Set SERD_HAS_NEWLINE and SERD_HAS_QUOTE flags when unescaped newlines or quotes are read.

Don't escape legal ASCII characters when writing long literals.


git-svn-id: http://svn.drobilla.net/serd/trunk@439 490d8e77-9747-427b-9fa3-0b8f29cee8a0

src/reader.c

@@ -436,9 +436,17 @@ read_utf8_character(SerdReader* reader, Ref dest, uint8_t c)
 // Read one character (possibly multi-byte)
 // The first byte, c, has already been eaten by caller
 static inline SerdStatus
-read_character(SerdReader* reader, Ref dest, uint8_t c)
+read_character(SerdReader* reader, Ref dest, SerdNodeFlags* flags, uint8_t c)
 {
 	if (!(c & 0x80)) {
+		switch (c) {
+		case 0xA: case 0xD:
+			*flags |= SERD_HAS_NEWLINE;
+			break;
+		case '"': case '\'':
+			*flags |= SERD_HAS_QUOTE;
+			break;
+		}
 		push_byte(reader, dest, c);
 		return SERD_SUCCESS;
 	} else {
@@ -533,10 +541,10 @@ read_STRING_LITERAL_LONG(SerdReader* reader, SerdNodeFlags* flags, uint8_t q)
 				} else {
 					*flags |= SERD_HAS_QUOTE;
 					push_byte(reader, ref, c);
-					read_character(reader, ref, q2);
+					read_character(reader, ref, flags, q2);
 				}
 			} else {
-				read_character(reader, ref, eat_byte_safe(reader, c));
+				read_character(reader, ref, flags, eat_byte_safe(reader, c));
 			}
 		}
 	}
@@ -568,7 +576,7 @@ read_STRING_LITERAL(SerdReader* reader, SerdNodeFlags* flags, uint8_t q)
 				eat_byte_check(reader, q);
 				return ref;
 			} else {
-				read_character(reader, ref, eat_byte_safe(reader, c));
+				read_character(reader, ref, flags, eat_byte_safe(reader, c));
 			}
 		}
 	}

src/writer.c

@@ -284,10 +284,20 @@ write_text(SerdWriter* writer, TextContext ctx,
 
 		uint8_t in = utf8[i++];
 		if (ctx == WRITE_LONG_STRING) {
-			if (in == '\\') {
-				len += sink("\\\\", 2, writer); continue;
-			} else if (in == '\"' && i == n_bytes) {
-				len += sink("\\\"", 2, writer); continue;  // '"' at string end
+			switch (in) {
+			case '\\': len += sink("\\\\", 2, writer); continue;
+			case '\b': len += sink("\\b", 2, writer);  continue;
+			case '\n': case '\r': case '\t': case '\f':
+				len += sink(&in, 1, writer);  // Write character as-is
+				continue;
+			case '\"':
+				if (i == n_bytes) {  // '"' at string end
+					len += sink("\\\"", 2, writer);
+				} else {
+					len += sink(&in, 1, writer);
+				}
+				continue;
+			default: break;
 			}
 		} else if (ctx == WRITE_STRING) {
 			switch (in) {

tests/good/test-backspace.nt

@@ -1,2 +1,3 @@
 <http://example.org/thing> <http://example.org/label> "\b" .
 <http://example.org/thing> <http://example.org/label> "\b" .
+<http://example.org/thing> <http://example.org/label> "\n\b\n" .

tests/good/test-backspace.ttl

@@ -1,2 +1,5 @@
 <http://example.org/thing> <http://example.org/label> "\u0008" .
-<http://example.org/thing> <http://example.org/label> "" .
+<http://example.org/thing> <http://example.org/label> "" .
+<http://example.org/thing> <http://example.org/label> """
+
+""" .