From 17b4d0a2b595d1504f3d957268e2085ae0f80db7 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Wed, 9 Aug 2017 15:53:02 +0200 Subject: [PATCH 1/9] EPUBGenerator: avoid container version roundtrip to double This will be an error for EPUB3: ERROR(RSC-005): test.epub/META-INF/container.xml(2,85): Error while parsing file: value of attribute "version" is invalid; must be equal to "1.0" But it does not hurt for EPUB2, either. --- src/lib/EPUBGenerator.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 3357cf2..1033c0f 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -117,7 +117,7 @@ void EPUBGenerator::writeContainer() EPUBXMLSink sink; RVNGPropertyList containerAttrs; - containerAttrs.insert("version", "1.0"); + containerAttrs.insert("version", RVNGPropertyFactory::newStringProp("1.0")); containerAttrs.insert("xmlns", "urn:oasis:names:tc:opendocument:xmlns:container"); sink.openElement("container", containerAttrs); -- 2.12.3 From 8ca1fe2b9db9bacd6e868e69a0909a441fb6a7f8 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Wed, 9 Aug 2017 16:00:05 +0200 Subject: [PATCH 2/9] EPUBGenerator: avoid opf:scheme attribute when writing the UUID This will be an error for EPUB3: ERROR(RSC-005): test.epub/OEBPS/content.opf(2,292): Error while parsing file: found attribute "opf:scheme", but no attributes allowed here But it's optional for EPUB2 already. --- src/lib/EPUBGenerator.cpp | 1 - 1 file changed, 1 deletion(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 1033c0f..14e3c58 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -199,7 +199,6 @@ void EPUBGenerator::writeRoot() RVNGPropertyList identifierAttrs; identifierAttrs.insert("id", uniqueId); - identifierAttrs.insert("opf:scheme", "UUID"); sink.openElement("dc:identifier", identifierAttrs); // The identifier element is required to have a unique character content. std::stringstream identifierStream("urn:uuid:"); -- 2.12.3 From aa71784fcee0404c2f136f035887ca4c52d0e756 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Wed, 9 Aug 2017 16:11:13 +0200 Subject: [PATCH 3/9] EPUBGenerator: avoid empty dc:title element This is a warning for EPUB2, but it'll be an error for EPUB3: ERROR(RSC-005): test.epub/OEBPS/content.opf(2,337): Error while parsing file: character content of element "dc:title" invalid; must be a string with length at least 1 (actual length was 0) The problem is that for ODF/librevenge this element is optional, so work it around by adding a zero-width space. A later commit should read the optional title of declared with a librevenge API call, though. --- src/lib/EPUBGenerator.cpp | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 14e3c58..75ccb5a 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -208,8 +208,12 @@ void EPUBGenerator::writeRoot() sink.insertCharacters(identifierCharactrs.c_str()); sink.closeElement("dc:identifier"); + // Zero-width space as it must be at least one character in length after + // white space has been trimmed. sink.openElement("dc:title"); + sink.insertCharacters("\u200b"); sink.closeElement("dc:title"); + sink.openElement("dc:language"); sink.closeElement("dc:language"); -- 2.12.3 From a4585b8f35c76472eb91688c9177b9f532c290d8 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Wed, 9 Aug 2017 16:15:12 +0200 Subject: [PATCH 4/9] EPUBGenerator: avoid empty dc:language element Same story as with dc:title: ERROR(RSC-005): test.epub/OEBPS/content.opf(2,362): Error while parsing file: character content of element "dc:language" invalid; must be an RFC 3066 language identifier (This is a warning for EPUB2, but it'll be an error for EPUB3; this is optional for ODF; later commit should read the this info, though.) --- src/lib/EPUBGenerator.cpp | 1 + 1 file changed, 1 insertion(+) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 75ccb5a..40ae0cc 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -215,6 +215,7 @@ void EPUBGenerator::writeRoot() sink.closeElement("dc:title"); sink.openElement("dc:language"); + sink.insertCharacters("en"); sink.closeElement("dc:language"); sink.closeElement("metadata"); -- 2.12.3 From 862ec6735c25760edadf05d83717daaf65f39f99 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Wed, 9 Aug 2017 16:47:12 +0200 Subject: [PATCH 5/9] [ABI CHANGE] optionally support generating EPUB3 output Versions are 20 and 30 to be consistent with . --- inc/libepubgen/EPUBDrawingGenerator.h | 6 +++++- inc/libepubgen/EPUBPresentationGenerator.h | 6 +++++- inc/libepubgen/EPUBTextGenerator.h | 6 +++++- src/lib/EPUBDrawingGenerator.cpp | 10 +++++----- src/lib/EPUBGenerator.cpp | 8 ++++++-- src/lib/EPUBGenerator.h | 4 +++- src/lib/EPUBPagedGenerator.cpp | 10 +++++----- src/lib/EPUBPagedGenerator.h | 2 +- src/lib/EPUBPresentationGenerator.cpp | 10 +++++----- src/lib/EPUBTextGenerator.cpp | 10 +++++----- 10 files changed, 45 insertions(+), 27 deletions(-) diff --git a/inc/libepubgen/EPUBDrawingGenerator.h b/inc/libepubgen/EPUBDrawingGenerator.h index 48bfc99..963e3b8 100644 --- a/inc/libepubgen/EPUBDrawingGenerator.h +++ b/inc/libepubgen/EPUBDrawingGenerator.h @@ -32,7 +32,11 @@ class EPUBGENAPI EPUBDrawingGenerator : public librevenge::RVNGDrawingInterface class Impl; public: - explicit EPUBDrawingGenerator(EPUBPackage *package, EPUBSplitMethod split = EPUB_SPLIT_METHOD_PAGE_BREAK); + /** Constructor. + * + * @param[in] version possible values: 20, 30. + */ + explicit EPUBDrawingGenerator(EPUBPackage *package, EPUBSplitMethod split = EPUB_SPLIT_METHOD_PAGE_BREAK, int version = 20); EPUBDrawingGenerator(EPUBEmbeddingContact &contact, EPUBPackage *package); ~EPUBDrawingGenerator() override; diff --git a/inc/libepubgen/EPUBPresentationGenerator.h b/inc/libepubgen/EPUBPresentationGenerator.h index 0a8152a..512c52d 100644 --- a/inc/libepubgen/EPUBPresentationGenerator.h +++ b/inc/libepubgen/EPUBPresentationGenerator.h @@ -32,7 +32,11 @@ class EPUBGENAPI EPUBPresentationGenerator: public librevenge::RVNGPresentationI class Impl; public: - explicit EPUBPresentationGenerator(EPUBPackage *package, EPUBSplitMethod split = EPUB_SPLIT_METHOD_PAGE_BREAK); + /** Constructor. + * + * @param[in] version possible values: 20, 30. + */ + explicit EPUBPresentationGenerator(EPUBPackage *package, EPUBSplitMethod split = EPUB_SPLIT_METHOD_PAGE_BREAK, int version = 20); EPUBPresentationGenerator(EPUBEmbeddingContact &contact, EPUBPackage *package); ~EPUBPresentationGenerator() override; diff --git a/inc/libepubgen/EPUBTextGenerator.h b/inc/libepubgen/EPUBTextGenerator.h index abf9e7f..664f673 100644 --- a/inc/libepubgen/EPUBTextGenerator.h +++ b/inc/libepubgen/EPUBTextGenerator.h @@ -32,7 +32,11 @@ class EPUBGENAPI EPUBTextGenerator : public librevenge::RVNGTextInterface struct Impl; public: - explicit EPUBTextGenerator(EPUBPackage *package, EPUBSplitMethod split = EPUB_SPLIT_METHOD_PAGE_BREAK); + /** Constructor. + * + * @param[in] version possible values: 20, 30. + */ + explicit EPUBTextGenerator(EPUBPackage *package, EPUBSplitMethod split = EPUB_SPLIT_METHOD_PAGE_BREAK, int version = 20); EPUBTextGenerator(EPUBEmbeddingContact &contact, EPUBPackage *package); ~EPUBTextGenerator() override; diff --git a/src/lib/EPUBDrawingGenerator.cpp b/src/lib/EPUBDrawingGenerator.cpp index e25a377..bcb4994 100644 --- a/src/lib/EPUBDrawingGenerator.cpp +++ b/src/lib/EPUBDrawingGenerator.cpp @@ -20,16 +20,16 @@ using librevenge::RVNGString; class EPUBDrawingGenerator::Impl : public EPUBPagedGenerator { public: - Impl(EPUBPackage *const package, const EPUBSplitMethod method); + Impl(EPUBPackage *const package, const EPUBSplitMethod method, int version); }; -EPUBDrawingGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod method) - : EPUBPagedGenerator(package, method) +EPUBDrawingGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod method, int version) + : EPUBPagedGenerator(package, method, version) { } -EPUBDrawingGenerator::EPUBDrawingGenerator(EPUBPackage *const package, EPUBSplitMethod split) - : m_impl(new Impl(package, split)) +EPUBDrawingGenerator::EPUBDrawingGenerator(EPUBPackage *const package, EPUBSplitMethod split, int version) + : m_impl(new Impl(package, split, version)) { } diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 40ae0cc..4888677 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -25,7 +25,7 @@ using librevenge::RVNGPropertyFactory; using librevenge::RVNGPropertyList; using librevenge::RVNGString; -EPUBGenerator::EPUBGenerator(EPUBPackage *const package, const EPUBSplitMethod split) +EPUBGenerator::EPUBGenerator(EPUBPackage *const package, const EPUBSplitMethod split, int version) : m_package(package) , m_manifest() , m_htmlManager(m_manifest) @@ -39,6 +39,7 @@ EPUBGenerator::EPUBGenerator(EPUBPackage *const package, const EPUBSplitMethod s , m_metadata() , m_currentHtml() , m_splitGuard(split) + , m_version(version) { } @@ -190,7 +191,10 @@ void EPUBGenerator::writeRoot() packageAttrs.insert("xmlns:dc", "http://purl.org/dc/elements/1.1/"); packageAttrs.insert("xmlns:dcterms", "http://purl.org/dc/terms/"); packageAttrs.insert("xmlns:opf", "http://www.idpf.org/2007/opf"); - packageAttrs.insert("version", RVNGPropertyFactory::newStringProp("2.0")); + if (m_version == 30) + packageAttrs.insert("version", RVNGPropertyFactory::newStringProp("3.0")); + else + packageAttrs.insert("version", RVNGPropertyFactory::newStringProp("2.0")); packageAttrs.insert("unique-identifier", uniqueId); sink.openElement("package", packageAttrs); diff --git a/src/lib/EPUBGenerator.h b/src/lib/EPUBGenerator.h index 51dd911..a0ef8ac 100644 --- a/src/lib/EPUBGenerator.h +++ b/src/lib/EPUBGenerator.h @@ -33,7 +33,7 @@ class EPUBGenerator EPUBGenerator &operator=(const EPUBGenerator &); public: - EPUBGenerator(EPUBPackage *package, EPUBSplitMethod method); + EPUBGenerator(EPUBPackage *package, EPUBSplitMethod method, int version); virtual ~EPUBGenerator(); void startDocument(const librevenge::RVNGPropertyList &props); @@ -75,6 +75,8 @@ private: EPUBHTMLGeneratorPtr_t m_currentHtml; EPUBSplitGuard m_splitGuard; + + int m_version; }; } diff --git a/src/lib/EPUBPagedGenerator.cpp b/src/lib/EPUBPagedGenerator.cpp index 913a592..6a3bff0 100644 --- a/src/lib/EPUBPagedGenerator.cpp +++ b/src/lib/EPUBPagedGenerator.cpp @@ -25,7 +25,7 @@ class EPUBPagedGenerator::Impl : public EPUBGenerator Impl &operator=(const Impl &); public: - Impl(EPUBPackage *const package, EPUBSplitMethod method); + Impl(EPUBPackage *const package, EPUBSplitMethod method, int version); private: void startHtmlFile() override; @@ -35,8 +35,8 @@ public: bool m_firstPage; }; -EPUBPagedGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod method) - : EPUBGenerator(package, method) +EPUBPagedGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod method, int version) + : EPUBGenerator(package, method, version) , m_firstPage(true) { } @@ -59,8 +59,8 @@ void EPUBPagedGenerator::Impl::endHtmlFile() { } -EPUBPagedGenerator::EPUBPagedGenerator(EPUBPackage *const package, const EPUBSplitMethod method) - : m_impl(new Impl(package, method)) +EPUBPagedGenerator::EPUBPagedGenerator(EPUBPackage *const package, const EPUBSplitMethod method, int version) + : m_impl(new Impl(package, method, version)) { } diff --git a/src/lib/EPUBPagedGenerator.h b/src/lib/EPUBPagedGenerator.h index f1d124f..74d70da 100644 --- a/src/lib/EPUBPagedGenerator.h +++ b/src/lib/EPUBPagedGenerator.h @@ -26,7 +26,7 @@ class EPUBPagedGenerator: public librevenge::RVNGPresentationInterface class Impl; public: - EPUBPagedGenerator(EPUBPackage *package, EPUBSplitMethod method); + EPUBPagedGenerator(EPUBPackage *package, EPUBSplitMethod method, int version); void setSplitHeadingLevel(unsigned level); void setSplitSize(unsigned size); diff --git a/src/lib/EPUBPresentationGenerator.cpp b/src/lib/EPUBPresentationGenerator.cpp index 5b2a2e9..80b5ac2 100644 --- a/src/lib/EPUBPresentationGenerator.cpp +++ b/src/lib/EPUBPresentationGenerator.cpp @@ -20,16 +20,16 @@ using librevenge::RVNGString; class EPUBPresentationGenerator::Impl : public EPUBPagedGenerator { public: - Impl(EPUBPackage *const package, EPUBSplitMethod method); + Impl(EPUBPackage *const package, EPUBSplitMethod method, int version); }; -EPUBPresentationGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod method) - : EPUBPagedGenerator(package, method) +EPUBPresentationGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod method, int version) + : EPUBPagedGenerator(package, method, version) { } -EPUBPresentationGenerator::EPUBPresentationGenerator(EPUBPackage *const package, EPUBSplitMethod method) - : m_impl(new Impl(package, method)) +EPUBPresentationGenerator::EPUBPresentationGenerator(EPUBPackage *const package, EPUBSplitMethod method, int version) + : m_impl(new Impl(package, method, version)) { (void) method; } diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index b3ca626..e8f785e 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -57,7 +57,7 @@ bool operator!=(const char *const left, const RVNGString &right) struct EPUBTextGenerator::Impl : public EPUBGenerator { - Impl(EPUBPackage *package, EPUBSplitMethod method); + Impl(EPUBPackage *package, EPUBSplitMethod method, int version); private: void startHtmlFile() override; @@ -83,8 +83,8 @@ private: Impl &operator=(const Impl &); }; -EPUBTextGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod method) - : EPUBGenerator(package, method) +EPUBTextGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod method, int version) + : EPUBGenerator(package, method, version) , m_inPageSpan(false) , m_inHeader(false) , m_inFooter(false) @@ -113,8 +113,8 @@ void EPUBTextGenerator::Impl::endHtmlFile() m_currentFooter->write(getHtml().get()); } -EPUBTextGenerator::EPUBTextGenerator(EPUBPackage *const package, const EPUBSplitMethod method) - : m_impl(new Impl(package, method)) +EPUBTextGenerator::EPUBTextGenerator(EPUBPackage *const package, const EPUBSplitMethod method, int version) + : m_impl(new Impl(package, method, version)) { (void) method; } -- 2.12.3 From cfd57f50c9d33781f90018d40902ccce68a13a5c Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Wed, 9 Aug 2017 17:19:06 +0200 Subject: [PATCH 6/9] EPUB3: write missing modification date ERROR(RSC-005): test.epub/OEBPS/content.opf(2,236): Error while parsing file: package dcterms:modified meta element must occur exactly once But not for EPUB2, which doesn't allow so. (Similar story as with dc:title: later commit should read the this optional info from librevenge.) --- src/lib/EPUBGenerator.cpp | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 4888677..1628a2b 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -7,6 +7,7 @@ * file, You can obtain one at http://mozilla.org/MPL/2.0/. */ +#include #include #include @@ -222,6 +223,23 @@ void EPUBGenerator::writeRoot() sink.insertCharacters("en"); sink.closeElement("dc:language"); + time_t now = 0; + time(&now); + const struct tm *local = localtime(&now); + if (m_version == 30 && local) + { + RVNGPropertyList metaAttrs; + metaAttrs.insert("property", "dcterms:modified"); + sink.openElement("meta", metaAttrs); + const int MAX_BUFFER = 1024; + char buffer[MAX_BUFFER]; + strftime(&buffer[0], MAX_BUFFER-1, "%Y-%m-%dT%H:%M:%SZ", local); + RVNGString result; + result.append(buffer); + sink.insertCharacters(result); + sink.closeElement("meta"); + } + sink.closeElement("metadata"); sink.openElement("manifest"); -- 2.12.3 From 575a09f637b5afe4d61387c7be3c8b2b67039ccb Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Wed, 9 Aug 2017 17:39:51 +0200 Subject: [PATCH 7/9] EPUB3: add missing nav property on first HTML stream ERROR(RSC-005): test30.epub/OEBPS/content.opf(2,459): Error while parsing file: Exactly one manifest item must declare the 'nav' property (number of 'nav' items: 0). This is the last error in OEBPS/content.opf in case of EPUB3. --- src/lib/EPUBGenerator.cpp | 11 ++++++++--- src/lib/EPUBGenerator.h | 1 + src/lib/EPUBHTMLManager.cpp | 9 +++++++-- src/lib/EPUBHTMLManager.h | 3 ++- src/lib/EPUBImageManager.cpp | 2 +- src/lib/EPUBManifest.cpp | 11 +++++++---- src/lib/EPUBManifest.h | 5 +++-- 7 files changed, 29 insertions(+), 13 deletions(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 1628a2b..f77aa6c 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -54,8 +54,8 @@ void EPUBGenerator::startDocument(const RVNGPropertyList &props) startNewHtmlFile(); - m_manifest.insert(EPUBPath("OEBPS/toc.ncx"), "application/x-dtbncx+xml", "toc.ncx"); - m_manifest.insert(m_stylesheetPath, "text/css", "stylesheet.css"); + m_manifest.insert(EPUBPath("OEBPS/toc.ncx"), "application/x-dtbncx+xml", "toc.ncx", ""); + m_manifest.insert(m_stylesheetPath, "text/css", "stylesheet.css", ""); } void EPUBGenerator::endDocument() @@ -90,7 +90,7 @@ void EPUBGenerator::startNewHtmlFile() m_splitGuard.onSplit(); - m_currentHtml = m_htmlManager.create(m_imageManager, m_listStyleManager, m_paragraphStyleManager, m_spanStyleManager, m_tableStyleManager, m_stylesheetPath); + m_currentHtml = m_htmlManager.create(m_imageManager, m_listStyleManager, m_paragraphStyleManager, m_spanStyleManager, m_tableStyleManager, m_stylesheetPath, *this); // restore state in the new file m_currentHtml->startDocument(m_documentProps); @@ -114,6 +114,11 @@ EPUBSplitGuard &EPUBGenerator::getSplitGuard() return m_splitGuard; } +int EPUBGenerator::getVersion() const +{ + return m_version; +} + void EPUBGenerator::writeContainer() { EPUBXMLSink sink; diff --git a/src/lib/EPUBGenerator.h b/src/lib/EPUBGenerator.h index a0ef8ac..1a67a88 100644 --- a/src/lib/EPUBGenerator.h +++ b/src/lib/EPUBGenerator.h @@ -47,6 +47,7 @@ public: const EPUBSplitGuard &getSplitGuard() const; EPUBSplitGuard &getSplitGuard(); + int getVersion() const; private: virtual void startHtmlFile() = 0; diff --git a/src/lib/EPUBHTMLManager.cpp b/src/lib/EPUBHTMLManager.cpp index 03dbf21..57636b9 100644 --- a/src/lib/EPUBHTMLManager.cpp +++ b/src/lib/EPUBHTMLManager.cpp @@ -13,6 +13,7 @@ #include "EPUBHTMLManager.h" #include "EPUBManifest.h" +#include "EPUBGenerator.h" namespace libepubgen { @@ -26,7 +27,7 @@ EPUBHTMLManager::EPUBHTMLManager(EPUBManifest &manifest) { } -const EPUBHTMLGeneratorPtr_t EPUBHTMLManager::create(EPUBImageManager &imageManager, EPUBListStyleManager &listStyleManager, EPUBParagraphStyleManager ¶graphStyleManager, EPUBSpanStyleManager &spanStyleManager, EPUBTableStyleManager &tableStyleManager, const EPUBPath &stylesheetPath) +const EPUBHTMLGeneratorPtr_t EPUBHTMLManager::create(EPUBImageManager &imageManager, EPUBListStyleManager &listStyleManager, EPUBParagraphStyleManager ¶graphStyleManager, EPUBSpanStyleManager &spanStyleManager, EPUBTableStyleManager &tableStyleManager, const EPUBPath &stylesheetPath, const EPUBGenerator &generator) { std::ostringstream nameBuf; nameBuf << "section" << std::setw(4) << std::setfill('0') << m_number.next(); @@ -35,7 +36,11 @@ const EPUBHTMLGeneratorPtr_t EPUBHTMLManager::create(EPUBImageManager &imageMana nameBuf << ".html"; m_paths.push_back(EPUBPath("OEBPS/sections") / nameBuf.str()); - m_manifest.insert(m_paths.back(), "application/xhtml+xml", m_ids.back()); + std::string properties; + if (m_number.current() == 1 && generator.getVersion() == 30) + // Only for the first HTML file. + properties = "nav"; + m_manifest.insert(m_paths.back(), "application/xhtml+xml", m_ids.back(), properties); m_contents.push_back(EPUBXMLSink()); diff --git a/src/lib/EPUBHTMLManager.h b/src/lib/EPUBHTMLManager.h index 84ecd2d..7dab33b 100644 --- a/src/lib/EPUBHTMLManager.h +++ b/src/lib/EPUBHTMLManager.h @@ -30,6 +30,7 @@ class EPUBSpanStyleManager; class EPUBTableStyleManager; class EPUBManifest; class EPUBPackage; +class EPUBGenerator; class EPUBHTMLManager { @@ -40,7 +41,7 @@ class EPUBHTMLManager public: explicit EPUBHTMLManager(EPUBManifest &manifest); - const EPUBHTMLGeneratorPtr_t create(EPUBImageManager &imageManager, EPUBListStyleManager &listStyleManager, EPUBParagraphStyleManager ¶graphStyleManager, EPUBSpanStyleManager &spanStyleManager, EPUBTableStyleManager &tableStyleManager, const EPUBPath &stylesheetPath); + const EPUBHTMLGeneratorPtr_t create(EPUBImageManager &imageManager, EPUBListStyleManager &listStyleManager, EPUBParagraphStyleManager ¶graphStyleManager, EPUBSpanStyleManager &spanStyleManager, EPUBTableStyleManager &tableStyleManager, const EPUBPath &stylesheetPath, const EPUBGenerator &generator); void writeTo(EPUBPackage &package); diff --git a/src/lib/EPUBImageManager.cpp b/src/lib/EPUBImageManager.cpp index 1b043d7..0179cad 100644 --- a/src/lib/EPUBImageManager.cpp +++ b/src/lib/EPUBImageManager.cpp @@ -86,7 +86,7 @@ const EPUBPath &EPUBImageManager::insert(const librevenge::RVNGBinaryData &data, const EPUBPath path(EPUBPath("OEBPS/images") / nameBuf.str()); - m_manifest.insert(path, mime, id); + m_manifest.insert(path, mime, id, ""); it = m_map.insert(MapType_t::value_type(data, path)).first; } diff --git a/src/lib/EPUBManifest.cpp b/src/lib/EPUBManifest.cpp index dcffdcc..813a097 100644 --- a/src/lib/EPUBManifest.cpp +++ b/src/lib/EPUBManifest.cpp @@ -21,9 +21,9 @@ EPUBManifest::EPUBManifest() { } -void EPUBManifest::insert(const EPUBPath &path, const std::string &mimetype, const std::string &id) +void EPUBManifest::insert(const EPUBPath &path, const std::string &mimetype, const std::string &id, const std::string &properties) { - if (!m_map.insert(MapType_t::value_type(path.relativeTo(EPUBPath("OEBPS/content.opf")).str(), ValueType_t(mimetype, id))).second) + if (!m_map.insert(MapType_t::value_type(path.relativeTo(EPUBPath("OEBPS/content.opf")).str(), ValueType_t(mimetype, id, properties))).second) { assert(!"duplicate entry!"); } @@ -35,8 +35,11 @@ void EPUBManifest::writeTo(EPUBXMLSink &sink) { librevenge::RVNGPropertyList attrs; attrs.insert("href", it->first.c_str()); - attrs.insert("media-type", it->second.first.c_str()); - attrs.insert("id", it->second.second.c_str()); + attrs.insert("media-type", std::get<0>(it->second).c_str()); + attrs.insert("id", std::get<1>(it->second).c_str()); + const std::string &properties = std::get<2>(it->second); + if (!properties.empty()) + attrs.insert("properties", properties.c_str()); sink.insertEmptyElement("item", attrs); } } diff --git a/src/lib/EPUBManifest.h b/src/lib/EPUBManifest.h index f2379cf..c4c9031 100644 --- a/src/lib/EPUBManifest.h +++ b/src/lib/EPUBManifest.h @@ -26,13 +26,14 @@ class EPUBManifest EPUBManifest(const EPUBManifest &); EPUBManifest &operator=(const EPUBManifest &); - typedef std::pair ValueType_t; + // media-type, id, properties + typedef std::tuple ValueType_t; typedef std::unordered_map MapType_t; public: EPUBManifest(); - void insert(const EPUBPath &path, const std::string &mimetype, const std::string &id); + void insert(const EPUBPath &path, const std::string &mimetype, const std::string &id, const std::string &properties); void writeTo(EPUBXMLSink &sink); -- 2.12.3 From 28090aa5d57162302122686cb020d4bf2231cab4 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Thu, 10 Aug 2017 15:05:46 +0200 Subject: [PATCH 8/9] EPUB3: implement navigation document As says, EPUB 3 defines a new grammar for navigation based on XHTML, which replaces the old NCX grammar -- so use that in EPUB3 mode. With this, a hello world input in EPUB3 mode results in 0 errors in the validator. --- src/lib/EPUBGenerator.cpp | 43 ++++++++++++++++++++++++++++++++++++++----- src/lib/EPUBHTMLManager.cpp | 28 +++++++++++++++++++++------- src/lib/EPUBHTMLManager.h | 4 ++-- 3 files changed, 61 insertions(+), 14 deletions(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index f77aa6c..7ec2a2b 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -54,7 +54,10 @@ void EPUBGenerator::startDocument(const RVNGPropertyList &props) startNewHtmlFile(); - m_manifest.insert(EPUBPath("OEBPS/toc.ncx"), "application/x-dtbncx+xml", "toc.ncx", ""); + if (m_version == 30) + m_manifest.insert(EPUBPath("OEBPS/toc.html"), "application/xhtml+xml", "toc.html", "nav"); + else + m_manifest.insert(EPUBPath("OEBPS/toc.ncx"), "application/x-dtbncx+xml", "toc.ncx", ""); m_manifest.insert(m_stylesheetPath, "text/css", "stylesheet.css", ""); } @@ -90,7 +93,7 @@ void EPUBGenerator::startNewHtmlFile() m_splitGuard.onSplit(); - m_currentHtml = m_htmlManager.create(m_imageManager, m_listStyleManager, m_paragraphStyleManager, m_spanStyleManager, m_tableStyleManager, m_stylesheetPath, *this); + m_currentHtml = m_htmlManager.create(m_imageManager, m_listStyleManager, m_paragraphStyleManager, m_spanStyleManager, m_tableStyleManager, m_stylesheetPath); // restore state in the new file m_currentHtml->startDocument(m_documentProps); @@ -144,9 +147,38 @@ void EPUBGenerator::writeContainer() void EPUBGenerator::writeNavigation() { - const EPUBPath path("OEBPS/toc.ncx"); EPUBXMLSink sink; + if (m_version == 30) + { + const EPUBPath path("OEBPS/toc.html"); + RVNGPropertyList htmlAttrs; + htmlAttrs.insert("xmlns", "http://www.w3.org/1999/xhtml"); + htmlAttrs.insert("xmlns:epub", "http://www.idpf.org/2007/ops"); + sink.openElement("html", htmlAttrs); + + sink.openElement("head"); + sink.closeElement("head"); + sink.openElement("body"); + + RVNGPropertyList navAttrs; + navAttrs.insert("epub:type", "toc"); + sink.openElement("nav", navAttrs); + + sink.openElement("ol"); + m_htmlManager.writeTocTo(sink, path, m_version); + sink.closeElement("ol"); + + sink.closeElement("nav"); + sink.closeElement("body"); + sink.closeElement("html"); + + sink.writeTo(*m_package, path.str().c_str()); + + return; + } + + const EPUBPath path("OEBPS/toc.ncx"); RVNGPropertyList