flowmetal/components/utokenize/test/python/test_tokenize.py

2651 lines
99 KiB
Python

from test import support
from test.support import os_helper
from tokenize import (tokenize, _tokenize, untokenize, NUMBER, NAME, OP,
STRING, ENDMARKER, ENCODING, tok_name, detect_encoding,
open as tokenize_open, Untokenizer, generate_tokens,
NEWLINE, _generate_tokens_from_c_tokenizer, DEDENT)
from io import BytesIO, StringIO
import unittest
from textwrap import dedent
from unittest import TestCase, mock
from test.test_grammar import (VALID_UNDERSCORE_LITERALS,
INVALID_UNDERSCORE_LITERALS)
from test.support import os_helper
from test.support.script_helper import run_test_script, make_script
import os
import token
# Converts a source string into a list of textual representation
# of the tokens such as:
# ` NAME 'if' (1, 0) (1, 2)`
# to make writing tests easier.
def stringify_tokens_from_source(token_generator, source_string):
result = []
num_lines = len(source_string.splitlines())
missing_trailing_nl = source_string[-1] not in '\r\n'
for type, token, start, end, line in token_generator:
if type == ENDMARKER:
break
# Ignore the new line on the last line if the input lacks one
if missing_trailing_nl and type == NEWLINE and end[0] == num_lines:
continue
type = tok_name[type]
result.append(f" {type:10} {token!r:13} {start} {end}")
return result
class TokenizeTest(TestCase):
# Tests for the tokenize module.
# The tests can be really simple. Given a small fragment of source
# code, print out a table with tokens. The ENDMARKER, ENCODING and
# final NEWLINE are omitted for brevity.
def check_tokenize(self, s, expected):
# Format the tokens in s in a table format.
# The ENDMARKER and final NEWLINE are omitted.
f = BytesIO(s.encode('utf-8'))
result = stringify_tokens_from_source(tokenize(f.readline), s)
self.assertEqual(result,
[" ENCODING 'utf-8' (0, 0) (0, 0)"] +
expected.rstrip().splitlines())
def test_implicit_newline(self):
# Make sure that the tokenizer puts in an implicit NEWLINE
# when the input lacks a trailing new line.
f = BytesIO("x".encode('utf-8'))
tokens = list(tokenize(f.readline))
self.assertEqual(tokens[-2].type, NEWLINE)
self.assertEqual(tokens[-1].type, ENDMARKER)
def test_basic(self):
self.check_tokenize("1 + 1", """\
NUMBER '1' (1, 0) (1, 1)
OP '+' (1, 2) (1, 3)
NUMBER '1' (1, 4) (1, 5)
""")
self.check_tokenize("if False:\n"
" # NL\n"
" \n"
" True = False # NEWLINE\n", """\
NAME 'if' (1, 0) (1, 2)
NAME 'False' (1, 3) (1, 8)
OP ':' (1, 8) (1, 9)
NEWLINE '\\n' (1, 9) (1, 10)
COMMENT '# NL' (2, 4) (2, 8)
NL '\\n' (2, 8) (2, 9)
NL '\\n' (3, 4) (3, 5)
INDENT ' ' (4, 0) (4, 4)
NAME 'True' (4, 4) (4, 8)
OP '=' (4, 9) (4, 10)
NAME 'False' (4, 11) (4, 16)
COMMENT '# NEWLINE' (4, 17) (4, 26)
NEWLINE '\\n' (4, 26) (4, 27)
DEDENT '' (5, 0) (5, 0)
""")
indent_error_file = b"""\
def k(x):
x += 2
x += 5
"""
readline = BytesIO(indent_error_file).readline
with self.assertRaisesRegex(IndentationError,
"unindent does not match any "
"outer indentation level"):
for tok in tokenize(readline):
pass
def test_int(self):
# Ordinary integers and binary operators
self.check_tokenize("0xff <= 255", """\
NUMBER '0xff' (1, 0) (1, 4)
OP '<=' (1, 5) (1, 7)
NUMBER '255' (1, 8) (1, 11)
""")
self.check_tokenize("0b10 <= 255", """\
NUMBER '0b10' (1, 0) (1, 4)
OP '<=' (1, 5) (1, 7)
NUMBER '255' (1, 8) (1, 11)
""")
self.check_tokenize("0o123 <= 0O123", """\
NUMBER '0o123' (1, 0) (1, 5)
OP '<=' (1, 6) (1, 8)
NUMBER '0O123' (1, 9) (1, 14)
""")
self.check_tokenize("1234567 > ~0x15", """\
NUMBER '1234567' (1, 0) (1, 7)
OP '>' (1, 8) (1, 9)
OP '~' (1, 10) (1, 11)
NUMBER '0x15' (1, 11) (1, 15)
""")
self.check_tokenize("2134568 != 1231515", """\
NUMBER '2134568' (1, 0) (1, 7)
OP '!=' (1, 8) (1, 10)
NUMBER '1231515' (1, 11) (1, 18)
""")
self.check_tokenize("(-124561-1) & 200000000", """\
OP '(' (1, 0) (1, 1)
OP '-' (1, 1) (1, 2)
NUMBER '124561' (1, 2) (1, 8)
OP '-' (1, 8) (1, 9)
NUMBER '1' (1, 9) (1, 10)
OP ')' (1, 10) (1, 11)
OP '&' (1, 12) (1, 13)
NUMBER '200000000' (1, 14) (1, 23)
""")
self.check_tokenize("0xdeadbeef != -1", """\
NUMBER '0xdeadbeef' (1, 0) (1, 10)
OP '!=' (1, 11) (1, 13)
OP '-' (1, 14) (1, 15)
NUMBER '1' (1, 15) (1, 16)
""")
self.check_tokenize("0xdeadc0de & 12345", """\
NUMBER '0xdeadc0de' (1, 0) (1, 10)
OP '&' (1, 11) (1, 12)
NUMBER '12345' (1, 13) (1, 18)
""")
self.check_tokenize("0xFF & 0x15 | 1234", """\
NUMBER '0xFF' (1, 0) (1, 4)
OP '&' (1, 5) (1, 6)
NUMBER '0x15' (1, 7) (1, 11)
OP '|' (1, 12) (1, 13)
NUMBER '1234' (1, 14) (1, 18)
""")
def test_long(self):
# Long integers
self.check_tokenize("x = 0", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '0' (1, 4) (1, 5)
""")
self.check_tokenize("x = 0xfffffffffff", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '0xfffffffffff' (1, 4) (1, 17)
""")
self.check_tokenize("x = 123141242151251616110", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '123141242151251616110' (1, 4) (1, 25)
""")
self.check_tokenize("x = -15921590215012591", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
OP '-' (1, 4) (1, 5)
NUMBER '15921590215012591' (1, 5) (1, 22)
""")
def test_float(self):
# Floating point numbers
self.check_tokenize("x = 3.14159", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '3.14159' (1, 4) (1, 11)
""")
self.check_tokenize("x = 314159.", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '314159.' (1, 4) (1, 11)
""")
self.check_tokenize("x = .314159", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '.314159' (1, 4) (1, 11)
""")
self.check_tokenize("x = 3e14159", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '3e14159' (1, 4) (1, 11)
""")
self.check_tokenize("x = 3E123", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '3E123' (1, 4) (1, 9)
""")
self.check_tokenize("x+y = 3e-1230", """\
NAME 'x' (1, 0) (1, 1)
OP '+' (1, 1) (1, 2)
NAME 'y' (1, 2) (1, 3)
OP '=' (1, 4) (1, 5)
NUMBER '3e-1230' (1, 6) (1, 13)
""")
self.check_tokenize("x = 3.14e159", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '3.14e159' (1, 4) (1, 12)
""")
def test_underscore_literals(self):
def number_token(s):
f = BytesIO(s.encode('utf-8'))
for toktype, token, start, end, line in tokenize(f.readline):
if toktype == NUMBER:
return token
return 'invalid token'
for lit in VALID_UNDERSCORE_LITERALS:
if '(' in lit:
# this won't work with compound complex inputs
continue
self.assertEqual(number_token(lit), lit)
for lit in INVALID_UNDERSCORE_LITERALS:
self.assertNotEqual(number_token(lit), lit)
def test_string(self):
# String literals
self.check_tokenize("x = ''; y = \"\"", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
STRING "''" (1, 4) (1, 6)
OP ';' (1, 6) (1, 7)
NAME 'y' (1, 8) (1, 9)
OP '=' (1, 10) (1, 11)
STRING '""' (1, 12) (1, 14)
""")
self.check_tokenize("x = '\"'; y = \"'\"", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
STRING '\\'"\\'' (1, 4) (1, 7)
OP ';' (1, 7) (1, 8)
NAME 'y' (1, 9) (1, 10)
OP '=' (1, 11) (1, 12)
STRING '"\\'"' (1, 13) (1, 16)
""")
self.check_tokenize("x = \"doesn't \"shrink\", does it\"", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
STRING '"doesn\\'t "' (1, 4) (1, 14)
NAME 'shrink' (1, 14) (1, 20)
STRING '", does it"' (1, 20) (1, 31)
""")
self.check_tokenize("x = 'abc' + 'ABC'", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
STRING "'abc'" (1, 4) (1, 9)
OP '+' (1, 10) (1, 11)
STRING "'ABC'" (1, 12) (1, 17)
""")
self.check_tokenize('y = "ABC" + "ABC"', """\
NAME 'y' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
STRING '"ABC"' (1, 4) (1, 9)
OP '+' (1, 10) (1, 11)
STRING '"ABC"' (1, 12) (1, 17)
""")
self.check_tokenize("x = r'abc' + r'ABC' + R'ABC' + R'ABC'", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
STRING "r'abc'" (1, 4) (1, 10)
OP '+' (1, 11) (1, 12)
STRING "r'ABC'" (1, 13) (1, 19)
OP '+' (1, 20) (1, 21)
STRING "R'ABC'" (1, 22) (1, 28)
OP '+' (1, 29) (1, 30)
STRING "R'ABC'" (1, 31) (1, 37)
""")
self.check_tokenize('y = r"abc" + r"ABC" + R"ABC" + R"ABC"', """\
NAME 'y' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
STRING 'r"abc"' (1, 4) (1, 10)
OP '+' (1, 11) (1, 12)
STRING 'r"ABC"' (1, 13) (1, 19)
OP '+' (1, 20) (1, 21)
STRING 'R"ABC"' (1, 22) (1, 28)
OP '+' (1, 29) (1, 30)
STRING 'R"ABC"' (1, 31) (1, 37)
""")
self.check_tokenize("u'abc' + U'abc'", """\
STRING "u'abc'" (1, 0) (1, 6)
OP '+' (1, 7) (1, 8)
STRING "U'abc'" (1, 9) (1, 15)
""")
self.check_tokenize('u"abc" + U"abc"', """\
STRING 'u"abc"' (1, 0) (1, 6)
OP '+' (1, 7) (1, 8)
STRING 'U"abc"' (1, 9) (1, 15)
""")
self.check_tokenize("b'abc' + B'abc'", """\
STRING "b'abc'" (1, 0) (1, 6)
OP '+' (1, 7) (1, 8)
STRING "B'abc'" (1, 9) (1, 15)
""")
self.check_tokenize('b"abc" + B"abc"', """\
STRING 'b"abc"' (1, 0) (1, 6)
OP '+' (1, 7) (1, 8)
STRING 'B"abc"' (1, 9) (1, 15)
""")
self.check_tokenize("br'abc' + bR'abc' + Br'abc' + BR'abc'", """\
STRING "br'abc'" (1, 0) (1, 7)
OP '+' (1, 8) (1, 9)
STRING "bR'abc'" (1, 10) (1, 17)
OP '+' (1, 18) (1, 19)
STRING "Br'abc'" (1, 20) (1, 27)
OP '+' (1, 28) (1, 29)
STRING "BR'abc'" (1, 30) (1, 37)
""")
self.check_tokenize('br"abc" + bR"abc" + Br"abc" + BR"abc"', """\
STRING 'br"abc"' (1, 0) (1, 7)
OP '+' (1, 8) (1, 9)
STRING 'bR"abc"' (1, 10) (1, 17)
OP '+' (1, 18) (1, 19)
STRING 'Br"abc"' (1, 20) (1, 27)
OP '+' (1, 28) (1, 29)
STRING 'BR"abc"' (1, 30) (1, 37)
""")
self.check_tokenize("rb'abc' + rB'abc' + Rb'abc' + RB'abc'", """\
STRING "rb'abc'" (1, 0) (1, 7)
OP '+' (1, 8) (1, 9)
STRING "rB'abc'" (1, 10) (1, 17)
OP '+' (1, 18) (1, 19)
STRING "Rb'abc'" (1, 20) (1, 27)
OP '+' (1, 28) (1, 29)
STRING "RB'abc'" (1, 30) (1, 37)
""")
self.check_tokenize('rb"abc" + rB"abc" + Rb"abc" + RB"abc"', """\
STRING 'rb"abc"' (1, 0) (1, 7)
OP '+' (1, 8) (1, 9)
STRING 'rB"abc"' (1, 10) (1, 17)
OP '+' (1, 18) (1, 19)
STRING 'Rb"abc"' (1, 20) (1, 27)
OP '+' (1, 28) (1, 29)
STRING 'RB"abc"' (1, 30) (1, 37)
""")
