From d51be9b8d97791bff796d046d10fe16fd4e41311 Mon Sep 17 00:00:00 2001
From: David Robillard <d@drobilla.net>
Date: Wed, 10 Dec 2014 20:55:14 +0000
Subject: Configurable strict parsing to tolerate invalid URIs.

git-svn-id: http://svn.drobilla.net/serd/trunk@480 490d8e77-9747-427b-9fa3-0b8f29cee8a0
---
 src/reader.c | 14 +++++++++++++-
 src/serdi.c  |  5 +++++
 2 files changed, 18 insertions(+), 1 deletion(-)

(limited to 'src')

diff --git a/src/reader.c b/src/reader.c
index 80f1c2ed..8dddb568 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -85,6 +85,7 @@ struct SerdReaderImpl {
 	uint8_t           read_byte;  ///< 1-byte 'buffer' used when not paging
 	bool              from_file;  ///< True iff reading from `fd`
 	bool              paging;     ///< True iff reading a page at a time
+	bool              strict;     ///< True iff strict parsing
 	bool              eof;
 	bool              seen_genid;
 #ifdef SERD_STACK_CHECK
@@ -810,7 +811,10 @@ read_IRIREF(SerdReader* reader)
 					r_err(reader, SERD_ERR_BAD_SYNTAX,
 					      "invalid IRI character (escape %%%02X)\n", c, c);
 				}
-				return pop_node(reader, ref);
+				if (reader->strict) {
+					return pop_node(reader, ref);
+				}
+				push_byte(reader, ref, eat_byte_safe(reader, c));
 			} else {
 				push_byte(reader, ref, eat_byte_safe(reader, c));
 			}
@@ -1473,6 +1477,7 @@ serd_reader_new(SerdSyntax        syntax,
 	me->read_buf         = 0;
 	me->file_buf         = 0;
 	me->read_head        = 0;
+	me->strict           = false;
 	me->eof              = false;
 	me->seen_genid       = false;
 #ifdef SERD_STACK_CHECK
@@ -1487,6 +1492,13 @@ serd_reader_new(SerdSyntax        syntax,
 	return me;
 }
 
+SERD_API
+void
+serd_reader_set_strict(SerdReader* reader, bool strict)
+{
+	reader->strict = strict;
+}
+
 SERD_API
 void
 serd_reader_set_error_sink(SerdReader*   reader,
diff --git a/src/serdi.c b/src/serdi.c
index 3543580a..7568378d 100644
--- a/src/serdi.c
+++ b/src/serdi.c
@@ -50,6 +50,7 @@ print_usage(const char* name, bool error)
 	fprintf(os, "  -f           Keep full URIs in input (don't qualify).\n");
 	fprintf(os, "  -h           Display this help and exit.\n");
 	fprintf(os, "  -i SYNTAX    Input syntax (`turtle' or `ntriples').\n");
+	fprintf(os, "  -l           Lax (non-strict) parsing.\n");
 	fprintf(os, "  -o SYNTAX    Output syntax (`turtle' or `ntriples').\n");
 	fprintf(os, "  -p PREFIX    Add PREFIX to blank node IDs.\n");
 	fprintf(os, "  -q           Suppress all output except data.\n");
@@ -100,6 +101,7 @@ main(int argc, char** argv)
 	bool           bulk_read     = true;
 	bool           bulk_write    = false;
 	bool           full_uris     = false;
+	bool           lax           = false;
 	bool           quiet         = false;
 	const uint8_t* in_name       = NULL;
 	const uint8_t* add_prefix    = NULL;
@@ -119,6 +121,8 @@ main(int argc, char** argv)
 			full_uris = true;
 		} else if (argv[a][1] == 'h') {
 			return print_usage(argv[0], false);
+		} else if (argv[a][1] == 'l') {
+			lax = true;
 		} else if (argv[a][1] == 'q') {
 			quiet = true;
 		} else if (argv[a][1] == 'v') {
@@ -215,6 +219,7 @@ main(int argc, char** argv)
 		(SerdStatementSink)serd_writer_write_statement,
 		(SerdEndSink)serd_writer_end_anon);
 
+	serd_reader_set_strict(reader, !lax);
 	if (quiet) {
 		serd_reader_set_error_sink(reader, quiet_error_sink, NULL);
 		serd_writer_set_error_sink(writer, quiet_error_sink, NULL);
-- 
cgit v1.2.1