Merge branch 'master' of https://github.com/odin-lang/Odin

core/os/os_windows.odin

@@ -282,8 +282,7 @@ get_current_directory :: proc() -> string {
 
 	intrinsics.atomic_store(&cwd_gate, false);
 
-	dir_utf8 := win32.utf16_to_utf8(dir_buf_wstr);
+	return win32.utf16_to_utf8(dir_buf_wstr);
-	return dir_utf8[:len(dir_utf8)-1]; // NOTE(tetra): Remove the NUL.
 }
 
 set_current_directory :: proc(path: string) -> (err: Errno) {

core/sys/win32/general.odin

@@ -813,9 +813,11 @@ wstring_to_utf8 :: proc(s: Wstring, N: int, allocator := context.temp_allocator)
 		return "";
 	}
 
-	// NOTE: If N == -1 the call to wide_char_to_multi_byte assumes the wide string is null terminated, 
+	// If N == -1 the call to wide_char_to_multi_byte assume the wide string is null terminated
-	//       and will scan it for the first null terminated character. The resulting string is also null terminated.
+	// and will scan it to find the first null terminated character. The resulting string will 
-	//       If N != -1 it assumes the wide string is not null terminated and the resulting string is not null terminated.
+	// also null terminated.
+	// If N != -1 it assumes the wide string is not null terminated and the resulting string 
+	// will not be null terminated, we therefore have to force it to be null terminated manually.
 	text := make([]byte, n+1 if N != -1 else n, allocator);
 
 	if n1 := wide_char_to_multi_byte(CP_UTF8, WC_ERR_INVALID_CHARS, s, i32(N), cstring(&text[0]), n, nil, nil); n1 == 0 {
@@ -823,14 +825,13 @@ wstring_to_utf8 :: proc(s: Wstring, N: int, allocator := context.temp_allocator)
 		return "";
 	}
 
-	if N > 0 {
+	for i in 0..<n {
-		// NOTE: The input string is not expected to be null terminated, so we strip excess zeros at the end. 
+		if text[i] == 0 {
-		text[n] = 0;
+			n = i;
-
+			break;
-		for n >= 1 && text[n-1] == 0 {
-			n -= 1;
 		}
 	}
+
 	return string(text[:n]);
 }
 

core/sys/win32/tests/general.odin

@@ -0,0 +1,44 @@
+package win32_tests
+
+import "core:fmt"
+import "core:sys/win32"
+
+main :: proc(){
+    test_utf16_to_utf8 :: proc(str: []u16, comparison: string, expected_result: bool, loc := #caller_location) {
+        result := win32.utf16_to_utf8(str[:]);
+        fmt.assertf((result == comparison) == expected_result, 
+                    "Incorrect utf16_to_utf8 conversion: %q %s %q\nloc = %#v\n", 
+                    result, "!=" if expected_result else "==", comparison, loc);
+    }
+
+    test_utf16_to_utf8([]u16{}, "", true); 
+    test_utf16_to_utf8([]u16{0}, "", true); 
+    test_utf16_to_utf8([]u16{0, 't', 'e', 's', 't'}, "", true); 
+    test_utf16_to_utf8([]u16{0, 't', 'e', 's', 't', 0}, "", true); 
+    test_utf16_to_utf8([]u16{'t', 'e', 's', 't'}, "test", true); 
+    test_utf16_to_utf8([]u16{'t', 'e', 's', 't', 0}, "test", true); 
+    test_utf16_to_utf8([]u16{'t', 'e', 0, 's', 't'}, "te", true); 
+    test_utf16_to_utf8([]u16{'t', 'e', 0, 's', 't', 0}, "te", true); 
+
+    test_wstring_to_utf8 :: proc(str: []u16, comparison: string, expected_result: bool, loc := #caller_location) {
+        result := win32.wstring_to_utf8(nil if len(str) == 0 else cast(win32.Wstring)&str[0], -1);
+        fmt.assertf((result == comparison) == expected_result, 
+                    "Incorrect wstring_to_utf8 conversion: %q %s %q\nloc = %#v\n", 
+                    result, "!=" if expected_result else "==", comparison, loc);
+    }
+
+    test_wstring_to_utf8([]u16{}, "", true); 
+    test_wstring_to_utf8([]u16{0}, "", true); 
+    test_wstring_to_utf8([]u16{0, 't', 'e', 's', 't'}, "", true); 
+    test_wstring_to_utf8([]u16{0, 't', 'e', 's', 't', 0}, "", true); 
+    test_wstring_to_utf8([]u16{'t', 'e', 's', 't', 0}, "test", true); 
+    test_wstring_to_utf8([]u16{'t', 'e', 0, 's', 't'}, "te", true); 
+    test_wstring_to_utf8([]u16{'t', 'e', 0, 's', 't', 0}, "te", true); 
+
+    // WARNING: Passing a non-zero-terminated string to wstring_to_utf8 is dangerous, 
+    //          as it will go out of bounds looking for a zero. 
+    //          It will "fail" or "succeed" by having a zero just after the end of the input string or not.
+    test_wstring_to_utf8([]u16{'t', 'e', 's', 't'}, "test", false); 
+    test_wstring_to_utf8([]u16{'t', 'e', 's', 't', 0}[:4], "test", true); 
+    test_wstring_to_utf8([]u16{'t', 'e', 's', 't', 'q'}[:4], "test", false); 
+}