# Check 0, 1, and 2 character string prefixes.
self.check_tokenize(r'"a\
de\
fg"', """\
STRING '"a\\\\\\nde\\\\\\nfg"\' (1, 0) (3, 3)
""")
self.check_tokenize(r'u"a\
de"', """\
STRING 'u"a\\\\\\nde"\' (1, 0) (2, 3)
""")
self.check_tokenize(r'rb"a\
d"', """\
STRING 'rb"a\\\\\\nd"\' (1, 0) (2, 2)
""")
self.check_tokenize(r'"""a\
b"""', """\
STRING '\"\""a\\\\\\nb\"\""' (1, 0) (2, 4)
""")
self.check_tokenize(r'u"""a\
b"""', """\
STRING 'u\"\""a\\\\\\nb\"\""' (1, 0) (2, 4)
""")
self.check_tokenize(r'rb"""a\
b\
c"""', """\
STRING 'rb"\""a\\\\\\nb\\\\\\nc"\""' (1, 0) (3, 4)
""")
self.check_tokenize('f"abc"', """\
STRING 'f"abc"' (1, 0) (1, 6)
""")
self.check_tokenize('fR"a{b}c"', """\
STRING 'fR"a{b}c"' (1, 0) (1, 9)
""")
self.check_tokenize('f"""abc"""', """\
STRING 'f\"\"\"abc\"\"\"' (1, 0) (1, 10)
""")
self.check_tokenize(r'f"abc\
def"', """\
STRING 'f"abc\\\\\\ndef"' (1, 0) (2, 4)
""")
self.check_tokenize(r'Rf"abc\
def"', """\
STRING 'Rf"abc\\\\\\ndef"' (1, 0) (2, 4)
""")
def test_function(self):
self.check_tokenize("def d22(a, b, c=2, d=2, *k): pass", """\
NAME 'def' (1, 0) (1, 3)
NAME 'd22' (1, 4) (1, 7)
OP '(' (1, 7) (1, 8)
NAME 'a' (1, 8) (1, 9)
OP ',' (1, 9) (1, 10)
NAME 'b' (1, 11) (1, 12)
OP ',' (1, 12) (1, 13)
NAME 'c' (1, 14) (1, 15)
OP '=' (1, 15) (1, 16)
NUMBER '2' (1, 16) (1, 17)
OP ',' (1, 17) (1, 18)
NAME 'd' (1, 19) (1, 20)
OP '=' (1, 20) (1, 21)
NUMBER '2' (1, 21) (1, 22)
OP ',' (1, 22) (1, 23)
OP '*' (1, 24) (1, 25)
NAME 'k' (1, 25) (1, 26)
OP ')' (1, 26) (1, 27)
OP ':' (1, 27) (1, 28)
NAME 'pass' (1, 29) (1, 33)
""")
self.check_tokenize("def d01v_(a=1, *k, **w): pass", """\
NAME 'def' (1, 0) (1, 3)
NAME 'd01v_' (1, 4) (1, 9)
OP '(' (1, 9) (1, 10)
NAME 'a' (1, 10) (1, 11)
OP '=' (1, 11) (1, 12)
NUMBER '1' (1, 12) (1, 13)
OP ',' (1, 13) (1, 14)
OP '*' (1, 15) (1, 16)
NAME 'k' (1, 16) (1, 17)
OP ',' (1, 17) (1, 18)
OP '**' (1, 19) (1, 21)
NAME 'w' (1, 21) (1, 22)
OP ')' (1, 22) (1, 23)
OP ':' (1, 23) (1, 24)
NAME 'pass' (1, 25) (1, 29)
""")
self.check_tokenize("def d23(a: str, b: int=3) -> int: pass", """\
NAME 'def' (1, 0) (1, 3)
NAME 'd23' (1, 4) (1, 7)
OP '(' (1, 7) (1, 8)
NAME 'a' (1, 8) (1, 9)
OP ':' (1, 9) (1, 10)
NAME 'str' (1, 11) (1, 14)
OP ',' (1, 14) (1, 15)
NAME 'b' (1, 16) (1, 17)
OP ':' (1, 17) (1, 18)
NAME 'int' (1, 19) (1, 22)
OP '=' (1, 22) (1, 23)
NUMBER '3' (1, 23) (1, 24)
OP ')' (1, 24) (1, 25)
OP '->' (1, 26) (1, 28)
NAME 'int' (1, 29) (1, 32)
OP ':' (1, 32) (1, 33)
NAME 'pass' (1, 34) (1, 38)
""")
def test_comparison(self):
# Comparison
self.check_tokenize("if 1 < 1 > 1 == 1 >= 5 <= 0x15 <= 0x12 != "
"1 and 5 in 1 not in 1 is 1 or 5 is not 1: pass", """\
NAME 'if' (1, 0) (1, 2)
NUMBER '1' (1, 3) (1, 4)
OP '<' (1, 5) (1, 6)
NUMBER '1' (1, 7) (1, 8)
OP '>' (1, 9) (1, 10)
NUMBER '1' (1, 11) (1, 12)
OP '==' (1, 13) (1, 15)
NUMBER '1' (1, 16) (1, 17)
OP '>=' (1, 18) (1, 20)
NUMBER '5' (1, 21) (1, 22)
OP '<=' (1, 23) (1, 25)
NUMBER '0x15' (1, 26) (1, 30)
OP '<=' (1, 31) (1, 33)
NUMBER '0x12' (1, 34) (1, 38)
OP '!=' (1, 39) (1, 41)
NUMBER '1' (1, 42) (1, 43)
NAME 'and' (1, 44) (1, 47)
NUMBER '5' (1, 48) (1, 49)
NAME 'in' (1, 50) (1, 52)
NUMBER '1' (1, 53) (1, 54)
NAME 'not' (1, 55) (1, 58)
NAME 'in' (1, 59) (1, 61)
NUMBER '1' (1, 62) (1, 63)
NAME 'is' (1, 64) (1, 66)
NUMBER '1' (1, 67) (1, 68)
NAME 'or' (1, 69) (1, 71)
NUMBER '5' (1, 72) (1, 73)
NAME 'is' (1, 74) (1, 76)
NAME 'not' (1, 77) (1, 80)
NUMBER '1' (1, 81) (1, 82)
OP ':' (1, 82) (1, 83)
NAME 'pass' (1, 84) (1, 88)
""")
def test_shift(self):
# Shift
self.check_tokenize("x = 1 << 1 >> 5", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '1' (1, 4) (1, 5)
OP '<<' (1, 6) (1, 8)
NUMBER '1' (1, 9) (1, 10)
OP '>>' (1, 11) (1, 13)
NUMBER '5' (1, 14) (1, 15)
""")
def test_additive(self):
# Additive
self.check_tokenize("x = 1 - y + 15 - 1 + 0x124 + z + a[5]", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '1' (1, 4) (1, 5)
OP '-' (1, 6) (1, 7)
NAME 'y' (1, 8) (1, 9)
OP '+' (1, 10) (1, 11)
NUMBER '15' (1, 12) (1, 14)
OP '-' (1, 15) (1, 16)
NUMBER '1' (1, 17) (1, 18)
OP '+' (1, 19) (1, 20)
NUMBER '0x124' (1, 21) (1, 26)
OP '+' (1, 27) (1, 28)
NAME 'z' (1, 29) (1, 30)
OP '+' (1, 31) (1, 32)
NAME 'a' (1, 33) (1, 34)
OP '[' (1, 34) (1, 35)
NUMBER '5' (1, 35) (1, 36)
OP ']' (1, 36) (1, 37)
""")
def test_multiplicative(self):
# Multiplicative
self.check_tokenize("x = 1//1*1/5*12%0x12@42", """\
NAME 'x' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
NUMBER '1' (1, 4) (1, 5)
OP '//' (1, 5) (1, 7)
NUMBER '1' (1, 7) (1, 8)
OP '*' (1, 8) (1, 9)
NUMBER '1' (1, 9) (1, 10)
OP '/' (1, 10) (1, 11)
NUMBER '5' (1, 11) (1, 12)
OP '*' (1, 12) (1, 13)
NUMBER '12' (1, 13) (1, 15)
OP '%' (1, 15) (1, 16)
NUMBER '0x12' (1, 16) (1, 20)
OP '@' (1, 20) (1, 21)
NUMBER '42' (1, 21) (1, 23)
""")
def test_unary(self):
# Unary
self.check_tokenize("~1 ^ 1 & 1 |1 ^ -1", """\
OP '~' (1, 0) (1, 1)
NUMBER '1' (1, 1) (1, 2)
OP '^' (1, 3) (1, 4)
NUMBER '1' (1, 5) (1, 6)
OP '&' (1, 7) (1, 8)
NUMBER '1' (1, 9) (1, 10)
OP '|' (1, 11) (1, 12)
NUMBER '1' (1, 12) (1, 13)
OP '^' (1, 14) (1, 15)
OP '-' (1, 16) (1, 17)
NUMBER '1' (1, 17) (1, 18)
""")
self.check_tokenize("-1*1/1+1*1//1 - ---1**1", """\
OP '-' (1, 0) (1, 1)
NUMBER '1' (1, 1) (1, 2)
OP '*' (1, 2) (1, 3)
NUMBER '1' (1, 3) (1, 4)
OP '/' (1, 4) (1, 5)
NUMBER '1' (1, 5) (1, 6)
OP '+' (1, 6) (1, 7)
NUMBER '1' (1, 7) (1, 8)
OP '*' (1, 8) (1, 9)
NUMBER '1' (1, 9) (1, 10)
OP '//' (1, 10) (1, 12)
NUMBER '1' (1, 12) (1, 13)
OP '-' (1, 14) (1, 15)
OP '-' (1, 16) (1, 17)
OP '-' (1, 17) (1, 18)
OP '-' (1, 18) (1, 19)
NUMBER '1' (1, 19) (1, 20)
OP '**' (1, 20) (1, 22)
NUMBER '1' (1, 22) (1, 23)
""")
def test_selector(self):
# Selector
self.check_tokenize("import sys, time\nx = sys.modules['time'].time()", """\
NAME 'import' (1, 0) (1, 6)
NAME 'sys' (1, 7) (1, 10)
OP ',' (1, 10) (1, 11)
NAME 'time' (1, 12) (1, 16)
NEWLINE '\\n' (1, 16) (1, 17)
NAME 'x' (2, 0) (2, 1)
OP '=' (2, 2) (2, 3)
NAME 'sys' (2, 4) (2, 7)
OP '.' (2, 7) (2, 8)
NAME 'modules' (2, 8) (2, 15)
OP '[' (2, 15) (2, 16)
STRING "'time'" (2, 16) (2, 22)
OP ']' (2, 22) (2, 23)
OP '.' (2, 23) (2, 24)
NAME 'time' (2, 24) (2, 28)
OP '(' (2, 28) (2, 29)
OP ')' (2, 29) (2, 30)
""")
def test_method(self):
# Methods
self.check_tokenize("@staticmethod\ndef foo(x,y): pass", """\
OP '@' (1, 0) (1, 1)
NAME 'staticmethod' (1, 1) (1, 13)
NEWLINE '\\n' (1, 13) (1, 14)
NAME 'def' (2, 0) (2, 3)
NAME 'foo' (2, 4) (2, 7)
OP '(' (2, 7) (2, 8)
NAME 'x' (2, 8) (2, 9)
OP ',' (2, 9) (2, 10)
NAME 'y' (2, 10) (2, 11)
OP ')' (2, 11) (2, 12)
OP ':' (2, 12) (2, 13)
NAME 'pass' (2, 14) (2, 18)
""")
def test_tabs(self):
# Evil tabs
self.check_tokenize("def f():\n"
"\tif x\n"
" \tpass", """\
NAME 'def' (1, 0) (1, 3)
NAME 'f' (1, 4) (1, 5)
OP '(' (1, 5) (1, 6)
OP ')' (1, 6) (1, 7)
OP ':' (1, 7) (1, 8)
NEWLINE '\\n' (1, 8) (1, 9)
INDENT '\\t' (2, 0) (2, 1)
NAME 'if' (2, 1) (2, 3)
NAME 'x' (2, 4) (2, 5)
NEWLINE '\\n' (2, 5) (2, 6)
INDENT ' \\t' (3, 0) (3, 9)
NAME 'pass' (3, 9) (3, 13)
DEDENT '' (4, 0) (4, 0)
DEDENT '' (4, 0) (4, 0)
""")
def test_non_ascii_identifiers(self):
# Non-ascii identifiers
self.check_tokenize("Örter = 'places'\ngrün = 'green'", """\
NAME 'Örter' (1, 0) (1, 5)
OP '=' (1, 6) (1, 7)
STRING "'places'" (1, 8) (1, 16)
NEWLINE '\\n' (1, 16) (1, 17)
NAME 'grün' (2, 0) (2, 4)
OP '=' (2, 5) (2, 6)
STRING "'green'" (2, 7) (2, 14)
""")
def test_unicode(self):
# Legacy unicode literals:
self.check_tokenize("Örter = u'places'\ngrün = U'green'", """\
NAME 'Örter' (1, 0) (1, 5)
OP '=' (1, 6) (1, 7)
STRING "u'places'" (1, 8) (1, 17)
NEWLINE '\\n' (1, 17) (1, 18)
NAME 'grün' (2, 0) (2, 4)
OP '=' (2, 5) (2, 6)
STRING "U'green'" (2, 7) (2, 15)
