aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--NEWS2
-rw-r--r--doc/serdi.12
-rw-r--r--include/serd/serd.h2
-rw-r--r--meson.build2
-rw-r--r--src/n3.c29
-rw-r--r--src/serd_config.h2
-rw-r--r--src/serdi.c11
-rw-r--r--src/string.c2
-rw-r--r--src/writer.c184
-rw-r--r--test/test_reader_writer.c97
-rw-r--r--test/test_string.c2
-rw-r--r--test/test_writer.c66
12 files changed, 312 insertions, 89 deletions
diff --git a/NEWS b/NEWS
index 364140a3..8b84142f 100644
--- a/NEWS
+++ b/NEWS
@@ -1,4 +1,4 @@
-serd (0.31.1) unstable; urgency=medium
+serd (0.31.3) unstable; urgency=medium
* Add Windows path separator support to serd_node_new_file_uri()
* Add serd_reader_skip_until_byte() to public API
diff --git a/doc/serdi.1 b/doc/serdi.1
index 782cfc23..998e008d 100644
--- a/doc/serdi.1
+++ b/doc/serdi.1
@@ -2,7 +2,7 @@
.\" # SPDX-License-Identifier: ISC
.Dd Jul 15, 2022
.Dt SERDI 1
-.Os Serd 0.31.1
+.Os Serd 0.31.3
.Sh NAME
.Nm serdi
.Nd read and write RDF syntax
diff --git a/include/serd/serd.h b/include/serd/serd.h
index ac9f2e06..2e7a13e6 100644
--- a/include/serd/serd.h
+++ b/include/serd/serd.h
@@ -121,6 +121,7 @@ typedef enum {
SERD_ERR_BAD_CURIE, ///< Invalid CURIE (e.g. prefix does not exist)
SERD_ERR_INTERNAL, ///< Unexpected internal error (should not happen)
SERD_ERR_BAD_WRITE, ///< Error writing to file/stream
+ SERD_ERR_BAD_TEXT, ///< Invalid text encoding
} SerdStatus;
/// Return a string describing a status code
@@ -848,6 +849,7 @@ typedef enum {
SERD_STYLE_RESOLVED = 1U << 2U, ///< Resolve URIs against base URI.
SERD_STYLE_CURIED = 1U << 3U, ///< Shorten URIs into CURIEs.
SERD_STYLE_BULK = 1U << 4U, ///< Write output in pages.
+ SERD_STYLE_STRICT = 1U << 5U, ///< Abort with error on lossy output
} SerdStyle;
/// Create a new RDF writer
diff --git a/meson.build b/meson.build
index c6da8288..74915a5a 100644
--- a/meson.build
+++ b/meson.build
@@ -13,7 +13,7 @@ project(
],
license: 'ISC',
meson_version: '>= 0.56.0',
- version: '0.31.1',
+ version: '0.31.3',
)
serd_src_root = meson.current_source_dir()
diff --git a/src/n3.c b/src/n3.c
index 573ffdda..011f57fa 100644
--- a/src/n3.c
+++ b/src/n3.c
@@ -608,8 +608,8 @@ read_PN_LOCAL(SerdReader* const reader, const Ref dest, bool* const ate_dot)
return r_err(reader, st, "bad escape\n");
}
- if (st != SERD_SUCCESS && read_PN_CHARS_BASE(reader, dest)) {
- return SERD_FAILURE;
+ if (st != SERD_SUCCESS && (st = read_PN_CHARS_BASE(reader, dest))) {
+ return st;
}
}
@@ -659,11 +659,9 @@ read_PN_PREFIX_tail(SerdReader* const reader, const Ref dest)
static SerdStatus
read_PN_PREFIX(SerdReader* const reader, const Ref dest)
{
- if (!read_PN_CHARS_BASE(reader, dest)) {
- return read_PN_PREFIX_tail(reader, dest);
- }
+ const SerdStatus st = read_PN_CHARS_BASE(reader, dest);
- return SERD_FAILURE;
+ return st ? st : read_PN_PREFIX_tail(reader, dest);
}
static SerdStatus
@@ -989,10 +987,11 @@ read_verb(SerdReader* const reader, Ref* const dest)
return SERD_SUCCESS;
}
- if (st > SERD_FAILURE || read_PrefixedName(reader, *dest, false, &ate_dot) ||
- ate_dot) {
+ if (st > SERD_FAILURE ||
