From fea0ed7754b49ba8775edb3de52ea60359b536f7 Mon Sep 17 00:00:00 2001
From: Simon Rozman <simon@rozman.si>
Date: Fri, 15 Sep 2023 15:32:14 +0200
Subject: [PATCH] unicode: add system charset detection

Signed-off-by: Simon Rozman <simon@rozman.si>
---
 include/stdex/unicode.hpp | 28 +++++++++++++++++++++++++++-
 1 file changed, 27 insertions(+), 1 deletion(-)

diff --git a/include/stdex/unicode.hpp b/include/stdex/unicode.hpp
index 2ecb42ed0..082fcd693 100644
--- a/include/stdex/unicode.hpp
+++ b/include/stdex/unicode.hpp
@@ -1,4 +1,4 @@
-/*
+/*
 	SPDX-License-Identifier: MIT
 	Copyright © 2023 Amebis
 */
@@ -329,6 +329,32 @@ namespace stdex
 #endif
 		}
 
+		static charset_id system_charset()
+		{
+#ifdef _WIN32
+			return static_cast<charset_id>(GetACP());
+#else
+			const char* lctype = nl_langinfo(LC_CTYPE);
+			if (strcmp(lctype, "UTF-8") == 0) return charset_id::utf8;
+			if (strcmp(lctype, "UTF-16") == 0) return charset_id::utf16;
+#if BYTE_ORDER == BIG_ENDIAN
+			if (strcmp(lctype, "UTF-16BE") == 0) return charset_id::utf16;
+#else
+			if (strcmp(lctype, "UTF-16LE") == 0) return charset_id::utf16;
+#endif
+			if (strcmp(lctype, "UTF-32") == 0) return charset_id::utf32;
+#if BYTE_ORDER == BIG_ENDIAN
+			if (strcmp(lctype, "UTF-32BE") == 0) return charset_id::utf32;
+#else
+			if (strcmp(lctype, "UTF-32LE") == 0) return charset_id::utf32;
+#endif
+			if (strcmp(lctype, "CP1250") == 0) return charset_id::windows1250;
+			if (strcmp(lctype, "CP1251") == 0) return charset_id::windows1251;
+			if (strcmp(lctype, "CP1252") == 0) return charset_id::windows1252;
+			return charset_id::system;
+#endif
+		}
+
 #ifdef _WIN32
 	protected:
 		static UINT to_encoding(_In_ charset_id charset)