bpo-34735: Fix a memory leak in Modules/timemodule.c

[ZackerySpytz]

The format string was not being freed when ValueError was raised for %y on certain platforms.

[pganssle]

The actual fix here looks good to me, but I am fairly sure that this code is not hit by any of the tests, so I recommend adding a test that hits there.

I think adding this to Lib/test/datetimetester.py under TestDate should work:

@unittest.skipif(not (sys.platform.startswith('aix') or
                      sys.platform.startswith('win'))
                 "%y on Windows and AIX doesn't support years < 1900")
def test_strftime_aixwin_pre1900(self):
    d = self.theclass(1, 1, 1)

    self.assertRaises(ValueError, d.strftime, '%y')

[ZackerySpytz]

@pganssle This code is hit in test_y_before_1900() in Lib/test/test_strftime.py.

[pganssle]

@ZackerySpytz Ah, tricky tricky. There's a bunch of strftime-related tests in test_time and datetimetester so I didn't notice test_strftime.py. My bad.

[brettcannon]

Thanks for the patch, @ZackerySpytz ! I'm not sure if my label adding and removing messed up our auto-merge. If it did I will try to merge this tomorrow.

[brettcannon]

merge approved PR

[miss-islington]

Thanks @ZackerySpytz for the PR, and @serhiy-storchaka for merging it 🌮🎉.. I'm working now to backport this PR to: 3.6, 3.7.
🐍🍒⛏🤖

[bedevere-bot]

GH-9469 is a backport of this pull request to the 3.7 branch.

[bedevere-bot]

GH-9470 is a backport of this pull request to the 3.6 branch.

[ZackerySpytz]

Thank you, @brettcannon and @serhiy-storchaka.