Cleanup pass on _curses and _curses_panel

[larryhastings]

BPO	1019
Nosy	@gvanrossum, @akuchling, @birkenfeld, @larryhastings
Files	lch.py3k.curses.module.diff.1.txt

^{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 2010-02-22.23:49:17.948>
created_at = <Date 2007-08-25.11:25:16.725>
labels = ['extension-modules', 'type-bug']
title = 'Cleanup pass on _curses and _curses_panel'
updated_at = <Date 2010-02-22.23:49:17.947>
user = 'https://github.com/larryhastings'

bugs.python.org fields:

activity = <Date 2010-02-22.23:49:17.947>
actor = 'larry'
assignee = 'none'
closed = True
closed_date = <Date 2010-02-22.23:49:17.948>
closer = 'larry'
components = ['Extension Modules']
creation = <Date 2007-08-25.11:25:16.725>
creator = 'larry'
dependencies = []
files = ['8316']
hgrepos = []
issue_num = 1019
keywords = ['patch']
message_count = 4.0
messages = ['55287', '55435', '55866', '99878']
nosy_count = 4.0
nosy_names = ['gvanrossum', 'akuchling', 'georg.brandl', 'larry']
pr_nums = []
priority = 'low'
resolution = 'rejected'
stage = None
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue1019'
versions = ['Python 3.0']

[larryhastings]

*BEFORE* APPLYING THIS PATCH, YOU MUST
% svn move Include/py_curses.h Modules/_py_curses.h
(A "patch" can't convey that sort of information.)

I made a pass cleaning up _curses and _curses_panel. Mainly I cleaned
up the header file, moving unnecessary things into the two .c files.
I also changed the way information is passed from _curses to
_curses_panel. Instead of a Py_CObject which points to functions--which
Neil Norwitz suspected could be a security hole--it now passes cleaner
information (one type, one int * pointer). Finally, I made a
code-cleanup pass as per PEP-7.

[gvanrossum]

I tried to run the various demos in Demo/curses/ with the new version.
They all work except for xmas.py. Before your patch, they all ran.

So I cannot apply this yet. (No time to look into it further, alas.)

[birkenfeld]

Unfortunately, the patch is hard to review with all the whitespace
cleanup distracting from semantic changes. Can you produce a patch with
only those changes?

[larryhastings]

The major reason I undertook this patch was to fix the security hole of the CObject. That entire class of security holes is now fixed by the capsule object. The rest of the diffs aren't interesting enough to worry about. So I'm abandoning my patch.