v2.0.0rc1 on pip doesn't contain the necessary files.

[UpstreamData]

v2.0.0rc1 on pip doesn't contain the necessary files.

[exhuma]

rc5 is released. Manually downloading and inspecting the contents looks fine now.

[UpstreamData]

Solution suggested here, attempting to migrate this thread off of the original PR into this issue.

@GhostofGoes

#100 (comment)

[exhuma]

That's weird... I will look into this.

[UpstreamData]

That's weird... I will look into this.

Also, typing_extensions doesn't install when pip installing puresnmp, that may need to be added as well (although I know nothing about pip installations or PyPi)

[exhuma]

It's fine. I do have quite some experience with this and will figure it out. It is definitely not properly picking up the package files when building the dist. I'm looking into it as we speak :)

[UpstreamData]

It's fine. I do have quite some experience with this and will figure it out. It is definitely not properly picking up the package files when building the dist. I'm looking into it as we speak :)

Awesome, much appreciated. If you need anything tested or any help let me know, I'm happy to help/learn as needed, especially when I "cause" the issues haha.

[exhuma]

Pipeline is triggered: https://github.com/exhuma/puresnmp/actions/runs/2359564353

Once that is through, v2.0.0rc2 should appear on pypi momentarily. The package looks more correct now. Hope I didn't miss anything... fingers crossed...

[exhuma]

I missed the issue with the incorrect initialisation of the protocol class. This was introduced with Python 3.8 support yesterday. At work we're still on 3.6 so that did not surface yet.

I'm confident that these are just minor kinks to be ironed out with the changes I had to do for Python 3.8+.

2.0.0rc3 should be available soon with that fix.

Let me know if it's fixed or not.

[UpstreamData]

I missed the issue with the incorrect initialisation of the protocol class. This was introduced with Python 3.8 support yesterday. At work we're still on 3.6 so that did not surface yet.

I'm confident that these are just minor kinks to be ironed out with the changes I had to do for Python 3.8+.

2.0.0rc3 should be available soon with that fix.

Let me know if it's fixed or not.

It fixed the original issue, but now plugins/mpm.py is raising a new issue.

    raise UnknownMessageProcessingModel(
puresnmp.exc.UnknownMessageProcessingModel: Namespace 'puresnmp_plugins.mpm' did not contain a plugin with identifier 1. 
Known identifiers: []. 
See the 'puresnmp' documentation on plugins.

I can't find anything in a quick search of the docs about plugins and I don't think it should raise this on initial install?

[UpstreamData]

I missed the issue with the incorrect initialisation of the protocol class. This was introduced with Python 3.8 support yesterday. At work we're still on 3.6 so that did not surface yet.

FYI I am on python 3.10 right now, so I will likely have a LOT more issues than you, I will try to give as much info as possible on what I do encounter, but I could see this being troublesome.

[exhuma]

oh ffs... I'm really sorry about this. I'm currently not at the office and have neither a proper dev-setup here for real in-depth testing, nor access to a network device for testing. So it's a bit of a whack-a-mole at the moment until all of this is through.

Please bear with me 😅

[exhuma]

This does all look like a packaging issue though. I will see if I can repro this in the automated pipeline tests... this would definitely help.

[UpstreamData]

No stress @exhuma, it's not a major rush at all for me, long weekend coming up and I will likely not get anything done on this until Tuesday, and if all else fails I can guess pretty well about how I want to implement it, so I can always work on that.

[exhuma]

Thanks for the patience :)

[GhostofGoes]

Yup, rc5 seems to have resolved both the ModuleNotFoundError and UnknownMessageProcessingModel errors for me (Python 3.8). Thanks for the rapid fixes 😸

[UpstreamData]

Yep, seems to work now for me. Appreciated.