gh-154726: Fix shutil.copyfile() for symlinks to special files with follow_symlinks=False

[lkk7]

Fixes #154726.

• Avoid checking a source symlink's target when follow_symlinks=False
• Add a regression test using a symlink to /dev/null.

• Issue: shutil.copyfile() rejects device symlinks with follow_symlinks=False #154726

[aisk]

fifo / socket have the same problem, and I think they are already fixed by this PR. We should update the PR's title to mention this, like using 'special file' in the news entry.

And maybe we can update the tests to test socket and fifo to ensure this and prevent future regressions. But it's up to you.

[lkk7]

Hi @aisk, done.

I just noticed that one test is not grouped with the others, but maybe it's too cosmetic to run the whole CI again for that. If any other change is needed, then I'll push that too.

[bedevere-bot]

🤖 New build scheduled with the buildbot fleet by @encukou for commit bf7e3d2 🤖

Results will be shown at:

https://buildbot.python.org/all/#/grid?branch=refs%2Fpull%2F154728%2Fmerge

If you want to schedule another build, you need to add the 🔨 test-with-buildbots label again.

[miss-islington-app]

Thanks @lkk7 for the PR, and @encukou for merging it 🌮🎉.. I'm working now to backport this PR to: 3.15.
🐍🍒⛏🤖

[miss-islington-app]

Thanks @lkk7 for the PR, and @encukou for merging it 🌮🎉.. I'm working now to backport this PR to: 3.15.
🐍🍒⛏🤖

[miss-islington-app]

Sorry, @lkk7 and @encukou, I could not cleanly backport this to 3.15 due to a conflict.
Please backport using cherry_picker on command line.

cherry_picker f5afa895ae322ea89c5942b0b21094c124293dd8 3.15

[miss-islington-app]

Sorry, @lkk7 and @encukou, I could not cleanly backport this to 3.15 due to a conflict.
Please backport using cherry_picker on command line.

cherry_picker f5afa895ae322ea89c5942b0b21094c124293dd8 3.15

[encukou]

Oh, #142693 is only in 3.16.