gh-148508: Add resilience to SSL preauth tests on iOS

[freakboy3742]

The test_ssl.TestPreHandshakeClose.test_preauth_data_to_tls_server test fails intermittently on the iOS buildbot because an SSL error doesn't raise the expected TLS handshake error.

This test already has an escape clause that allows a skip under certain error conditions (raising a ConnectionError, an OSError, or an SSL error with "wrong version number"). Running locally, I reliably hit the OSError handling case; the buildbot also hits the OSError case most of the time, but sometimes it raises SSLError with "[SSL: RECORD_LAYER_FAILURE] record layer failure".

This adds a check for "record layer failure" as an additional failure mode that will be accepted.

• Issue: Intermittent iOS buildbot failure in test_preauth_data_to_tls_server #148508

[freakboy3742]

Tagging @vstinner for review on this since you added the "non_linux_skip_if_other_okay_error" handling.

[freakboy3742]

!buildbot iOS

[bedevere-bot]

🤖 New build scheduled with the buildbot fleet by @freakboy3742 for commit 37b278f 🤖

Results will be shown at:

https://buildbot.python.org/all/#/grid?branch=refs%2Fpull%2F148536%2Fmerge

The command will test the builders whose names match following regular expression: iOS

The builders matched are:

• iOS ARM64 Simulator PR

[gpshead]

could use a (a|b) regex instead, works either way. this is readable.

[miss-islington-app]

Thanks @freakboy3742 for the PR 🌮🎉.. I'm working now to backport this PR to: 3.13, 3.14.
🐍🍒⛏🤖

[bedevere-app]

GH-148539 is a backport of this pull request to the 3.14 branch.

[bedevere-app]

GH-148540 is a backport of this pull request to the 3.13 branch.

[vstinner]

Tagging @vstinner for review on this since you added the "non_linux_skip_if_other_okay_error" handling.

I only made a small change to break reference cycles. non_linux_skip_if_other_okay_error() was added by @ambv with the commit 0cb0c23.

This function is quite complicated. I don't understand well why OpenSSL behaves so differently depending on the platform. I'm glad that @gpshead reviewed this change :-)