commit c5a1bbbee3061c2526af7e87b7688e5c8fd55b8b
Author: Juergen Spitzmueller <[email protected]>
Date: Tue Nov 25 09:36:31 2025 +0100
tex2lyx: support for language-specific charstyles
---
src/tests/CMakeLists.txt | 4 +-
src/tex2lyx/CMakeLists.txt | 4 +-
src/tex2lyx/Makefile.am | 1 +
src/tex2lyx/dummy_impl.cpp | 20 +++++++++
src/tex2lyx/tex2lyx.cpp | 34 +++++++++++++-
src/tex2lyx/tex2lyx.h | 1 +
src/tex2lyx/text.cpp | 109 ++++++++++++++++++++++++++-------------------
7 files changed, 120 insertions(+), 53 deletions(-)
diff --git a/src/tests/CMakeLists.txt b/src/tests/CMakeLists.txt
index 476e6470e9..de466c0b01 100644
--- a/src/tests/CMakeLists.txt
+++ b/src/tests/CMakeLists.txt
@@ -6,7 +6,7 @@
set(check_layout_SOURCES)
foreach(_f insets/InsetLayout.cpp CiteEnginesList.cpp Color.cpp Counters.cpp
Floating.cpp
- FloatList.cpp FontInfo.cpp Layout.cpp LayoutFile.cpp support/Lexer.cpp
+ FloatList.cpp FontInfo.cpp Language.cpp Layout.cpp LayoutFile.cpp
support/Lexer.cpp
ModuleList.cpp Spacing.cpp SpecialChar.cpp TextClass.cpp
tests/check_layout.cpp
tests/dummy_functions.cpp tests/dummy4checklayout.cpp)
list(APPEND check_layout_SOURCES "${TOP_SRC_DIR}/src/${_f}")
@@ -120,7 +120,7 @@ if (${LYX_DEBUG_SANITIZE} MATCHES "ADDRESS")
handle_source_option("CXX" ${TOP_SRC_DIR}/src/${_base}.cpp
-Wno-maybe-uninitialized)
endforeach()
endif()
-foreach(_base tests/dummy4checklayout Counters FontInfo Layout LayoutFile
+foreach(_base tests/dummy4checklayout Counters FontInfo Language Layout
LayoutFile
SpecialChar TextClass insets/InsetLayout tests/check_layout)
handle_source_option("CXX" ${TOP_SRC_DIR}/src/${_base}.cpp -Wno-shadow)
endforeach()
diff --git a/src/tex2lyx/CMakeLists.txt b/src/tex2lyx/CMakeLists.txt
index fbabf07314..5fee354821 100644
--- a/src/tex2lyx/CMakeLists.txt
+++ b/src/tex2lyx/CMakeLists.txt
@@ -12,7 +12,7 @@ set(LINKED_headers)
foreach(_src graphics/GraphicsParams insets/ExternalTemplate
insets/ExternalTransforms insets/InsetLayout Author CiteEnginesList
Color Counters
- Encoding FloatList Floating FontInfo LaTeXPackages Layout
+ Encoding FloatList Floating FontInfo Language LaTeXPackages Layout
LayoutFile LayoutModuleList support/Lexer ModuleList TextClass
LaTeXColors Spacing SpecialChar version)
list(APPEND LINKED_sources ${TOP_SRC_DIR}/src/${_src}.cpp)
@@ -30,7 +30,7 @@ if (${LYX_DEBUG_SANITIZE} MATCHES "ADDRESS")
endforeach()
endif()
foreach(_base tex2lyx/Context tex2lyx/Parser tex2lyx/Preamble tex2lyx/text
- Counters FontInfo Layout LayoutFile LayoutModuleList SpecialChar
TextClass
+ Counters FontInfo Language Layout LayoutFile LayoutModuleList
SpecialChar TextClass
insets/InsetLayout tex2lyx/dummy_impl tex2lyx/math tex2lyx/table
tex2lyx/tex2lyx
)
handle_source_option("CXX" ${TOP_SRC_DIR}/src/${_base}.cpp -Wno-shadow)
diff --git a/src/tex2lyx/Makefile.am b/src/tex2lyx/Makefile.am
index 3a1db4ca76..a3c231f1b2 100644
--- a/src/tex2lyx/Makefile.am
+++ b/src/tex2lyx/Makefile.am
@@ -99,6 +99,7 @@ LYX_OBJS = \
../FloatList.o \
../Floating.o \
../FontInfo.o \
+ ../Language.o \
../LaTeXPackages.o \
../LaTeXColors.o \
../Layout.o \
diff --git a/src/tex2lyx/dummy_impl.cpp b/src/tex2lyx/dummy_impl.cpp
