aboutsummaryrefslogtreecommitdiffstats
path: root/src/reader.c
diff options
context:
space:
mode:
authorDavid Robillard <d@drobilla.net>2011-12-27 17:30:36 +0000
committerDavid Robillard <d@drobilla.net>2011-12-27 17:30:36 +0000
commite0f18e34021004a19709f0c627db51af1a27afcf (patch)
tree7d6f7750da0868b1ee47559d7ac10bdb8d73fea8 /src/reader.c
parentf4b25c02971b82cd841d3f6f7afd020b97825226 (diff)
downloadserd-e0f18e34021004a19709f0c627db51af1a27afcf.tar.gz
serd-e0f18e34021004a19709f0c627db51af1a27afcf.tar.bz2
serd-e0f18e34021004a19709f0c627db51af1a27afcf.zip
Warn on invalid unicode character and replace (fix #796).
git-svn-id: http://svn.drobilla.net/serd/trunk@279 490d8e77-9747-427b-9fa3-0b8f29cee8a0
Diffstat (limited to 'src/reader.c')
-rw-r--r--src/reader.c20
1 files changed, 13 insertions, 7 deletions
diff --git a/src/reader.c b/src/reader.c
index 44f03997..836aa349 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -223,6 +223,14 @@ push_byte(SerdReader* reader, Ref ref, const uint8_t c)
}
static inline void
+push_replacement(SerdReader* reader, Ref dest)
+{
+ push_byte(reader, dest, 0xEF);
+ push_byte(reader, dest, 0xBF);
+ push_byte(reader, dest, 0xBD);
+}
+
+static inline void
append_string(SerdReader* reader, Ref ref, const uint8_t* suffix, size_t len)
{
#ifdef SERD_STACK_CHECK
@@ -312,10 +320,12 @@ read_hex_escape(SerdReader* reader, unsigned length, Ref dest)
size = 2;
} else if (c < 0x00010000) {
size = 3;
- } else if (c < 0x00200000) {
+ } else if (c < 0x00110000) {
size = 4;
} else {
- return false;
+ error(reader, "unicode character 0x%X out of range\n", c);
+ push_replacement(reader, dest);
+ return true;
}
// Build output in buf
@@ -414,11 +424,7 @@ static inline SerdStatus
bad_char(SerdReader* reader, Ref dest, const char* fmt, uint8_t c)
{
warn(reader, fmt, c);
-
- // Emit replacement character
- push_byte(reader, dest, 0xEF);
- push_byte(reader, dest, 0xBF);
- push_byte(reader, dest, 0xBD);
+ push_replacement(reader, dest);
// Skip bytes until the next start byte
for (uint8_t c = peek_byte(reader); (c & 0x80);) {