GeezerVison

I didn't expect that finding the location of the Python source code and checking if it's actually available or not would be so complicated!

At least, it's good, we are finding and fixing issues :-)

For now, I'm ok to keep the code in test_support.py, but later we might share more code in support\__init__.py.

test_freeze and test_venv copy the stdlib. test_venv only cares about Python code, whereas test_freeze wants the C code.

The last change makes test_copy_python_src_ignore fail rather than skip on Windows installs,

Oh. How did you notice that? Is Azure Pipelines running tests which such use cases, or did you discover the issue manually?

I didn't expect that finding the location of the Python source code and checking if it's actually available or not would be so complicated!

Yeah, there's a reason I listed this on my wishlist for a new build system. Would be great if we just always built an install-like layout, rather than trying to intermingle it with the source tree.

The pybuilddir.txt file should always be there in a build, even on Windows, adjacent to the main executable (approximately, os.path.realpath(sys.orig_argv[0])). The official check is in getpath.py, and it sets the _is_python_build config value. I forget where that ends up precisely.

Your PR now includes unrelated change.

Nope. That's what I get for trying to fix OpenSSL stuff 😆

Oh. How did you notice that? Is Azure Pipelines running tests which such use cases, or did you discover the issue manually?

It should, but they seem to have been passing for some reason. I haven't looked closely, but possibly they got disabled recently because people want to move to GHA.

I also run my own daily tests at work, and it was found in the "layout test", which generates an install layout (python PC/layout --preset-default --copy <DIR>), copies it to a brand new machine, and then runs the test suite. That was where I noticed it first.

Thanks @zooba for the PR 🌮🎉.. I'm working now to backport this PR to: 3.11, 3.12.
🐍🍒⛏🤖

Sorry, @zooba, I could not cleanly backport this to 3.12 due to a conflict.
Please backport using cherry_picker on command line.
cherry_picker 20bc5f7c28a6f8a2e156c4a748ffabb5efc7c761 3.12

GH-110110 is a backport of this pull request to the 3.11 branch.

I'm guessing 3.12 has conflicts because the earlier PRs haven't been merged yet? I'll wait until that's done and then trigger it again

Sorry, @zooba, I could not cleanly backport this to 3.12 due to a conflict.

Last days, I gave up on backports since 3.12 is blocked. I plan to try to backport my Tests and Docs changes once 3.12 will be reopened again next week. Until that, it's a backport nightmare :-(

Thanks @zooba for the PR 🌮🎉.. I'm working now to backport this PR to: 3.12.
🐍🍒⛏🤖

Sorry, @zooba, I could not cleanly backport this to 3.12 due to a conflict.
Please backport using cherry_picker on command line.

cherry_picker 20bc5f7c28a6f8a2e156c4a748ffabb5efc7c761 3.12

Ah, this fix was backported to 3.12 as part of larger backport: commit 80f9585.