Make boost::regex optional

CMakeLists.txt

@@ -96,7 +96,6 @@ else()
 	message("Not trying to recreate SymOpTable_data.hpp since CCP4 is not defined")
 endif()
 
-# set(CMAKE_DEBUG_POSTFIX d)
 if(MSVC)
 	# make msvc standards compliant...
 	add_compile_options(/permissive-)
@@ -148,11 +147,26 @@ else()
 endif()
 
 # Libraries
+
+# Start by finding out if std:regex is usable. Note that the current
+# implementation in GCC is not acceptable, it crashes on long lines.
+try_run(STD_REGEX_RUNNING STD_REGEX_COMPILING
+	${CMAKE_CURRENT_BINARY_DIR}/test ${CMAKE_SOURCE_DIR}/cmake/test-rx.cpp)
+
+if(STD_REGEX_RUNNING STREQUAL FAILED_TO_RUN)
+	message(WARNING "You are probably trying to compile using the g++ standard library which contains a crashing std::regex implementation. Will try to use boost::regex instead")
+
+	set(NEED_BOOST_REGEX ON)
+endif()
+
 set(CMAKE_THREAD_PREFER_PTHREAD)
 set(THREADS_PREFER_PTHREAD_FLAG)
 find_package(Threads)
 
-find_package(Boost 1.70.0 REQUIRED COMPONENTS system regex program_options)
+if(NEED_BOOST_REGEX)
+	find_package(Boost 1.70.0 REQUIRED COMPONENTS system regex)
+endif()
+
 find_package(gxrio REQUIRED)
 
 include(FindFilesystem)
@@ -170,11 +184,8 @@ if(CIFPP_RECREATE_SYMOP_DATA)
 	# The tool to create the table
 	add_executable(symop-map-generator "${CMAKE_SOURCE_DIR}/tools/symop-map-generator.cpp")
 
-	target_link_libraries(symop-map-generator Threads::Threads ${Boost_LIBRARIES} ${CIFPP_REQUIRED_LIBRARIES})
-
-	if(Boost_INCLUDE_DIR)
-		target_include_directories(symop-map-generator PUBLIC ${Boost_INCLUDE_DIR})
-	endif()
+	target_link_libraries(symop-map-generator Threads::Threads
+		Boost::regex Boost::program_options ${CIFPP_REQUIRED_LIBRARIES})
 
 	set($ENV{CLIBD} ${CLIBD})
 
@@ -250,6 +261,10 @@ set(project_headers
 add_library(cifpp ${project_sources} ${project_headers} ${CMAKE_SOURCE_DIR}/src/structure/SymOpTable_data.hpp)
 set_target_properties(cifpp PROPERTIES POSITION_INDEPENDENT_CODE ON)
 
+if(NEED_BOOST_REGEX)
+	target_compile_definitions(cifpp PUBLIC USE_BOOST_REGEX=1)
+endif()
+
 target_include_directories(cifpp
 	PUBLIC
 	"$<BUILD_INTERFACE:${PROJECT_SOURCE_DIR}/include>"

cmake/test-rx.cpp

+// See: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86164
+
+#include <iostream>
+#include <regex>
+
+int main()
+{
+	std::string s(100'000, '*');
+	std::smatch m;
+	std::regex r("^(.*?)$");
+
+	std::regex_search(s, m, r);
+
+	std::cout << s.substr(0, 10) << std::endl;
+	std::cout << m.str(1).substr(0, 10) << std::endl;
+
+	return 0;
+}

include/cif++/cif/validate.hpp

@@ -27,12 +27,9 @@
 #pragma once
 
 #include <filesystem>
+#include <list>
 #include <mutex>
 
-// duh.. https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86164
-// #include <regex>
-#include <boost/regex.hpp>
-
 #include <cif++/utilities.hpp>
 
 namespace cif
@@ -63,12 +60,36 @@ enum class DDL_PrimitiveType
 
 DDL_PrimitiveType map_to_primitive_type(std::string_view s);
 
