Leave statement caret at the start of literals

This allows a precise location to be reported for errors within literals, by
adding the offset of the error in the literal to the caret.  This will be used
to report nice errors for things like regular expressions and supported XSD
datatypes.

4 files changed, 73 insertions, 5 deletions

diff --git a/src/n3.c b/src/n3.c
index 80c7b9fe..5940c010 100644
--- a/src/n3.c
+++ b/src/n3.c
@@ -15,6 +15,7 @@
 */
 
 #include "byte_source.h"
+#include "caret.h"
 #include "env.h"
 #include "namespaces.h"
 #include "node.h"
@@ -732,6 +733,7 @@ read_object(SerdReader* const  reader,
             bool* const        ate_dot)
 {
   const size_t orig_stack_size = reader->stack.size;
+  SerdCaret    orig_caret      = reader->source->caret;
 
   assert(ctx->subject);
 
@@ -782,6 +784,7 @@ read_object(SerdReader* const  reader,
     break;
   case '\"':
   case '\'':
+    ++orig_caret.col;
     st = read_literal(reader, &o, ate_dot);
     break;
   default:
@@ -790,7 +793,7 @@ read_object(SerdReader* const  reader,
   }
 
   if (!st && simple && o) {
-    st = emit_statement(reader, *ctx, o);
+    st = emit_statement_at(reader, *ctx, o, &orig_caret);
   }
 
   serd_stack_pop_to(&reader->stack, orig_stack_size);
diff --git a/src/reader.c b/src/reader.c
index 6ce40452..5afc28ff 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -154,9 +154,10 @@ push_node(SerdReader* const  reader,
 }
 
 SerdStatus
-emit_statement(SerdReader* const reader,
-               const ReadContext ctx,
-               SerdNode* const   o)
+emit_statement_at(SerdReader* const reader,
+                  const ReadContext ctx,
+                  SerdNode* const   o,
+                  SerdCaret* const  caret)
 {
   if (reader->stack.size + (2 * sizeof(SerdNode)) > reader->stack.buf_size) {
     return SERD_ERR_OVERFLOW;
@@ -167,7 +168,7 @@ emit_statement(SerdReader* const reader,
   serd_node_zero_pad(o);
 
   const SerdStatement statement = {{ctx.subject, ctx.predicate, o, ctx.graph},
-                                   &reader->source->caret};
+                                   caret};
 
   const SerdStatus st =
     serd_sink_write_statement(reader->sink, *ctx.flags, &statement);
@@ -176,6 +177,14 @@ emit_statement(SerdReader* const reader,
   return st;
 }
 
+SerdStatus
+emit_statement(SerdReader* const reader,
+               const ReadContext ctx,
+               SerdNode* const   o)
+{
+  return emit_statement_at(reader, ctx, o, &reader->source->caret);
+}
+
 static SerdStatus
 read_statement(SerdReader* const reader)
 {
diff --git a/src/reader.h b/src/reader.h
index 30fd69a9..23d38009 100644
--- a/src/reader.h
+++ b/src/reader.h
@@ -95,6 +95,12 @@ void
 set_blank_id(SerdReader* reader, SerdNode* node, size_t buf_size);
 
 SerdStatus
+emit_statement_at(SerdReader* reader,
+                  ReadContext ctx,
+                  SerdNode*   o,
+                  SerdCaret*  caret);
+
+SerdStatus
 emit_statement(SerdReader* reader, ReadContext ctx, SerdNode* o);
 
 SerdStatus
diff --git a/test/test_reader.c b/test/test_reader.c
index 3651e014..32b475f9 100644
--- a/test/test_reader.c
+++ b/test/test_reader.c
@@ -368,6 +368,55 @@ test_read_empty(void)
   serd_world_free(world);
 }
 
+static SerdStatus
+check_cursor(void* handle, const SerdEvent* event)
+{
+  bool* const called = (bool*)handle;
+
+  if (event->type == SERD_STATEMENT) {
+    const SerdCaret* const caret =
+      serd_statement_caret(event->statement.statement);
+    assert(caret);
+
+    assert(!strcmp(serd_node_string(serd_caret_name(caret)), "string"));
+    assert(serd_caret_line(caret) == 1);
+    assert(serd_caret_column(caret) == 47);
+  }
+
+  *called = true;
+  return SERD_SUCCESS;
+}
+
+static void
+test_error_cursor(void)
+{
+  SerdWorld*        world  = serd_world_new();
+  bool              called = false;
+  SerdSink*         sink   = serd_sink_new(&called, check_cursor, NULL);
+  SerdEnv* const    env    = serd_env_new(SERD_EMPTY_STRING());
+  SerdReader* const reader =
+    serd_reader_new(world, SERD_TURTLE, 0, env, sink, 4096);
+
+  assert(sink);
+  assert(reader);
+
+  SerdByteSource* byte_source =
+    serd_byte_source_new_string("<http://example.org/s> <http://example.org/p> "
+                                "<http://example.org/o> .",
+                                NULL);
+
+  SerdStatus st = serd_reader_start(reader, byte_source);
+  assert(!st);
+  assert(serd_reader_read_document(reader) == SERD_SUCCESS);
+  assert(called);
+
+  serd_byte_source_free(byte_source);
+  serd_reader_free(reader);
+  serd_env_free(env);
+  serd_sink_free(sink);
+  serd_world_free(world);
+}
+
 int
 main(void)
 {
@@ -377,5 +426,6 @@ main(void)
   test_read_eof_by_byte();
   test_read_chunks();
   test_read_empty();
+  test_error_cursor();
   return 0;
 }