reading an NCX file with a URI-escaped 'src' attribute in a 'content' XML element should succeed")]
+        public async void ReadEpub2NcxAsyncWithEscapedContentSrcTest()
+        {
+            Epub2Ncx expectedEpub2Ncx = new
+            (
+                filePath: NCX_FILE_PATH,
+                head: new Epub2NcxHead(),
+                docTitle: null,
+                docAuthors: null,
+                navMap: new Epub2NcxNavigationMap
+                (
+                    items: new List<Epub2NcxNavigationPoint>()
+                    {
+                        new Epub2NcxNavigationPoint
+                        (
+                            id: "navpoint-1",
+                            navigationLabels: new List<Epub2NcxNavigationLabel>()
+                            {
+                                new Epub2NcxNavigationLabel
+                                (
+                                    text: "Chapter 1"
+                                )
+                            },
+                            content: new Epub2NcxContent
+                            (
+                                source: "chapter1.html"
+                            )
+                        )
+                    }
+                ),
+                pageList: null,
+                navLists: null
+            );
+            await TestSuccessfulReadOperation(NCX_FILE_WITH_ESCAPED_CONTENT_SRC_ATTRIBUTE, expectedEpub2Ncx);
+        }
+
         [Fact(DisplayName = "ReadEpub2NcxAsync should throw Epub2NcxException if a 'content' XML element has no 'src' attribute")]
         public async void ReadEpub2NcxAsyncWithoutContentSrcTest()
         {

diff --git a/Source/VersOne.Epub.Test/Unit/Readers/Epub3NavDocumentReaderTests.cs b/Source/VersOne.Epub.Test/Unit/Readers/Epub3NavDocumentReaderTests.cs
@@ -145,6 +145,20 @@ public class Epub3NavDocumentReaderTests
             </html>
             """;
 
+        private const string NAV_FILE_WITH_ESCAPED_HREF_IN_A_ELEMENT = """
+            <html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops">
+              <body>
+                <nav epub:type="toc">
+                  <ol>
+                    <li>
+                      <a href="chapter%31.html">Chapter 1</a>
+                    </li>
+                  </ol>
+                </nav>
+              </body>
+            </html>
+            """;
+
         private static EpubPackage MinimalEpubPackageWithNav =>
             new
             (
@@ -402,6 +416,37 @@ public async void ReadEpub3NavDocumentAsyncWithEmptyLiElement()
             await TestFailingReadOperation(NAV_FILE_WITH_EMPTY_LI_ELEMENT);
         }
 
+        [Fact(DisplayName = "Reading a NAV file with a URI-escaped 'href' attribute in an 'a' XML element should succeed")]
+        public async void ReadEpub3NavDocumentAsyncWithEscapedAHrefTest()
+        {
+            Epub3NavDocument expectedEpub3NavDocument = new
+            (
+                filePath: NAV_FILE_PATH,
+                navs: new List<Epub3Nav>()
+                {
+                    new Epub3Nav
+                    (
+                        type: Epub3StructuralSemanticsProperty.TOC,
+                        ol: new Epub3NavOl
+                        (
+                            lis: new List<Epub3NavLi>()
+                            {
+                                new Epub3NavLi
+                                (
+                                    anchor: new Epub3NavAnchor
+                                    (
+                                        href: "chapter1.html",
+                                        text: "Chapter 1"
+                                    )
+                                )
+                            }
+                        )
+                    )
+                }
+            );
+            await TestSuccessfulReadOperation(NAV_FILE_WITH_ESCAPED_HREF_IN_A_ELEMENT, expectedEpub3NavDocument);
+        }
+
         private static async Task TestSuccessfulReadOperation(string navFileContent, Epub3NavDocument expectedEpub3NavDocument, EpubReaderOptions? epubReaderOptions = null)
         {
             TestZipFile testZipFile = CreateTestZipFileWithNavFile(navFileContent);

diff --git a/Source/VersOne.Epub.Test/Unit/Readers/PackageReaderTests.cs b/Source/VersOne.Epub.Test/Unit/Readers/PackageReaderTests.cs
@@ -165,6 +165,17 @@ public class PackageReaderTests
             </package>
             """;
 
