Repository navigation
Refactor str tests to reflect that str and unicode are merged in Python 3 #81005
Description
| BPO | 36824 |
|---|---|
| Nosy | @ncoghlan, @serhiy-storchaka, @asqui, @tirkarthi |
| PRs |
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 = None
created_at = <Date 2019-05-07.06:37:53.164>
labels = ['3.8', 'type-feature', 'tests']
title = 'Refactor str tests to reflect that str and unicode are merged in Python 3'
updated_at = <Date 2019-05-07.20:05:56.998>
user = 'https://github.com/asqui'bugs.python.org fields:
activity = <Date 2019-05-07.20:05:56.998>
actor = 'dfortunov'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Tests']
creation = <Date 2019-05-07.06:37:53.164>
creator = 'dfortunov'
dependencies = []
files = []
hgrepos = []
issue_num = 36824
keywords = ['patch']
message_count = 4.0
messages = ['341685', '341687', '341689', '341809']
nosy_count = 4.0
nosy_names = ['ncoghlan', 'serhiy.storchaka', 'dfortunov', 'xtreak']
pr_nums = ['13172']
priority = 'normal'
resolution = None
stage = 'patch review'
status = 'open'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue36824'
versions = ['Python 3.8']
Activity
asqui commentedon May 7, 2019
Unit tests of
strand related types (e.g.UserString) contain nomenclature and structure that dates back to the Python 2 distinction betweenstrandunicode.Previously it was undesirable to disturb the structure of these tests too much as this would complicate the merging of bug fixes back to Python 2, however with Python 2 drawing near to end-of-life this is perhaps less of a concern.
I would propose the following changes as a start:
UnicodeTestclass toStrTest(to reflect the type that is now being tested)
MixinStrUnicodeUserStringTestclass and move its tests intoCommonTest(The comment for this class says
additional tests that only work for
stringlike objects, i.e. str, UserString
but in the absence ofunicodetheCommonTestclass is also only used for these two types now.)
UnicodeTestclass toCommonTestwhere it makes sense to do so; remove checks that no longer make sense.
(e.g. checks around mixed
str/unicodearguments that are all juststrnow)Maybe the duplicative
checkequalnofix()method can also go away now?BadSeq1helper class is not used because the corresponding check wascommented out in 2007. Either reinstate the check, or remove this class.
I'm happy to submit a PR for this, but just wanted to get some feedback before making the changes.
serhiy-storchaka commentedon May 7, 2019
There is still the functionality common to str, bytes and bytearray. So it make sense to have the common class for testing str, bytes, bytearray and UserList and other common class (subclass of the former one) for testing str and UserList.
asqui commentedon May 7, 2019
Agreed. This functionality is in
BaseTest(which is the base forCommonTest) and I don't propose to change this.asqui commentedon May 7, 2019
PS opened here: #13172
I've tried to break down the changes into individual steps, with justification in commit messages. Happy to collapse these down into fewer commits before merge if preferred.
I haven't done the "Promote tests from the current
UnicodeTestclass toCommonTest" portion yet, but I'm running out of sprint time so I wanted to submit what I have.I believe these changes are suitable for merge to master, but keen to hear feedback :-)
gh-81005: Refactor str tests to reflect that str and unicode are merg…