+ (st = read_PrefixedName(reader, *dest, false, &ate_dot)) || ate_dot) {
*dest = pop_node(reader, *dest);
- return r_err(reader, SERD_ERR_BAD_SYNTAX, "bad verb\n");
+ st = st > SERD_FAILURE ? st : SERD_ERR_BAD_SYNTAX;
+ return r_err(reader, st, "bad verb\n");
}
return SERD_SUCCESS;
@@ -1125,8 +1124,9 @@ read_anon(SerdReader* const reader,
*ctx.flags = old_flags;
}
- return (eat_byte_check(reader, ']') == ']') ? SERD_SUCCESS
- : SERD_ERR_BAD_SYNTAX;
+ return st > SERD_FAILURE ? st
+ : (eat_byte_check(reader, ']') == ']') ? SERD_SUCCESS
+ : SERD_ERR_BAD_SYNTAX;
}
/* If emit is true: recurses, calling statement_sink for every statement
@@ -1394,7 +1394,7 @@ read_subject(SerdReader* const reader,
bool ate_dot = false;
switch ((*s_type = peek_byte(reader))) {
case '[':
- read_anon(reader, ctx, true, dest);
+ st = read_anon(reader, ctx, true, dest);
break;
case '(':
st = read_collection(reader, ctx, dest);
@@ -1567,9 +1567,8 @@ read_wrappedGraph(SerdReader* const reader, ReadContext* const ctx)
return r_err(reader, SERD_ERR_BAD_SYNTAX, "bad subject\n");
}
- if (read_triples(reader, *ctx, &ate_dot) && s_type != '[') {
- return r_err(
- reader, SERD_ERR_BAD_SYNTAX, "missing predicate object list\n");
+ if ((st = read_triples(reader, *ctx, &ate_dot)) && s_type != '[') {
+ return r_err(reader, st, "bad predicate object list\n");
}
ctx->subject = pop_node(reader, ctx->subject);
diff --git a/src/serd_config.h b/src/serd_config.h
index 2a2c9b27..8dbfe28b 100644
--- a/src/serd_config.h
+++ b/src/serd_config.h
@@ -36,7 +36,7 @@
#define SERD_SRC_SERD_CONFIG_H
// Define version unconditionally so a warning will catch a mismatch
-#define SERD_VERSION "0.31.1"
+#define SERD_VERSION "0.31.3"
#if !defined(SERD_NO_DEFAULT_CONFIG)
diff --git a/src/serdi.c b/src/serdi.c
index d82198ad..f6a329bb 100644
--- a/src/serdi.c
+++ b/src/serdi.c
@@ -144,7 +144,8 @@ choose_style(const SerdSyntax input_syntax,
const SerdSyntax output_syntax,
const bool ascii,
const bool bulk_write,
- const bool full_uris)
+ const bool full_uris,
+ const bool lax)
{
unsigned output_style = 0U;
if (output_syntax == SERD_NTRIPLES || ascii) {
@@ -166,6 +167,10 @@ choose_style(const SerdSyntax input_syntax,
output_style |= SERD_STYLE_BULK;
}
+ if (!lax) {
+ output_style |= SERD_STYLE_STRICT;
+ }
+
return (SerdStyle)output_style;
}
@@ -301,8 +306,8 @@ main(int argc, char** argv)
: SERD_NQUADS);
}
- const SerdStyle output_style =
- choose_style(input_syntax, output_syntax, ascii, bulk_write, full_uris);
+ const SerdStyle output_style = choose_style(
+ input_syntax, output_syntax, ascii, bulk_write, full_uris, lax);
SerdURI base_uri = SERD_URI_NULL;
SerdNode base = SERD_NODE_NULL;
diff --git a/src/string.c b/src/string.c
index 07513739..ecba3463 100644
--- a/src/string.c
+++ b/src/string.c
@@ -39,6 +39,8 @@ serd_strerror(const SerdStatus status)
return (const uint8_t*)"Internal error";
case SERD_ERR_BAD_WRITE:
return (const uint8_t*)"Error writing to file/stream";
+ case SERD_ERR_BAD_TEXT:
+ return (const uint8_t*)"Invalid text encoding";
}
return (const uint8_t*)"Unknown error"; // never reached
}
diff --git a/src/writer.c b/src/writer.c
