gh-110259: Fix f-strings with multiline expressions and format specs

Signed-off-by: Pablo Galindo <pablogsal@gmail.com>

Lib/ast.py

@@ -1270,13 +1270,15 @@ def visit_JoinedStr(self, node):
  quote_type = quote_types[0]
  self.write(f"{quote_type}{value}{quote_type}")
 
- def _write_fstring_inner(self, node):
+ def _write_fstring_inner(self, node, scape_newlines=False):
  if isinstance(node, JoinedStr):
  # for both the f-string itself, and format_spec
  for value in node.values:
- self._write_fstring_inner(value)
+ self._write_fstring_inner(value, scape_newlines=scape_newlines)
  elif isinstance(node, Constant) and isinstance(node.value, str):
  value = node.value.replace("{", "{{").replace("}", "}}")
+ if scape_newlines:
+ value = value.replace("\n", "\\n")
  self.write(value)
  elif isinstance(node, FormattedValue):
  self.visit_FormattedValue(node)
@@ -1299,7 +1301,10 @@ def unparse_inner(inner):
  self.write(f"!{chr(node.conversion)}")
  if node.format_spec:
  self.write(":")
- self._write_fstring_inner(node.format_spec)
+ self._write_fstring_inner(
+ node.format_spec,
+ scape_newlines=True
+ )
 
  def visit_Name(self, node):
  self.write(node.id)

Lib/test/test_tokenize.py

@@ -566,6 +566,38 @@ def test_string(self):
  OP '=' (3, 0) (3, 1)
  OP '}' (3, 1) (3, 2)
  FSTRING_END "'''" (3, 2) (3, 5)
+ """)
+ self.check_tokenize("""\
+f'''__{
+ x:a
+}__'''""", """\
+ FSTRING_START "f'''" (1, 0) (1, 4)
+ FSTRING_MIDDLE '__' (1, 4) (1, 6)
+ OP '{' (1, 6) (1, 7)
+ NL '\\n' (1, 7) (1, 8)
+ NAME 'x' (2, 4) (2, 5)
+ OP ':' (2, 5) (2, 6)
+ FSTRING_MIDDLE 'a' (2, 6) (2, 7)
+ NL '\\n' (2, 7) (2, 8)
+ OP '}' (3, 0) (3, 1)
+ FSTRING_MIDDLE '__' (3, 1) (3, 3)
+ FSTRING_END "'''" (3, 3) (3, 6)
+ """)
+ self.check_tokenize("""\
+f'__{
+ x:d
+}__'""", """\
+ FSTRING_START "f'" (1, 0) (1, 2)
+ FSTRING_MIDDLE '__' (1, 2) (1, 4)
+ OP '{' (1, 4) (1, 5)
+ NL '\\n' (1, 5) (1, 6)
+ NAME 'x' (2, 4) (2, 5)
+ OP ':' (2, 5) (2, 6)
+ FSTRING_MIDDLE 'd' (2, 6) (2, 7)
+ NL '\\n' (2, 7) (2, 8)
+ OP '}' (3, 0) (3, 1)
+ FSTRING_MIDDLE '__' (3, 1) (3, 3)
+ FSTRING_END "'" (3, 3) (3, 4)
  """)
 
  def test_function(self):
@@ -2277,6 +2309,35 @@ def test_string(self):
  FSTRING_START \'f"\' (1, 0) (1, 2)
  FSTRING_MIDDLE 'hola\\\\\\\\\\\\r\\\\ndfgf' (1, 2) (1, 16)
  FSTRING_END \'"\' (1, 16) (1, 17)
+ """)
+
+ self.check_tokenize("""\
+f'''__{
+ x:a
+}__'''""", """\
+ FSTRING_START "f'''" (1, 0) (1, 4)
+ FSTRING_MIDDLE '__' (1, 4) (1, 6)
+ LBRACE '{' (1, 6) (1, 7)
+ NAME 'x' (2, 4) (2, 5)
+ COLON ':' (2, 5) (2, 6)
+ FSTRING_MIDDLE 'a' (2, 6) (2, 7)
+ RBRACE '}' (3, 0) (3, 1)
+ FSTRING_MIDDLE '__' (3, 1) (3, 3)
+ FSTRING_END "'''" (3, 3) (3, 6)
+ """)
+ self.check_tokenize("""\
+f'__{
+ x:d
+}__'""", """\
+ FSTRING_START "f'" (1, 0) (1, 2)
+ FSTRING_MIDDLE '__' (1, 2) (1, 4)
+ LBRACE '{' (1, 4) (1, 5)
+ NAME 'x' (2, 4) (2, 5)
+ COLON ':' (2, 5) (2, 6)
+ FSTRING_MIDDLE 'd' (2, 6) (2, 7)
+ RBRACE '}' (3, 0) (3, 1)
+ FSTRING_MIDDLE '__' (3, 1) (3, 3)
+ FSTRING_END "'" (3, 3) (3, 4)
  """)
 
  def test_function(self):

Lib/test/test_unparse.py

@@ -730,7 +730,8 @@ class DirectoryTestCase(ASTTestCase):
  test_directories = (lib_dir, lib_dir / "test")
  run_always_files = {"test_grammar.py", "test_syntax.py", "test_compile.py",
  "test_ast.py", "test_asdl_parser.py", "test_fstring.py",
- "test_patma.py", "test_type_alias.py", "test_type_params.py"}
+ "test_patma.py", "test_type_alias.py", "test_type_params.py",
+ "test_tokenize.py"}
 
  _files_to_test = None
 

Misc/NEWS.d/next/Core and Builtins/2023-10-03-11-43-48.gh-issue-110259.ka93x5.rst

@@ -0,0 +1,3 @@
+Correctly identify the format spec in f-strings (with single or triple
+quotes) that have multiple lines in the expression part and include a
+formatting spec. Patch by Pablo Galindo

Parser/tokenizer.c

@@ -2690,6 +2690,22 @@ tok_get_fstring_mode(struct tok_state *tok, tokenizer_mode* current_tok, struct
  if (tok->done == E_ERROR) {
  return MAKE_TOKEN(ERRORTOKEN);
  }
+ int in_format_spec = (
+ current_tok->last_expr_end != -1
+ &&
+ INSIDE_FSTRING_EXPR(current_tok)
+ );
+
+ // If we are in a format spec and we found a newline,
+ // it means that the format spec ends here and we should
+ // return to the regular mode.
+ if (in_format_spec && c == '\n') {
+ tok_backup(tok, c);
+ TOK_GET_MODE(tok)->kind = TOK_REGULAR_MODE;
+ p_start = tok->start;
+ p_end = tok->cur;
+ return MAKE_TOKEN(FSTRING_MIDDLE);
+ }
  if (c == EOF || (current_tok->f_string_quote_size == 1 && c == '\n')) {
  if (tok->decoding_erred) {
  return MAKE_TOKEN(ERRORTOKEN);
@@ -2726,11 +2742,6 @@ tok_get_fstring_mode(struct tok_state *tok, tokenizer_mode* current_tok, struct
  end_quote_size = 0;
  }
 
- int in_format_spec = (
- current_tok->last_expr_end != -1
- &&
- INSIDE_FSTRING_EXPR(current_tok)
- );
  if (c == '{') {
  int peek = tok_nextc(tok);
  if (peek != '{' || in_format_spec) {