xref: /third_party/python/Lib/tomllib/_parser.py

74     src = s.replace("\r\n", "\n")
84         pos = skip_chars(src, pos, TOML_WS)
95             char = src[pos]
102             pos = key_value_rule(src, pos, out, header, parse_float)
103             pos = skip_chars(src, pos, TOML_WS)
106                 second_char: str | None = src[pos + 1]
111                 pos, header = create_list_rule(src, pos, out)
113                 pos, header = create_dict_rule(src, pos, out)
114             pos = skip_chars(src, pos, TOML_WS)
116             raise suffixed_err(src, pos, "Invalid statement")
119         pos = skip_comment(src, pos)
123             char = src[pos]
128                 src, pos, "Expected newline or end of document after a statement"
232 def skip_chars(src: str, pos: Pos, chars: Iterable[str]) -> Pos:
234         while src[pos] in chars:
242     src: str,
250         new_pos = src.index(expect, pos)
252         new_pos = len(src)
254             raise suffixed_err(src, new_pos, f"Expected {expect!r}") from None
256     if not error_on.isdisjoint(src[pos:new_pos]):
257         while src[pos] not in error_on:
259         raise suffixed_err(src, pos, f"Found invalid character {src[pos]!r}")
263 def skip_comment(src: str, pos: Pos) -> Pos:
265         char: str | None = src[pos]
270             src, pos + 1, "\n", error_on=ILLEGAL_COMMENT_CHARS, error_on_eof=False
275 def skip_comments_and_array_ws(src: str, pos: Pos) -> Pos:
278         pos = skip_chars(src, pos, TOML_WS_AND_NEWLINE)
279         pos = skip_comment(src, pos)
284 def create_dict_rule(src: str, pos: Pos, out: Output) -> tuple[Pos, Key]:
286     pos = skip_chars(src, pos, TOML_WS)
287     pos, key = parse_key(src, pos)
290         raise suffixed_err(src, pos, f"Cannot declare {key} twice")
295         raise suffixed_err(src, pos, "Cannot overwrite a value") from None
297     if not src.startswith("]", pos):
298         raise suffixed_err(src, pos, "Expected ']' at the end of a table declaration")
302 def create_list_rule(src: str, pos: Pos, out: Output) -> tuple[Pos, Key]:
304     pos = skip_chars(src, pos, TOML_WS)
305     pos, key = parse_key(src, pos)
308         raise suffixed_err(src, pos, f"Cannot mutate immutable namespace {key}")
316         raise suffixed_err(src, pos, "Cannot overwrite a value") from None
318     if not src.startswith("]]", pos):
319         raise suffixed_err(src, pos, "Expected ']]' at the end of an array declaration")
324     src: str, pos: Pos, out: Output, header: Key, parse_float: ParseFloat
326     pos, key, value = parse_key_value_pair(src, pos, parse_float)
334             raise suffixed_err(src, pos, f"Cannot redefine namespace {cont_key}")
341             src, pos, f"Cannot mutate immutable namespace {abs_key_parent}"
347         raise suffixed_err(src, pos, "Cannot overwrite a value") from None
349         raise suffixed_err(src, pos, "Cannot overwrite a value")
358     src: str, pos: Pos, parse_float: ParseFloat
360     pos, key = parse_key(src, pos)
362         char: str | None = src[pos]
366         raise suffixed_err(src, pos, "Expected '=' after a key in a key/value pair")
368     pos = skip_chars(src, pos, TOML_WS)
369     pos, value = parse_value(src, pos, parse_float)
373 def parse_key(src: str, pos: Pos) -> tuple[Pos, Key]:
374     pos, key_part = parse_key_part(src, pos)
376     pos = skip_chars(src, pos, TOML_WS)
379             char: str | None = src[pos]
385         pos = skip_chars(src, pos, TOML_WS)
386         pos, key_part = parse_key_part(src, pos)
388         pos = skip_chars(src, pos, TOML_WS)
391 def parse_key_part(src: str, pos: Pos) -> tuple[Pos, str]:
393         char: str | None = src[pos]
398         pos = skip_chars(src, pos, BARE_KEY_CHARS)
399         return pos, src[start_pos:pos]
401         return parse_literal_str(src, pos)
403         return parse_one_line_basic_str(src, pos)
404     raise suffixed_err(src, pos, "Invalid initial character for a key part")
407 def parse_one_line_basic_str(src: str, pos: Pos) -> tuple[Pos, str]:
409     return parse_basic_str(src, pos, multiline=False)
412 def parse_array(src: str, pos: Pos, parse_float: ParseFloat) -> tuple[Pos, list]:
416     pos = skip_comments_and_array_ws(src, pos)
417     if src.startswith("]", pos):
420         pos, val = parse_value(src, pos, parse_float)
