[issue23136] BUG in how _strptime() handles week 0

Jim Carroll added the comment: Thanks for all your work ! -- ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: ht

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: Probably not. -- ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https:

[issue23136] BUG in how _strptime() handles week 0

Serhiy Storchaka added the comment: Is it possible to fix the commit message without breaking any of cloned repositories that already have pulled it? -- ___ Python tracker ___ _

[issue23136] BUG in how _strptime() handles week 0

Roundup Robot added the comment: New changeset d9f3dd1d6b48 by Serhiy Storchaka in branch '2.7': Fixed Misc/NEWS entry for issue #23136. https://hg.python.org/cpython/rev/d9f3dd1d6b48 New changeset afca9867e1bb by Serhiy Storchaka in branch '3.4': Fixed Misc/NEWS entry for issue #23136. https://

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: I don't think it is worth the effort to try to fix the commit message, but it would be nice to have the NEWS text corrected. -- ___ Python tracker __

[issue23136] BUG in how _strptime() handles week 0

Serhiy Storchaka added the comment: Oh, I meant Dec 30. -- ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: http

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: "including Jan 30" - did you mean Dec 31? -- ___ Python tracker ___ ___ Python-bugs-list maili

[issue23136] BUG in how _strptime() handles week 0

Serhiy Storchaka added the comment: Thank you for the review Alexander. Thank you for your contribution Jim. -- resolution: -> fixed stage: commit review -> resolved status: open -> closed ___ Python tracker _

[issue23136] BUG in how _strptime() handles week 0

Roundup Robot added the comment: New changeset 350ea309953e by Serhiy Storchaka in branch '2.7': Issue #23136: _strptime now uniformly handles all days in week 0, including https://hg.python.org/cpython/rev/350ea309953e New changeset 93e1da502338 by Serhiy Storchaka in branch '3.4': Issue #23136

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: Serhiy, Your latest patch LGTM. -- stage: patch review -> commit review type: -> behavior ___ Python tracker ___ ___

[issue23136] BUG in how _strptime() handles week 0

Changes by Serhiy Storchaka : Added file: http://bugs.python.org/file37674/strptime_julian_none_2.patch ___ Python tracker ___ ___ Python-bugs

[issue23136] BUG in how _strptime() handles week 0

Jim Carroll added the comment: All the proposed patches are acceptable from my point of view, but it would be really great if we could get the fix in the 2.x line. It seems unlikely there could be any legacy code that would depend on the bug to exist (ie: to only be wrong when then date reque

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: Note that round-trip also fails in weeks 52-53. For example, >>> datetime.strptime('2014 53 6', '%Y %W %w') datetime.datetime(2015, 1, 10, 0, 0) >>> datetime.strptime('2014 53 6', '%Y %W %w').strftime('%Y %W %w') '2015 01 6' If we decide to make "2015 0

[issue23136] BUG in how _strptime() handles week 0

Changes by Alexander Belopolsky : -- nosy: +Erik.Cederstrand, aganders3 ___ Python tracker ___ ___ Python-bugs-list mailing list Unsub

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: Also, given that "2015 0 0", "2015 0 1", and "2015 0 3", all currently work with '%Y %W %w' format specification and produce dates in the year 2014, I think raising an error in all those cases is more likely to break user code than correcting the value

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: I would prefer to accept "denormalized" "%Y %W %w" combinations. Note that while Mac OSX and glibc versions of strptime produce different results, neither implementation signals an error by returning NULL. > In C the strptime function doesn't return val

[issue23136] BUG in how _strptime() handles week 0

Saimadhav Heblikar added the comment: >>Not necessarily. String to datetime mapping implemented by strptime can be >>many to one. For example, I was referring to C strptime and strftime. But thanks for posting the round trip example. I was unaware of it. --

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: > For any valid string, strptime followed by strftime should return > the same string. Not necessarily. String to datetime mapping implemented by strptime can be many to one. For example, >>> datetime.strptime("2014 366", "%Y %j") == datetime.strptime(

[issue23136] BUG in how _strptime() handles week 0

Jim Carroll added the comment: I understand. Actually, raising an exception would be perfectly acceptable as well (possibly even more desirable). I too experimented with the c-lib strptime() and discovered the negative values of tm_mday. These results are good too -- as they clearly show the

[issue23136] BUG in how _strptime() handles week 0

Serhiy Storchaka added the comment: Here is alternative patch which rejects week-weekday combinations out of specified year. -- Added file: http://bugs.python.org/file37575/strptime_check_valid_week.patch ___ Python tracker

[issue23136] BUG in how _strptime() handles week 0

Changes by Tomasz Ryczkowski : -- nosy: +torm ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.pyth

[issue23136] BUG in how _strptime() handles week 0

Serhiy Storchaka added the comment: Thank you Jim for your patch. That all right and the patch fixes the issue. But the code looks a little tricky, and it would be more robust to use other signal value for julian (e.g. None). However I'm not sure that strptime should accept weekdays before the

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: See also #12006. -- ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: htt

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: In a comment for #23134, Serhiy suggested that this may not be a bug (msg233215) because C strptime behavior for this case is undefined. I disagree. On Mac OSX, strptime man page states: """ The %U and %W format specifiers accept any value within t

[issue23136] BUG in how _strptime() handles week 0

Saimadhav Heblikar added the comment: Hi, Here is my understanding of it. I have used the following C example of strptime and strftime to show that the string '0 2015 2' along with the format specifier '%W %Y %w' is invalid. For any valid string, strptime followed by strftime should return th

[issue23136] BUG in how _strptime() handles week 0

Alexander Belopolsky added the comment: What's a New Year without fixing a calendar bug! :-) Jim's logic looks right to me. The patch needs a test before it can be applied. -- ___ Python tracker _

[issue23136] BUG in how _strptime() handles week 0

Changes by Saimadhav Heblikar : -- nosy: +sahutd ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.p

[issue23136] BUG in how _strptime() handles week 0

Changes by Antoine Pitrou : -- nosy: +serhiy.storchaka ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://

[issue23136] BUG in how _strptime() handles week 0

Changes by Antoine Pitrou : -- nosy: +belopolsky, lemburg stage: -> patch review versions: +Python 3.4, Python 3.5 ___ Python tracker ___ ___

[issue23136] BUG in how _strptime() handles week 0

New submission from Jim Carroll: The following bit of code demonstrates a bug in how _strptime handles week 0. The bug is an edge condition that specifically affects how it handles two days before the first day of the new year >>> for i in range(7): ... datetime.strptime('%s %s %s' % (0, 2