index fa3aef2628..3214334b3e 100644
--- a/src/tex2lyx/dummy_impl.cpp
+++ b/src/tex2lyx/dummy_impl.cpp
@@ -21,6 +21,7 @@
#include "tex2lyx.h"
#include "LaTeXColors.h"
#include "LaTeXFeatures.h"
+#include "LaTeXFonts.h"
#include "LyXRC.h"
#include "xml.h"
@@ -63,6 +64,25 @@ LaTeXColors & theLaTeXColors()
return * lc;
}
+//
+// Dummy definitions needed by Language
+//
+
+LaTeXFonts & theLaTeXFonts()
+{
+ LaTeXFonts * lc = new LaTeXFonts;
+ return * lc;
+}
+
+LaTeXFont LaTeXFonts::getLaTeXFont(docstring const &)
+{
+ return LaTeXFont();
+}
+
+bool LaTeXFont::hasFontenc(std::string const &) const
+{
+ return false;
+}
//
// Dummy translation support (needed at many places)
diff --git a/src/tex2lyx/tex2lyx.cpp b/src/tex2lyx/tex2lyx.cpp
index 0787c2bdec..e7870936a2 100644
--- a/src/tex2lyx/tex2lyx.cpp
+++ b/src/tex2lyx/tex2lyx.cpp
@@ -17,6 +17,7 @@
#include "Context.h"
#include "Encoding.h"
+#include "Language.h"
#include "Layout.h"
#include "LayoutFile.h"
#include "LayoutModuleList.h"
@@ -492,8 +493,8 @@ bool isProvided(string const & name)
}
-bool isKnownInsetSpecialChar(string const & latex, string & lyxname, bool
const only_protected,
- bool const partof)
+bool isKnownInsetSpecialChar(string const & latex, string & lyxname, string
const & language,
+ bool const only_protected, bool const partof)
{
for (auto const & [name, sc] : textclass.specialChars()) {
if (partof && prefixIs(to_ascii(ltrim(sc.latex_output, "\\")),
latex)) {
@@ -505,6 +506,19 @@ bool isKnownInsetSpecialChar(string const & latex, string
& lyxname, bool const
return only_protected ? sc.need_protect : true;
}
}
+ Language * lang =
const_cast<Language*>(languages.getLanguage(language));
+ if (!lang)
+ return false;
+ for (auto const & [name, sc] : lang->specialChars()) {
+ if (partof && prefixIs(to_ascii(ltrim(sc.latex_output, "\\")),
latex)) {
+ lyxname = name;
+ return only_protected ? sc.need_protect : true;
+ }
+ if (to_ascii(ltrim(sc.latex_output, "\\")) == latex) {
+ lyxname = name;
+ return only_protected ? sc.need_protect : true;
+ }
+ }
return false;
}
@@ -938,6 +952,18 @@ bool roundtripMode()
return roundtrip;
}
+bool readLanguagesFile(string const & name)
+{
+ LYXERR(Debug::INIT, "About to read " << name << "...");
+
+ FileName const lang_path = libFileSearch(string(), name);
+ if (lang_path.empty())
+ return false;
+
+ languages.read(lang_path);
+ return true;
+}
+
namespace {
@@ -965,6 +991,10 @@ bool tex2lyx(idocstream & is, ostream & os, string const &
encoding,
return false;
}
+ // Load language definitions if not already done
+ if (languages.size() == 0 && !readLanguagesFile("languages"))
+ error_message("Could not load languages file.");
+
// Load preloaded modules.
// This needs to be done after the preamble is parsed, since the text
// class may not be known before. It needs to be done before parsing
diff --git a/src/tex2lyx/tex2lyx.h b/src/tex2lyx/tex2lyx.h
index 598923e4c3..1b853ef79b 100644
--- a/src/tex2lyx/tex2lyx.h
+++ b/src/tex2lyx/tex2lyx.h
@@ -130,6 +130,7 @@ extern bool checkModule(std::string const & name, bool
command);
extern bool isProvided(std::string const & name);
/// Is this a special character supported by InsetSpecialChar?
