Respect csv.field_size_limit

aiocsv/_parser.c

@@ -232,6 +232,9 @@ typedef struct {
     /// C-friendly representation of the csv dialect.
     Dialect dialect;
 
+    /// Limit for the field size
+    long field_size_limit;
+
     /// Zero-based line number of the current position, which is equivalent to
     /// a one-based line number of the last-encountered line.
     unsigned int line_num;
@@ -313,6 +316,20 @@ static PyObject* Parser_new(PyObject* module, PyObject* args, PyObject* kwargs)
     Py_INCREF(reader);
     self->reader = reader;
 
+    PyObject* field_size_limit_obj =
+        PyObject_CallObject(module_get_state(module)->csv_field_size_limit, NULL);
+    if (!field_size_limit_obj) {
+        Py_DECREF(self);
+        return NULL;
+    }
+
+    self->field_size_limit = PyLong_AsLong(field_size_limit_obj);
+    Py_DECREF(field_size_limit_obj);
+    if (PyErr_Occurred()) {
+        Py_DECREF(self);
+        return NULL;
+    }
+
     self->current_read = NULL;
     self->record_so_far = NULL;
     self->field_so_far = NULL;
@@ -331,9 +348,11 @@ static PyObject* Parser_new(PyObject* module, PyObject* args, PyObject* kwargs)
 }
 
 static int Parser_add_char(Parser* self, Py_UCS4 c) {
-    // TODO: Check against csv.field_size_limit
-
-    if (self->field_so_far_len >= self->field_so_far_capacity) {
+    if (self->field_so_far_len == self->field_size_limit) {
+        PyObject* err = module_get_state(self->module)->csv_error;
+        PyErr_Format(err, "field larger than field limit (%ld)", self->field_size_limit);
+        return 0;
+    } else if (self->field_so_far_len >= self->field_so_far_capacity) {
         Py_ssize_t new_capacity =
             self->field_so_far_capacity ? self->field_so_far_capacity * 2 : 4096;
         Py_UCS4* new_buffer = self->field_so_far;

aiocsv/parser.py

@@ -41,6 +41,7 @@ def __init__(self, reader: WithAsyncRead, dialect: DialectLike) -> None:
         self.state = ParserState.START_RECORD
         self.record_so_far: list[str] = []
         self.field_so_far: list[str] = []
+        self.field_limit: int = csv.field_size_limit()
         self.field_was_numeric: bool = False
         self.last_char_was_cr: bool = False
 
@@ -231,7 +232,8 @@ def process_char_in_eat_newline(self, c: str) -> Decision:
         return Decision.DONE if c == "\n" else Decision.DONE_WITHOUT_CONSUMING
 
     def add_char(self, c: str) -> None:
-        # TODO: Check against field_limit
+        if len(self.field_so_far) == self.field_limit:
+            raise csv.Error(f"field larger than field limit ({self.field_limit})")
         self.field_so_far.append(c)
 
     def save_field(self) -> None:

tests/test_parser.py

@@ -247,3 +247,19 @@ async def test_parsing_line_num(parser: Type[Parser]):
 
     assert csv_result == expected_result
     assert custom_result == expected_result
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize("parser", PARSERS, ids=PARSER_NAMES)
+async def test_parsing_field_size_limit(parser: Type[Parser]):
+    csv.field_size_limit(64)
+
+    data = "a" * 65 + "\r\n"
+
+    csv_parser = csv.reader(io.StringIO(data, newline=""), strict=True)
+
+    with pytest.raises(csv.Error, match=r"field larger than field limit \(64\)"):
+        list(csv_parser)
+
+    with pytest.raises(csv.Error, match=r"field larger than field limit \(64\)"):
+        [r async for r in parser(AsyncStringIO(data), csv_parser.dialect)]