bump crengine: text typography by language (#1082)

Includes: - Fix a few clang-tidy warnings - Add support for <img src="data:image/png;base64,...> - XML parsing: add more HTML5 named entities, optimize search - Text: fix standalone BR not making an empty line - Fix BR with "display: block" not making an empty line - Fix hyphens from soft-hyphens not part of highlighted segments - Use libunibreak for line breaking - Adds TextLangMan for text typography by language
koreader · Apr 18, 2020 · fed5fa8 · fed5fa8
1 parent 3091ad3
commit fed5fa8
Show file tree

Hide file tree

Showing 4 changed files with 43 additions and 4 deletions.
diff --git a/Makefile.third b/Makefile.third
@@ -177,7 +177,7 @@ endif
 
 # crengine, fetched via GIT as a submodule
 $(CRENGINE_LIB): $(ZLIB) $(PNG_LIB) $(FREETYPE_LIB) $(HARFBUZZ_LIB) $(FRIBIDI_LIB) \
-		$(UTF8PROC_LIB) $(JPEG_LIB) $(NANOSVG_HEADERS) \
+		$(LIBUNIBREAK_LIB) $(UTF8PROC_LIB) $(JPEG_LIB) $(NANOSVG_HEADERS) \
 		$(CRENGINE_SRC_FILES) $(THIRDPARTY_DIR)/kpvcrlib/*.*
 	install -d $(CRENGINE_BUILD_DIR)
 	cd $(CRENGINE_BUILD_DIR) && \
@@ -186,6 +186,7 @@ $(CRENGINE_LIB): $(ZLIB) $(PNG_LIB) $(FREETYPE_LIB) $(HARFBUZZ_LIB) $(FRIBIDI_LI
 		FREETYPE_LIB="$(CURDIR)/$(FREETYPE_LIB)" \
 		HARFBUZZ_LIB="$(CURDIR)/$(HARFBUZZ_LIB)" \
 		FRIBIDI_LIB="$(CURDIR)/$(FRIBIDI_LIB)" \
+		LIBUNIBREAK_LIB="$(CURDIR)/$(LIBUNIBREAK_LIB)" \
 		UTF8PROC_LIB="$(CURDIR)/$(UTF8PROC_LIB)" \
 		ZLIB="$(CURDIR)/$(ZLIB)" \
 		LIBS_DIR="$(CURDIR)/$(OUTPUT_DIR)/libs" \
@@ -198,6 +199,7 @@ $(CRENGINE_LIB): $(ZLIB) $(PNG_LIB) $(FREETYPE_LIB) $(HARFBUZZ_LIB) $(FRIBIDI_LI
 		-DFREETYPE_INCLUDE_DIR="$(FREETYPE_DIR)/include/freetype2" \
 		-DHARFBUZZ_INCLUDE_DIR="$(HARFBUZZ_DIR)/include/harfbuzz" \
 		-DFRIBIDI_INCLUDE_DIR="$(FRIBIDI_DIR)/include" \
+		-DLIBUNIBREAK_INCLUDE_DIR="$(LIBUNIBREAK_DIR)/include" \
 		-DUTF8PROC_INCLUDE_DIR="$(UTF8PROC_DIR)" \
 		-DCMAKE_BUILD_TYPE=$(CMAKE_BUILD_TYPE) \
 		$(if $(WIN32),-DCMAKE_SYSTEM_NAME=Windows,) \

diff --git a/cre.cpp b/cre.cpp
@@ -623,6 +623,35 @@ static int setHyphDictionary(lua_State *L) {
 	return 0;
 }
 