""")
def test_async(self):
# Async/await extension:
self.check_tokenize("async = 1", """\
NAME 'async' (1, 0) (1, 5)
OP '=' (1, 6) (1, 7)
NUMBER '1' (1, 8) (1, 9)
""")
self.check_tokenize("a = (async = 1)", """\
NAME 'a' (1, 0) (1, 1)
OP '=' (1, 2) (1, 3)
OP '(' (1, 4) (1, 5)
NAME 'async' (1, 5) (1, 10)
OP '=' (1, 11) (1, 12)
NUMBER '1' (1, 13) (1, 14)
OP ')' (1, 14) (1, 15)
""")
self.check_tokenize("async()", """\
NAME 'async' (1, 0) (1, 5)
OP '(' (1, 5) (1, 6)
OP ')' (1, 6) (1, 7)
""")
self.check_tokenize("class async(Bar):pass", """\
NAME 'class' (1, 0) (1, 5)
NAME 'async' (1, 6) (1, 11)
OP '(' (1, 11) (1, 12)
NAME 'Bar' (1, 12) (1, 15)
OP ')' (1, 15) (1, 16)
OP ':' (1, 16) (1, 17)
NAME 'pass' (1, 17) (1, 21)
""")
self.check_tokenize("class async:pass", """\
NAME 'class' (1, 0) (1, 5)
NAME 'async' (1, 6) (1, 11)
OP ':' (1, 11) (1, 12)
NAME 'pass' (1, 12) (1, 16)
""")
self.check_tokenize("await = 1", """\
NAME 'await' (1, 0) (1, 5)
OP '=' (1, 6) (1, 7)
NUMBER '1' (1, 8) (1, 9)
""")
self.check_tokenize("foo.async", """\
NAME 'foo' (1, 0) (1, 3)
OP '.' (1, 3) (1, 4)
NAME 'async' (1, 4) (1, 9)
""")
self.check_tokenize("async for a in b: pass", """\
NAME 'async' (1, 0) (1, 5)
NAME 'for' (1, 6) (1, 9)
NAME 'a' (1, 10) (1, 11)
NAME 'in' (1, 12) (1, 14)
NAME 'b' (1, 15) (1, 16)
OP ':' (1, 16) (1, 17)
NAME 'pass' (1, 18) (1, 22)
""")
self.check_tokenize("async with a as b: pass", """\
NAME 'async' (1, 0) (1, 5)
NAME 'with' (1, 6) (1, 10)
NAME 'a' (1, 11) (1, 12)
NAME 'as' (1, 13) (1, 15)
NAME 'b' (1, 16) (1, 17)
OP ':' (1, 17) (1, 18)
NAME 'pass' (1, 19) (1, 23)
""")
self.check_tokenize("async.foo", """\
NAME 'async' (1, 0) (1, 5)
OP '.' (1, 5) (1, 6)
NAME 'foo' (1, 6) (1, 9)
""")
self.check_tokenize("async", """\
NAME 'async' (1, 0) (1, 5)
""")
self.check_tokenize("async\n#comment\nawait", """\
NAME 'async' (1, 0) (1, 5)
NEWLINE '\\n' (1, 5) (1, 6)
COMMENT '#comment' (2, 0) (2, 8)
NL '\\n' (2, 8) (2, 9)
NAME 'await' (3, 0) (3, 5)
""")
self.check_tokenize("async\n...\nawait", """\
NAME 'async' (1, 0) (1, 5)
NEWLINE '\\n' (1, 5) (1, 6)
OP '...' (2, 0) (2, 3)
NEWLINE '\\n' (2, 3) (2, 4)
NAME 'await' (3, 0) (3, 5)
""")
self.check_tokenize("async\nawait", """\
NAME 'async' (1, 0) (1, 5)
NEWLINE '\\n' (1, 5) (1, 6)
NAME 'await' (2, 0) (2, 5)
""")
self.check_tokenize("foo.async + 1", """\
NAME 'foo' (1, 0) (1, 3)
OP '.' (1, 3) (1, 4)
NAME 'async' (1, 4) (1, 9)
OP '+' (1, 10) (1, 11)
NUMBER '1' (1, 12) (1, 13)
""")
self.check_tokenize("async def foo(): pass", """\
NAME 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'foo' (1, 10) (1, 13)
OP '(' (1, 13) (1, 14)
OP ')' (1, 14) (1, 15)
OP ':' (1, 15) (1, 16)
NAME 'pass' (1, 17) (1, 21)
""")
self.check_tokenize('''\
async def foo():
def foo(await):
await = 1
if 1:
await
async += 1
''', """\
NAME 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'foo' (1, 10) (1, 13)
OP '(' (1, 13) (1, 14)
OP ')' (1, 14) (1, 15)
OP ':' (1, 15) (1, 16)
NEWLINE '\\n' (1, 16) (1, 17)
INDENT ' ' (2, 0) (2, 2)
NAME 'def' (2, 2) (2, 5)
NAME 'foo' (2, 6) (2, 9)
OP '(' (2, 9) (2, 10)
NAME 'await' (2, 10) (2, 15)
OP ')' (2, 15) (2, 16)
OP ':' (2, 16) (2, 17)
NEWLINE '\\n' (2, 17) (2, 18)
INDENT ' ' (3, 0) (3, 4)
NAME 'await' (3, 4) (3, 9)
OP '=' (3, 10) (3, 11)
NUMBER '1' (3, 12) (3, 13)
NEWLINE '\\n' (3, 13) (3, 14)
DEDENT '' (4, 2) (4, 2)
NAME 'if' (4, 2) (4, 4)
NUMBER '1' (4, 5) (4, 6)
OP ':' (4, 6) (4, 7)
NEWLINE '\\n' (4, 7) (4, 8)
INDENT ' ' (5, 0) (5, 4)
NAME 'await' (5, 4) (5, 9)
NEWLINE '\\n' (5, 9) (5, 10)
DEDENT '' (6, 0) (6, 0)
DEDENT '' (6, 0) (6, 0)
NAME 'async' (6, 0) (6, 5)
OP '+=' (6, 6) (6, 8)
NUMBER '1' (6, 9) (6, 10)
NEWLINE '\\n' (6, 10) (6, 11)
""")
self.check_tokenize('''\
async def foo():
async for i in 1: pass''', """\
NAME 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'foo' (1, 10) (1, 13)
OP '(' (1, 13) (1, 14)
OP ')' (1, 14) (1, 15)
OP ':' (1, 15) (1, 16)
NEWLINE '\\n' (1, 16) (1, 17)
INDENT ' ' (2, 0) (2, 2)
NAME 'async' (2, 2) (2, 7)
NAME 'for' (2, 8) (2, 11)
NAME 'i' (2, 12) (2, 13)
NAME 'in' (2, 14) (2, 16)
NUMBER '1' (2, 17) (2, 18)
OP ':' (2, 18) (2, 19)
NAME 'pass' (2, 20) (2, 24)
DEDENT '' (3, 0) (3, 0)
""")
self.check_tokenize('''async def foo(async): await''', """\
NAME 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'foo' (1, 10) (1, 13)
OP '(' (1, 13) (1, 14)
NAME 'async' (1, 14) (1, 19)
OP ')' (1, 19) (1, 20)
OP ':' (1, 20) (1, 21)
NAME 'await' (1, 22) (1, 27)
""")
self.check_tokenize('''\
def f():
def baz(): pass
async def bar(): pass
await = 2''', """\
NAME 'def' (1, 0) (1, 3)
NAME 'f' (1, 4) (1, 5)
OP '(' (1, 5) (1, 6)
OP ')' (1, 6) (1, 7)
OP ':' (1, 7) (1, 8)
NEWLINE '\\n' (1, 8) (1, 9)
NL '\\n' (2, 0) (2, 1)
INDENT ' ' (3, 0) (3, 2)
NAME 'def' (3, 2) (3, 5)
NAME 'baz' (3, 6) (3, 9)
OP '(' (3, 9) (3, 10)
OP ')' (3, 10) (3, 11)
OP ':' (3, 11) (3, 12)
NAME 'pass' (3, 13) (3, 17)
NEWLINE '\\n' (3, 17) (3, 18)
NAME 'async' (4, 2) (4, 7)
NAME 'def' (4, 8) (4, 11)
NAME 'bar' (4, 12) (4, 15)
OP '(' (4, 15) (4, 16)
OP ')' (4, 16) (4, 17)
OP ':' (4, 17) (4, 18)
NAME 'pass' (4, 19) (4, 23)
NEWLINE '\\n' (4, 23) (4, 24)
NL '\\n' (5, 0) (5, 1)
NAME 'await' (6, 2) (6, 7)
OP '=' (6, 8) (6, 9)
NUMBER '2' (6, 10) (6, 11)
DEDENT '' (7, 0) (7, 0)
""")
self.check_tokenize('''\
async def f():
def baz(): pass
async def bar(): pass
await = 2''', """\
NAME 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'f' (1, 10) (1, 11)
OP '(' (1, 11) (1, 12)
OP ')' (1, 12) (1, 13)
OP ':' (1, 13) (1, 14)
NEWLINE '\\n' (1, 14) (1, 15)
NL '\\n' (2, 0) (2, 1)
INDENT ' ' (3, 0) (3, 2)
NAME 'def' (3, 2) (3, 5)
NAME 'baz' (3, 6) (3, 9)
OP '(' (3, 9) (3, 10)
OP ')' (3, 10) (3, 11)
OP ':' (3, 11) (3, 12)
NAME 'pass' (3, 13) (3, 17)
NEWLINE '\\n' (3, 17) (3, 18)
NAME 'async' (4, 2) (4, 7)
NAME 'def' (4, 8) (4, 11)
NAME 'bar' (4, 12) (4, 15)
OP '(' (4, 15) (4, 16)
OP ')' (4, 16) (4, 17)
OP ':' (4, 17) (4, 18)
NAME 'pass' (4, 19) (4, 23)
NEWLINE '\\n' (4, 23) (4, 24)
NL '\\n' (5, 0) (5, 1)
NAME 'await' (6, 2) (6, 7)
OP '=' (6, 8) (6, 9)
NUMBER '2' (6, 10) (6, 11)
DEDENT '' (7, 0) (7, 0)
""")
class GenerateTokensTest(TokenizeTest):
def check_tokenize(self, s, expected):
# Format the tokens in s in a table format.
# The ENDMARKER and final NEWLINE are omitted.
f = StringIO(s)
result = stringify_tokens_from_source(generate_tokens(f.readline), s)
self.assertEqual(result, expected.rstrip().splitlines())
def decistmt(s):
result = []
g = tokenize(BytesIO(s.encode('utf-8')).readline) # tokenize the string
for toknum, tokval, _, _, _ in g:
if toknum == NUMBER and '.' in tokval: # replace NUMBER tokens
result.extend([
(NAME, 'Decimal'),
(OP, '('),
(STRING, repr(tokval)),
(OP, ')')
])
else:
result.append((toknum, tokval))
return untokenize(result).decode('utf-8')
class TestMisc(TestCase):
def test_decistmt(self):
# Substitute Decimals for floats in a string of statements.
# This is an example from the docs.
from decimal import Decimal
s = '+21.3e-5*-.1234/81.7'
self.assertEqual(decistmt(s),
"+Decimal ('21.3e-5')*-Decimal ('.1234')/Decimal ('81.7')")
# The format of the exponent is inherited from the platform C library.
# Known cases are "e-007" (Windows) and "e-07" (not Windows). Since
# we're only showing 11 digits, and the 12th isn't close to 5, the
# rest of the output should be platform-independent.
self.assertRegex(repr(eval(s)), '-3.2171603427[0-9]*e-0+7')
# Output from calculations with Decimal should be identical across all
# platforms.
self.assertEqual(eval(decistmt(s)),
Decimal('-3.217160342717258261933904529E-7'))
class TestTokenizerAdheresToPep0263(TestCase):
"""
Test that tokenizer adheres to the coding behaviour stipulated in PEP 0263.