index bc7146c1..07edc7f4 100644
--- a/src/writer.c
+++ b/src/writer.c
@@ -11,6 +11,7 @@
#include "serd/serd.h"
+#include <errno.h>
#include <stdarg.h>
#include <stdbool.h>
#include <stdint.h>
@@ -141,13 +142,13 @@ write_node(SerdWriter* writer,
Field field,
SerdStatementFlags flags);
-static bool
+SERD_NODISCARD static bool
supports_abbrev(const SerdWriter* writer)
{
return writer->syntax == SERD_TURTLE || writer->syntax == SERD_TRIG;
}
-static bool
+SERD_NODISCARD static bool
supports_uriref(const SerdWriter* writer)
{
return writer->syntax == SERD_TURTLE || writer->syntax == SERD_TRIG;
@@ -227,10 +228,20 @@ pop_context(SerdWriter* writer)
serd_stack_pop(&writer->anon_stack, sizeof(WriteContext));
}
-static size_t
+SERD_NODISCARD static size_t
sink(const void* buf, size_t len, SerdWriter* writer)
{
- return serd_byte_sink_write(buf, len, &writer->byte_sink);
+ const size_t written = serd_byte_sink_write(buf, len, &writer->byte_sink);
+ if (written != len) {
+ if (errno) {
+ const char* const message = strerror(errno);
+ w_err(writer, SERD_ERR_BAD_WRITE, "write error (%s)\n", message);
+ } else {
+ w_err(writer, SERD_ERR_BAD_WRITE, "write error\n");
+ }
+ }
+
+ return written;
}
SERD_NODISCARD static inline SerdStatus
@@ -242,13 +253,17 @@ esink(const void* buf, size_t len, SerdWriter* writer)
// Write a single character, as an escape for single byte characters
// (Caller prints any single byte characters that don't need escaping)
static size_t
-write_character(SerdWriter* writer, const uint8_t* utf8, size_t* size)
+write_character(SerdWriter* writer,
+ const uint8_t* utf8,
+ size_t* size,
+ SerdStatus* st)
{
char escape[11] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
const uint32_t c = parse_utf8_char(utf8, size);
switch (*size) {
case 0:
- w_err(writer, SERD_ERR_BAD_ARG, "invalid UTF-8 start: %X\n", utf8[0]);
+ *st =
+ w_err(writer, SERD_ERR_BAD_TEXT, "invalid UTF-8 start: %X\n", utf8[0]);
return 0;
case 1:
snprintf(escape, sizeof(escape), "\\u%04X", utf8[0]);
@@ -271,7 +286,7 @@ write_character(SerdWriter* writer, const uint8_t* utf8, size_t* size)
return sink(escape, 10, writer);
}
-static bool
+SERD_NODISCARD static bool
uri_must_escape(const uint8_t c)
{
switch (c) {
@@ -292,7 +307,10 @@ uri_must_escape(const uint8_t c)
}
static size_t
-write_uri(SerdWriter* writer, const uint8_t* utf8, size_t n_bytes)
+write_uri(SerdWriter* writer,
+ const uint8_t* utf8,
+ size_t n_bytes,
+ SerdStatus* st)
{
size_t len = 0;
for (size_t i = 0; i < n_bytes;) {
@@ -304,15 +322,25 @@ write_uri(SerdWriter* writer, const uint8_t* utf8, size_t n_bytes)
}
// Bulk write all characters up to this special one
- len += sink(&utf8[i], j - i, writer);
+ const size_t n_bulk = sink(&utf8[i], j - i, writer);
+ len += n_bulk;
+ if (n_bulk != j - i) {
+ *st = SERD_ERR_BAD_WRITE;
+ return len;
+ }
+
if ((i = j) == n_bytes) {
break; // Reached end
}
// Write UTF-8 character
size_t size = 0;
- len += write_character(writer, utf8 + i, &size);
+ len += write_character(writer, utf8 + i, &size, st);
i += size;
+ if (*st && (writer->style & SERD_STYLE_STRICT)) {
+ break;
+ }
+
if (size == 0) {
