Make invalid UTF-8 characters in strings non-fatal

STRLEN and STRSUB report the erroneous bytes

Fixes #848

src/asm/parser.y

@@ -82,16 +82,24 @@ static char *strrstr(char *s1, char *s2)
 	return NULL;
 }
 
+static void errorInvalidUTF8Byte(uint8_t byte, char const *functionName)
+{
+	error("%s: Invalid UTF-8 byte 0x%02hhX\n", functionName, byte);
+}
+
 static size_t strlenUTF8(char const *s)
 {
 	size_t len = 0;
 	uint32_t state = 0;
 
 	for (uint32_t codep = 0; *s; s++) {
-		switch (decode(&state, &codep, *s)) {
+		uint8_t byte = *s;
+
+		switch (decode(&state, &codep, byte)) {
 		case 1:
-			fatalerror("STRLEN: Invalid UTF-8 character\n");
+			errorInvalidUTF8Byte(byte, "STRLEN");
-			break;
+			state = 0;
+			/* fallthrough */
 		case 0:
 			len++;
 			break;
@@ -100,7 +108,7 @@ static size_t strlenUTF8(char const *s)
 
 	/* Check for partial code point. */
 	if (state != 0)
-		fatalerror("STRLEN: Invalid UTF-8 character\n");
+		error("STRLEN: Incomplete UTF-8 character\n");
 
 	return len;
 }
@@ -116,14 +124,16 @@ static void strsubUTF8(char *dest, size_t destLen, char const *src, uint32_t pos
 
 	/* Advance to starting position in source string. */
 	while (src[srcIndex] && curPos < pos) {
-		switch (decode(&state, &codep, src[srcIndex++])) {
+		switch (decode(&state, &codep, src[srcIndex])) {
 		case 1:
-			fatalerror("STRSUB: Invalid UTF-8 character\n");
+			errorInvalidUTF8Byte(src[srcIndex], "STRSUB");
-			break;
+			state = 0;
+			/* fallthrough */
 		case 0:
 			curPos++;
 			break;
 		}
+		srcIndex++;
 	}
 
 	/*
@@ -138,8 +148,9 @@ static void strsubUTF8(char *dest, size_t destLen, char const *src, uint32_t pos
 	while (src[srcIndex] && destIndex < destLen - 1 && curLen < len) {
 		switch (decode(&state, &codep, src[srcIndex])) {
 		case 1:
-			fatalerror("STRSUB: Invalid UTF-8 character\n");
+			errorInvalidUTF8Byte(src[srcIndex], "STRSUB");
-			break;
+			state = 0;
+			/* fallthrough */
 		case 0:
 			curLen++;
 			break;
@@ -152,7 +163,7 @@ static void strsubUTF8(char *dest, size_t destLen, char const *src, uint32_t pos
 
 	/* Check for partial code point. */
 	if (state != 0)
-		fatalerror("STRSUB: Invalid UTF-8 character\n");
+		error("STRSUB: Incomplete UTF-8 character\n");
 
 	dest[destIndex] = '\0';
 }

test/asm/invalid-utf-8-strings.asm

@@ -0,0 +1,23 @@
+; characters:
+; 1: U+0061 a
+; 2: U+00E4 a with diaresis (0xC3 0xA4)
+; 3: U+0062 b
+; 4: U+6F22 kanji (0xE6 0xBC 0xA2)
+; 5: U+002C ,
+; 6: U+0061 a
+; 7: invalid byte 0xA3
+; 8: invalid byte 0xA4
+; 9: U+0062 b
+; 10: invalid bytes 0xE6 0xF0
+; 11: invalid byte 0xA2
+; 12: U+0021 !
+invalid EQUS "aäb漢,a<><61>b<EFBFBD><62><EFBFBD>!"
+
+n = STRLEN("{invalid}")
+copy EQUS STRSUB("{invalid}", 1)
+
+println "\"{invalid}\" == \"{copy}\" ({d:n})"
+
+mid1 EQUS STRSUB("{invalid}", 5, 2)
+mid2 EQUS STRSUB("{invalid}", 9, 1)
+println "\"{mid2}{mid1}\""

test/asm/invalid-utf-8-strings.err

@@ -0,0 +1,45 @@
+ERROR: invalid-utf-8-strings.asm(16):
+    STRLEN: Invalid UTF-8 byte 0xA3
+ERROR: invalid-utf-8-strings.asm(16):
+    STRLEN: Invalid UTF-8 byte 0xA4
+ERROR: invalid-utf-8-strings.asm(16):
+    STRLEN: Invalid UTF-8 byte 0xF0
+ERROR: invalid-utf-8-strings.asm(16):
+    STRLEN: Invalid UTF-8 byte 0xA2
+ERROR: invalid-utf-8-strings.asm(17):
+    STRLEN: Invalid UTF-8 byte 0xA3
+ERROR: invalid-utf-8-strings.asm(17):
+    STRLEN: Invalid UTF-8 byte 0xA4
+ERROR: invalid-utf-8-strings.asm(17):
+    STRLEN: Invalid UTF-8 byte 0xF0
+ERROR: invalid-utf-8-strings.asm(17):
+    STRLEN: Invalid UTF-8 byte 0xA2
+ERROR: invalid-utf-8-strings.asm(17):
+    STRSUB: Invalid UTF-8 byte 0xA3
+ERROR: invalid-utf-8-strings.asm(17):
+    STRSUB: Invalid UTF-8 byte 0xA4
+ERROR: invalid-utf-8-strings.asm(17):
+    STRSUB: Invalid UTF-8 byte 0xF0
+ERROR: invalid-utf-8-strings.asm(17):
+    STRSUB: Invalid UTF-8 byte 0xA2
+ERROR: invalid-utf-8-strings.asm(21):
+    STRLEN: Invalid UTF-8 byte 0xA3
+ERROR: invalid-utf-8-strings.asm(21):
+    STRLEN: Invalid UTF-8 byte 0xA4
+ERROR: invalid-utf-8-strings.asm(21):
+    STRLEN: Invalid UTF-8 byte 0xF0
+ERROR: invalid-utf-8-strings.asm(21):
+    STRLEN: Invalid UTF-8 byte 0xA2
+ERROR: invalid-utf-8-strings.asm(22):
+    STRLEN: Invalid UTF-8 byte 0xA3
+ERROR: invalid-utf-8-strings.asm(22):
+    STRLEN: Invalid UTF-8 byte 0xA4
+ERROR: invalid-utf-8-strings.asm(22):
+    STRLEN: Invalid UTF-8 byte 0xF0
+ERROR: invalid-utf-8-strings.asm(22):
+    STRLEN: Invalid UTF-8 byte 0xA2
+ERROR: invalid-utf-8-strings.asm(22):
+    STRSUB: Invalid UTF-8 byte 0xA3
+ERROR: invalid-utf-8-strings.asm(22):
+    STRSUB: Invalid UTF-8 byte 0xA4
+error: Assembly aborted (22 errors)!

test/asm/invalid-utf-8-strings.out

@@ -0,0 +1,2 @@
+"aäb漢,a<><61>b<EFBFBD><62><EFBFBD>!" == "aäb漢,a<><61>b<EFBFBD><62><EFBFBD>!" (12)
+"b,a"

test/asm/invalid-utf-8.asm

@@ -1,5 +1,5 @@
 ; This test tries to pass invalid UTF-8 through a macro argument
-; to exercise the lexer's reportGarbageChar
+; to exercise the lexer's unknown character reporting
 m:MACRO
 	\1
 ENDM