IndentationError for unexpected indent

[rcmiller]

BPO	1475845
Nosy	@loewis, @birkenfeld
Files	patch_indent_error: patch for IndendationError

^{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 = 'https://github.com/loewis'
closed_at = <Date 2006-05-04.05:51:45.000>
created_at = <Date 2006-04-24.23:12:04.000>
labels = ['interpreter-core']
title = 'IndentationError for unexpected indent'
updated_at = <Date 2006-05-04.05:51:45.000>
user = 'https://bugs.python.org/rcmiller'

bugs.python.org fields:

activity = <Date 2006-05-04.05:51:45.000>
actor = 'loewis'
assignee = 'loewis'
closed = True
closed_date = None
closer = None
components = ['Interpreter Core']
creation = <Date 2006-04-24.23:12:04.000>
creator = 'rcmiller'
dependencies = []
files = ['7198']
hgrepos = []
issue_num = 1475845
keywords = ['patch']
message_count = 3.0
messages = ['50106', '50107', '50108']
nosy_count = 3.0
nosy_names = ['loewis', 'georg.brandl', 'rcmiller']
pr_nums = []
priority = 'normal'
resolution = 'accepted'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue1475845'
versions = ['Python 2.5']

[rcmiller]

This patch raises an IndentationError rather than a
generic "invalid syntax" error for unexpected
indentation. Code to do this was already in
pythonrun.c:err_input() but was not being reached due
to a failure to pass the INDENT token in the perrdetail
structure. The patch also adds tests for the 3 kinds
of indentation errors (unexpected indent, no indent
where required, invalid outdent level) to test_syntax.py .

[birkenfeld]

Logged In: YES
user_id=849994

Martin, do we want to change this? I myself have always
wondered what IndentationError was for if it was not raised
in these cases.

[loewis]

Logged In: YES
user_id=21627

IndentationError is already raised for bad indentation, e.g.
for

"def f():\nreturn"

or

if 1:\nfoo()" (which is the test_no_indent)

However, the patch is right in filling the token in this
case, also; I accepted it as r45897. As it changes the
exceptio behaviour, I don't think it should be backported.