// Corrupt input, write percent-encoded bytes and scan to next start
char escape[4] = {0, 0, 0, 0};
@@ -326,6 +354,23 @@ write_uri(SerdWriter* writer, const uint8_t* utf8, size_t n_bytes)
return len;
}
+SERD_NODISCARD static SerdStatus
+ewrite_uri(SerdWriter* writer, const uint8_t* utf8, size_t n_bytes)
+{
+ SerdStatus st = SERD_SUCCESS;
+ write_uri(writer, utf8, n_bytes, &st);
+
+ return (st == SERD_ERR_BAD_WRITE || (writer->style & SERD_STYLE_STRICT))
+ ? st
+ : SERD_SUCCESS;
+}
+
+SERD_NODISCARD static SerdStatus
+write_uri_from_node(SerdWriter* writer, const SerdNode* node)
+{
+ return ewrite_uri(writer, node->buf, node->n_bytes);
+}
+
static bool
lname_must_escape(const uint8_t c)
{
@@ -363,10 +408,10 @@ lname_must_escape(const uint8_t c)
return false;
}
-static size_t
+SERD_NODISCARD static SerdStatus
write_lname(SerdWriter* writer, const uint8_t* utf8, size_t n_bytes)
{
- size_t len = 0;
+ SerdStatus st = SERD_SUCCESS;
for (size_t i = 0; i < n_bytes; ++i) {
size_t j = i; // Index of next character that must be escaped
for (; j < n_bytes; ++j) {
@@ -376,28 +421,28 @@ write_lname(SerdWriter* writer, const uint8_t* utf8, size_t n_bytes)
}
// Bulk write all characters up to this special one
- len += sink(&utf8[i], j - i, writer);
+ TRY(st, esink(&utf8[i], j - i, writer));
if ((i = j) == n_bytes) {
break; // Reached end
}
// Write escape
- len += sink("\\", 1, writer);
- len += sink(&utf8[i], 1, writer);
+ TRY(st, esink("\\", 1, writer));
+ TRY(st, esink(&utf8[i], 1, writer));
}
- return len;
+ return st;
}
-static size_t
+SERD_NODISCARD static SerdStatus
write_text(SerdWriter* writer,
TextContext ctx,
const uint8_t* utf8,
size_t n_bytes)
{
- size_t len = 0;
- size_t n_consecutive_quotes = 0;
- for (size_t i = 0; i < n_bytes;) {
+ size_t n_consecutive_quotes = 0;
+ SerdStatus st = SERD_SUCCESS;
+ for (size_t i = 0; !st && i < n_bytes;) {
if (utf8[i] != '"') {
n_consecutive_quotes = 0;
}
@@ -411,7 +456,7 @@ write_text(SerdWriter* writer,
}
}
- len += sink(&utf8[i], j - i, writer);
+ st = esink(&utf8[i], j - i, writer);
if ((i = j) == n_bytes) {
break; // Reached end
}
@@ -422,44 +467,44 @@ write_text(SerdWriter* writer,
switch (in) {
case '\\':
- len += sink("\\\\", 2, writer);
+ st = esink("\\\\", 2, writer);
continue;
case '\b':
- len += sink("\\b", 2, writer);
+ st = esink("\\b", 2, writer);
continue;
case '\n':
case '\r':
case '\t':
case '\f':
- len += sink(&in, 1, writer); // Write character as-is
+ st = esink(&in, 1, writer); // Write character as-is
continue;
case '\"':
if (n_consecutive_quotes >= 3 || i == n_bytes) {
// Two quotes in a row, or quote at string end, escape
- len += sink("\\\"", 2, writer);
+ st = esink("\\\"", 2, writer);
} else {
- len += sink(&in, 1, writer);
+ st = esink(&in, 1, writer);
}
continue;
default:
break;
}
- } else if (ctx == WRITE_STRING) {
+ } else {
switch (in) {
case '\\':
- len += sink("\\\\", 2, writer);
+ st = esink("\\\\", 2, writer);
continue;
case '\n':
- len += sink("\\n", 2, writer);
+ st = esink("\\n", 2, writer);
continue;
case '\r':
- len += sink("\\r", 2, writer);
+ st = esink("\\r", 2, writer);