"""
def _testFile(self, filename):
path = os.path.join(os.path.dirname(__file__), filename)
TestRoundtrip.check_roundtrip(self, open(path, 'rb'))
def test_utf8_coding_cookie_and_no_utf8_bom(self):
f = 'tokenize_tests-utf8-coding-cookie-and-no-utf8-bom-sig.txt'
self._testFile(f)
def test_latin1_coding_cookie_and_utf8_bom(self):
"""
As per PEP 0263, if a file starts with a utf-8 BOM signature, the only
allowed encoding for the comment is 'utf-8'. The text file used in
this test starts with a BOM signature, but specifies latin1 as the
coding, so verify that a SyntaxError is raised, which matches the
behaviour of the interpreter when it encounters a similar condition.
"""
f = 'tokenize_tests-latin1-coding-cookie-and-utf8-bom-sig.txt'
self.assertRaises(SyntaxError, self._testFile, f)
def test_no_coding_cookie_and_utf8_bom(self):
f = 'tokenize_tests-no-coding-cookie-and-utf8-bom-sig-only.txt'
self._testFile(f)
def test_utf8_coding_cookie_and_utf8_bom(self):
f = 'tokenize_tests-utf8-coding-cookie-and-utf8-bom-sig.txt'
self._testFile(f)
def test_bad_coding_cookie(self):
self.assertRaises(SyntaxError, self._testFile, 'bad_coding.py')
self.assertRaises(SyntaxError, self._testFile, 'bad_coding2.py')
class Test_Tokenize(TestCase):
def test__tokenize_decodes_with_specified_encoding(self):
literal = '"ЉЊЈЁЂ"'
line = literal.encode('utf-8')
first = False
def readline():
nonlocal first
if not first:
first = True
return line
else:
return b''
# skip the initial encoding token and the end tokens
tokens = list(_tokenize(readline, encoding='utf-8'))[1:-2]
expected_tokens = [(3, '"ЉЊЈЁЂ"', (1, 0), (1, 7), '"ЉЊЈЁЂ"')]
self.assertEqual(tokens, expected_tokens,
"bytes not decoded with encoding")
def test__tokenize_does_not_decode_with_encoding_none(self):
literal = '"ЉЊЈЁЂ"'
first = False
def readline():
nonlocal first
if not first:
first = True
return literal
else:
return b''
# skip the end tokens
tokens = list(_tokenize(readline, encoding=None))[:-2]
expected_tokens = [(3, '"ЉЊЈЁЂ"', (1, 0), (1, 7), '"ЉЊЈЁЂ"')]
self.assertEqual(tokens, expected_tokens,
"string not tokenized when encoding is None")
class TestDetectEncoding(TestCase):
def get_readline(self, lines):
index = 0
def readline():
nonlocal index
if index == len(lines):
raise StopIteration
line = lines[index]
index += 1
return line
return readline
def test_no_bom_no_encoding_cookie(self):
lines = (
b'# something\n',
b'print(something)\n',
b'do_something(else)\n'
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'utf-8')
self.assertEqual(consumed_lines, list(lines[:2]))
def test_bom_no_cookie(self):
lines = (
b'\xef\xbb\xbf# something\n',
b'print(something)\n',
b'do_something(else)\n'
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'utf-8-sig')
self.assertEqual(consumed_lines,
[b'# something\n', b'print(something)\n'])
def test_cookie_first_line_no_bom(self):
lines = (
b'# -*- coding: latin-1 -*-\n',
b'print(something)\n',
b'do_something(else)\n'
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'iso-8859-1')
self.assertEqual(consumed_lines, [b'# -*- coding: latin-1 -*-\n'])
def test_matched_bom_and_cookie_first_line(self):
lines = (
b'\xef\xbb\xbf# coding=utf-8\n',
b'print(something)\n',
b'do_something(else)\n'
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'utf-8-sig')
self.assertEqual(consumed_lines, [b'# coding=utf-8\n'])
def test_mismatched_bom_and_cookie_first_line_raises_syntaxerror(self):
lines = (
b'\xef\xbb\xbf# vim: set fileencoding=ascii :\n',
b'print(something)\n',
b'do_something(else)\n'
)
readline = self.get_readline(lines)
self.assertRaises(SyntaxError, detect_encoding, readline)
def test_cookie_second_line_no_bom(self):
lines = (
b'#! something\n',
b'# vim: set fileencoding=ascii :\n',
b'print(something)\n',
b'do_something(else)\n'
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'ascii')
expected = [b'#! something\n', b'# vim: set fileencoding=ascii :\n']
self.assertEqual(consumed_lines, expected)
def test_matched_bom_and_cookie_second_line(self):
lines = (
b'\xef\xbb\xbf#! something\n',
b'f# coding=utf-8\n',
b'print(something)\n',
b'do_something(else)\n'
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'utf-8-sig')
self.assertEqual(consumed_lines,
[b'#! something\n', b'f# coding=utf-8\n'])
def test_mismatched_bom_and_cookie_second_line_raises_syntaxerror(self):
lines = (
b'\xef\xbb\xbf#! something\n',
b'# vim: set fileencoding=ascii :\n',
b'print(something)\n',
b'do_something(else)\n'
)
readline = self.get_readline(lines)
self.assertRaises(SyntaxError, detect_encoding, readline)
def test_cookie_second_line_noncommented_first_line(self):
lines = (
b"print('\xc2\xa3')\n",
b'# vim: set fileencoding=iso8859-15 :\n',
b"print('\xe2\x82\xac')\n"
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'utf-8')
expected = [b"print('\xc2\xa3')\n"]
self.assertEqual(consumed_lines, expected)
def test_cookie_second_line_commented_first_line(self):
lines = (
b"#print('\xc2\xa3')\n",
b'# vim: set fileencoding=iso8859-15 :\n',
b"print('\xe2\x82\xac')\n"
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'iso8859-15')
expected = [b"#print('\xc2\xa3')\n", b'# vim: set fileencoding=iso8859-15 :\n']
self.assertEqual(consumed_lines, expected)
def test_cookie_second_line_empty_first_line(self):
lines = (
b'\n',
b'# vim: set fileencoding=iso8859-15 :\n',
b"print('\xe2\x82\xac')\n"
)
encoding, consumed_lines = detect_encoding(self.get_readline(lines))
self.assertEqual(encoding, 'iso8859-15')
expected = [b'\n', b'# vim: set fileencoding=iso8859-15 :\n']
self.assertEqual(consumed_lines, expected)
def test_latin1_normalization(self):
# See get_normal_name() in tokenizer.c.
encodings = ("latin-1", "iso-8859-1", "iso-latin-1", "latin-1-unix",
"iso-8859-1-unix", "iso-latin-1-mac")
for encoding in encodings:
for rep in ("-", "_"):
enc = encoding.replace("-", rep)
lines = (b"#!/usr/bin/python\n",
b"# coding: " + enc.encode("ascii") + b"\n",
b"print(things)\n",
b"do_something += 4\n")
rl = self.get_readline(lines)
found, consumed_lines = detect_encoding(rl)
self.assertEqual(found, "iso-8859-1")
def test_syntaxerror_latin1(self):
# Issue 14629: need to raise SyntaxError if the first
# line(s) have non-UTF-8 characters
lines = (
b'print("\xdf")', # Latin-1: LATIN SMALL LETTER SHARP S
)
readline = self.get_readline(lines)
self.assertRaises(SyntaxError, detect_encoding, readline)
def test_utf8_normalization(self):
# See get_normal_name() in tokenizer.c.
encodings = ("utf-8", "utf-8-mac", "utf-8-unix")
for encoding in encodings:
for rep in ("-", "_"):
enc = encoding.replace("-", rep)
lines = (b"#!/usr/bin/python\n",
b"# coding: " + enc.encode("ascii") + b"\n",
b"1 + 3\n")
rl = self.get_readline(lines)
found, consumed_lines = detect_encoding(rl)
self.assertEqual(found, "utf-8")
def test_short_files(self):
readline = self.get_readline((b'print(something)\n',))
encoding, consumed_lines = detect_encoding(readline)
self.assertEqual(encoding, 'utf-8')
self.assertEqual(consumed_lines, [b'print(something)\n'])
encoding, consumed_lines = detect_encoding(self.get_readline(()))
self.assertEqual(encoding, 'utf-8')
self.assertEqual(consumed_lines, [])
readline = self.get_readline((b'\xef\xbb\xbfprint(something)\n',))
encoding, consumed_lines = detect_encoding(readline)
self.assertEqual(encoding, 'utf-8-sig')
self.assertEqual(consumed_lines, [b'print(something)\n'])
readline = self.get_readline((b'\xef\xbb\xbf',))
encoding, consumed_lines = detect_encoding(readline)
self.assertEqual(encoding, 'utf-8-sig')
self.assertEqual(consumed_lines, [])
readline = self.get_readline((b'# coding: bad\n',))
self.assertRaises(SyntaxError, detect_encoding, readline)
def test_false_encoding(self):
# Issue 18873: "Encoding" detected in non-comment lines
readline = self.get_readline((b'print("#coding=fake")',))
encoding, consumed_lines = detect_encoding(readline)
self.assertEqual(encoding, 'utf-8')
self.assertEqual(consumed_lines, [b'print("#coding=fake")'])
def test_open(self):
filename = os_helper.TESTFN + '.py'
self.addCleanup(os_helper.unlink, filename)
# test coding cookie
for encoding in ('iso-8859-15', 'utf-8'):
with open(filename, 'w', encoding=encoding) as fp:
print("# coding: %s" % encoding, file=fp)
print("print('euro:\u20ac')", file=fp)
with tokenize_open(filename) as fp:
self.assertEqual(fp.encoding, encoding)
self.assertEqual(fp.mode, 'r')
# test BOM (no coding cookie)
with open(filename, 'w', encoding='utf-8-sig') as fp:
print("print('euro:\u20ac')", file=fp)
with tokenize_open(filename) as fp:
self.assertEqual(fp.encoding, 'utf-8-sig')
self.assertEqual(fp.mode, 'r')
def test_filename_in_exception(self):
# When possible, include the file name in the exception.
path = 'some_file_path'
lines = (
b'print("\xdf")', # Latin-1: LATIN SMALL LETTER SHARP S
)
class Bunk:
def __init__(self, lines, path):
self.name = path
self._lines = lines
self._index = 0
def readline(self):
if self._index == len(lines):
raise StopIteration
line = lines[self._index]
self._index += 1
return line
with self.assertRaises(SyntaxError):
ins = Bunk(lines, path)