422         pos = skip_comments_and_array_ws(src, pos)
424         c = src[pos : pos + 1]
428             raise suffixed_err(src, pos, "Unclosed array")
431         pos = skip_comments_and_array_ws(src, pos)
432         if src.startswith("]", pos):
436 def parse_inline_table(src: str, pos: Pos, parse_float: ParseFloat) -> tuple[Pos, dict]:
441     pos = skip_chars(src, pos, TOML_WS)
442     if src.startswith("}", pos):
445         pos, key, value = parse_key_value_pair(src, pos, parse_float)
448             raise suffixed_err(src, pos, f"Cannot mutate immutable namespace {key}")
452             raise suffixed_err(src, pos, "Cannot overwrite a value") from None
454             raise suffixed_err(src, pos, f"Duplicate inline table key {key_stem!r}")
456         pos = skip_chars(src, pos, TOML_WS)
457         c = src[pos : pos + 1]
461             raise suffixed_err(src, pos, "Unclosed inline table")
465         pos = skip_chars(src, pos, TOML_WS)
469     src: str, pos: Pos, *, multiline: bool = False
471     escape_id = src[pos : pos + 2]
477             pos = skip_chars(src, pos, TOML_WS)
479                 char = src[pos]
483                 raise suffixed_err(src, pos, "Unescaped '\\' in a string")
485         pos = skip_chars(src, pos, TOML_WS_AND_NEWLINE)
488         return parse_hex_char(src, pos, 4)
490         return parse_hex_char(src, pos, 8)
494         raise suffixed_err(src, pos, "Unescaped '\\' in a string") from None
497 def parse_basic_str_escape_multiline(src: str, pos: Pos) -> tuple[Pos, str]:
498     return parse_basic_str_escape(src, pos, multiline=True)
501 def parse_hex_char(src: str, pos: Pos, hex_len: int) -> tuple[Pos, str]:
502     hex_str = src[pos : pos + hex_len]
504         raise suffixed_err(src, pos, "Invalid hex value")
508         raise suffixed_err(src, pos, "Escaped character is not a Unicode scalar value")
512 def parse_literal_str(src: str, pos: Pos) -> tuple[Pos, str]:
516         src, pos, "'", error_on=ILLEGAL_LITERAL_STR_CHARS, error_on_eof=True
518     return pos + 1, src[start_pos:pos]  # Skip ending apostrophe
521 def parse_multiline_str(src: str, pos: Pos, *, literal: bool) -> tuple[Pos, str]:
523     if src.startswith("\n", pos):
529             src,
535         result = src[pos:end_pos]
539         pos, result = parse_basic_str(src, pos, multiline=True)
543     if not src.startswith(delim, pos):
546     if not src.startswith(delim, pos):
552 def parse_basic_str(src: str, pos: Pos, *, multiline: bool) -> tuple[Pos, str]:
563             char = src[pos]
565             raise suffixed_err(src, pos, "Unterminated string") from None
568                 return pos + 1, result + src[start_pos:pos]
569             if src.startswith('"""', pos):
570                 return pos + 3, result + src[start_pos:pos]
574             result += src[start_pos:pos]
575             pos, parsed_escape = parse_escapes(src, pos)
580             raise suffixed_err(src, pos, f"Illegal character {char!r}")
585     src: str, pos: Pos, parse_float: ParseFloat
588         char: str | None = src[pos]
596         if src.startswith('"""', pos):
597             return parse_multiline_str(src, pos, literal=False)
598         return parse_one_line_basic_str(src, pos)
602         if src.startswith("'''", pos):
603             return parse_multiline_str(src, pos, literal=True)
604         return parse_literal_str(src, pos)
608         if src.startswith("true", pos):
611         if src.startswith("false", pos):
616         return parse_array(src, pos, parse_float)
620         return parse_inline_table(src, pos, parse_float)
623     datetime_match = RE_DATETIME.match(src, pos)
628             raise suffixed_err(src, pos, "Invalid date or datetime") from e
630     localtime_match = RE_LOCALTIME.match(src, pos)
637     number_match = RE_NUMBER.match(src, pos)
642     first_three = src[pos : pos + 3]
645     first_four = src[pos : pos + 4]
649     raise suffixed_err(src, pos, "Invalid value")
652 def suffixed_err(src: str, pos: Pos, msg: str) -> TOMLDecodeError:
656     def coord_repr(src: str, pos: Pos) -> str:
657         if pos >= len(src):
659         line = src.count("\n", 0, pos) + 1
663             column = pos - src.rindex("\n", 0, pos)
666     return TOMLDecodeError(f"{msg} (at {coord_repr(src, pos)})")

Indexes created Thu Nov 07 10:32:03 CST 2024