continue;
case '\t':
- len += sink("\\t", 2, writer);
+ st = esink("\\t", 2, writer);
continue;
case '"':
- len += sink("\\\"", 2, writer);
+ st = esink("\\\"", 2, writer);
continue;
default:
break;
@@ -467,10 +512,10 @@ write_text(SerdWriter* writer,
if (writer->syntax == SERD_TURTLE) {
switch (in) {
case '\b':
- len += sink("\\b", 2, writer);
+ st = esink("\\b", 2, writer);
continue;
case '\f':
- len += sink("\\f", 2, writer);
+ st = esink("\\f", 2, writer);
continue;
default:
break;
@@ -480,10 +525,14 @@ write_text(SerdWriter* writer,
// Write UTF-8 character
size_t size = 0;
- len += write_character(writer, utf8 + i - 1, &size);
+ write_character(writer, utf8 + i - 1, &size, &st);
+ if (st && (writer->style & SERD_STYLE_STRICT)) {
+ return st;
+ }
+
if (size == 0) {
// Corrupt input, write replacement character and scan to the next start
- len += sink(replacement_char, sizeof(replacement_char), writer);
+ st = esink(replacement_char, sizeof(replacement_char), writer);
for (; i < n_bytes && (utf8[i] & 0x80); ++i) {
}
} else {
@@ -491,13 +540,21 @@ write_text(SerdWriter* writer,
}
}
- return len;
+ return (writer->style & SERD_STYLE_STRICT) ? st : SERD_SUCCESS;
}
-static size_t
+typedef struct {
+ SerdWriter* writer;
+ SerdStatus status;
+} UriSinkContext;
+
+SERD_NODISCARD static size_t
uri_sink(const void* buf, size_t len, void* stream)
{
- return write_uri((SerdWriter*)stream, (const uint8_t*)buf, len);
+ UriSinkContext* const context = (UriSinkContext*)stream;
+ SerdWriter* const writer = context->writer;
+
+ return write_uri(writer, (const uint8_t*)buf, len, &context->status);
}
SERD_NODISCARD static SerdStatus
@@ -596,14 +653,6 @@ reset_context(SerdWriter* writer, const unsigned flags)
return SERD_SUCCESS;
}
-static bool
-is_inline_start(const SerdWriter* writer, Field field, SerdStatementFlags flags)
-{
- return (supports_abbrev(writer) &&
- ((field == FIELD_SUBJECT && (flags & SERD_ANON_S_BEGIN)) ||
- (field == FIELD_OBJECT && (flags & SERD_ANON_O_BEGIN))));
-}
-
SERD_NODISCARD static SerdStatus
write_literal(SerdWriter* writer,
const SerdNode* node,
@@ -636,11 +685,11 @@ write_literal(SerdWriter* writer,
if (supports_abbrev(writer) &&
(node->flags & (SERD_HAS_NEWLINE | SERD_HAS_QUOTE))) {
TRY(st, esink("\"\"\"", 3, writer));
- write_text(writer, WRITE_LONG_STRING, node->buf, node->n_bytes);
+ TRY(st, write_text(writer, WRITE_LONG_STRING, node->buf, node->n_bytes));
TRY(st, esink("\"\"\"", 3, writer));
} else {
TRY(st, esink("\"", 1, writer));
- write_text(writer, WRITE_STRING, node->buf, node->n_bytes);
+ TRY(st, write_text(writer, WRITE_STRING, node->buf, node->n_bytes));
TRY(st, esink("\"", 1, writer));
}
if (lang && lang->buf) {
@@ -650,7 +699,8 @@ write_literal(SerdWriter* writer,
TRY(st, esink("^^", 2, writer));
return write_node(writer, datatype, NULL, NULL, FIELD_NONE, flags);
}
- return SERD_SUCCESS;
+
+ return st;
}
// Return true iff `buf` is a valid prefixed name prefix or suffix
@@ -691,10 +741,9 @@ write_uri_node(SerdWriter* const writer,
serd_env_qualify(writer->env, node, &prefix, &suffix) &&
is_name(prefix.buf, prefix.n_bytes) &&