# Make sure lacking a name isn't an issue.
del ins.name
detect_encoding(ins.readline)
with self.assertRaisesRegex(SyntaxError, '.*{}'.format(path)):
ins = Bunk(lines, path)
detect_encoding(ins.readline)
def test_open_error(self):
# Issue #23840: open() must close the binary file on error
m = BytesIO(b'#coding:xxx')
with mock.patch('tokenize._builtin_open', return_value=m):
self.assertRaises(SyntaxError, tokenize_open, 'foobar')
self.assertTrue(m.closed)
class TestTokenize(TestCase):
def test_tokenize(self):
import tokenize as tokenize_module
encoding = object()
encoding_used = None
def mock_detect_encoding(readline):
return encoding, [b'first', b'second']
def mock__tokenize(readline, encoding):
nonlocal encoding_used
encoding_used = encoding
out = []
while True:
next_line = readline()
if next_line:
out.append(next_line)
continue
return out
counter = 0
def mock_readline():
nonlocal counter
counter += 1
if counter == 5:
return b''
return str(counter).encode()
orig_detect_encoding = tokenize_module.detect_encoding
orig__tokenize = tokenize_module._tokenize
tokenize_module.detect_encoding = mock_detect_encoding
tokenize_module._tokenize = mock__tokenize
try:
results = tokenize(mock_readline)
self.assertEqual(list(results),
[b'first', b'second', b'1', b'2', b'3', b'4'])
finally:
tokenize_module.detect_encoding = orig_detect_encoding
tokenize_module._tokenize = orig__tokenize
self.assertEqual(encoding_used, encoding)
def test_oneline_defs(self):
buf = []
for i in range(500):
buf.append('def i{i}(): return {i}'.format(i=i))
buf.append('OK')
buf = '\n'.join(buf)
# Test that 500 consequent, one-line defs is OK
toks = list(tokenize(BytesIO(buf.encode('utf-8')).readline))
self.assertEqual(toks[-3].string, 'OK') # [-1] is always ENDMARKER
# [-2] is always NEWLINE
def assertExactTypeEqual(self, opstr, *optypes):
tokens = list(tokenize(BytesIO(opstr.encode('utf-8')).readline))
num_optypes = len(optypes)
self.assertEqual(len(tokens), 3 + num_optypes)
self.assertEqual(tok_name[tokens[0].exact_type],
tok_name[ENCODING])
for i in range(num_optypes):
self.assertEqual(tok_name[tokens[i + 1].exact_type],
tok_name[optypes[i]])
self.assertEqual(tok_name[tokens[1 + num_optypes].exact_type],
tok_name[token.NEWLINE])
self.assertEqual(tok_name[tokens[2 + num_optypes].exact_type],
tok_name[token.ENDMARKER])
def test_exact_type(self):
self.assertExactTypeEqual('()', token.LPAR, token.RPAR)
self.assertExactTypeEqual('[]', token.LSQB, token.RSQB)
self.assertExactTypeEqual(':', token.COLON)
self.assertExactTypeEqual(',', token.COMMA)
self.assertExactTypeEqual(';', token.SEMI)
self.assertExactTypeEqual('+', token.PLUS)
self.assertExactTypeEqual('-', token.MINUS)
self.assertExactTypeEqual('*', token.STAR)
self.assertExactTypeEqual('/', token.SLASH)
self.assertExactTypeEqual('|', token.VBAR)
self.assertExactTypeEqual('&', token.AMPER)
self.assertExactTypeEqual('<', token.LESS)
self.assertExactTypeEqual('>', token.GREATER)
self.assertExactTypeEqual('=', token.EQUAL)
self.assertExactTypeEqual('.', token.DOT)
self.assertExactTypeEqual('%', token.PERCENT)
self.assertExactTypeEqual('{}', token.LBRACE, token.RBRACE)
self.assertExactTypeEqual('==', token.EQEQUAL)
self.assertExactTypeEqual('!=', token.NOTEQUAL)
self.assertExactTypeEqual('<=', token.LESSEQUAL)
self.assertExactTypeEqual('>=', token.GREATEREQUAL)
self.assertExactTypeEqual('~', token.TILDE)
self.assertExactTypeEqual('^', token.CIRCUMFLEX)
self.assertExactTypeEqual('<<', token.LEFTSHIFT)
self.assertExactTypeEqual('>>', token.RIGHTSHIFT)
self.assertExactTypeEqual('**', token.DOUBLESTAR)
self.assertExactTypeEqual('+=', token.PLUSEQUAL)
self.assertExactTypeEqual('-=', token.MINEQUAL)
self.assertExactTypeEqual('*=', token.STAREQUAL)
self.assertExactTypeEqual('/=', token.SLASHEQUAL)
self.assertExactTypeEqual('%=', token.PERCENTEQUAL)
self.assertExactTypeEqual('&=', token.AMPEREQUAL)
self.assertExactTypeEqual('|=', token.VBAREQUAL)
self.assertExactTypeEqual('^=', token.CIRCUMFLEXEQUAL)
self.assertExactTypeEqual('^=', token.CIRCUMFLEXEQUAL)
self.assertExactTypeEqual('<<=', token.LEFTSHIFTEQUAL)
self.assertExactTypeEqual('>>=', token.RIGHTSHIFTEQUAL)
self.assertExactTypeEqual('**=', token.DOUBLESTAREQUAL)
self.assertExactTypeEqual('//', token.DOUBLESLASH)
self.assertExactTypeEqual('//=', token.DOUBLESLASHEQUAL)
self.assertExactTypeEqual(':=', token.COLONEQUAL)
self.assertExactTypeEqual('...', token.ELLIPSIS)
self.assertExactTypeEqual('->', token.RARROW)
self.assertExactTypeEqual('@', token.AT)
self.assertExactTypeEqual('@=', token.ATEQUAL)
self.assertExactTypeEqual('a**2+b**2==c**2',
NAME, token.DOUBLESTAR, NUMBER,
token.PLUS,
NAME, token.DOUBLESTAR, NUMBER,
token.EQEQUAL,
NAME, token.DOUBLESTAR, NUMBER)
self.assertExactTypeEqual('{1, 2, 3}',
token.LBRACE,
token.NUMBER, token.COMMA,
token.NUMBER, token.COMMA,
token.NUMBER,
token.RBRACE)
self.assertExactTypeEqual('^(x & 0x1)',
token.CIRCUMFLEX,
token.LPAR,
token.NAME, token.AMPER, token.NUMBER,
token.RPAR)
def test_pathological_trailing_whitespace(self):
# See http://bugs.python.org/issue16152
self.assertExactTypeEqual('@ ', token.AT)
def test_comment_at_the_end_of_the_source_without_newline(self):
# See http://bugs.python.org/issue44667
source = 'b = 1\n\n#test'
expected_tokens = [token.NAME, token.EQUAL, token.NUMBER, token.NEWLINE, token.NL, token.COMMENT]
tokens = list(tokenize(BytesIO(source.encode('utf-8')).readline))
self.assertEqual(tok_name[tokens[0].exact_type], tok_name[ENCODING])
for i in range(6):
self.assertEqual(tok_name[tokens[i + 1].exact_type], tok_name[expected_tokens[i]])
self.assertEqual(tok_name[tokens[-1].exact_type], tok_name[token.ENDMARKER])
class UntokenizeTest(TestCase):
def test_bad_input_order(self):
# raise if previous row
u = Untokenizer()
u.prev_row = 2
u.prev_col = 2
with self.assertRaises(ValueError) as cm:
u.add_whitespace((1,3))
self.assertEqual(cm.exception.args[0],
'start (1,3) precedes previous end (2,2)')
# raise if previous column in row
self.assertRaises(ValueError, u.add_whitespace, (2,1))
def test_backslash_continuation(self):
# The problem is that <whitespace>\<newline> leaves no token
u = Untokenizer()
u.prev_row = 1
u.prev_col = 1
u.tokens = []
u.add_whitespace((2, 0))
self.assertEqual(u.tokens, ['\\\n'])
u.prev_row = 2
u.add_whitespace((4, 4))
self.assertEqual(u.tokens, ['\\\n', '\\\n\\\n', ' '])
TestRoundtrip.check_roundtrip(self, 'a\n b\n c\n \\\n c\n')
def test_iter_compat(self):
u = Untokenizer()
token = (NAME, 'Hello')
tokens = [(ENCODING, 'utf-8'), token]
u.compat(token, iter([]))
self.assertEqual(u.tokens, ["Hello "])
u = Untokenizer()
self.assertEqual(u.untokenize(iter([token])), 'Hello ')
u = Untokenizer()
self.assertEqual(u.untokenize(iter(tokens)), 'Hello ')
self.assertEqual(u.encoding, 'utf-8')
self.assertEqual(untokenize(iter(tokens)), b'Hello ')
class TestRoundtrip(TestCase):
def check_roundtrip(self, f):
"""
Test roundtrip for `untokenize`. `f` is an open file or a string.
The source code in f is tokenized to both 5- and 2-tuples.
Both sequences are converted back to source code via
tokenize.untokenize(), and the latter tokenized again to 2-tuples.
The test fails if the 3 pair tokenizations do not match.
When untokenize bugs are fixed, untokenize with 5-tuples should
reproduce code that does not contain a backslash continuation
following spaces. A proper test should test this.
"""
# Get source code and original tokenizations
if isinstance(f, str):
code = f.encode('utf-8')
else:
code = f.read()
f.close()
readline = iter(code.splitlines(keepends=True)).__next__
tokens5 = list(tokenize(readline))
tokens2 = [tok[:2] for tok in tokens5]
# Reproduce tokens2 from pairs
bytes_from2 = untokenize(tokens2)
readline2 = iter(bytes_from2.splitlines(keepends=True)).__next__
tokens2_from2 = [tok[:2] for tok in tokenize(readline2)]
self.assertEqual(tokens2_from2, tokens2)
# Reproduce tokens2 from 5-tuples
bytes_from5 = untokenize(tokens5)
readline5 = iter(bytes_from5.splitlines(keepends=True)).__next__
tokens2_from5 = [tok[:2] for tok in tokenize(readline5)]
self.assertEqual(tokens2_from5, tokens2)
def test_roundtrip(self):
# There are some standard formatting practices that are easy to get right.
self.check_roundtrip("if x == 1:\n"
" print(x)\n")
self.check_roundtrip("# This is a comment\n"
"# This also\n")
# Some people use different formatting conventions, which makes
# untokenize a little trickier. Note that this test involves trailing
# whitespace after the colon. Note that we use hex escapes to make the
# two trailing blanks apparent in the expected output.
self.check_roundtrip("if x == 1 : \n"
" print(x)\n")
fn = support.findfile("tokenize_tests.txt")
with open(fn, 'rb') as f:
self.check_roundtrip(f)
self.check_roundtrip("if x == 1:\n"
" # A comment by itself.\n"
" print(x) # Comment here, too.\n"
" # Another comment.\n"
"after_if = True\n")
self.check_roundtrip("if (x # The comments need to go in the right place\n"
" == 1):\n"
" print('x==1')\n")
self.check_roundtrip("class Test: # A comment here\n"
" # A comment with weird indent\n"
" after_com = 5\n"
" def x(m): return m*5 # a one liner\n"
" def y(m): # A whitespace after the colon\n"
" return y*4 # 3-space indent\n")
# Some error-handling code
self.check_roundtrip("try: import somemodule\n"
"except ImportError: # comment\n"
" print('Can not import' # comment2\n)"
"else: print('Loaded')\n")
def test_continuation(self):
# Balancing continuation
self.check_roundtrip("a = (3,4, \n"
"5,6)\n"
"y = [3, 4,\n"
"5]\n"
"z = {'a': 5,\n"
"'b':15, 'c':True}\n"
"x = len(y) + 5 - a[\n"
"3] - a[2]\n"
"+ len(z) - z[\n"
"'b']\n")
def test_backslash_continuation(self):
# Backslash means line continuation, except for comments
self.check_roundtrip("x=1+\\\n"
"1\n"
"# This is a comment\\\n"
"# This also\n")
self.check_roundtrip("# Comment \\\n"
"x = 0")
def test_string_concatenation(self):
# Two string literals on the same line
self.check_roundtrip("'' ''")
def test_random_files(self):
# Test roundtrip on random python modules.
# pass the '-ucpu' option to process the full directory.
import glob, random
fn = support.findfile("tokenize_tests.txt")
tempdir = os.path.dirname(fn) or os.curdir
testfiles = glob.glob(os.path.join(glob.escape(tempdir), "test*.py"))
# Tokenize is broken on test_pep3131.py because regular expressions are
# broken on the obscure unicode identifiers in it. *sigh*
# With roundtrip extended to test the 5-tuple mode of untokenize,
# 7 more testfiles fail. Remove them also until the failure is diagnosed.
testfiles.remove(os.path.join(tempdir, "test_unicode_identifiers.py"))
for f in ('buffer', 'builtin', 'fileio', 'inspect', 'os', 'platform', 'sys'):
testfiles.remove(os.path.join(tempdir, "test_%s.py") % f)
if not support.is_resource_enabled("cpu"):
testfiles = random.sample(testfiles, 10)
for testfile in testfiles:
if support.verbose >= 2:
print('tokenize', testfile)
with open(testfile, 'rb') as f:
with self.subTest(file=testfile):
self.check_roundtrip(f)
def roundtrip(self, code):
if isinstance(code, str):
code = code.encode('utf-8')
return untokenize(tokenize(BytesIO(code).readline)).decode('utf-8')
def test_indentation_semantics_retained(self):
"""
Ensure that although whitespace might be mutated in a roundtrip,
the semantic meaning of the indentation remains consistent.