+static int getTextLangStatus(lua_State *L) {
+	lua_pushstring(L, UnicodeToLocal(TextLangMan::getMainLang()).c_str());
+	lua_pushstring(L, UnicodeToLocal(TextLangMan::getMainLangHyphMethod()->getId()).c_str());
+	lua_newtable(L);
+	LVPtrVector<TextLangCfg> *list = TextLangMan::getLangCfgList();
+	for(int i = 0; i < list->length(); i++) {
+                TextLangCfg * lang_cfg = list->get(i);
+		// Key
+		lua_pushstring(L, UnicodeToLocal(lang_cfg->getLangTag()).c_str());
+		// Value: table
+		lua_newtable(L);
+
+		lua_pushstring(L, "hyph_dict_name");
+		lua_pushstring(L, UnicodeToLocal(lang_cfg->getDefaultHyphMethod()->getId()).c_str());
+		lua_settable(L, -3);
+
+		lua_pushstring(L, "hyph_nb_patterns");
+		lua_pushinteger(L, lang_cfg->getDefaultHyphMethod()->getCount());
+		lua_settable(L, -3);
+
+		lua_pushstring(L, "hyph_mem_size");
+		lua_pushinteger(L, lang_cfg->getDefaultHyphMethod()->getSize());
+		lua_settable(L, -3);
+
+		lua_settable(L, -3);
+	}
+	return 3;
+}
+
 static int loadDocument(lua_State *L) {
 	CreDocument *doc = (CreDocument*) luaL_checkudata(L, 1, "credocument");
 	const char *file_name = luaL_checkstring(L, 2);
@@ -3284,6 +3313,7 @@ static const struct luaL_Reg cre_func[] = {
     {"getHyphDictList", getHyphDictList},
     {"getSelectedHyphDict", getSelectedHyphDict},
     {"setHyphDictionary", setHyphDictionary},
+    {"getTextLangStatus", getTextLangStatus},
     {"getLatestDomVersion", getLatestDomVersion},
     {"getDomVersionWithNormalizedXPointers", getDomVersionWithNormalizedXPointers},
     {"requestDomVersion", requestDomVersion},