is_name(suffix.buf, suffix.len)) {
- write_uri(writer, prefix.buf, prefix.n_bytes);
+ TRY(st, write_uri_from_node(writer, &prefix));
TRY(st, esink(":", 1, writer));
- write_uri(writer, suffix.buf, suffix.len);
- return SERD_SUCCESS;
+ return ewrite_uri(writer, suffix.buf, suffix.len);
}
}
@@ -707,6 +756,7 @@ write_uri_node(SerdWriter* const writer,
}
TRY(st, esink("<", 1, writer));
+
if (writer->style & SERD_STYLE_RESOLVED) {
SerdURI in_base_uri;
SerdURI uri;
@@ -714,21 +764,21 @@ write_uri_node(SerdWriter* const writer,
serd_env_get_base_uri(writer->env, &in_base_uri);
serd_uri_parse(node->buf, &uri);
serd_uri_resolve(&uri, &in_base_uri, &abs_uri);
- bool rooted = uri_is_under(&writer->base_uri, &writer->root_uri);
- SerdURI* root = rooted ? &writer->root_uri : &writer->base_uri;
+ bool rooted = uri_is_under(&writer->base_uri, &writer->root_uri);
+ SerdURI* root = rooted ? &writer->root_uri : &writer->base_uri;
+ UriSinkContext ctx = {writer, SERD_SUCCESS};
if (!uri_is_under(&abs_uri, root) || writer->syntax == SERD_NTRIPLES ||
writer->syntax == SERD_NQUADS) {
- serd_uri_serialise(&abs_uri, uri_sink, writer);
+ serd_uri_serialise(&abs_uri, uri_sink, &ctx);
} else {
serd_uri_serialise_relative(
- &uri, &writer->base_uri, root, uri_sink, writer);
+ &uri, &writer->base_uri, root, uri_sink, &ctx);
}
} else {
- write_uri(writer, node->buf, node->n_bytes);
+ TRY(st, write_uri_from_node(writer, node));
}
- TRY(st, esink(">", 1, writer));
- return SERD_SUCCESS;
+ return esink(">", 1, writer);
}
SERD_NODISCARD static SerdStatus
@@ -750,11 +800,11 @@ write_curie(SerdWriter* const writer, const SerdNode* const node)
if (!supports_abbrev(writer)) {
TRY(st, esink("<", 1, writer));
- write_uri(writer, prefix.buf, prefix.len);
- write_uri(writer, suffix.buf, suffix.len);
+ TRY(st, ewrite_uri(writer, prefix.buf, prefix.len));
+ TRY(st, ewrite_uri(writer, suffix.buf, suffix.len));
TRY(st, esink(">", 1, writer));
} else {
- write_lname(writer, node->buf, node->n_bytes);
+ TRY(st, write_lname(writer, node->buf, node->n_bytes));
}
return st;
@@ -769,7 +819,8 @@ write_blank(SerdWriter* const writer,
SerdStatus st = SERD_SUCCESS;
if (supports_abbrev(writer)) {
- if (is_inline_start(writer, field, flags)) {
+ if ((field == FIELD_SUBJECT && (flags & SERD_ANON_S_BEGIN)) ||
+ (field == FIELD_OBJECT && (flags & SERD_ANON_O_BEGIN))) {
return write_sep(writer, SEP_ANON_BEGIN);
}
@@ -808,6 +859,7 @@ write_node(SerdWriter* writer,
SerdStatementFlags flags)
{
SerdStatus st = SERD_SUCCESS;
+
switch (node->type) {
case SERD_NOTHING:
break;
@@ -1130,7 +1182,7 @@ serd_writer_set_base_uri(SerdWriter* writer, const SerdNode* uri)
serd_env_get_base_uri(writer->env, &writer->base_uri);
- if (writer->syntax == SERD_TURTLE || writer->syntax == SERD_TRIG) {
+ if (uri && (writer->syntax == SERD_TURTLE || writer->syntax == SERD_TRIG)) {
TRY(st, terminate_context(writer));
TRY(st, esink("@base <", 7, writer));
TRY(st, esink(uri->buf, uri->n_bytes, writer));
@@ -1172,7 +1224,7 @@ serd_writer_set_prefix(SerdWriter* writer,
TRY(st, esink("@prefix ", 8, writer));
TRY(st, esink(name->buf, name->n_bytes, writer));
