publish wheels

[dimbleby]

publish wheels

rather than make every consumer build the wheel themselves, publish it once and for all

[robsdedude]

Thanks for taking the time to craft a PR.

As you can see on PyPI the wheel for neo4j are already published. For the legacy package neo4j-driver I didn't go for uploading wheels for these reasons:

• Support for neo4j-driver will be dropped at the end of version 5. neo4j is a package alias and can be swapped in without code changes. So we expect users to migrate to the new package name. This brings me to point 2:
• Less comforts can increase migration velocity.
• Finally, the installation of the legacy package emits a deprecation warning. This isn't possible with pre-build wheels. Granted, pip doesn't show the warning. However for users running their env setup (for instance in CI) with -W error, the installation will fail and let them know that they should change the name of the package.

[dimbleby]

I don't find any of that convincing to be honest, but I don't care to argue about it - this isn't very important.

If you want a deprecation warning that people will actually see, perhaps putting it in __init__.py so that it is shown on package import would make sense.

[robsdedude]

If you want a deprecation warning that people will actually see, perhaps putting it in init.py so that it is shown on package import would make sense.

That's in place as well 😅

neo4j-python-driver/src/neo4j/__init__.py

Lines 232 to 238 in 2a1a772

	if _deprecated_package:
	_deprecation_warn(
	"The neo4j driver was installed under the package name `noe4j-driver` "
	"which is deprecated and will stop receiving updates starting with "
	"version 6.0.0. Please install `neo4j` instead (which is an alias, "
	"i.e., a drop-in replacement). See https://pypi.org/project/neo4j/ ."
	)

I don't find any of that convincing to be honest, but I don't care to argue about it - this isn't very important.

I'm still curious to hear your view. Do you have or see a use-case where switching over to neo4j is not an option?

[dimbleby]

I am not a user of neo4j and have no views on switching to or from one package or the other.

I just spotted that this package is published without a wheel, that is usually undesirable, and so I propose to fix it.

I am sceptical that the minor barrier of having users build their own wheel is doing anything at all to help your deprecation, I would guess it it is more likely just making life ever so slightly worse for the people who are not yet ready to move

But as I say this is not important - "minor" and "ever so slightly worse" by my own reasoning! If you are deliberately choosing not to publish a wheel and think that is right for you - then I am happy to respectfully disagree and leave it there.

[robsdedude]

Thanks for you for taking the time to share your view. After thinking about it for longer, I came to the conclusion that you are right.

Before I can review the PR, however, I have to ask you to sign our Contributor License Agreement (CLA): https://neo4j.com/developer/cla/

[dimbleby]

I will not be signing the contributor license agreement. You are welcome to this change if you want it.

[robsdedude]

Thanks 😅 I was almost expecting such reply. I will close this PR and take care of publishing the wheels myself. It's not a big change anyway.

Again: thank you very much for taking the time to open the PR and to discuss the matter with me. This is much appreciated 🙇