diff --git a/thirdparty/kpvcrlib/CMakeLists.txt b/thirdparty/kpvcrlib/CMakeLists.txt
@@ -23,6 +23,7 @@ set(CREGINE_INCLUDE_DIR ${CRE_DIR}/include)
 assert_var_defined(FREETYPE_INCLUDE_DIR)
 assert_var_defined(HARFBUZZ_INCLUDE_DIR)
 assert_var_defined(FRIBIDI_INCLUDE_DIR)
+assert_var_defined(LIBUNIBREAK_INCLUDE_DIR)
 assert_var_defined(UTF8PROC_INCLUDE_DIR)
 assert_var_defined(JPEGLIB_INCLUDE_DIR)
 assert_var_defined(PNG_INCLUDE_DIR)
@@ -34,6 +35,7 @@ include_directories(
     ${FREETYPE_INCLUDE_DIR}
     ${HARFBUZZ_INCLUDE_DIR}
     ${FRIBIDI_INCLUDE_DIR}
+    ${LIBUNIBREAK_INCLUDE_DIR}
     ${UTF8PROC_INCLUDE_DIR}
     ${ANTIWORD_INCLUDE_DIR}
     ${CHM_INCLUDE_DIRS}
@@ -55,6 +57,7 @@ add_definitions(
     -DUSE_FREETYPE=1
     -DUSE_HARFBUZZ=1
     -DUSE_FRIBIDI=1
+    -DUSE_LIBUNIBREAK=1
     -DUSE_UTF8PROC=1
     -DUSE_NANOSVG=1
     -DALLOW_KERNING=1
@@ -78,9 +81,10 @@ if(DEFINED ENV{WIN32})
     find_library(FREETYPE_LIB NAMES "freetype-6" PATHS $ENV{LIBS_DIR})
     find_library(HARFBUZZ_LIB NAMES "harfbuzz-0" PATHS $ENV{LIBS_DIR})
     find_library(FRIBIDI_LIB NAMES "fribidi-0" PATHS $ENV{LIBS_DIR})
+    find_library(LIBUNIBREAK_LIB NAMES "libunibreak-3" PATHS $ENV{LIBS_DIR})
     find_library(UTF8PROC_LIB NAMES "utf8proc-2" PATHS $ENV{LIBS_DIR})
     find_library(JPEG_LIB NAMES "jpeg-8" PATHS $ENV{LIBS_DIR})
-    set(THIRDPARTY_LIBS ${ZLIB} ${FREETYPE_LIB} ${HARFBUZZ_LIB} ${FRIBIDI_LIB} ${UTF8PROC_LIB} ${JPEG_LIB} ${PNG_LIB})
+    set(THIRDPARTY_LIBS ${ZLIB} ${FREETYPE_LIB} ${HARFBUZZ_LIB} ${FRIBIDI_LIB} ${LIBUNIBREAK_LIB} ${UTF8PROC_LIB} ${JPEG_LIB} ${PNG_LIB})
 else()
     add_library(PNG_LIB SHARED IMPORTED)
     set_target_properties(PNG_LIB PROPERTIES IMPORTED_LOCATION $ENV{PNG_LIB})
@@ -90,13 +94,15 @@ else()
     set_target_properties(HARFBUZZ_LIB PROPERTIES IMPORTED_LOCATION $ENV{HARFBUZZ_LIB})
     add_library(FRIBIDI_LIB SHARED IMPORTED)
     set_target_properties(FRIBIDI_LIB PROPERTIES IMPORTED_LOCATION $ENV{FRIBIDI_LIB})
+    add_library(LIBUNIBREAK_LIB SHARED IMPORTED)
+    set_target_properties(LIBUNIBREAK_LIB PROPERTIES IMPORTED_LOCATION $ENV{LIBUNIBREAK_LIB})
     add_library(UTF8PROC_LIB SHARED IMPORTED)
     set_target_properties(UTF8PROC_LIB PROPERTIES IMPORTED_LOCATION $ENV{UTF8PROC_LIB})
     add_library(JPEG_LIB SHARED IMPORTED)
     set_target_properties(JPEG_LIB PROPERTIES IMPORTED_LOCATION $ENV{JPEG_LIB})
     add_library(ZLIB SHARED IMPORTED)
     set_target_properties(ZLIB PROPERTIES IMPORTED_LOCATION $ENV{ZLIB})
-    set(THIRDPARTY_LIBS ZLIB FREETYPE_LIB HARFBUZZ_LIB FRIBIDI_LIB UTF8PROC_LIB JPEG_LIB PNG_LIB)
+    set(THIRDPARTY_LIBS ZLIB FREETYPE_LIB HARFBUZZ_LIB FRIBIDI_LIB LIBUNIBREAK_LIB UTF8PROC_LIB JPEG_LIB PNG_LIB)
 endif()
 
 message("Will build LIBCHM library")
@@ -141,6 +147,7 @@ set (CRENGINE_SOURCES
     ${CRE_DIR}/src/lvbmpbuf.cpp
     ${CRE_DIR}/src/lvfnt.cpp
     ${CRE_DIR}/src/hyphman.cpp
+    ${CRE_DIR}/src/textlang.cpp
     ${CRE_DIR}/src/lvfntman.cpp
     ${CRE_DIR}/src/lvimg.cpp
     ${CRE_DIR}/src/crskin.cpp

diff --git a/thirdparty/kpvcrlib/crengine b/thirdparty/kpvcrlib/crengine
+47 −20		crengine/include/hyphman.h
+13 −4		crengine/include/lvdocviewprops.h
+12 −4		crengine/include/lvfntman.h
+4 −2		crengine/include/lvrend.h
+15 −12		crengine/include/lvstyles.h
+8 −1		crengine/include/lvtextfm.h
+2 −0		crengine/include/lvtinydom.h
+44 −0		crengine/include/lvxml.h
+152 −0		crengine/include/textlang.h
+125 −11		crengine/src/hyphman.cpp
+34 −3		crengine/src/lvdocview.cpp
+44 −17		crengine/src/lvfntman.cpp
+221 −72		crengine/src/lvrend.cpp
+166 −10		crengine/src/lvtextfm.cpp
+60 −15		crengine/src/lvtinydom.cpp
+2,314 −450		crengine/src/lvxml.cpp
+467 −0		crengine/src/textlang.cpp