TRY(st, esink(": <", 3, writer));
- write_uri(writer, uri->buf, uri->n_bytes);
+ TRY(st, ewrite_uri(writer, uri->buf, uri->n_bytes));
TRY(st, esink(">", 1, writer));
writer->last_sep = SEP_NODE;
TRY(st, write_sep(writer, SEP_END_DIRECT));
diff --git a/test/test_reader_writer.c b/test/test_reader_writer.c
index 32367bf7..83ee7b2f 100644
--- a/test/test_reader_writer.c
+++ b/test/test_reader_writer.c
@@ -10,6 +10,7 @@
#endif
#include <assert.h>
+#include <errno.h>
#include <stdbool.h>
#include <stdint.h>
#include <stdio.h>
@@ -19,6 +20,11 @@
#define USTR(s) ((const uint8_t*)(s))
typedef struct {
+ size_t n_written;
+ size_t error_offset;
+} ErrorContext;
+
+typedef struct {
int n_base;
int n_prefix;
int n_statement;
@@ -26,6 +32,29 @@ typedef struct {
const SerdNode* graph;
} ReaderTest;
+static const char* const doc_string =
+ "@base <http://drobilla.net/> .\n"
+ "@prefix eg: <http://example.org/> .\n"
+ "eg:g {\n"
+ "<http://example.com/s> eg:p \"l\\n\\\"it\" ,\n"
+ " \"\"\"long\"\"\" ,\n"
+ " \"lang\"@en ;\n"
+ " eg:p <http://example.com/o> .\n"
+ "}\n"
+ "eg:s\n"
+ " <http://example.org/p> [\n"
+ " eg:p 3.0 ,\n"
+ " 4 ,\n"
+ " \"lit\" ,\n"
+ " _:n42 ,\n"
+ " \"t\"^^eg:T\n"
+ " ] ;\n"
+ " eg:p () ;\n"
+ " eg:p\\!q (\"s\" 1 2.0 \"l\"@en eg:o) .\n"
+ "[] eg:p eg:o .\n"
+ "[ eg:p eg:o ] eg:q eg:r .\n"
+ "( eg:o ) eg:t eg:u .\n";
+
static SerdStatus
test_base_sink(void* handle, const SerdNode* uri)
{
@@ -250,6 +279,73 @@ test_read_string(void)
serd_reader_free(reader);
}
+static size_t
+faulty_sink(const void* const buf, const size_t len, void* const stream)
+{
+ (void)buf;
+ (void)len;
+
+ ErrorContext* const ctx = (ErrorContext*)stream;
+ const size_t new_n_written = ctx->n_written + len;
+ if (new_n_written >= ctx->error_offset) {
+ errno = EINVAL;
+ return 0U;
+ }
+
+ ctx->n_written += len;
+ errno = 0;
+ return len;
+}
+
+static SerdStatus
+quiet_error_sink(void* const handle, const SerdError* const e)
+{
+ (void)handle;
+ (void)e;
+ return SERD_SUCCESS;
+}
+
+static void
+test_write_errors(void)
+{
+ ErrorContext ctx = {0U, 0U};
+ const SerdStyle style = (SerdStyle)(SERD_STYLE_STRICT | SERD_STYLE_CURIED);
+
+ const size_t max_offsets[] = {0, 386, 1911, 2003, 386};
+
+ // Test errors at different offsets to hit different code paths
+ for (unsigned s = 1; s <= (unsigned)SERD_TRIG; ++s) {
+ const SerdSyntax syntax = (SerdSyntax)s;
+ for (size_t o = 0; o < max_offsets[s]; ++o) {
+ ctx.n_written = 0;
+ ctx.error_offset = o;
+
+ SerdEnv* const env = serd_env_new(NULL);
+ SerdWriter* const writer =
+ serd_writer_new(syntax, style, env, NULL, faulty_sink, &ctx);
+
+ SerdReader* const reader =
+ serd_reader_new(SERD_TRIG,
+ writer,
+ NULL,
+ (SerdBaseSink)serd_writer_set_base_uri,
+ (SerdPrefixSink)serd_writer_set_prefix,
+ (SerdStatementSink)serd_writer_write_statement,
+ (SerdEndSink)serd_writer_end_anon);
+
+ serd_reader_set_error_sink(reader, quiet_error_sink, NULL);
+ serd_writer_set_error_sink(writer, quiet_error_sink, NULL);
+
+ const SerdStatus st = serd_reader_read_string(reader, USTR(doc_string));
+ assert(st == SERD_ERR_BAD_WRITE);