"""
code = "if False:\n\tx=3\n\tx=3\n"
codelines = self.roundtrip(code).split('\n')
self.assertEqual(codelines[1], codelines[2])
self.check_roundtrip(code)
class CTokenizeTest(TestCase):
def check_tokenize(self, s, expected):
# Format the tokens in s in a table format.
# The ENDMARKER and final NEWLINE are omitted.
with self.subTest(source=s):
result = stringify_tokens_from_source(
_generate_tokens_from_c_tokenizer(s), s
)
self.assertEqual(result, expected.rstrip().splitlines())
def test_int(self):
self.check_tokenize('0xff <= 255', """\
NUMBER '0xff' (1, 0) (1, 4)
LESSEQUAL '<=' (1, 5) (1, 7)
NUMBER '255' (1, 8) (1, 11)
""")
self.check_tokenize('0b10 <= 255', """\
NUMBER '0b10' (1, 0) (1, 4)
LESSEQUAL '<=' (1, 5) (1, 7)
NUMBER '255' (1, 8) (1, 11)
""")
self.check_tokenize('0o123 <= 0O123', """\
NUMBER '0o123' (1, 0) (1, 5)
LESSEQUAL '<=' (1, 6) (1, 8)
NUMBER '0O123' (1, 9) (1, 14)
""")
self.check_tokenize('1234567 > ~0x15', """\
NUMBER '1234567' (1, 0) (1, 7)
GREATER '>' (1, 8) (1, 9)
TILDE '~' (1, 10) (1, 11)
NUMBER '0x15' (1, 11) (1, 15)
""")
self.check_tokenize('2134568 != 1231515', """\
NUMBER '2134568' (1, 0) (1, 7)
NOTEQUAL '!=' (1, 8) (1, 10)
NUMBER '1231515' (1, 11) (1, 18)
""")
self.check_tokenize('(-124561-1) & 200000000', """\
LPAR '(' (1, 0) (1, 1)
MINUS '-' (1, 1) (1, 2)
NUMBER '124561' (1, 2) (1, 8)
MINUS '-' (1, 8) (1, 9)
NUMBER '1' (1, 9) (1, 10)
RPAR ')' (1, 10) (1, 11)
AMPER '&' (1, 12) (1, 13)
NUMBER '200000000' (1, 14) (1, 23)
""")
self.check_tokenize('0xdeadbeef != -1', """\
NUMBER '0xdeadbeef' (1, 0) (1, 10)
NOTEQUAL '!=' (1, 11) (1, 13)
MINUS '-' (1, 14) (1, 15)
NUMBER '1' (1, 15) (1, 16)
""")
self.check_tokenize('0xdeadc0de & 12345', """\
NUMBER '0xdeadc0de' (1, 0) (1, 10)
AMPER '&' (1, 11) (1, 12)
NUMBER '12345' (1, 13) (1, 18)
""")
self.check_tokenize('0xFF & 0x15 | 1234', """\
NUMBER '0xFF' (1, 0) (1, 4)
AMPER '&' (1, 5) (1, 6)
NUMBER '0x15' (1, 7) (1, 11)
VBAR '|' (1, 12) (1, 13)
NUMBER '1234' (1, 14) (1, 18)
""")
def test_float(self):
self.check_tokenize('x = 3.14159', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
NUMBER '3.14159' (1, 4) (1, 11)
""")
self.check_tokenize('x = 314159.', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
NUMBER '314159.' (1, 4) (1, 11)
""")
self.check_tokenize('x = .314159', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
NUMBER '.314159' (1, 4) (1, 11)
""")
self.check_tokenize('x = 3e14159', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
NUMBER '3e14159' (1, 4) (1, 11)
""")
self.check_tokenize('x = 3E123', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
NUMBER '3E123' (1, 4) (1, 9)
""")
self.check_tokenize('x+y = 3e-1230', """\
NAME 'x' (1, 0) (1, 1)
PLUS '+' (1, 1) (1, 2)
NAME 'y' (1, 2) (1, 3)
EQUAL '=' (1, 4) (1, 5)
NUMBER '3e-1230' (1, 6) (1, 13)
""")
self.check_tokenize('x = 3.14e159', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
NUMBER '3.14e159' (1, 4) (1, 12)
""")
def test_string(self):
self.check_tokenize('x = \'\'; y = ""', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
STRING "''" (1, 4) (1, 6)
SEMI ';' (1, 6) (1, 7)
NAME 'y' (1, 8) (1, 9)
EQUAL '=' (1, 10) (1, 11)
STRING '""' (1, 12) (1, 14)
""")
self.check_tokenize('x = \'"\'; y = "\'"', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
STRING '\\'"\\'' (1, 4) (1, 7)
SEMI ';' (1, 7) (1, 8)
NAME 'y' (1, 9) (1, 10)
EQUAL '=' (1, 11) (1, 12)
STRING '"\\'"' (1, 13) (1, 16)
""")
self.check_tokenize('x = "doesn\'t "shrink", does it"', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
STRING '"doesn\\'t "' (1, 4) (1, 14)
NAME 'shrink' (1, 14) (1, 20)
STRING '", does it"' (1, 20) (1, 31)
""")
self.check_tokenize("x = 'abc' + 'ABC'", """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
STRING "'abc'" (1, 4) (1, 9)
PLUS '+' (1, 10) (1, 11)
STRING "'ABC'" (1, 12) (1, 17)
""")
self.check_tokenize('y = "ABC" + "ABC"', """\
NAME 'y' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
STRING '"ABC"' (1, 4) (1, 9)
PLUS '+' (1, 10) (1, 11)
STRING '"ABC"' (1, 12) (1, 17)
""")
self.check_tokenize("x = r'abc' + r'ABC' + R'ABC' + R'ABC'", """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
STRING "r'abc'" (1, 4) (1, 10)
PLUS '+' (1, 11) (1, 12)
STRING "r'ABC'" (1, 13) (1, 19)
PLUS '+' (1, 20) (1, 21)
STRING "R'ABC'" (1, 22) (1, 28)
PLUS '+' (1, 29) (1, 30)
STRING "R'ABC'" (1, 31) (1, 37)
""")
self.check_tokenize('y = r"abc" + r"ABC" + R"ABC" + R"ABC"', """\
NAME 'y' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
STRING 'r"abc"' (1, 4) (1, 10)
PLUS '+' (1, 11) (1, 12)
STRING 'r"ABC"' (1, 13) (1, 19)
PLUS '+' (1, 20) (1, 21)
STRING 'R"ABC"' (1, 22) (1, 28)
PLUS '+' (1, 29) (1, 30)
STRING 'R"ABC"' (1, 31) (1, 37)
""")
self.check_tokenize("u'abc' + U'abc'", """\
STRING "u'abc'" (1, 0) (1, 6)
PLUS '+' (1, 7) (1, 8)
STRING "U'abc'" (1, 9) (1, 15)
""")
self.check_tokenize('u"abc" + U"abc"', """\
STRING 'u"abc"' (1, 0) (1, 6)
PLUS '+' (1, 7) (1, 8)
STRING 'U"abc"' (1, 9) (1, 15)
""")
self.check_tokenize("b'abc' + B'abc'", """\
STRING "b'abc'" (1, 0) (1, 6)
PLUS '+' (1, 7) (1, 8)
STRING "B'abc'" (1, 9) (1, 15)
""")
self.check_tokenize('b"abc" + B"abc"', """\
STRING 'b"abc"' (1, 0) (1, 6)
PLUS '+' (1, 7) (1, 8)
STRING 'B"abc"' (1, 9) (1, 15)
""")
self.check_tokenize("br'abc' + bR'abc' + Br'abc' + BR'abc'", """\
STRING "br'abc'" (1, 0) (1, 7)
PLUS '+' (1, 8) (1, 9)
STRING "bR'abc'" (1, 10) (1, 17)
PLUS '+' (1, 18) (1, 19)
STRING "Br'abc'" (1, 20) (1, 27)
PLUS '+' (1, 28) (1, 29)
STRING "BR'abc'" (1, 30) (1, 37)
""")
self.check_tokenize('br"abc" + bR"abc" + Br"abc" + BR"abc"', """\
STRING 'br"abc"' (1, 0) (1, 7)
PLUS '+' (1, 8) (1, 9)
STRING 'bR"abc"' (1, 10) (1, 17)
PLUS '+' (1, 18) (1, 19)
STRING 'Br"abc"' (1, 20) (1, 27)
PLUS '+' (1, 28) (1, 29)
STRING 'BR"abc"' (1, 30) (1, 37)
""")
self.check_tokenize("rb'abc' + rB'abc' + Rb'abc' + RB'abc'", """\
STRING "rb'abc'" (1, 0) (1, 7)
PLUS '+' (1, 8) (1, 9)
STRING "rB'abc'" (1, 10) (1, 17)
PLUS '+' (1, 18) (1, 19)
STRING "Rb'abc'" (1, 20) (1, 27)
PLUS '+' (1, 28) (1, 29)
STRING "RB'abc'" (1, 30) (1, 37)
""")
self.check_tokenize('rb"abc" + rB"abc" + Rb"abc" + RB"abc"', """\
STRING 'rb"abc"' (1, 0) (1, 7)
PLUS '+' (1, 8) (1, 9)
STRING 'rB"abc"' (1, 10) (1, 17)
PLUS '+' (1, 18) (1, 19)
STRING 'Rb"abc"' (1, 20) (1, 27)
PLUS '+' (1, 28) (1, 29)
STRING 'RB"abc"' (1, 30) (1, 37)
""")
self.check_tokenize('"a\\\nde\\\nfg"', """\
STRING '"a\\\\\\nde\\\\\\nfg"\' (1, 0) (3, 3)
""")
self.check_tokenize('u"a\\\nde"', """\
STRING 'u"a\\\\\\nde"\' (1, 0) (2, 3)
""")
self.check_tokenize('rb"a\\\nd"', """\
STRING 'rb"a\\\\\\nd"\' (1, 0) (2, 2)
""")
self.check_tokenize(r'"""a\
b"""', """\
STRING '\"\""a\\\\\\nb\"\""' (1, 0) (2, 4)
""")
self.check_tokenize(r'u"""a\
b"""', """\
STRING 'u\"\""a\\\\\\nb\"\""' (1, 0) (2, 4)
""")
self.check_tokenize(r'rb"""a\
b\
c"""', """\
STRING 'rb"\""a\\\\\\nb\\\\\\nc"\""' (1, 0) (3, 4)
""")
self.check_tokenize('f"abc"', """\
STRING 'f"abc"' (1, 0) (1, 6)
""")
self.check_tokenize('fR"a{b}c"', """\
STRING 'fR"a{b}c"' (1, 0) (1, 9)
""")
self.check_tokenize('f"""abc"""', """\
STRING 'f\"\"\"abc\"\"\"' (1, 0) (1, 10)
""")
self.check_tokenize(r'f"abc\
def"', """\
STRING 'f"abc\\\\\\ndef"' (1, 0) (2, 4)
""")
self.check_tokenize(r'Rf"abc\
def"', """\
STRING 'Rf"abc\\\\\\ndef"' (1, 0) (2, 4)
""")
def test_function(self):
self.check_tokenize('def d22(a, b, c=2, d=2, *k): pass', """\
NAME 'def' (1, 0) (1, 3)
NAME 'd22' (1, 4) (1, 7)
LPAR '(' (1, 7) (1, 8)
NAME 'a' (1, 8) (1, 9)
COMMA ',' (1, 9) (1, 10)
NAME 'b' (1, 11) (1, 12)
COMMA ',' (1, 12) (1, 13)
NAME 'c' (1, 14) (1, 15)
EQUAL '=' (1, 15) (1, 16)
NUMBER '2' (1, 16) (1, 17)
COMMA ',' (1, 17) (1, 18)
NAME 'd' (1, 19) (1, 20)
EQUAL '=' (1, 20) (1, 21)
NUMBER '2' (1, 21) (1, 22)
COMMA ',' (1, 22) (1, 23)
STAR '*' (1, 24) (1, 25)
NAME 'k' (1, 25) (1, 26)
RPAR ')' (1, 26) (1, 27)
COLON ':' (1, 27) (1, 28)
NAME 'pass' (1, 29) (1, 33)
""")
self.check_tokenize('def d01v_(a=1, *k, **w): pass', """\
NAME 'def' (1, 0) (1, 3)
NAME 'd01v_' (1, 4) (1, 9)
LPAR '(' (1, 9) (1, 10)
NAME 'a' (1, 10) (1, 11)