ncxAttrs; ncxAttrs.insert("xmlns", "http://www.daisy.org/z3986/2005/ncx/"); ncxAttrs.insert("version", "2005-1"); @@ -166,7 +198,7 @@ void EPUBGenerator::writeNavigation() sink.closeElement("docTitle"); sink.openElement("navMap"); - m_htmlManager.writeTocTo(sink, path); + m_htmlManager.writeTocTo(sink, path, m_version); sink.closeElement("navMap"); sink.closeElement("ncx"); @@ -252,7 +284,8 @@ void EPUBGenerator::writeRoot() sink.closeElement("manifest"); RVNGPropertyList spineAttrs; - spineAttrs.insert("toc", "toc.ncx"); + if (m_version == 20) + spineAttrs.insert("toc", "toc.ncx"); sink.openElement("spine", spineAttrs); m_htmlManager.writeSpineTo(sink); diff --git a/src/lib/EPUBHTMLManager.cpp b/src/lib/EPUBHTMLManager.cpp index 57636b9..7b17304 100644 --- a/src/lib/EPUBHTMLManager.cpp +++ b/src/lib/EPUBHTMLManager.cpp @@ -27,7 +27,7 @@ EPUBHTMLManager::EPUBHTMLManager(EPUBManifest &manifest) { } -const EPUBHTMLGeneratorPtr_t EPUBHTMLManager::create(EPUBImageManager &imageManager, EPUBListStyleManager &listStyleManager, EPUBParagraphStyleManager ¶graphStyleManager, EPUBSpanStyleManager &spanStyleManager, EPUBTableStyleManager &tableStyleManager, const EPUBPath &stylesheetPath, const EPUBGenerator &generator) +const EPUBHTMLGeneratorPtr_t EPUBHTMLManager::create(EPUBImageManager &imageManager, EPUBListStyleManager &listStyleManager, EPUBParagraphStyleManager ¶graphStyleManager, EPUBSpanStyleManager &spanStyleManager, EPUBTableStyleManager &tableStyleManager, const EPUBPath &stylesheetPath) { std::ostringstream nameBuf; nameBuf << "section" << std::setw(4) << std::setfill('0') << m_number.next(); @@ -36,11 +36,7 @@ const EPUBHTMLGeneratorPtr_t EPUBHTMLManager::create(EPUBImageManager &imageMana nameBuf << ".html"; m_paths.push_back(EPUBPath("OEBPS/sections") / nameBuf.str()); - std::string properties; - if (m_number.current() == 1 && generator.getVersion() == 30) - // Only for the first HTML file. - properties = "nav"; - m_manifest.insert(m_paths.back(), "application/xhtml+xml", m_ids.back(), properties); + m_manifest.insert(m_paths.back(), "application/xhtml+xml", m_ids.back(), ""); m_contents.push_back(EPUBXMLSink()); @@ -71,8 +67,26 @@ void EPUBHTMLManager::writeSpineTo(EPUBXMLSink &sink) } } -void EPUBHTMLManager::writeTocTo(EPUBXMLSink &sink, const EPUBPath &tocPath) +void EPUBHTMLManager::writeTocTo(EPUBXMLSink &sink, const EPUBPath &tocPath, int version) { + if (version == 30) + { + for (std::vector::size_type i = 0; m_paths.size() != i; ++i) + { + sink.openElement("li"); + librevenge::RVNGPropertyList anchorAttrs; + anchorAttrs.insert("href", m_paths[i].relativeTo(tocPath).str().c_str()); + sink.openElement("a", anchorAttrs); + std::ostringstream label; + label << "Section " << (i + 1); + sink.insertCharacters(label.str().c_str()); + sink.closeElement("a"); + sink.closeElement("li"); + } + + return; + } + librevenge::RVNGPropertyList navPointAttrs; for (std::vector::size_type i = 0; m_paths.size() != i; ++i) { diff --git a/src/lib/EPUBHTMLManager.h b/src/lib/EPUBHTMLManager.h index 7dab33b..2ec7bb7 100644 --- a/src/lib/EPUBHTMLManager.h +++ b/src/lib/EPUBHTMLManager.h @@ -41,12 +41,12 @@ class EPUBHTMLManager public: explicit EPUBHTMLManager(EPUBManifest &manifest); - const EPUBHTMLGeneratorPtr_t create(EPUBImageManager &imageManager, EPUBListStyleManager &listStyleManager, EPUBParagraphStyleManager ¶graphStyleManager, EPUBSpanStyleManager &spanStyleManager, EPUBTableStyleManager &tableStyleManager, const EPUBPath &stylesheetPath, const EPUBGenerator &generator); + const EPUBHTMLGeneratorPtr_t create(EPUBImageManager &imageManager, EPUBListStyleManager &listStyleManager, EPUBParagraphStyleManager ¶graphStyleManager, EPUBSpanStyleManager &spanStyleManager, EPUBTableStyleManager &tableStyleManager, const EPUBPath &stylesheetPath); void writeTo(EPUBPackage &package); void writeSpineTo(EPUBXMLSink &sink); - void writeTocTo(EPUBXMLSink &sink, const EPUBPath &tocPath); + void writeTocTo(EPUBXMLSink &sink, const EPUBPath &tocPath, int version); private: EPUBManifest &m_manifest; -- 2.12.3 From 2e9e9af9f49a78cca75d3c862c8dd4b5f7cc7eb2 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Thu, 10 Aug 2017 15:11:49 +0200 Subject: [PATCH 9/9] Use .xhtml for XHTML content The EPUB3 validator warns: WARNING(HTM-014a): test30.epub/OEBPS/content.opf(2,718): XHTML Content Document file name 'OEBPS/sections/section0001.html' should have the extension '.xhtml'. And it does not hurt for EPUB2, either. --- src/lib/EPUBGenerator.cpp | 4 ++-- src/lib/EPUBHTMLManager.cpp | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 7ec2a2b..ca05ea7 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -55,7 +55,7 @@ void EPUBGenerator::startDocument(const RVNGPropertyList &props) startNewHtmlFile(); if (m_version == 30) - m_manifest.insert(EPUBPath("OEBPS/toc.html"), "application/xhtml+xml", "toc.html", "nav"); + m_manifest.insert(EPUBPath("OEBPS/toc.xhtml"), "application/xhtml+xml", "toc.xhtml", "nav"); else m_manifest.insert(EPUBPath("OEBPS/toc.ncx"), "application/x-dtbncx+xml", "toc.ncx", ""); m_manifest.insert(m_stylesheetPath, "text/css", "stylesheet.css", ""); @@ -151,7 +151,7 @@ void EPUBGenerator::writeNavigation() if (m_version == 30) { - const EPUBPath path("OEBPS/toc.html"); + const EPUBPath path("OEBPS/toc.xhtml"); RVNGPropertyList htmlAttrs; htmlAttrs.insert("xmlns", "http://www.w3.org/1999/xhtml"); htmlAttrs.insert("xmlns:epub", "http://www.idpf.org/2007/ops"); diff --git a/src/lib/EPUBHTMLManager.cpp b/src/lib/EPUBHTMLManager.cpp index 7b17304..be56cc7 100644 --- a/src/lib/EPUBHTMLManager.cpp +++ b/src/lib/EPUBHTMLManager.cpp @@ -33,7 +33,7 @@ const EPUBHTMLGeneratorPtr_t EPUBHTMLManager::create(EPUBImageManager &imageMana nameBuf << "section" << std::setw(4) << std::setfill('0') << m_number.next(); m_ids.push_back(nameBuf.str()); - nameBuf << ".html"; + nameBuf << ".xhtml"; m_paths.push_back(EPUBPath("OEBPS/sections") / nameBuf.str()); m_manifest.insert(m_paths.back(), "application/xhtml+xml", m_ids.back(), ""); -- 2.12.3 From c30bc184c18837203e9f249386711e9cd616c9f0 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Fri, 11 Aug 2017 10:50:47 +0200 Subject: [PATCH] EPUB3: write the deprecated NCX navication as well "3.2 Navigation" says: EPUB 3 Publications may include the EPUB 2 NCX for EPUB 2 Reading System forward compatibility purposes. --- src/lib/EPUBGenerator.cpp | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index ca05ea7..8017ffe 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -56,8 +56,7 @@ void EPUBGenerator::startDocument(const RVNGPropertyList &props) if (m_version == 30) m_manifest.insert(EPUBPath("OEBPS/toc.xhtml"), "application/xhtml+xml", "toc.xhtml", "nav"); - else - m_manifest.insert(EPUBPath("OEBPS/toc.ncx"), "application/x-dtbncx+xml", "toc.ncx", ""); + m_manifest.insert(EPUBPath("OEBPS/toc.ncx"), "application/x-dtbncx+xml", "toc.ncx", ""); m_manifest.insert(m_stylesheetPath, "text/css", "stylesheet.css", ""); } @@ -147,10 +146,10 @@ void EPUBGenerator::writeContainer() void EPUBGenerator::writeNavigation() { - EPUBXMLSink sink; - if (m_version == 30) { + EPUBXMLSink sink; + const EPUBPath path("OEBPS/toc.xhtml"); RVNGPropertyList htmlAttrs; htmlAttrs.insert("xmlns", "http://www.w3.org/1999/xhtml"); @@ -174,10 +173,10 @@ void EPUBGenerator::writeNavigation() sink.closeElement("html"); sink.writeTo(*m_package, path.str().c_str()); - - return; } + EPUBXMLSink sink; + const EPUBPath path("OEBPS/toc.ncx"); RVNGPropertyList ncxAttrs; ncxAttrs.insert("xmlns", "http://www.daisy.org/z3986/2005/ncx/"); @@ -198,7 +197,9 @@ void EPUBGenerator::writeNavigation() sink.closeElement("docTitle"); sink.openElement("navMap"); - m_htmlManager.writeTocTo(sink, path, m_version); + // In case of EPUB3 the (deprecated, but valid) EPUB2 markup is wanted, so + // the version is unconditional here. + m_htmlManager.writeTocTo(sink, path, /*version=*/20); sink.closeElement("navMap"); sink.closeElement("ncx"); @@ -284,8 +285,7 @@ void EPUBGenerator::writeRoot() sink.closeElement("manifest"); RVNGPropertyList spineAttrs; - if (m_version == 20) - spineAttrs.insert("toc", "toc.ncx"); + spineAttrs.insert("toc", "toc.ncx"); sink.openElement("spine", spineAttrs); m_htmlManager.writeSpineTo(sink); -- 2.12.3 From d855721de6b4e55fb0b2a02a6a1132802dba5f63 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Fri, 11 Aug 2017 15:40:49 +0200 Subject: [PATCH] EPUBGenerator: write title into content.opf --- src/lib/EPUBGenerator.cpp | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 8017ffe..056c48c 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -253,8 +253,11 @@ void EPUBGenerator::writeRoot() // Zero-width space as it must be at least one character in length after // white space has been trimmed. + RVNGString title("\u200b"); + if (m_metadata["dc:title"] && !m_metadata["dc:title"]->getStr().empty()) + title = m_metadata["dc:title"]->getStr(); sink.openElement("dc:title"); - sink.insertCharacters("\u200b"); + sink.insertCharacters(title); sink.closeElement("dc:title"); sink.openElement("dc:language"); -- 2.12.3 From 14da42819f1f96b87b2337da395e5ad6639dcebe Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Fri, 11 Aug 2017 16:43:10 +0200 Subject: [PATCH] EPUBGenerator: write language and modification date into content.opf With this, all metadata which had a hardcoded value is now parsed and the old values are just defaults. --- src/lib/EPUBGenerator.cpp | 32 +++++++++++++++++++++----------- 1 file changed, 21 insertions(+), 11 deletions(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 056c48c..fc6c848 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -260,24 +260,34 @@ void EPUBGenerator::writeRoot() sink.insertCharacters(title); sink.closeElement("dc:title"); + RVNGString language("en"); + if (m_metadata["dc:language"] && !m_metadata["dc:language"]->getStr().empty()) + language = m_metadata["dc:language"]->getStr(); sink.openElement("dc:language"); - sink.insertCharacters("en"); + sink.insertCharacters(language); sink.closeElement("dc:language"); - time_t now = 0; - time(&now); - const struct tm *local = localtime(&now); - if (m_version == 30 && local) + if (m_version == 30) { + RVNGString date; + time_t now = 0; + time(&now); + const struct tm *local = localtime(&now); + if (local) + { + const int MAX_BUFFER = 1024; + char buffer[MAX_BUFFER]; + strftime(&buffer[0], MAX_BUFFER-1, "%Y-%m-%dT%H:%M:%SZ", local); + date.append(buffer); + } + + if (m_metadata["dc:date"] && !m_metadata["dc:date"]->getStr().empty()) + date = m_metadata["dc:date"]->getStr(); + RVNGPropertyList metaAttrs; metaAttrs.insert("property", "dcterms:modified"); sink.openElement("meta", metaAttrs); - const int MAX_BUFFER = 1024; - char buffer[MAX_BUFFER]; - strftime(&buffer[0], MAX_BUFFER-1, "%Y-%m-%dT%H:%M:%SZ", local); - RVNGString result; - result.append(buffer); - sink.insertCharacters(result); + sink.insertCharacters(date); sink.closeElement("meta"); } -- 2.12.3 From c8cba54a7025ee0f1129233912f6e9ceda254c64 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Fri, 11 Aug 2017 17:40:10 +0200 Subject: [PATCH] EPUB3: fix validation error with doc modified date format ERROR(RSC-005): test.epub/OEBPS/content.opf(2,420): Error while parsing file: dcterms:modified illegal syntax (expecting: 'CCYY-MM-DDThh:mm:ssZ') Both of them looked like XML date formats, but actually librevenge is more precise, so truncate it. --- src/lib/EPUBGenerator.cpp | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index fc6c848..4bffb8b 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -282,7 +282,11 @@ void EPUBGenerator::writeRoot() } if (m_metadata["dc:date"] && !m_metadata["dc:date"]->getStr().empty()) - date = m_metadata["dc:date"]->getStr(); + { + // Expecting CCYY-MM-DDThh:mm:ssZ, librevenge provides CCYY-MM-DDThh:mm:ss.sssssssss + date = std::string(m_metadata["dc:date"]->getStr().cstr()).substr(0, 19).c_str(); + date.append("Z"); + } RVNGPropertyList metaAttrs; metaAttrs.insert("property", "dcterms:modified"); -- 2.12.3 From c28f02f21a6d80ad258cf8f052705508567e2418 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Fri, 11 Aug 2017 18:19:15 +0200 Subject: [PATCH 1/3] Fix image mime-type key libepubgen expected librevenge:mimetype, but: 1) LO's ODF output has loext:mime-type 2) libabw generates librevenge:mime-type 3) libodfgen expects librevenge:mime-type So probably this one has to be adjusted. --- src/lib/EPUBHTMLGenerator.cpp | 2 +- src/lib/EPUBTextGenerator.cpp | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/lib/EPUBHTMLGenerator.cpp b/src/lib/EPUBHTMLGenerator.cpp index 5ef5e14..40c507e 100644 --- a/src/lib/EPUBHTMLGenerator.cpp +++ b/src/lib/EPUBHTMLGenerator.cpp @@ -866,7 +866,7 @@ void EPUBHTMLGenerator::insertBinaryObject(const RVNGPropertyList &propList) const EPUBPath &path = m_impl->m_imageManager.insert( RVNGBinaryData(propList["office:binary-data"]->getStr()), - propList["librevenge:mimetype"]->getStr()); + propList["librevenge:mime-type"]->getStr()); RVNGPropertyList attrs; attrs.insert("src", path.relativeTo(m_impl->m_path).str().c_str()); diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index e8f785e..cb557b2 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -596,7 +596,7 @@ void EPUBTextGenerator::insertBinaryObject(const librevenge::RVNGPropertyList &p for (RVNGPropertyList::Iter iter(propList); !iter.last(); iter.next()) { - if (RVNGString("librevenge:mimetype") == iter.key()) + if (RVNGString("librevenge:mime-type") == iter.key()) mimetype.reset(iter()->clone()); else if (RVNGString("office:binary-data") == iter.key()) data.reset(iter()->clone()); @@ -627,7 +627,7 @@ void EPUBTextGenerator::insertBinaryObject(const librevenge::RVNGPropertyList &p } } - newPropList.insert("librevenge:mimetype", mimetype->clone()); + newPropList.insert("librevenge:mime-type", mimetype->clone()); newPropList.insert("office:binary-data", data->clone()); if (m_impl->m_inHeader || m_impl->m_inFooter) -- 2.12.3 From 39470cf360cfc67f2dd078646162a63168a84c05 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Tue, 15 Aug 2017 12:12:12 +0200 Subject: [PATCH 2/3] EPUBSplitGuard: fix tracking current size In case incrementing size and split affects the reference size, the current size won't be ever greather than zero in EPUBSplitGuard::canSplit(), and it will always return false. With this, EPUB_SPLIT_METHOD_PAGE_BREAK works again. --- src/lib/EPUBSplitGuard.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/lib/EPUBSplitGuard.cpp b/src/lib/EPUBSplitGuard.cpp index 8c279c8..15ba20f 100644 --- a/src/lib/EPUBSplitGuard.cpp +++ b/src/lib/EPUBSplitGuard.cpp @@ -46,7 +46,7 @@ void EPUBSplitGuard::closeLevel() void EPUBSplitGuard::incrementSize(const unsigned size) { - m_size += size; + m_currentSize += size; } bool EPUBSplitGuard::splitOnPageBreak() const @@ -66,7 +66,7 @@ bool EPUBSplitGuard::splitOnSize() const void EPUBSplitGuard::onSplit() { - m_size = 0; + m_currentSize = 0; } bool EPUBSplitGuard::canSplit(const EPUBSplitMethod method) const -- 2.12.3 From 3155cb6164f04fa8170dd9912c579ad90586c8a8 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Tue, 15 Aug 2017 13:53:16 +0200 Subject: [PATCH 3/3] EPUBTextGenerator: handle EPUB_SPLIT_METHOD_HEADING It seems to me this was unimplemented so far. --- src/lib/EPUBTextGenerator.cpp | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index cb557b2..aba8827 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -255,6 +255,12 @@ void EPUBTextGenerator::openParagraph(const librevenge::RVNGPropertyList &propLi m_impl->m_breakAfterPara = breakAfter && ("column" != breakAfter->getStr()); if (m_impl->getSplitGuard().splitOnSize()) m_impl->startNewHtmlFile(); + + // Handle split by chapters. + const RVNGProperty *const outlineLevel = propList["text:outline-level"]; + if (outlineLevel && m_impl->getSplitGuard().splitOnHeading(outlineLevel->getInt())) + m_impl->startNewHtmlFile(); + m_impl->getSplitGuard().openLevel(); if (m_impl->m_inHeader || m_impl->m_inFooter) -- 2.12.3 From 576c2472e384ea1a71739b15f42561cd34de5bba Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Wed, 16 Aug 2017 17:52:37 +0200 Subject: [PATCH] EPUB_SPLIT_METHOD_HEADING: try to use real chapter names, not Section N What users call "chapter title" is the text of a paragraph with an outline level set. To keep this simple just handle the text after opening such a paragraph, but no other paragraph is opened, i.e. assume that in: ABC only A is interesting, but not B, neither C. (Which could happen with an at-character anchored frame inside a heading text e.g.) --- src/lib/EPUBGenerator.cpp | 5 +++++ src/lib/EPUBGenerator.h | 2 ++ src/lib/EPUBHTMLManager.cpp | 26 ++++++++++++++++++++++++-- src/lib/EPUBHTMLManager.h | 3 +++ src/lib/EPUBPath.cpp | 11 +++++++++++ src/lib/EPUBPath.h | 4 ++++ src/lib/EPUBTextGenerator.cpp | 6 ++++++ 7 files changed, 55 insertions(+), 2 deletions(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 4bffb8b..4ce2964 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -106,6 +106,11 @@ const EPUBHTMLGeneratorPtr_t &EPUBGenerator::getHtml() const return m_currentHtml; } +EPUBHTMLManager &EPUBGenerator::getHtmlManager() +{ + return m_htmlManager; +} + const EPUBSplitGuard &EPUBGenerator::getSplitGuard() const { return m_splitGuard; diff --git a/src/lib/EPUBGenerator.h b/src/lib/EPUBGenerator.h index 1a67a88..f98c80f 100644 --- a/src/lib/EPUBGenerator.h +++ b/src/lib/EPUBGenerator.h @@ -45,6 +45,8 @@ public: const EPUBHTMLGeneratorPtr_t &getHtml() const; + EPUBHTMLManager &getHtmlManager(); + const EPUBSplitGuard &getSplitGuard() const; EPUBSplitGuard &getSplitGuard(); int getVersion() const; diff --git a/src/lib/EPUBHTMLManager.cpp b/src/lib/EPUBHTMLManager.cpp index be56cc7..5141f31 100644 --- a/src/lib/EPUBHTMLManager.cpp +++ b/src/lib/EPUBHTMLManager.cpp @@ -18,6 +18,20 @@ namespace libepubgen { +namespace +{ + +/// Extracts a title string from a path and provides a fallback if it would be empty. +void getPathTitle(std::ostringstream &label, const EPUBPath &path, std::vector::size_type index) +{ + if (path.getTitle().empty()) + label << "Section " << (index + 1); + else + label << path.getTitle(); +} + +} + EPUBHTMLManager::EPUBHTMLManager(EPUBManifest &manifest) : m_manifest(manifest) , m_paths() @@ -78,7 +92,7 @@ void EPUBHTMLManager::writeTocTo(EPUBXMLSink &sink, const EPUBPath &tocPath, int anchorAttrs.insert("href", m_paths[i].relativeTo(tocPath).str().c_str()); sink.openElement("a", anchorAttrs); std::ostringstream label; - label << "Section " << (i + 1); + getPathTitle(label, m_paths[i], i); sink.insertCharacters(label.str().c_str()); sink.closeElement("a"); sink.closeElement("li"); @@ -101,7 +115,7 @@ void EPUBHTMLManager::writeTocTo(EPUBXMLSink &sink, const EPUBPath &tocPath, int sink.openElement("navLabel"); sink.openElement("text"); std::ostringstream label; - label << "Section " << (i + 1); + getPathTitle(label, m_paths[i], i); sink.insertCharacters(label.str().c_str()); sink.closeElement("text"); sink.closeElement("navLabel"); @@ -112,6 +126,14 @@ void EPUBHTMLManager::writeTocTo(EPUBXMLSink &sink, const EPUBPath &tocPath, int } } +void EPUBHTMLManager::insertHeadingText(const std::string &text) +{ + if (m_paths.empty()) + return; + + m_paths.back().appendTitle(text); +} + } /* vim:set shiftwidth=2 softtabstop=2 expandtab: */ diff --git a/src/lib/EPUBHTMLManager.h b/src/lib/EPUBHTMLManager.h index 2ec7bb7..6b480c4 100644 --- a/src/lib/EPUBHTMLManager.h +++ b/src/lib/EPUBHTMLManager.h @@ -48,6 +48,9 @@ public: void writeSpineTo(EPUBXMLSink &sink); void writeTocTo(EPUBXMLSink &sink, const EPUBPath &tocPath, int version); + /// Appends text to the title of the current heading. + void insertHeadingText(const std::string &text); + private: EPUBManifest &m_manifest; std::vector m_paths; diff --git a/src/lib/EPUBPath.cpp b/src/lib/EPUBPath.cpp index 9def6f6..e1c05ed 100644 --- a/src/lib/EPUBPath.cpp +++ b/src/lib/EPUBPath.cpp @@ -53,6 +53,7 @@ EPUBPath::Relative::Relative(const std::vector &components) EPUBPath::EPUBPath(const std::string &path) : m_components() + , m_title() { const std::string trimmed(algorithm::trim_left_copy_if(path, algorithm::is_any_of("/"))); algorithm::split(m_components, trimmed, algorithm::is_any_of("/"), algorithm::token_compress_on); @@ -110,6 +111,16 @@ const EPUBPath::Relative EPUBPath::relativeTo(const EPUBPath &base) const return Relative(components); } +void EPUBPath::appendTitle(const std::string &title) +{ + m_title += title; +} + +std::string EPUBPath::getTitle() const +{ + return m_title; +} + bool operator==(const EPUBPath &left, const EPUBPath &right) { return left.m_components == right.m_components; diff --git a/src/lib/EPUBPath.h b/src/lib/EPUBPath.h index 18bf058..12b8f25 100644 --- a/src/lib/EPUBPath.h +++ b/src/lib/EPUBPath.h @@ -48,8 +48,12 @@ public: const Relative relativeTo(const EPUBPath &base) const; + void appendTitle(const std::string &title); + std::string getTitle() const; + private: std::vector m_components; + std::string m_title; }; bool operator==(const EPUBPath &left, const EPUBPath &right); diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index aba8827..b1e33f8 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -67,6 +67,7 @@ public: bool m_inPageSpan; bool m_inHeader; bool m_inFooter; + bool m_inHeading; RVNGPropertyList m_pageSpanProps; shared_ptr m_currentHeader; @@ -88,6 +89,7 @@ EPUBTextGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod , m_inPageSpan(false) , m_inHeader(false) , m_inFooter(false) + , m_inHeading(false) , m_pageSpanProps() , m_currentHeader() , m_currentFooter() @@ -260,6 +262,7 @@ void EPUBTextGenerator::openParagraph(const librevenge::RVNGPropertyList &propLi const RVNGProperty *const outlineLevel = propList["text:outline-level"]; if (outlineLevel && m_impl->getSplitGuard().splitOnHeading(outlineLevel->getInt())) m_impl->startNewHtmlFile(); + m_impl->m_inHeading = outlineLevel != nullptr; m_impl->getSplitGuard().openLevel(); @@ -366,6 +369,9 @@ void EPUBTextGenerator::insertText(const librevenge::RVNGString &text) if (m_impl->m_inHeader || m_impl->m_inFooter) m_impl->m_currentHeaderOrFooter->addInsertText(text); + if (m_impl->m_inHeading) + m_impl->getHtmlManager().insertHeadingText(text.cstr()); + m_impl->getSplitGuard().incrementSize(text.len()); m_impl->getHtml()->insertText(text); -- 2.12.3 From 8a3cc9733f567864ad0658cf1a25c60d5e7205a6 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Thu, 17 Aug 2017 12:01:44 +0200 Subject: [PATCH] EPUB_SPLIT_METHOD_HEADING: fix unexpected text in chapter names We used to add text for all headings for the chapter name, but if we split on heading 1, then text of heading2+ should not be included. --- src/lib/EPUBSplitGuard.cpp | 14 ++++++++++++++ src/lib/EPUBSplitGuard.h | 3 +++ src/lib/EPUBTextGenerator.cpp | 6 ++---- 3 files changed, 19 insertions(+), 4 deletions(-) diff --git a/src/lib/EPUBSplitGuard.cpp b/src/lib/EPUBSplitGuard.cpp index 15ba20f..25eae87 100644 --- a/src/lib/EPUBSplitGuard.cpp +++ b/src/lib/EPUBSplitGuard.cpp @@ -18,6 +18,7 @@ static const unsigned DEFAULT_SPLIT_SIZE = 1 << 16; EPUBSplitGuard::EPUBSplitGuard(const EPUBSplitMethod method) : m_method(method) , m_headingLevel(DEFAULT_SPLIT_HEADING_LEVEL) + , m_currentHeadingLevel(0) , m_size(DEFAULT_SPLIT_SIZE) , m_currentSize(0) , m_nestingLevel(0) @@ -29,6 +30,11 @@ void EPUBSplitGuard::setSplitHeadingLevel(const unsigned level) m_headingLevel = level; } +void EPUBSplitGuard::setCurrentHeadingLevel(const unsigned level) +{ + m_currentHeadingLevel = level; +} + void EPUBSplitGuard::setSplitSize(const unsigned size) { m_size = size; @@ -59,6 +65,14 @@ bool EPUBSplitGuard::splitOnHeading(const unsigned level) const return canSplit(EPUB_SPLIT_METHOD_HEADING) && (m_headingLevel >= level); } +bool EPUBSplitGuard::inHeading() const +{ + if (!m_currentHeadingLevel) + return false; + + return m_headingLevel >= m_currentHeadingLevel; +} + bool EPUBSplitGuard::splitOnSize() const { return canSplit(EPUB_SPLIT_METHOD_SIZE) && (m_size <= m_currentSize); diff --git a/src/lib/EPUBSplitGuard.h b/src/lib/EPUBSplitGuard.h index 17613ac..7bc53ce 100644 --- a/src/lib/EPUBSplitGuard.h +++ b/src/lib/EPUBSplitGuard.h @@ -21,6 +21,7 @@ public: explicit EPUBSplitGuard(EPUBSplitMethod method); void setSplitHeadingLevel(unsigned level); + void setCurrentHeadingLevel(unsigned level); void setSplitSize(unsigned size); void openLevel(); @@ -29,6 +30,7 @@ public: bool splitOnPageBreak() const; bool splitOnHeading(unsigned level) const; + bool inHeading() const; bool splitOnSize() const; void onSplit(); @@ -39,6 +41,7 @@ private: private: const EPUBSplitMethod m_method; unsigned m_headingLevel; + unsigned m_currentHeadingLevel; unsigned m_size; unsigned m_currentSize; unsigned m_nestingLevel; diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index b1e33f8..5206b37 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -67,7 +67,6 @@ public: bool m_inPageSpan; bool m_inHeader; bool m_inFooter; - bool m_inHeading; RVNGPropertyList m_pageSpanProps; shared_ptr m_currentHeader; @@ -89,7 +88,6 @@ EPUBTextGenerator::Impl::Impl(EPUBPackage *const package, const EPUBSplitMethod , m_inPageSpan(false) , m_inHeader(false) , m_inFooter(false) - , m_inHeading(false) , m_pageSpanProps() , m_currentHeader() , m_currentFooter() @@ -262,7 +260,7 @@ void EPUBTextGenerator::openParagraph(const librevenge::RVNGPropertyList &propLi const RVNGProperty *const outlineLevel = propList["text:outline-level"]; if (outlineLevel && m_impl->getSplitGuard().splitOnHeading(outlineLevel->getInt())) m_impl->startNewHtmlFile(); - m_impl->m_inHeading = outlineLevel != nullptr; + m_impl->getSplitGuard().setCurrentHeadingLevel(outlineLevel ? outlineLevel->getInt() : 0); m_impl->getSplitGuard().openLevel(); @@ -369,7 +367,7 @@ void EPUBTextGenerator::insertText(const librevenge::RVNGString &text) if (m_impl->m_inHeader || m_impl->m_inFooter) m_impl->m_currentHeaderOrFooter->addInsertText(text); - if (m_impl->m_inHeading) + if (m_impl->getSplitGuard().inHeading()) m_impl->getHtmlManager().insertHeadingText(text.cstr()); m_impl->getSplitGuard().incrementSize(text.len()); -- 2.12.3 From a4e85e191813e7c8f4e6b5bcf2458504f9d06aeb Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Tue, 22 Aug 2017 12:05:28 +0200 Subject: [PATCH] EPUB_SPLIT_METHOD_HEADING: fix unexpected first chapter name If the structure of the document is like this: Heading 2 Heading 1 Heading 2 Heading 1 Heading 2 Then this resulted in 3 chapters: Section 1, Heading 1 and Heading 1. The first one is unexpected; so in case we don't have a heading 1 paragraph for the first section, then fall back to any other heading. --- src/lib/EPUBHTMLManager.cpp | 8 ++++++++ src/lib/EPUBHTMLManager.h | 3 +++ src/lib/EPUBSplitGuard.cpp | 5 ++++- src/lib/EPUBSplitGuard.h | 2 +- src/lib/EPUBTextGenerator.cpp | 2 +- 5 files changed, 17 insertions(+), 3 deletions(-) diff --git a/src/lib/EPUBHTMLManager.cpp b/src/lib/EPUBHTMLManager.cpp index 5141f31..7753160 100644 --- a/src/lib/EPUBHTMLManager.cpp +++ b/src/lib/EPUBHTMLManager.cpp @@ -134,6 +134,14 @@ void EPUBHTMLManager::insertHeadingText(const std::string &text) m_paths.back().appendTitle(text); } +bool EPUBHTMLManager::hasHeadingText() const +{ + if (m_paths.empty()) + return false; + + return !m_paths.back().getTitle().empty(); +} + } /* vim:set shiftwidth=2 softtabstop=2 expandtab: */ diff --git a/src/lib/EPUBHTMLManager.h b/src/lib/EPUBHTMLManager.h index 6b480c4..158b466 100644 --- a/src/lib/EPUBHTMLManager.h +++ b/src/lib/EPUBHTMLManager.h @@ -51,6 +51,9 @@ public: /// Appends text to the title of the current heading. void insertHeadingText(const std::string &text); + /// If the current heading has a title. + bool hasHeadingText() const; + private: EPUBManifest &m_manifest; std::vector m_paths; diff --git a/src/lib/EPUBSplitGuard.cpp b/src/lib/EPUBSplitGuard.cpp index 25eae87..890500b 100644 --- a/src/lib/EPUBSplitGuard.cpp +++ b/src/lib/EPUBSplitGuard.cpp @@ -65,11 +65,14 @@ bool EPUBSplitGuard::splitOnHeading(const unsigned level) const return canSplit(EPUB_SPLIT_METHOD_HEADING) && (m_headingLevel >= level); } -bool EPUBSplitGuard::inHeading() const +bool EPUBSplitGuard::inHeading(bool any) const { if (!m_currentHeadingLevel) return false; + if (any) + return true; + return m_headingLevel >= m_currentHeadingLevel; } diff --git a/src/lib/EPUBSplitGuard.h b/src/lib/EPUBSplitGuard.h index 7bc53ce..1a74079 100644 --- a/src/lib/EPUBSplitGuard.h +++ b/src/lib/EPUBSplitGuard.h @@ -30,7 +30,7 @@ public: bool splitOnPageBreak() const; bool splitOnHeading(unsigned level) const; - bool inHeading() const; + bool inHeading(bool any) const; bool splitOnSize() const; void onSplit(); diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index 5206b37..a39f266 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -367,7 +367,7 @@ void EPUBTextGenerator::insertText(const librevenge::RVNGString &text) if (m_impl->m_inHeader || m_impl->m_inFooter) m_impl->m_currentHeaderOrFooter->addInsertText(text); - if (m_impl->getSplitGuard().inHeading()) + if (m_impl->getSplitGuard().inHeading(!m_impl->getHtmlManager().hasHeadingText())) m_impl->getHtmlManager().insertHeadingText(text.cstr()); m_impl->getSplitGuard().incrementSize(text.len()); -- 2.12.3 From 1376b91046ad50f3a443b6fd4887252c1922870c Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Tue, 22 Aug 2017 16:23:55 +0200 Subject: [PATCH] EPUBHTMLGenerator: fix validation problem with non-page anchored images In most cases (except for at-page anchored images) there is a paragraph already opened, and writing

inside results in a validation error. So just write

in case we're not in paragraph already. --- src/lib/EPUBHTMLGenerator.cpp | 35 +++++++++++++++++++++++++++++------ 1 file changed, 29 insertions(+), 6 deletions(-) diff --git a/src/lib/EPUBHTMLGenerator.cpp b/src/lib/EPUBHTMLGenerator.cpp index 40c507e..019404f 100644 --- a/src/lib/EPUBHTMLGenerator.cpp +++ b/src/lib/EPUBHTMLGenerator.cpp @@ -360,6 +360,7 @@ struct EPUBHTMLGeneratorImpl , m_stylesheetPath(stylesheetPath) , m_actualPage(0) , m_ignore(false) + , m_frameAnchorTypes() , m_actualSink() , m_sinkStack() { @@ -446,6 +447,8 @@ struct EPUBHTMLGeneratorImpl int m_actualPage; bool m_ignore; + std::stack m_frameAnchorTypes; + protected: std::unique_ptr m_actualSink; std::stack> m_sinkStack; @@ -846,8 +849,32 @@ void EPUBHTMLGenerator::closeTable() m_impl->m_tableManager.closeTable(); } -void EPUBHTMLGenerator::openFrame(const RVNGPropertyList & /* propList */) {} -void EPUBHTMLGenerator::closeFrame() {} +void EPUBHTMLGenerator::openFrame(const RVNGPropertyList &propList) +{ + librevenge::RVNGPropertyList::Iter i(propList); + std::string anchorType; + for (i.rewind(); i.next();) + { + if (std::string("text:anchor-type") == i.key()) + anchorType = i()->getStr().cstr(); + } + + if (anchorType == "page") + // Other anchor types are already inside a paragraph. + m_impl->output().openElement("p", RVNGPropertyList()); + m_impl->m_frameAnchorTypes.push(anchorType); +} + +void EPUBHTMLGenerator::closeFrame() +{ + if (m_impl->m_frameAnchorTypes.empty()) + return; + + if (m_impl->m_frameAnchorTypes.top() == "page") + m_impl->output().closeElement("p"); + + m_impl->m_frameAnchorTypes.pop(); +} void EPUBHTMLGenerator::openGroup(const librevenge::RVNGPropertyList & /* propList */) {} void EPUBHTMLGenerator::closeGroup() {} @@ -862,8 +889,6 @@ void EPUBHTMLGenerator::drawConnector(const librevenge::RVNGPropertyList & /* pr void EPUBHTMLGenerator::insertBinaryObject(const RVNGPropertyList &propList) { - m_impl->output().openElement("p", RVNGPropertyList()); - const EPUBPath &path = m_impl->m_imageManager.insert( RVNGBinaryData(propList["office:binary-data"]->getStr()), propList["librevenge:mime-type"]->getStr()); @@ -873,8 +898,6 @@ void EPUBHTMLGenerator::insertBinaryObject(const RVNGPropertyList &propList) // FIXME: use alternative repr. if available attrs.insert("alt", path.str().c_str()); m_impl->output().insertEmptyElement("img", attrs); - - m_impl->output().closeElement("p"); } void EPUBHTMLGenerator::insertEquation(const RVNGPropertyList & /* propList */) {} -- 2.12.3 From 89ae3e392890b9360d271f1c1796cb27e36be26f Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Thu, 24 Aug 2017 17:11:17 +0200 Subject: [PATCH] EPUBTextGenerator: empty mime type is the same as no mime type epubcheck says: ERROR(RSC-005): image.epub/OEBPS/content.opf(11,69): Error while parsing file: value of attribute "media-type" is invalid; must be a string matching the regular expression "[a-zA-Z0-9!#$&+\-\^_]+/[a-zA-Z0-9!#$&+\-\^_]+.*" --- src/lib/EPUBTextGenerator.cpp | 2 +- src/test/EPUBTextGeneratorTest.cpp | 24 ++++++++++++++++++++++++ 2 files changed, 25 insertions(+), 1 deletion(-) diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index a39f266..0f7f1e0 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -614,7 +614,7 @@ void EPUBTextGenerator::insertBinaryObject(const librevenge::RVNGPropertyList &p newPropList.insert(iter.key(), iter()->clone()); } - if (!mimetype || !data) + if (!mimetype || mimetype->getStr().empty() || !data) { EPUBGEN_DEBUG_MSG(("invalid binary object dropped")); return; -- 2.12.3 From 28e5e30c20aba54dff6505df4c03d6a3da0ee0f3 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Thu, 24 Aug 2017 17:41:49 +0200 Subject: [PATCH] EPUBHTMLGenerator: sanitize URLs a bit in openLink() epubcheck warns on this: WARNING(RSC-023): large.epub/OEBPS/sections/section0018.xhtml(2,5171): The URL 'https:///www.fsf.org' is missing 1 slash(es) '/' after the protocol 'https:' --- configure.ac | 1 + src/lib/EPUBHTMLGenerator.cpp | 12 +++++++++++- src/test/EPUBTextGeneratorTest.cpp | 40 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 52 insertions(+), 1 deletion(-) diff --git a/src/lib/EPUBHTMLGenerator.cpp b/src/lib/EPUBHTMLGenerator.cpp index 019404f..aa09332 100644 --- a/src/lib/EPUBHTMLGenerator.cpp +++ b/src/lib/EPUBHTMLGenerator.cpp @@ -12,6 +12,8 @@ #include #include +#include + #include "EPUBHTMLGenerator.h" #include "EPUBImageManager.h" #include "EPUBListStyleManager.h" @@ -625,7 +627,15 @@ void EPUBHTMLGenerator::openLink(const RVNGPropertyList &propList) } RVNGPropertyList attrs; if (propList["xlink:href"]) - attrs.insert("href", propList["xlink:href"]->getStr().cstr()); + { + std::string href(propList["xlink:href"]->getStr().cstr()); + + // Basic URL sanitization. + boost::replace_all(href, "http:///", "http://"); + boost::replace_all(href, "https:///", "https://"); + + attrs.insert("href", href.c_str()); + } m_impl->output(false).openElement("a", attrs); } -- 2.12.3 From 143a470f190aaf2f420d2f84b5f08e9b01b40473 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Fri, 25 Aug 2017 17:44:38 +0200 Subject: [PATCH] Export generator --- src/lib/EPUBGenerator.cpp | 28 ++++++++++++++++++++++++++++ src/test/EPUBTextGeneratorTest.cpp | 17 +++++++++++++++++ 2 files changed, 45 insertions(+) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 4ce2964..1661064 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -7,6 +7,10 @@ * file, You can obtain one at http://mozilla.org/MPL/2.0/. */ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + #include #include @@ -298,6 +302,30 @@ void EPUBGenerator::writeRoot() sink.openElement("meta", metaAttrs); sink.insertCharacters(date); sink.closeElement("meta"); + +#ifdef VERSION + const std::string version(VERSION); +#else + const std::string version("unknown"); +#endif + std::string generator; + if (m_metadata["meta:generator"]) + generator = m_metadata["meta:generator"]->getStr().cstr(); + + if (generator.empty()) + generator = "libepubgen/" + version; + else + { + generator += " ("; + generator += "libepubgen/" + version; + generator += ")"; + } + + metaAttrs.clear(); + metaAttrs.insert("name", "generator"); + metaAttrs.insert("content", generator.c_str()); + sink.openElement("meta", metaAttrs); + sink.closeElement("meta"); } sink.closeElement("metadata"); -- 2.12.3 From 7e3b5186616326534b1ae95c6d2d188c5e522c7f Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Mon, 4 Sep 2017 17:18:49 +0200 Subject: [PATCH] EPUBGenerator: always write author and title Some EPUB3 readers categorize books by author and then title, so if these are empty, then it's next to impossible to reach the export result there. --- src/lib/EPUBGenerator.cpp | 11 ++++++++--- src/test/EPUBTextGeneratorTest.cpp | 11 +++++++++++ 2 files changed, 19 insertions(+), 3 deletions(-) diff --git a/src/lib/EPUBGenerator.cpp b/src/lib/EPUBGenerator.cpp index 1661064..3340643 100644 --- a/src/lib/EPUBGenerator.cpp +++ b/src/lib/EPUBGenerator.cpp @@ -260,15 +260,20 @@ void EPUBGenerator::writeRoot() sink.insertCharacters(identifierCharactrs.c_str()); sink.closeElement("dc:identifier"); - // Zero-width space as it must be at least one character in length after - // white space has been trimmed. - RVNGString title("\u200b"); + RVNGString title("Unknown Title"); if (m_metadata["dc:title"] && !m_metadata["dc:title"]->getStr().empty()) title = m_metadata["dc:title"]->getStr(); sink.openElement("dc:title"); sink.insertCharacters(title); sink.closeElement("dc:title"); + RVNGString creator("Unknown Author"); + if (m_metadata["meta:initial-creator"] && !m_metadata["meta:initial-creator"]->getStr().empty()) + creator = m_metadata["meta:initial-creator"]->getStr(); + sink.openElement("dc:creator"); + sink.insertCharacters(creator); + sink.closeElement("dc:creator"); + RVNGString language("en"); if (m_metadata["dc:language"] && !m_metadata["dc:language"]->getStr().empty()) language = m_metadata["dc:language"]->getStr(); -- 2.12.3 From bce7c05a18a4c5089d5ac77bc61b9f6978e7224b Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Fri, 8 Sep 2017 11:21:32 +0200 Subject: [PATCH] EPUBHTMLGenerator: write un-escaped NBSP Package implementations are supposed to take care of escaping, like it was already a requirement for normal text. --- src/lib/EPUBHTMLGenerator.cpp | 3 ++- src/test/EPUBTextGeneratorTest.cpp | 18 ++++++++++++++++++ 2 files changed, 20 insertions(+), 1 deletion(-) diff --git a/src/lib/EPUBHTMLGenerator.cpp b/src/lib/EPUBHTMLGenerator.cpp index aa09332..ed968bf 100644 --- a/src/lib/EPUBHTMLGenerator.cpp +++ b/src/lib/EPUBHTMLGenerator.cpp @@ -681,7 +681,8 @@ void EPUBHTMLGenerator::insertSpace() { if (m_impl->m_ignore) return; - m_impl->output().insertCharacters(" "); + // NBSP. + m_impl->output().insertCharacters("\xc2\xa0"); } void EPUBHTMLGenerator::openOrderedListLevel(const RVNGPropertyList &propList) -- 2.12.3 From 7cf3c6541a61f9d74cd51fb2a01344df783cb26d Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Fri, 8 Sep 2017 16:25:36 +0200 Subject: [PATCH] EPUBTextGenerator: ignore invalid media types This assumes there is no image type that is valid in EPUB2, but not in EPUB3. --- src/lib/EPUBTextGenerator.cpp | 26 +++++++++++++++++++++++++- src/test/EPUBTextGeneratorTest.cpp | 23 +++++++++++++++++++++++ 2 files changed, 48 insertions(+), 1 deletion(-) diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index 0f7f1e0..1bd1e16 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -595,6 +595,27 @@ void EPUBTextGenerator::closeFrame() m_impl->getHtml()->closeFrame(); } +/// Checks if the media type is an EPUB 3 Core Media Type or not. +static bool isValidMimeType(const RVNGString& mediaType) +{ + // Defined at . + static char const *(types[])= + { + "image/gif", + "image/png", + "image/jpeg", + "image/svg+xml" + }; + + for (const auto& i : types) + { + if (mediaType == i) + return true; + } + + return false; +} + void EPUBTextGenerator::insertBinaryObject(const librevenge::RVNGPropertyList &propList) { if (m_impl->getSplitGuard().splitOnSize()) @@ -607,7 +628,10 @@ void EPUBTextGenerator::insertBinaryObject(const librevenge::RVNGPropertyList &p for (RVNGPropertyList::Iter iter(propList); !iter.last(); iter.next()) { if (RVNGString("librevenge:mime-type") == iter.key()) - mimetype.reset(iter()->clone()); + { + if (isValidMimeType(iter()->getStr())) + mimetype.reset(iter()->clone()); + } else if (RVNGString("office:binary-data") == iter.key()) data.reset(iter()->clone()); else -- 2.12.3 From 7be89d1881e175182039ca93a1546d79933cab85 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Mon, 11 Sep 2017 11:03:03 +0200 Subject: [PATCH] EPUBSpanStyleManager: ignore line though type/style 'none' In CSS this is a boolean property, but in ODF it's an enum, so map none to false and everything else to true. --- src/lib/EPUBSpanStyleManager.cpp | 11 +++++++- src/lib/EPUBTextGenerator.cpp | 4 +-- src/test/EPUBTextGeneratorTest.cpp | 58 ++++++++++++++++++++++++++++++++++++-- 3 files changed, 68 insertions(+), 5 deletions(-) diff --git a/src/lib/EPUBSpanStyleManager.cpp b/src/lib/EPUBSpanStyleManager.cpp index 4887858..211946c 100644 --- a/src/lib/EPUBSpanStyleManager.cpp +++ b/src/lib/EPUBSpanStyleManager.cpp @@ -137,8 +137,17 @@ void EPUBSpanStyleManager::extractDecorations(RVNGPropertyList const &pList, EPU // replaceme by text-decoration-line when its implementation will appear in browser std::stringstream s; - if (pList["style:text-line-through-style"] || pList["style:text-line-through-type"]) + // line-though style or type 'none' is not line-though, everything else is. + const librevenge::RVNGProperty *textLineThoughStyle = pList["style:text-line-through-style"]; + bool lineThough = textLineThoughStyle && textLineThoughStyle->getStr() != "none"; + if (!lineThough) + { + const librevenge::RVNGProperty *textLineThoughType = pList["style:text-line-through-type"]; + lineThough = textLineThoughType && textLineThoughType->getStr() != "none"; + } + if (lineThough) s << " line-through"; + if (pList["style:text-overline-style"] || pList["style:text-overline-type"]) s << " overline"; if (pList["style:text-underline-style"] || pList["style:text-underline-type"]) diff --git a/src/lib/EPUBTextGenerator.cpp b/src/lib/EPUBTextGenerator.cpp index 1bd1e16..07aa50d 100644 --- a/src/lib/EPUBTextGenerator.cpp +++ b/src/lib/EPUBTextGenerator.cpp @@ -596,7 +596,7 @@ void EPUBTextGenerator::closeFrame() } /// Checks if the media type is an EPUB 3 Core Media Type or not. -static bool isValidMimeType(const RVNGString& mediaType) +static bool isValidMimeType(const RVNGString &mediaType) { // Defined at . static char const *(types[])= @@ -607,7 +607,7 @@ static bool isValidMimeType(const RVNGString& mediaType) "image/svg+xml" }; - for (const auto& i : types) + for (const auto &i : types) { if (mediaType == i) return true; -- 2.12.3