+struct regex_impl;
+
 struct type_validator
 {
 	std::string m_name;
 	DDL_PrimitiveType m_primitive_type;
-	// std::regex m_rx;
-	boost::regex m_rx;
+	regex_impl *m_rx;
+
+	type_validator() = delete;
+	type_validator(std::string_view name, DDL_PrimitiveType type, std::string_view rx);
+
+	type_validator(const type_validator &) = delete;
+	type_validator(type_validator &&rhs)
+		: m_name(std::move(rhs.m_name))
+		, m_primitive_type(rhs.m_primitive_type)
+	{
+		m_rx = std::exchange(rhs.m_rx, nullptr);
+	}
+
+	type_validator &operator=(const type_validator &) = delete;
+	type_validator &operator=(type_validator &&rhs)
+	{
+		m_name = std::move(rhs.m_name);
+		m_primitive_type = rhs.m_primitive_type;
+		m_rx = std::exchange(rhs.m_rx, nullptr);
+
+		return *this;
+	}
+
+	~type_validator();
 
 	bool operator<(const type_validator &rhs) const
 	{

src/cif/dictionary_parser.cpp

@@ -450,7 +450,7 @@ class dictionary_parser : public parser
 			try
 			{
 				type_validator v = {
-					code, map_to_primitive_type(primitiveCode), boost::regex(construct, boost::regex::extended | boost::regex::optimize)};
+					code, map_to_primitive_type(primitiveCode), construct};
 
 				m_validator.add_type_validator(std::move(v));
 			}

src/cif/validate.cpp

@@ -28,7 +28,20 @@
 #include <fstream>
 #include <iostream>
 
-#include <gzstream/gzstream.hpp>
+// The validator depends on regular expressions. Unfortunately,
+// the implementation of std::regex in g++ is buggy and crashes
+// on reading the pdbx dictionary. Therefore, in case g++ is used
+// the code will use boost::regex instead. 
+
+#if USE_BOOST_REGEX
+#include <boost/regex.hpp>
+using boost::regex;
+#else
+#include <regex>
+using std::regex;
+#endif
+
+#include <gxrio.hpp>
 
 #include <cif++/cif/dictionary_parser.hpp>
 #include <cif++/cif/validate.hpp>
@@ -37,13 +50,14 @@
 
 namespace cif
 {
-extern int VERBOSE;
-}
 
-namespace cif
+struct regex_impl : public regex
 {
-
-using cif::VERBOSE;
+	regex_impl(std::string_view rx)
+		: regex(rx.begin(), rx.end(), regex::extended | regex::optimize)
+	{
+	}
+};
 
 validation_error::validation_error(const std::string &msg)
 	: m_msg(msg)
@@ -73,6 +87,18 @@ DDL_PrimitiveType map_to_primitive_type(std::string_view s)
 
 // --------------------------------------------------------------------
 
+type_validator::type_validator(std::string_view name, DDL_PrimitiveType type, std::string_view rx)
+	: m_name(name)
+	, m_primitive_type(type)
+	, m_rx(new regex_impl(rx))
+{
+}
+
+type_validator::~type_validator()
+{
+	delete m_rx;
+}
+
 int type_validator::compare(std::string_view a, std::string_view b) const
 {
 	int result = 0;
@@ -191,7 +217,7 @@ void item_validator::operator()(std::string_view value) const
 {
 	if (not value.empty() and value != "?" and value != ".")
 	{
-		if (m_type != nullptr and not regex_match(value.begin(), value.end(), m_type->m_rx))
+		if (m_type != nullptr and not regex_match(value.begin(), value.end(), *m_type->m_rx))
 			throw validation_error(m_category->m_name, m_tag, "Value '" + std::string{value} + "' does not match type expression for type " + m_type->m_name);
 
 		if (not m_enums.empty())
@@ -240,7 +266,7 @@ const type_validator *validator::get_validator_for_type(std::string_view typeCod
 {
 	const type_validator *result = nullptr;
 
-	auto i = m_type_validators.find(type_validator{std::string(typeCode), DDL_PrimitiveType::Char, boost::regex()});
+	auto i = m_type_validators.find(type_validator{std::string(typeCode), DDL_PrimitiveType::Char, {}});
 	if (i != m_type_validators.end())
 		result = &*i;
 	else if (VERBOSE > 4)
@@ -409,7 +435,7 @@ const validator &validator_factory::operator[](std::string_view dictionary_name)
 
 		if (std::filesystem::exists(p, ec) and not ec)
 		{
-			gzstream::ifstream in(p);
+			gxrio::ifstream in(p);
 
 			if (not in.is_open())
 				throw std::runtime_error("Could not open dictionary (" + p.string() + ")");