gh-99892: test_unicodedata: skip test on download failure

[vstinner]

Skip test_normalization() of test_unicodedata if it fails to download NormalizationTest.txt file from pythontest.net.

• Issue: test_unicodedata: test_normalization() fails randomly with IncompleteRead on PPC64LE Fedora buildbots #99892

[miss-islington]

Thanks @vstinner for the PR 🌮🎉.. I'm working now to backport this PR to: 3.10, 3.11.
🐍🍒⛏🤖

[miss-islington]

Sorry @vstinner, I had trouble checking out the 3.11 backport branch.
Please retry by removing and re-adding the "needs backport to 3.11" label.
Alternatively, you can backport using cherry_picker on the command line.
cherry_picker 2488c1e1b66366a3a933ff248eff080fabd2351c 3.11

[bedevere-bot]

GH-100012 is a backport of this pull request to the 3.10 branch.

[miss-islington]

Thanks @vstinner for the PR 🌮🎉.. I'm working now to backport this PR to: 3.11.
🐍🍒⛏🤖

[bedevere-bot]

GH-100013 is a backport of this pull request to the 3.11 branch.

[benjaminp]

Why not add retries to the fetch? I'm afraid this will silently disable the test forever if fetching the file becomes a permanent error.

[vstinner]

Hi @benjaminp,

Why not add retries to the fetch?

My reason: I want to limit time spent on this issue :-) I only worked on it because I'm getting 1 to 5 emails per day related to test_unicodedata failures, but only on some workers which have other network issues. If someone else is volunteer to enhance the code, please go ahead ;-) A generic function "download a file from the Internet, retry on error, and skip the test after a timeout" sounds like an useful feature ;-)

[benjaminp]

I very much understand not wanting to spend time. But, as I said, I'm afraid this could have scary side-effects.