EQUAL '=' (1, 11) (1, 12)
NUMBER '1' (1, 12) (1, 13)
COMMA ',' (1, 13) (1, 14)
STAR '*' (1, 15) (1, 16)
NAME 'k' (1, 16) (1, 17)
COMMA ',' (1, 17) (1, 18)
DOUBLESTAR '**' (1, 19) (1, 21)
NAME 'w' (1, 21) (1, 22)
RPAR ')' (1, 22) (1, 23)
COLON ':' (1, 23) (1, 24)
NAME 'pass' (1, 25) (1, 29)
""")
self.check_tokenize('def d23(a: str, b: int=3) -> int: pass', """\
NAME 'def' (1, 0) (1, 3)
NAME 'd23' (1, 4) (1, 7)
LPAR '(' (1, 7) (1, 8)
NAME 'a' (1, 8) (1, 9)
COLON ':' (1, 9) (1, 10)
NAME 'str' (1, 11) (1, 14)
COMMA ',' (1, 14) (1, 15)
NAME 'b' (1, 16) (1, 17)
COLON ':' (1, 17) (1, 18)
NAME 'int' (1, 19) (1, 22)
EQUAL '=' (1, 22) (1, 23)
NUMBER '3' (1, 23) (1, 24)
RPAR ')' (1, 24) (1, 25)
RARROW '->' (1, 26) (1, 28)
NAME 'int' (1, 29) (1, 32)
COLON ':' (1, 32) (1, 33)
NAME 'pass' (1, 34) (1, 38)
""")
def test_comparison(self):
self.check_tokenize("if 1 < 1 > 1 == 1 >= 5 <= 0x15 <= 0x12 != "
"1 and 5 in 1 not in 1 is 1 or 5 is not 1: pass", """\
NAME 'if' (1, 0) (1, 2)
NUMBER '1' (1, 3) (1, 4)
LESS '<' (1, 5) (1, 6)
NUMBER '1' (1, 7) (1, 8)
GREATER '>' (1, 9) (1, 10)
NUMBER '1' (1, 11) (1, 12)
EQEQUAL '==' (1, 13) (1, 15)
NUMBER '1' (1, 16) (1, 17)
GREATEREQUAL '>=' (1, 18) (1, 20)
NUMBER '5' (1, 21) (1, 22)
LESSEQUAL '<=' (1, 23) (1, 25)
NUMBER '0x15' (1, 26) (1, 30)
LESSEQUAL '<=' (1, 31) (1, 33)
NUMBER '0x12' (1, 34) (1, 38)
NOTEQUAL '!=' (1, 39) (1, 41)
NUMBER '1' (1, 42) (1, 43)
NAME 'and' (1, 44) (1, 47)
NUMBER '5' (1, 48) (1, 49)
NAME 'in' (1, 50) (1, 52)
NUMBER '1' (1, 53) (1, 54)
NAME 'not' (1, 55) (1, 58)
NAME 'in' (1, 59) (1, 61)
NUMBER '1' (1, 62) (1, 63)
NAME 'is' (1, 64) (1, 66)
NUMBER '1' (1, 67) (1, 68)
NAME 'or' (1, 69) (1, 71)
NUMBER '5' (1, 72) (1, 73)
NAME 'is' (1, 74) (1, 76)
NAME 'not' (1, 77) (1, 80)
NUMBER '1' (1, 81) (1, 82)
COLON ':' (1, 82) (1, 83)
NAME 'pass' (1, 84) (1, 88)
""")
def test_additive(self):
self.check_tokenize('x = 1 - y + 15 - 1 + 0x124 + z + a[5]', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
NUMBER '1' (1, 4) (1, 5)
MINUS '-' (1, 6) (1, 7)
NAME 'y' (1, 8) (1, 9)
PLUS '+' (1, 10) (1, 11)
NUMBER '15' (1, 12) (1, 14)
MINUS '-' (1, 15) (1, 16)
NUMBER '1' (1, 17) (1, 18)
PLUS '+' (1, 19) (1, 20)
NUMBER '0x124' (1, 21) (1, 26)
PLUS '+' (1, 27) (1, 28)
NAME 'z' (1, 29) (1, 30)
PLUS '+' (1, 31) (1, 32)
NAME 'a' (1, 33) (1, 34)
LSQB '[' (1, 34) (1, 35)
NUMBER '5' (1, 35) (1, 36)
RSQB ']' (1, 36) (1, 37)
""")
def test_multiplicative(self):
self.check_tokenize('x = 1//1*1/5*12%0x12@42', """\
NAME 'x' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
NUMBER '1' (1, 4) (1, 5)
DOUBLESLASH '//' (1, 5) (1, 7)
NUMBER '1' (1, 7) (1, 8)
STAR '*' (1, 8) (1, 9)
NUMBER '1' (1, 9) (1, 10)
SLASH '/' (1, 10) (1, 11)
NUMBER '5' (1, 11) (1, 12)
STAR '*' (1, 12) (1, 13)
NUMBER '12' (1, 13) (1, 15)
PERCENT '%' (1, 15) (1, 16)
NUMBER '0x12' (1, 16) (1, 20)
AT '@' (1, 20) (1, 21)
NUMBER '42' (1, 21) (1, 23)
""")
def test_unary(self):
self.check_tokenize('~1 ^ 1 & 1 |1 ^ -1', """\
TILDE '~' (1, 0) (1, 1)
NUMBER '1' (1, 1) (1, 2)
CIRCUMFLEX '^' (1, 3) (1, 4)
NUMBER '1' (1, 5) (1, 6)
AMPER '&' (1, 7) (1, 8)
NUMBER '1' (1, 9) (1, 10)
VBAR '|' (1, 11) (1, 12)
NUMBER '1' (1, 12) (1, 13)
CIRCUMFLEX '^' (1, 14) (1, 15)
MINUS '-' (1, 16) (1, 17)
NUMBER '1' (1, 17) (1, 18)
""")
self.check_tokenize('-1*1/1+1*1//1 - ---1**1', """\
MINUS '-' (1, 0) (1, 1)
NUMBER '1' (1, 1) (1, 2)
STAR '*' (1, 2) (1, 3)
NUMBER '1' (1, 3) (1, 4)
SLASH '/' (1, 4) (1, 5)
NUMBER '1' (1, 5) (1, 6)
PLUS '+' (1, 6) (1, 7)
NUMBER '1' (1, 7) (1, 8)
STAR '*' (1, 8) (1, 9)
NUMBER '1' (1, 9) (1, 10)
DOUBLESLASH '//' (1, 10) (1, 12)
NUMBER '1' (1, 12) (1, 13)
MINUS '-' (1, 14) (1, 15)
MINUS '-' (1, 16) (1, 17)
MINUS '-' (1, 17) (1, 18)
MINUS '-' (1, 18) (1, 19)
NUMBER '1' (1, 19) (1, 20)
DOUBLESTAR '**' (1, 20) (1, 22)
NUMBER '1' (1, 22) (1, 23)
""")
def test_selector(self):
self.check_tokenize("import sys, time\nx = sys.modules['time'].time()", """\
NAME 'import' (1, 0) (1, 6)
NAME 'sys' (1, 7) (1, 10)
COMMA ',' (1, 10) (1, 11)
NAME 'time' (1, 12) (1, 16)
NEWLINE '' (1, 16) (1, 16)
NAME 'x' (2, 0) (2, 1)
EQUAL '=' (2, 2) (2, 3)
NAME 'sys' (2, 4) (2, 7)
DOT '.' (2, 7) (2, 8)
NAME 'modules' (2, 8) (2, 15)
LSQB '[' (2, 15) (2, 16)
STRING "'time'" (2, 16) (2, 22)
RSQB ']' (2, 22) (2, 23)
DOT '.' (2, 23) (2, 24)
NAME 'time' (2, 24) (2, 28)
LPAR '(' (2, 28) (2, 29)
RPAR ')' (2, 29) (2, 30)
""")
def test_method(self):
self.check_tokenize('@staticmethod\ndef foo(x,y): pass', """\
AT '@' (1, 0) (1, 1)
NAME 'staticmethod' (1, 1) (1, 13)
NEWLINE '' (1, 13) (1, 13)
NAME 'def' (2, 0) (2, 3)
NAME 'foo' (2, 4) (2, 7)
LPAR '(' (2, 7) (2, 8)
NAME 'x' (2, 8) (2, 9)
COMMA ',' (2, 9) (2, 10)
NAME 'y' (2, 10) (2, 11)
RPAR ')' (2, 11) (2, 12)
COLON ':' (2, 12) (2, 13)
NAME 'pass' (2, 14) (2, 18)
""")
def test_tabs(self):
self.check_tokenize('@staticmethod\ndef foo(x,y): pass', """\
AT '@' (1, 0) (1, 1)
NAME 'staticmethod' (1, 1) (1, 13)
NEWLINE '' (1, 13) (1, 13)
NAME 'def' (2, 0) (2, 3)
NAME 'foo' (2, 4) (2, 7)
LPAR '(' (2, 7) (2, 8)
NAME 'x' (2, 8) (2, 9)
COMMA ',' (2, 9) (2, 10)
NAME 'y' (2, 10) (2, 11)
RPAR ')' (2, 11) (2, 12)
COLON ':' (2, 12) (2, 13)
NAME 'pass' (2, 14) (2, 18)
""")
def test_async(self):
self.check_tokenize('async = 1', """\
ASYNC 'async' (1, 0) (1, 5)
EQUAL '=' (1, 6) (1, 7)
NUMBER '1' (1, 8) (1, 9)
""")
self.check_tokenize('a = (async = 1)', """\
NAME 'a' (1, 0) (1, 1)
EQUAL '=' (1, 2) (1, 3)
LPAR '(' (1, 4) (1, 5)
ASYNC 'async' (1, 5) (1, 10)
EQUAL '=' (1, 11) (1, 12)
NUMBER '1' (1, 13) (1, 14)
RPAR ')' (1, 14) (1, 15)
""")
self.check_tokenize('async()', """\
ASYNC 'async' (1, 0) (1, 5)
LPAR '(' (1, 5) (1, 6)
RPAR ')' (1, 6) (1, 7)
""")
self.check_tokenize('class async(Bar):pass', """\
NAME 'class' (1, 0) (1, 5)
ASYNC 'async' (1, 6) (1, 11)
LPAR '(' (1, 11) (1, 12)
NAME 'Bar' (1, 12) (1, 15)
RPAR ')' (1, 15) (1, 16)
COLON ':' (1, 16) (1, 17)
NAME 'pass' (1, 17) (1, 21)
""")
self.check_tokenize('class async:pass', """\
NAME 'class' (1, 0) (1, 5)
ASYNC 'async' (1, 6) (1, 11)
COLON ':' (1, 11) (1, 12)
NAME 'pass' (1, 12) (1, 16)
""")
self.check_tokenize('await = 1', """\
AWAIT 'await' (1, 0) (1, 5)
EQUAL '=' (1, 6) (1, 7)
NUMBER '1' (1, 8) (1, 9)
""")
self.check_tokenize('foo.async', """\
NAME 'foo' (1, 0) (1, 3)
DOT '.' (1, 3) (1, 4)
ASYNC 'async' (1, 4) (1, 9)
""")
self.check_tokenize('async for a in b: pass', """\
ASYNC 'async' (1, 0) (1, 5)
NAME 'for' (1, 6) (1, 9)
NAME 'a' (1, 10) (1, 11)
NAME 'in' (1, 12) (1, 14)
NAME 'b' (1, 15) (1, 16)
COLON ':' (1, 16) (1, 17)
NAME 'pass' (1, 18) (1, 22)
""")
self.check_tokenize('async with a as b: pass', """\
ASYNC 'async' (1, 0) (1, 5)
NAME 'with' (1, 6) (1, 10)
NAME 'a' (1, 11) (1, 12)
NAME 'as' (1, 13) (1, 15)
NAME 'b' (1, 16) (1, 17)
COLON ':' (1, 17) (1, 18)
NAME 'pass' (1, 19) (1, 23)
""")
self.check_tokenize('async.foo', """\
ASYNC 'async' (1, 0) (1, 5)
DOT '.' (1, 5) (1, 6)
NAME 'foo' (1, 6) (1, 9)
""")
self.check_tokenize('async', """\
ASYNC 'async' (1, 0) (1, 5)
""")
self.check_tokenize('async\n#comment\nawait', """\
ASYNC 'async' (1, 0) (1, 5)
NEWLINE '' (1, 5) (1, 5)
AWAIT 'await' (3, 0) (3, 5)
""")
self.check_tokenize('async\n...\nawait', """\
ASYNC 'async' (1, 0) (1, 5)
NEWLINE '' (1, 5) (1, 5)
ELLIPSIS '...' (2, 0) (2, 3)
NEWLINE '' (2, 3) (2, 3)
AWAIT 'await' (3, 0) (3, 5)
""")
self.check_tokenize('async\nawait', """\
ASYNC 'async' (1, 0) (1, 5)
NEWLINE '' (1, 5) (1, 5)
AWAIT 'await' (2, 0) (2, 5)
""")
self.check_tokenize('foo.async + 1', """\
NAME 'foo' (1, 0) (1, 3)
DOT '.' (1, 3) (1, 4)
ASYNC 'async' (1, 4) (1, 9)
PLUS '+' (1, 10) (1, 11)
NUMBER '1' (1, 12) (1, 13)
""")
self.check_tokenize('async def foo(): pass', """\
ASYNC 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'foo' (1, 10) (1, 13)
LPAR '(' (1, 13) (1, 14)
RPAR ')' (1, 14) (1, 15)
COLON ':' (1, 15) (1, 16)
NAME 'pass' (1, 17) (1, 21)
""")
self.check_tokenize('''\
async def foo():
def foo(await):
await = 1
if 1:
await
async += 1
''', """\
ASYNC 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'foo' (1, 10) (1, 13)