+
+ serd_reader_free(reader);
+ serd_writer_free(writer);
+ serd_env_free(env);
+ }
+ }
+}
+
static void
test_writer(const char* const path)
{
@@ -470,6 +566,7 @@ main(void)
test_read_chunks(path);
test_read_string();
+ test_write_errors();
test_writer(path);
test_reader(path);
diff --git a/test/test_string.c b/test/test_string.c
index 86ccf982..6767e5ae 100644
--- a/test/test_string.c
+++ b/test/test_string.c
@@ -31,7 +31,7 @@ test_strerror(void)
{
const uint8_t* msg = serd_strerror(SERD_SUCCESS);
assert(!strcmp((const char*)msg, "Success"));
- for (int i = SERD_FAILURE; i <= SERD_ERR_BAD_WRITE; ++i) {
+ for (int i = SERD_FAILURE; i <= SERD_ERR_BAD_TEXT; ++i) {
msg = serd_strerror((SerdStatus)i);
assert(strcmp((const char*)msg, "Success"));
}
diff --git a/test/test_writer.c b/test/test_writer.c
index 356dbfcb..d236d7e6 100644
--- a/test/test_writer.c
+++ b/test/test_writer.c
@@ -84,16 +84,82 @@ test_writer_cleanup(void)
// Finish writing without terminating nodes
assert(!(st = serd_writer_finish(writer)));
+ // Set the base to an empty URI
+ assert(!(st = serd_writer_set_base_uri(writer, NULL)));
+
// Free (which could leak if the writer doesn't clean up the stack properly)
serd_writer_free(writer);
serd_env_free(env);
}
+static void
+test_strict_write(void)
+{
+ const char* const path = "serd_strict_write_test.ttl";
+ FILE* const fd = fopen(path, "wb");
+ assert(fd);
+
+ SerdEnv* const env = serd_env_new(NULL);
+ SerdWriter* const writer = serd_writer_new(
+ SERD_TURTLE, (SerdStyle)SERD_STYLE_STRICT, env, NULL, null_sink, fd);
+
+ assert(writer);
+
+ const uint8_t bad_str[] = {0xFF, 0x90, 'h', 'i', 0};
+
+ SerdNode s = serd_node_from_string(SERD_URI, USTR("http://example.org/s"));
+ SerdNode p = serd_node_from_string(SERD_URI, USTR("http://example.org/p"));
+
+ SerdNode bad_lit = serd_node_from_string(SERD_LITERAL, bad_str);
+ SerdNode bad_uri = serd_node_from_string(SERD_URI, bad_str);
+
+ assert(serd_writer_write_statement(
+ writer, 0, NULL, &s, &p, &bad_lit, NULL, NULL) == SERD_ERR_BAD_TEXT);
+
+ assert(serd_writer_write_statement(
+ writer, 0, NULL, &s, &p, &bad_uri, NULL, NULL) == SERD_ERR_BAD_TEXT);
+
+ serd_writer_free(writer);
+ serd_env_free(env);
+ fclose(fd);
+}
+
+// Produce a write error without setting errno
+static size_t
+error_sink(const void* const buf, const size_t len, void* const stream)
+{
+ (void)buf;
+ (void)len;
+ (void)stream;
+ return 0U;
+}
+
+static void
+test_write_error(void)
+{
+ SerdEnv* const env = serd_env_new(NULL);
+ SerdWriter* writer = NULL;
+ SerdStatus st = SERD_SUCCESS;
+
+ SerdNode u = serd_node_from_string(SERD_URI, USTR("http://example.com/u"));
+
+ writer =
+ serd_writer_new(SERD_TURTLE, (SerdStyle)0, env, NULL, error_sink, NULL);
+ assert(writer);
+ st = serd_writer_write_statement(writer, 0U, NULL, &u, &u, &u, NULL, NULL);
+ assert(st == SERD_ERR_BAD_WRITE);
+ serd_writer_free(writer);
+
+ serd_env_free(env);
+}
+
int
main(void)
{
test_write_long_literal();
test_writer_cleanup();
+ test_strict_write();
+ test_write_error();
return 0;
}
generated by cgit v1.2.1 (git 2.18.0) at 2024-11-06 03:01:22 +0000