list/generator comprehension parser doesn't match spec

[kousu]

BPO	2529
Nosy	@amauryfa, @kousu

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2008-04-01.20:56:04.957>
created_at = <Date 2008-04-01.19:19:09.269>
labels = ['interpreter-core', 'type-bug', 'invalid']
title = "list/generator comprehension parser doesn't match spec"
updated_at = <Date 2008-04-02.15:33:02.236>
user = 'https://github.com/kousu'

bugs.python.org fields:

activity = <Date 2008-04-02.15:33:02.236>
actor = 'kousu'
assignee = 'none'
closed = True
closed_date = <Date 2008-04-01.20:56:04.957>
closer = 'amaury.forgeotdarc'
components = ['Interpreter Core']
creation = <Date 2008-04-01.19:19:09.269>
creator = 'kousu'
dependencies = []
files = []
hgrepos = []
issue_num = 2529
keywords = []
message_count = 4.0
messages = ['64818', '64819', '64823', '64861']
nosy_count = 3.0
nosy_names = ['amaury.forgeotdarc', 'kousu', 'lehmannro']
pr_nums = []
priority = 'normal'
resolution = 'not a bug'
stage = None
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue2529'
versions = ['Python 2.5']

[kousu]

I think I've found a bug in python's list comprehension parser. Observe:

>>> [e for i in j in ['a','b','c']]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'j' is not defined

Now, according to the grammar at http://docs.python.org/ref/lists.html,
a list comprehension is (condensed for clarity):
list_comprehension ::= expression list_for
list_for ::= "for" target_list "in" old_expression_list [list_for]

So a list comprehension should always be
[.... for ... in .... for ... in ... for ... in ...]
(that is, alternating 'for's and 'in's) but here I have a test case that
python happily tries to run that looks like
[... for ... in ... in ....]

[lehmannro]

Your example is parsed as [e for i in (j in ['a','b','c'])] and since
j is not defined, you get a NameError. If it was defined, you would
still be iterating a boolean (which is not defined).

Grammatically, this is the following (just the important parts, again):
list_comprehension ::= expression list_for
list_for ::= "for" target_list "in" old_expression_list
old_expression_list ::= old_expression
old_expression ::= <stripped test hierarchy...> comparison
comparison ::= or_expr ( comp_operator or_expr )*
comp_operator ::= "in"

So your basic misconception is that both in keywords are belonging to
the list comprehension syntax -- the former does while the latter is
simply an operator.

[amauryfa]

Indeed; your sample is equivalent to:

temp = (j in ['a','b','c'])        # the "contains" operator
[e for i in temp]                  # basic list comprehension

Even if not meaningful, this code is syntactically correct.

[kousu]

Oh, okay. That's really confusing because I expect "in" to always return
a bool, but in the spirit of python there's no reason it should I guess.