Check for iconv() TRANSLIT support

Some iconv() implementations don't support transliteration. When iconv with //TRANSLIT is first attempted and fails, try again without it, and remember the result. Fixes #364
akrennmair · Sep 29, 2016 · d41345f · d41345f
1 parent fcbe4fd
commit d41345f
Show file tree

Hide file tree

Showing 3 changed files with 46 additions and 4 deletions.
diff --git a/include/utils.h b/include/utils.h
@@ -46,6 +46,7 @@ class utils {
 		static bool try_fs_lock(const std::string& lock_file, pid_t & pid);
 		static void remove_fs_lock(const std::string& lock_file);
 
+		static std::string translit(const std::string& tocode, const std::string& fromcode);
 		static std::string convert_text(const std::string& text, const std::string& tocode, const std::string& fromcode);
 
 		static std::string get_command_output(const std::string& cmd);

diff --git a/src/stflpp.cpp b/src/stflpp.cpp
@@ -4,6 +4,7 @@
 #include <cerrno>
 
 #include <langinfo.h>
+#include <utils.h>
 
 namespace newsbeuter {
 
@@ -16,7 +17,7 @@ namespace newsbeuter {
  */
 
 stfl::form::form(const std::string& text) : f(0) {
-	ipool = stfl_ipool_create((std::string(nl_langinfo(CODESET)) + "//TRANSLIT").c_str());
+	ipool = stfl_ipool_create(utils::translit(std::string(nl_langinfo(CODESET)), "WCHAR_T").c_str());
 	if (!ipool) {
 		throw exception(errno);
 	}
@@ -82,7 +83,7 @@ static std::mutex quote_mtx;
 
 std::string stfl::quote(const std::string& text) {
 	std::lock_guard<std::mutex> lock(quote_mtx);
-	stfl_ipool * ipool = stfl_ipool_create((std::string(nl_langinfo(CODESET)) + "//TRANSLIT").c_str());
+	stfl_ipool * ipool = stfl_ipool_create(utils::translit(std::string(nl_langinfo(CODESET)), "WCHAR_T").c_str());
 	std::string retval = stfl_ipool_fromwc(ipool,stfl_quote(stfl_ipool_towc(ipool,text.c_str())));
 	stfl_ipool_destroy(ipool);
 	return retval;

diff --git a/src/utils.cpp b/src/utils.cpp
@@ -256,14 +256,54 @@ bool utils::try_fs_lock(const std::string& lock_file, pid_t & pid) {
 	return false;
 }
 
+std::string utils::translit(const std::string& tocode, const std::string& fromcode)
+{
+	std::string tlit = "//TRANSLIT";
+
+	typedef enum translit_state {
+		UNKNOWN,
+		SUPPORTED,
+		UNSUPPORTED
+	} translit_state_t;
+
+	static translit_state_t state = UNKNOWN;
+
+	// TRANSLIT is not needed when converting to unicode encodings
+	if (tocode == "utf-8" || tocode == "WCHAR_T") return tocode;
+
+	if (state == UNKNOWN) {
+		iconv_t cd = ::iconv_open((tocode + "//TRANSLIT").c_str(), fromcode.c_str());
+
+		if (cd == reinterpret_cast<iconv_t>(-1)) {
+			if (errno == EINVAL) {
+				iconv_t cd = ::iconv_open(tocode.c_str(), fromcode.c_str());
+				if (cd != reinterpret_cast<iconv_t>(-1)) {
+					state = UNSUPPORTED;
+				} else {
+					fprintf(stderr, "iconv_open('%s', '%s') failed: %s", tocode.c_str(), fromcode.c_str(), strerror(errno));
+					abort();
+				}
+			} else {
+				fprintf(stderr, "iconv_open('%s//TRANSLIT', '%s') failed: %s", tocode.c_str(), fromcode.c_str(), strerror(errno));
+				abort();
+			}
+		} else {
+			state = SUPPORTED;
+		}
+
+		iconv_close(cd);
+	}
+
+	return ((state == SUPPORTED) ? (tocode + tlit) : (tocode));
+}
 
 std::string utils::convert_text(const std::string& text, const std::string& tocode, const std::string& fromcode) {
 	std::string result;
 
 	if (strcasecmp(tocode.c_str(), fromcode.c_str())==0)
 		return text;
 
-	iconv_t cd = ::iconv_open((tocode + "//TRANSLIT").c_str(), fromcode.c_str());
+	iconv_t cd = ::iconv_open(translit(tocode, fromcode).c_str(), fromcode.c_str());
 
 	if (cd == reinterpret_cast<iconv_t>(-1))
 		return result;
@@ -499,7 +539,7 @@ std::wstring utils::str2wstr(const std::string& str) {
 
 std::string utils::wstr2str(const std::wstring& wstr) {
 	std::string codeset = nl_langinfo(CODESET);
-	codeset.append("//TRANSLIT");
+	codeset = translit(codeset, "WCHAR_T");
 	struct stfl_ipool * ipool = stfl_ipool_create(codeset.c_str());
 	std::string result = stfl_ipool_fromwc(ipool, wstr.c_str());
 	stfl_ipool_destroy(ipool);