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, 80 insertions, 5 deletions

diff --git a/src/read_turtle.c b/src/read_turtle.c
index 027f0918..c3970a1e 100644
--- a/src/read_turtle.c
+++ b/src/read_turtle.c
@@ -13,6 +13,7 @@
 #include "try.h"
 #include "turtle.h"
 
+#include "serd/caret.h"
 #include "serd/node.h"
 #include "serd/reader.h"
 #include "serd/sink.h"
@@ -610,6 +611,7 @@ 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);
 
@@ -660,6 +662,7 @@ read_object(SerdReader* const  reader,
     break;
   case '\"':
   case '\'':
+    ++orig_caret.col;
     st = read_literal(reader, &o, ate_dot);
     break;
   default:
@@ -668,7 +671,7 @@ 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 f34f7974..7e0864cc 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -162,9 +162,10 @@ tokcmp(const SerdNode* const node, const char* const tok, const size_t n)
 }
 
 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_BAD_STACK;
@@ -175,7 +176,7 @@ 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);
@@ -185,6 +186,14 @@ emit_statement(SerdReader* const reader,
 }
 
 SerdStatus
+emit_statement(SerdReader* const reader,
+               const ReadContext ctx,
+               SerdNode* const   o)
+{
+  return emit_statement_at(reader, ctx, o, &reader->source->caret);
+}
+
+SerdStatus
 serd_reader_read_document(SerdReader* const reader)
 {
   assert(reader);
diff --git a/src/reader.h b/src/reader.h
index 601453b0..132d038e 100644
--- a/src/reader.h
+++ b/src/reader.h
@@ -10,6 +10,7 @@
 #include "try.h"
 
 #include "serd/attributes.h"
+#include "serd/caret.h"
 #include "serd/node.h"
 #include "serd/reader.h"
 #include "serd/sink.h"
@@ -85,6 +86,12 @@ 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);
 
 static inline int
diff --git a/test/test_reader.c b/test/test_reader.c
index dd090d83..b5cefbe1 100644
--- a/test/test_reader.c
+++ b/test/test_reader.c
@@ -3,12 +3,16 @@
 
 #undef NDEBUG
 
+#include "serd/caret.h"
 #include "serd/event.h"
 #include "serd/input_stream.h"
+#include "serd/node.h"
 #include "serd/reader.h"
 #include "serd/sink.h"
+#include "serd/statement.h"
 #include "serd/status.h"
 #include "serd/stream.h"
+#include "serd/string_view.h"
 #include "serd/syntax.h"
 #include "serd/world.h"
 #include "zix/allocator.h"
@@ -21,6 +25,7 @@
 #endif
 
 #include <assert.h>
+#include <stdbool.h>
 #include <stdio.h>
 #include <string.h>
 
@@ -486,6 +491,56 @@ test_read_empty(const char* const path)
   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_document(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);
+  SerdReader* const reader = serd_reader_new(world, SERD_TURTLE, 0, sink);
+  assert(sink);
+  assert(reader);
+
+  static const char* const string =
+    "<http://example.org/s> <http://example.org/p> "
+    "<http://example.org/o> .";
+
+  SerdNode* const string_name = serd_new_string(serd_string("string"));
+  const char*     position    = string;
+  SerdInputStream in          = serd_open_input_string(&position);
+
+  SerdStatus st = serd_reader_start(reader, &in, string_name, 1);
+  assert(!st);
+  assert(serd_reader_read_document(reader) == SERD_SUCCESS);
+  assert(!serd_reader_finish(reader));
+  assert(called);
+  assert(!serd_close_input(&in));
+
+  serd_node_free(string_name);
+  serd_reader_free(reader);
+  serd_sink_free(sink);
+  serd_world_free(world);
+}
+
 int
 main(void)
 {
@@ -504,6 +559,7 @@ main(void)
   test_read_nquads_chunks(nq_path);
   test_read_turtle_chunks(ttl_path);
   test_read_empty(ttl_path);
+  test_error_cursor();
 
   assert(!zix_remove(dir));