extern bool isKnownInsetSpecialChar(std::string const & latex, std::string &
lyxname,
+ std::string const & language,
bool const only_protected = false,
bool const partof = false);
// Access to environment stack
diff --git a/src/tex2lyx/text.cpp b/src/tex2lyx/text.cpp
index f8e19205d3..919e1bcbd1 100644
--- a/src/tex2lyx/text.cpp
+++ b/src/tex2lyx/text.cpp
@@ -3254,6 +3254,59 @@ void fix_child_filename(string & name)
}
+bool parseSpecialChar(ostream & os, Token const & t, Parser & p, Context &
context)
+{
+ // (Single) special chars supported via InsetSpecialChar
+ string lyxname;
+ if (isKnownInsetSpecialChar(t.cs(), lyxname, context.font.language)
+ || (t.cs() == "protect"
+ && p.next_token().cat() == catEscape
+ && isKnownInsetSpecialChar(p.next_token().cs(), lyxname,
+ context.font.language, true))) {
+ // LyX sometimes puts a \protect in front, so we have to ignore
it
+ if (t.cs() == "protect")
+ p.get_token();
+ context.check_layout(os);
+ os << "\\SpecialChar " << lyxname << '\n';
+ skip_spaces_braces(p);
+ return true;
+ }
+ // And multi-token special chars
+ if (isKnownInsetSpecialChar(t.cs(), lyxname, context.font.language,
false, true)
+ || (t.cs() == "protect"
+ && p.next_token().cat() == catEscape
+ && isKnownInsetSpecialChar(p.next_token().cs(), lyxname,
+ context.font.language, true, true)))
{
+ p.pushPosition();
+ string latex;
+ if (t.cs() == "protect") {
+ // ignore \protect
+ latex = p.next_token().cs();
+ p.get_token();
+ } else
+ latex = t.cs();
+ // Try to the complete token sequence as long as it is known
+ // as part of a known special char sequence
+ while (isKnownInsetSpecialChar(latex + p.next_token().cs(),
lyxname,
+ context.font.language, false,
true)) {
+ latex += p.next_token().cs();
+ p.get_token();
+ }
+ // It this a known complete sequence?
+ if (isKnownInsetSpecialChar(latex, lyxname,
context.font.language)) {
+ context.check_layout(os);
+ os << "\\SpecialChar " << lyxname << '\n';
+ skip_braces(p);
+ p.dropPosition();
+ return true;
+ }
+ // If not, go back and fall through
+ p.popPosition();
+ }
+ return false;
+}
+
+
void parse_text(Parser & p, ostream & os, unsigned flags, bool outer,
Context & context, string const & rdelim, string const &
rdelimesc)
{
@@ -3465,12 +3518,16 @@ void parse_text(Parser & p, ostream & os, unsigned
flags, bool outer,
}
// babel shorthands (also used by polyglossia)
- // Since these can have different meanings for different
languages
- // we import them as ERT (but they must be put in ERT to get
output
- // verbatim).
if (t.asInput() == "\"") {
+ // First check if shorthand is supported
+ // via language-specific SpecialChar
+ if (parseSpecialChar(os, t, p, context))
+ continue;
+ // Unsupported shorthands: Since these can have
different
+ // meanings for different languages,
+ // we import them as ERT (to get output verbatim).
string s = "\"";
- // We put the known shorthand pairs together in
+ // We put known shorthand pairs together in
// one ERT inset. In other cases (such as "a), only
// the quotation mark is ERTed.
if (is_known(p.next_token().asInput(),
known_babel_shorthands)) {
@@ -6688,50 +6745,8 @@ void parse_text(Parser & p, ostream & os, unsigned
flags, bool outer,
name += '{' + p.verbatim_item() + '}';
}
- // (Single) special chars supported via InsetSpecialChar
- string lyxname;
- if (isKnownInsetSpecialChar(t.cs(), lyxname)
- || (t.cs() == "protect"
- && p.next_token().cat() == catEscape
- && isKnownInsetSpecialChar(p.next_token().cs(),
lyxname, true))) {
- // LyX sometimes puts a \protect in front, so we have
to ignore it
- if (t.cs() == "protect")
- p.get_token();
- context.check_layout(os);
- os << "\\SpecialChar " << lyxname << '\n';
- skip_spaces_braces(p);
+ if (parseSpecialChar(os, t, p, context))
continue;
- }
- // And multi-token special chars
- if (isKnownInsetSpecialChar(t.cs(), lyxname, false, true)
- || (t.cs() == "protect"
- && p.next_token().cat() == catEscape
- && isKnownInsetSpecialChar(p.next_token().cs(),
lyxname, true, true))) {
- p.pushPosition();
- string latex;
- if (t.cs() == "protect") {
- // ignore \protect
- latex = p.next_token().cs();
- p.get_token();
- } else
- latex = t.cs();
- // Try to the complete token sequence as long as it is
known
- // as part of a known special char sequence
- while (isKnownInsetSpecialChar(latex +
p.next_token().cs(), lyxname, false, true)) {
- latex += p.next_token().cs();
- p.get_token();
- }
- // It this a known complete sequence?
- if (isKnownInsetSpecialChar(latex, lyxname)) {
- context.check_layout(os);
- os << "\\SpecialChar " << lyxname << '\n';
- skip_braces(p);
- p.dropPosition();
- continue;
- }
- // If not, go back and fall through
- p.popPosition();
- }
// now get the character from unicodesymbols
bool termination;
--
lyx-cvs mailing list
[email protected]
https://lists.lyx.org/mailman/listinfo/lyx-cvs