+        private const string OPF_FILE_WITH_ESCAPED_HREF_IN_MANIFEST_ITEM = $"""
+            <?xml version='1.0' encoding='UTF-8'?>
+            <package xmlns="http://www.idpf.org/2007/opf" version="3.0">
+              <metadata />
+              <manifest>
+                <item id="item-1" href="chapter%31.html" media-type="application/xhtml+xml" />
+              </manifest>
+              <spine />
+            </package>
+            """;
+
         private const string OPF_FILE_WITHOUT_HREF_IN_MANIFEST_ITEM = $"""
             <?xml version='1.0' encoding='UTF-8'?>
             <package xmlns="http://www.idpf.org/2007/opf" version="3.0">
@@ -743,6 +754,33 @@ public async void ReadPackageWithoutManifestItemIdWithSkippingInvalidManifestIte
             await TestSuccessfulReadOperationWithSkippingInvalidManifestItems(OPF_FILE_WITHOUT_ID_IN_MANIFEST_ITEM, MinimalEpub3Package);
         }
 
+        [Fact(DisplayName = "Read an OPF package with a URI-escaped 'href' attribute in a manifest item XML node should succeed")]
+        public async void ReadPackageWithEscapedManifestItemHrefTest()
+        {
+            EpubPackage expectedPackage = new
+            (
+                uniqueIdentifier: null,
+                epubVersion: EpubVersion.EPUB_3,
+                metadata: new EpubMetadata(),
+                manifest: new EpubManifest
+                (
+                    id: null,
+                    items: new List<EpubManifestItem>()
+                    {
+                        new EpubManifestItem
+                        (
+                            id: "item-1",
+                            href: "chapter1.html",
+                            mediaType: "application/xhtml+xml"
+                        )
+                    }
+                ),
+                spine: new EpubSpine(),
+                guide: null
+            );
+            await TestSuccessfulReadOperationWithSkippingInvalidManifestItems(OPF_FILE_WITH_ESCAPED_HREF_IN_MANIFEST_ITEM, expectedPackage);
+        }
+
         [Fact(DisplayName = "Trying to read OPF package without 'href' attribute in a manifest item XML node should fail with EpubPackageException")]
         public async void ReadPackageWithoutManifestItemHrefTest()
         {

diff --git a/Source/VersOne.Epub/Readers/Epub2NcxReader.cs b/Source/VersOne.Epub/Readers/Epub2NcxReader.cs
@@ -241,7 +241,7 @@ private static Epub2NcxContent ReadNavigationContent(XElement navigationContentN
                         id = attributeValue;
                         break;
                     case "src":
-                        source = attributeValue;
+                        source = Uri.UnescapeDataString(attributeValue);
                         break;
                 }
             }

diff --git a/Source/VersOne.Epub/Readers/Epub3NavDocumentReader.cs b/Source/VersOne.Epub/Readers/Epub3NavDocumentReader.cs
@@ -170,7 +170,7 @@ private static Epub3NavAnchor ReadEpub3NavAnchor(XElement epub3NavAnchorNode)
                 switch (navAnchorNodeAttribute.GetLowerCaseLocalName())
                 {
                     case "href":
-                        href = attributeValue;
+                        href = Uri.UnescapeDataString(attributeValue);
                         break;
                     case "title":
                         title = attributeValue;

diff --git a/Source/VersOne.Epub/Readers/NavigationReader.cs b/Source/VersOne.Epub/Readers/NavigationReader.cs
@@ -41,7 +41,7 @@ private static List<EpubNavigationItemRef> GetNavigationItems(EpubSchema epubSch
                     Epub2NcxNavigationLabel? firstNavigationLabel = navigationPoint.NavigationLabels.FirstOrDefault() ??
                         throw new Epub2NcxException($"Incorrect EPUB 2 NCX: navigation point \"{navigationPoint.Id}\" should contain at least one navigation label.");
                     string title = firstNavigationLabel.Text;
-                    string source = Uri.UnescapeDataString(navigationPoint.Content.Source);
+                    string source = navigationPoint.Content.Source;
                     if (!ContentPathUtils.IsLocalPath(source))
                     {
                         throw new Epub2NcxException($"Incorrect EPUB 2 NCX: content source \"{source}\" cannot be a remote resource.");
@@ -100,7 +100,7 @@ private static List<EpubNavigationItemRef> GetNavigationItems(EpubSchema epubSch
                         List<EpubNavigationItemRef> nestedItems = GetNavigationItems(epubSchema, epubContentRef, epub3NavLi.ChildOl, epub3NavigationBaseDirectoryPath);
                         if (navAnchor.Href != null)
                         {
-                            string href = Uri.UnescapeDataString(navAnchor.Href);
+                            string href = navAnchor.Href;
                             if (!ContentPathUtils.IsLocalPath(href))
                             {
                                 throw new Epub3NavException($"Incorrect EPUB 3 navigation document: anchor href \"{href}\" cannot be a remote resource.");