For my own edification, have we ever, or would we ever consider dropping Pants support of Python 2? I've always been musing using LibCST to parse the AST for dep inference as it is much more powerful/robust. But it's Python 3 only. We could use different parsers for Py2 v Py3, but documentation+maintanance of that seems ripe for staleness/errors

I've encouraged it for a while. Python 2 users can still use Pants 2.15 for example

Maybe we ought to get this on the meetup? Then next step would likely be a public Discussion?

I'd be down for that. I think the most helpful thing would be identifying how it helps the project. Dropping Py2 support only for the sake of dropping Py2 support is capricious if we don't have actual benefits

Are there other places Py2 support is holding us back? If not, I don't think this alone is a forcing function. We could freeze Py2 dep inference, or offload the inference onto users

Not that I can think of. The main thing is Pex code still has to work with Python 2, although John has been okay with that in the past to help out users. And he wouldn't want to drop Py2 support until Pex 3.0

Definitely something to discuss and socialize at length

Hmm maybe a more fruitful discussion would be how we can expand our telemetry

The telemetry that very few opt in to...

I opt out because I have no clue what it does. It might just be a matter of user expectation

We document it, but yeah

In fact we can have Pants itself pop up a question about that!

Perhaps one thing to do is add a question to the Community Roadmap Survey this year, so we can begin to establish for the sake of the roadmap whether Python 2 support is something the community wants prioritized? If not, cool beans.

The only recent almost use case I am aware of was from Foursquare. Pants 2.7 support could have helped them transition to Pants v2. They migrated to Bazel instead though and sucked up migrating from Python 2.7 to 3 1st IIUC.

@dry-analyst-73584 might have some light in that domain 😛

Dark Sky or Black Sky?