New Python release to avoid Deprecation Warning

[MarcoGorelli]

The latest Python release shows a DeprecationWarning:

E   DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses

This was fixed in #6362 - any chance a new release could be made, so the warning will go away?

---

Edit: apologies for having asked, I now understand that it's rude to ask for new releases

[aardappel]

We don't typically make releases just because something improved for one language, we'd end up with a release each commit :)

I recommend using the latest master instead of a last release.

[MarcoGorelli]

sure, thanks

[cool-RR]

@aardappel It's been 6 months, and this DeprecationWarning is still a nuisance. Could you make a new release soon please?

[aardappel]

@cool-RR see my last comment.

[johnthagen]

The imp module will be removed in Python 3.12.

Given that Python 3.11 recently entered beta, this means that very soon flatbuffers may not work on the latest Python development branch.

[teunFTW]

@aardappel I agree with with @cool-RR that it would be really nice if a new release could be made soon! :)

[oscarbenjamin]

It's been over 12 months since #6362 was merged. Are there any plans to make a release soon?

[NeilGirdhar]

@aardappel

We don't typically make releases just because something improved for one language, we'd end up with a release each commit :)

Your point is reasonable. However, no one is asking for a release for each commit. I think it's more than fair to expect a release every year to solve deprecation warnings.

We do want to be respectful of your valuable time. However, I invested my time in the pull request in the hopes that it would solve this problem for me and other users. #6362

I recommend using the latest master instead of a last release.

That's not a reasonable suggestion since pinning a requirement to master is bad practice.

[aardappel]

That's not a reasonable suggestion since pinning a requirement to master is bad practice.

I didn't get that memo.

Either way, up to @dbaileychess when a new release happens.

[NeilGirdhar]

I didn't get that memo.

The reason pinning to master is usually considered bad is because it exposes you to untested combinations. How should you know that a future version master is compatible with your current version? You can't. Typically, you pin to a minor version, with reasonable assurance that everything will keep working within the same minor version.

Another problem with pinning to master is that it is three times the work for everyone. In this case, we would have to convince tensorflow to pin flatbuffers to master, and wait for their release. And then when flatbuffers releases tensorflow will have to restore the flatbuffers pin. That's three steps instead of flatbuffers releasing.

[tbenthompson]

I get a DeprecationWarning whenever I import jax. The underlying reason is that this release hasn't happened yet. I would be very grateful for a release!