LPAR '(' (1, 13) (1, 14)
RPAR ')' (1, 14) (1, 15)
COLON ':' (1, 15) (1, 16)
NEWLINE '' (1, 16) (1, 16)
INDENT '' (2, -1) (2, -1)
NAME 'def' (2, 2) (2, 5)
NAME 'foo' (2, 6) (2, 9)
LPAR '(' (2, 9) (2, 10)
AWAIT 'await' (2, 10) (2, 15)
RPAR ')' (2, 15) (2, 16)
COLON ':' (2, 16) (2, 17)
NEWLINE '' (2, 17) (2, 17)
INDENT '' (3, -1) (3, -1)
AWAIT 'await' (3, 4) (3, 9)
EQUAL '=' (3, 10) (3, 11)
NUMBER '1' (3, 12) (3, 13)
NEWLINE '' (3, 13) (3, 13)
DEDENT '' (4, -1) (4, -1)
NAME 'if' (4, 2) (4, 4)
NUMBER '1' (4, 5) (4, 6)
COLON ':' (4, 6) (4, 7)
NEWLINE '' (4, 7) (4, 7)
INDENT '' (5, -1) (5, -1)
AWAIT 'await' (5, 4) (5, 9)
NEWLINE '' (5, 9) (5, 9)
DEDENT '' (6, -1) (6, -1)
DEDENT '' (6, -1) (6, -1)
ASYNC 'async' (6, 0) (6, 5)
PLUSEQUAL '+=' (6, 6) (6, 8)
NUMBER '1' (6, 9) (6, 10)
NEWLINE '' (6, 10) (6, 10)
""")
self.check_tokenize('async def foo():\n async for i in 1: pass', """\
ASYNC 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'foo' (1, 10) (1, 13)
LPAR '(' (1, 13) (1, 14)
RPAR ')' (1, 14) (1, 15)
COLON ':' (1, 15) (1, 16)
NEWLINE '' (1, 16) (1, 16)
INDENT '' (2, -1) (2, -1)
ASYNC 'async' (2, 2) (2, 7)
NAME 'for' (2, 8) (2, 11)
NAME 'i' (2, 12) (2, 13)
NAME 'in' (2, 14) (2, 16)
NUMBER '1' (2, 17) (2, 18)
COLON ':' (2, 18) (2, 19)
NAME 'pass' (2, 20) (2, 24)
DEDENT '' (2, -1) (2, -1)
""")
self.check_tokenize('async def foo(async): await', """\
ASYNC 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'foo' (1, 10) (1, 13)
LPAR '(' (1, 13) (1, 14)
ASYNC 'async' (1, 14) (1, 19)
RPAR ')' (1, 19) (1, 20)
COLON ':' (1, 20) (1, 21)
AWAIT 'await' (1, 22) (1, 27)
""")
self.check_tokenize('''\
def f():
def baz(): pass
async def bar(): pass
await = 2''', """\
NAME 'def' (1, 0) (1, 3)
NAME 'f' (1, 4) (1, 5)
LPAR '(' (1, 5) (1, 6)
RPAR ')' (1, 6) (1, 7)
COLON ':' (1, 7) (1, 8)
NEWLINE '' (1, 8) (1, 8)
INDENT '' (3, -1) (3, -1)
NAME 'def' (3, 2) (3, 5)
NAME 'baz' (3, 6) (3, 9)
LPAR '(' (3, 9) (3, 10)
RPAR ')' (3, 10) (3, 11)
COLON ':' (3, 11) (3, 12)
NAME 'pass' (3, 13) (3, 17)
NEWLINE '' (3, 17) (3, 17)
ASYNC 'async' (4, 2) (4, 7)
NAME 'def' (4, 8) (4, 11)
NAME 'bar' (4, 12) (4, 15)
LPAR '(' (4, 15) (4, 16)
RPAR ')' (4, 16) (4, 17)
COLON ':' (4, 17) (4, 18)
NAME 'pass' (4, 19) (4, 23)
NEWLINE '' (4, 23) (4, 23)
AWAIT 'await' (6, 2) (6, 7)
EQUAL '=' (6, 8) (6, 9)
NUMBER '2' (6, 10) (6, 11)
DEDENT '' (6, -1) (6, -1)
""")
self.check_tokenize('''\
async def f():
def baz(): pass
async def bar(): pass
await = 2''', """\
ASYNC 'async' (1, 0) (1, 5)
NAME 'def' (1, 6) (1, 9)
NAME 'f' (1, 10) (1, 11)
LPAR '(' (1, 11) (1, 12)
RPAR ')' (1, 12) (1, 13)
COLON ':' (1, 13) (1, 14)
NEWLINE '' (1, 14) (1, 14)
INDENT '' (3, -1) (3, -1)
NAME 'def' (3, 2) (3, 5)
NAME 'baz' (3, 6) (3, 9)
LPAR '(' (3, 9) (3, 10)
RPAR ')' (3, 10) (3, 11)
COLON ':' (3, 11) (3, 12)
NAME 'pass' (3, 13) (3, 17)
NEWLINE '' (3, 17) (3, 17)
ASYNC 'async' (4, 2) (4, 7)
NAME 'def' (4, 8) (4, 11)
NAME 'bar' (4, 12) (4, 15)
LPAR '(' (4, 15) (4, 16)
RPAR ')' (4, 16) (4, 17)
COLON ':' (4, 17) (4, 18)
NAME 'pass' (4, 19) (4, 23)
NEWLINE '' (4, 23) (4, 23)
AWAIT 'await' (6, 2) (6, 7)
EQUAL '=' (6, 8) (6, 9)
NUMBER '2' (6, 10) (6, 11)
DEDENT '' (6, -1) (6, -1)
""")
def test_unicode(self):
self.check_tokenize("Örter = u'places'\ngrün = U'green'", """\
NAME 'Örter' (1, 0) (1, 6)
EQUAL '=' (1, 7) (1, 8)
STRING "u'places'" (1, 9) (1, 18)
NEWLINE '' (1, 18) (1, 18)
NAME 'grün' (2, 0) (2, 5)
EQUAL '=' (2, 6) (2, 7)
STRING "U'green'" (2, 8) (2, 16)
""")
def test_invalid_syntax(self):
def get_tokens(string):
return list(_generate_tokens_from_c_tokenizer(string))
self.assertRaises(SyntaxError, get_tokens, "(1+2]")
self.assertRaises(SyntaxError, get_tokens, "(1+2}")
self.assertRaises(SyntaxError, get_tokens, "{1+2]")
self.assertRaises(SyntaxError, get_tokens, "1_")
self.assertRaises(SyntaxError, get_tokens, "1.2_")
self.assertRaises(SyntaxError, get_tokens, "1e2_")
self.assertRaises(SyntaxError, get_tokens, "1e+")
self.assertRaises(SyntaxError, get_tokens, "\xa0")
self.assertRaises(SyntaxError, get_tokens, "")
self.assertRaises(SyntaxError, get_tokens, "0b12")
self.assertRaises(SyntaxError, get_tokens, "0b1_2")
self.assertRaises(SyntaxError, get_tokens, "0b2")
self.assertRaises(SyntaxError, get_tokens, "0b1_")
self.assertRaises(SyntaxError, get_tokens, "0b")
self.assertRaises(SyntaxError, get_tokens, "0o18")
self.assertRaises(SyntaxError, get_tokens, "0o1_8")
self.assertRaises(SyntaxError, get_tokens, "0o8")
self.assertRaises(SyntaxError, get_tokens, "0o1_")
self.assertRaises(SyntaxError, get_tokens, "0o")
self.assertRaises(SyntaxError, get_tokens, "0x1_")
self.assertRaises(SyntaxError, get_tokens, "0x")
self.assertRaises(SyntaxError, get_tokens, "1_")
self.assertRaises(SyntaxError, get_tokens, "012")
self.assertRaises(SyntaxError, get_tokens, "1.2_")
self.assertRaises(SyntaxError, get_tokens, "1e2_")
self.assertRaises(SyntaxError, get_tokens, "1e+")
self.assertRaises(SyntaxError, get_tokens, "'sdfsdf")
self.assertRaises(SyntaxError, get_tokens, "'''sdfsdf''")
self.assertRaises(SyntaxError, get_tokens, "("*1000+"a"+")"*1000)
self.assertRaises(SyntaxError, get_tokens, "]")
def test_max_indent(self):
MAXINDENT = 100
def generate_source(indents):
source = ''.join((' ' * x) + 'if True:\n' for x in range(indents))
source += ' ' * indents + 'pass\n'
return source
valid = generate_source(MAXINDENT - 1)
tokens = list(_generate_tokens_from_c_tokenizer(valid))
self.assertEqual(tokens[-1].type, DEDENT)
compile(valid, "<string>", "exec")
invalid = generate_source(MAXINDENT)
tokens = list(_generate_tokens_from_c_tokenizer(invalid))
self.assertEqual(tokens[-1].type, NEWLINE)
self.assertRaises(
IndentationError, compile, invalid, "<string>", "exec"
)
def test_continuation_lines_indentation(self):
def get_tokens(string):
return [(kind, string) for (kind, string, *_) in _generate_tokens_from_c_tokenizer(string)]
code = dedent("""
def fib(n):
\\
'''Print a Fibonacci series up to n.'''
\\
a, b = 0, 1
""")
self.check_tokenize(code, """\
NAME 'def' (2, 0) (2, 3)
NAME 'fib' (2, 4) (2, 7)
LPAR '(' (2, 7) (2, 8)
NAME 'n' (2, 8) (2, 9)
RPAR ')' (2, 9) (2, 10)
COLON ':' (2, 10) (2, 11)
NEWLINE '' (2, 11) (2, 11)
INDENT '' (4, -1) (4, -1)
STRING "'''Print a Fibonacci series up to n.'''" (4, 0) (4, 39)
NEWLINE '' (4, 39) (4, 39)
NAME 'a' (6, 0) (6, 1)
COMMA ',' (6, 1) (6, 2)
NAME 'b' (6, 3) (6, 4)
EQUAL '=' (6, 5) (6, 6)
NUMBER '0' (6, 7) (6, 8)
COMMA ',' (6, 8) (6, 9)
NUMBER '1' (6, 10) (6, 11)
NEWLINE '' (6, 11) (6, 11)
DEDENT '' (6, -1) (6, -1)
""")
code_no_cont = dedent("""
def fib(n):
'''Print a Fibonacci series up to n.'''
a, b = 0, 1
""")
self.assertEqual(get_tokens(code), get_tokens(code_no_cont))
code = dedent("""
pass
\\
pass
""")
self.check_tokenize(code, """\
NAME 'pass' (2, 0) (2, 4)
NEWLINE '' (2, 4) (2, 4)
NAME 'pass' (5, 0) (5, 4)
NEWLINE '' (5, 4) (5, 4)
""")
code_no_cont = dedent("""
pass
pass
""")
self.assertEqual(get_tokens(code), get_tokens(code_no_cont))
code = dedent("""
if x:
y = 1
\\
\\
\\
\\
foo = 1
""")
self.check_tokenize(code, """\
NAME 'if' (2, 0) (2, 2)
NAME 'x' (2, 3) (2, 4)
COLON ':' (2, 4) (2, 5)
NEWLINE '' (2, 5) (2, 5)
INDENT '' (3, -1) (3, -1)
NAME 'y' (3, 4) (3, 5)
EQUAL '=' (3, 6) (3, 7)
NUMBER '1' (3, 8) (3, 9)
NEWLINE '' (3, 9) (3, 9)
NAME 'foo' (8, 4) (8, 7)
EQUAL '=' (8, 8) (8, 9)
NUMBER '1' (8, 10) (8, 11)
NEWLINE '' (8, 11) (8, 11)
DEDENT '' (8, -1) (8, -1)
""")
code_no_cont = dedent("""
if x:
y = 1
foo = 1
""")
self.assertEqual(get_tokens(code), get_tokens(code_no_cont))
class CTokenizerBufferTests(unittest.TestCase):
def test_newline_at_the_end_of_buffer(self):
# See issue 99581: Make sure that if we need to add a new line at the
# end of the buffer, we have enough space in the buffer, specially when
# the current line is as long as the buffer space available.
test_script = f"""\
#coding: latin-1
#{"a"*10000}
#{"a"*10002}"""
with os_helper.temp_dir() as temp_dir:
file_name = make_script(temp_dir, 'foo', test_script)
run_test_script(file_name)
if __name__ == "__main__":
unittest.main()