From 187a60b1e0e80e44b5be52a7054585f0be35ca9f Mon Sep 17 00:00:00 2001
From: David Robillard <d@drobilla.net>
Date: Fri, 21 Jan 2011 22:58:53 +0000
Subject: Full rount-trip test coverage for Turtle reading and writing. Fix
 writing UTF-8 URIs. Improved test coverage.

git-svn-id: http://svn.drobilla.net/serd/trunk@28 490d8e77-9747-427b-9fa3-0b8f29cee8a0
---
 src/reader.c | 36 ++++++++++++++++++++++++++++--------
 src/serdi.c  |  3 +--
 src/writer.c | 26 ++++++++++++++++++++++----
 3 files changed, 51 insertions(+), 14 deletions(-)

(limited to 'src')

diff --git a/src/reader.c b/src/reader.c
index bfdfa8ba..a769fe55 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -449,7 +449,7 @@ read_character(SerdReader parser, Ref dest)
 	case '\0':
 		error(parser, "unexpected end of file\n", peek_byte(parser));
 		return SERD_ERROR;
-	case '\\':
+	case '\\':  // 0x5C
 		eat_byte(parser, '\\');
 		if (read_character_escape(parser, dest)) {
 			return SERD_SUCCESS;
@@ -458,11 +458,32 @@ read_character(SerdReader parser, Ref dest)
 			return SERD_ERROR;
 		}
 	default:
-		if (in_range(c, 0x20, 0x5B) || in_range(c, 0x5D, 0x10FFF)) {
+		if (c < 0x20) {  // ASCII control character
+			error(parser, "unexpected control character\n");
+			return SERD_ERROR;
+		} else if (c <= 0x7E) {  // Printable ASCII
 			push_byte(parser, dest, eat_byte(parser, c));
 			return SERD_SUCCESS;
-		} else {
-			return SERD_FAILURE;
+		} else {  // Wide UTF-8 character
+			unsigned size = 1;
+			if ((c & 0xE0) == 0xC0) {  // Starts with `110'
+				size = 2;
+			} else if ((c & 0xF0) == 0xE0) {  // Starts with `1110'
+				size = 3;
+			} else if ((c & 0xF8) == 0xF0) {  // Starts with `11110'
+				size = 4;
+			} else if ((c & 0xFC) == 0xF8) {  // Starts with `111110'
+				size = 5;
+			} else if ((c & 0xFE) == 0xFC) {  // Starts with `1111110'
+				size = 6;
+			} else {
+				error(parser, "invalid character\n");
+				return SERD_ERROR;
+			}
+			for (unsigned i = 0; i < size; ++i) {
+				push_byte(parser, dest, eat_byte(parser, peek_byte(parser)));
+			}
+			return SERD_SUCCESS;
 		}
 	}
 }
@@ -663,7 +684,8 @@ read_relativeURI(SerdReader parser)
 	if (st != SERD_ERROR) {
 		return str;
 	}
-	return st;
+	pop_string(parser, str);
+	return 0;
 }
 
 // [30] nameStartChar ::= [A-Z] | "_" | [a-z]
@@ -769,11 +791,9 @@ read_uriref(SerdReader parser)
 {
 	TRY_RET(eat_byte(parser, '<'));
 	Ref const str = read_relativeURI(parser);
-	if (str) {
-		TRY_THROW(eat_byte(parser, '>'));
+	if (str && eat_byte(parser, '>')) {
 		return str;
 	}
-except:
 	pop_string(parser, str);
 	return 0;
 }
diff --git a/src/serdi.c b/src/serdi.c
index b14df3d3..00e687de 100644
--- a/src/serdi.c
+++ b/src/serdi.c
@@ -46,7 +46,6 @@ event_base(void*             handle,
 		SerdURI abs_base_uri;
 		if (!serd_uri_resolve(&uri, &state->base_uri, &abs_base_uri)) {
 			fprintf(stderr, "error: failed to resolve new base URI\n");
-			assert(false);
 			return false;
 		}
 		base_uri_str = serd_string_new_from_uri(&abs_base_uri, &base_uri);
@@ -173,7 +172,7 @@ main(int argc, char** argv)
 		if (strncmp((const char*)in_filename, "file:", 5)) {
 			fprintf(stderr, "unsupported URI scheme `%s'\n", in_filename);
 			return 1;
-		} else if (!strncmp((const char*)in_filename, "file:///", 7)) {
+		} else if (!strncmp((const char*)in_filename, "file://", 7)) {
 			in_filename += 7;
 		} else {
 			in_filename += 5;
diff --git a/src/writer.c b/src/writer.c
index eb6310a7..17179c60 100644
--- a/src/writer.c
+++ b/src/writer.c
@@ -115,10 +115,10 @@ write_text(SerdWriter writer, WriteContext ctx,
 			return false;
 		}
 
-		if (!(writer->style & SERD_STYLE_ASCII)) {
-			// Write UTF-8 input directly to UTF-8 output
-			writer->sink(utf8, n_bytes, writer->stream);
-			i += n_bytes - 1;
+		if (ctx == WRITE_STRING && !(writer->style & SERD_STYLE_ASCII)) {
+			// Write UTF-8 character directly to UTF-8 output
+			writer->sink(utf8 + i - 1, size, writer->stream);
+			i += size - 1;
 			continue;
 		}
 
@@ -370,6 +370,17 @@ serd_writer_set_base_uri(SerdWriter     writer,
                          const SerdURI* uri)
 {
 	writer->base_uri = *uri;
+	if (writer->syntax != SERD_NTRIPLES) {
+		if (writer->prev_g || writer->prev_s) {
+			writer->sink(" .\n\n", 4, writer->stream);
+			writer->prev_g = writer->prev_s =
+				writer->prev_p = writer->prev_o = 0;
+		}
+		writer->sink("@base ", 6, writer->stream);
+		writer->sink(" <", 2, writer->stream);
+		serd_uri_serialise(uri, writer->sink, writer->stream);
+		writer->sink("> .\n", 4, writer->stream);
+	}
 }
 
 SERD_API
@@ -379,6 +390,13 @@ serd_writer_set_prefix(SerdWriter        writer,
                        const SerdString* uri)
 {
 	if (writer->syntax != SERD_NTRIPLES) {
+		if (writer->prev_g || writer->prev_s) {
+			writer->sink(" .\n\n", 4, writer->stream);
+			writer->prev_g = 0;
+			writer->prev_s = 0;
+			writer->prev_p = 0;
+			writer->prev_o = 0;
+		}
 		writer->sink("@prefix ", 8, writer->stream);
 		writer->sink(name->buf, name->n_bytes - 1, writer->stream);
 		writer->sink(": <", 3, writer->stream);
-- 
cgit v1.2.1