Add sanity-check else case to regrtest option parsing

[bitdancer]

BPO	7324
Nosy	@brettcannon, @pitrou, @bitdancer
Files	regrtest-detect-bad-option.patch

^{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/bitdancer'
closed_at = <Date 2009-11-15.00:41:56.811>
created_at = <Date 2009-11-14.18:43:09.458>
labels = ['type-bug', 'tests']
title = 'Add sanity-check else case to regrtest option parsing'
updated_at = <Date 2009-11-15.00:41:56.809>
user = 'https://github.com/bitdancer'

bugs.python.org fields:

activity = <Date 2009-11-15.00:41:56.809>
actor = 'r.david.murray'
assignee = 'r.david.murray'
closed = True
closed_date = <Date 2009-11-15.00:41:56.811>
closer = 'r.david.murray'
components = ['Tests']
creation = <Date 2009-11-14.18:43:09.458>
creator = 'r.david.murray'
dependencies = []
files = ['15330']
hgrepos = []
issue_num = 7324
keywords = ['patch']
message_count = 4.0
messages = ['95249', '95254', '95255', '95267']
nosy_count = 3.0
nosy_names = ['brett.cannon', 'pitrou', 'r.david.murray']
pr_nums = []
priority = 'low'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue7324'
versions = ['Python 2.7', 'Python 3.2']

[bitdancer]

In forward porting a patch to py3k I noticed that there is a 'g' option
in the optparse argument list in regrtest in 2.x that is not present in
3.x. But the surprising thing was that there are no docs for this
option, nor any option handler in the 2.x regrtest.

I propose to add an 'else' case to the option parsing loop that asks the
user to report a bug if it is handed an unknown option. Patch attached.

My one question is whether this might have been intentional for backward
compatibility reasons: let -g be passed and ignore it silently. I'm
guessing it was just a deletion oversight, though.

[bitdancer]

That should have been 'getopt option list'.

[brettcannon]

I bet it was an option oversight. Since regrtest is an internal tool we
don't really need to fret about backwards-compatibility for anyone.

[bitdancer]

Committed in r76276 through r76281, along with removing 'g' from the
getopt list in 2.6. 3.1 still has other traces of the -g option; I
haven't cleaned that up.