project requirements

@nicoalt What would be your intended target interpreters?

With briar_repl I targeted Python3.6 and above, to enjoy f-strings and the asyncio syntax.

With the python death clock https://pythonclock.org/ in mind I would not consider python2 at all.

But we could also target Python3.7 which is standard in debian buster and bring the amazing dataclasses.

What are your thoughts on this?

added discussion label

From my perspective, there's absolutely no need to support Python 2

Regarding Python 3, I never found myself in the position to having to decide between 3.6 and 3.7. If it's standard in Debian stable, I'm pretty sure we can make it a requirement. Feel free to decide this yourself!

(For a general perspective: for me, Briar is a long-term project and I think we can choose reasonably "new" dependencies now because in the long-term they will become standard.)

I just skimmed over The Ultimate Guide to Data Classes in Python 3.7 by Real Python and wow, this is really cool. Do you plan to use this for storing objects like messages and contacts? That's the first thing coming to my mind. I think we can absolutely require Python 3.7 for these features, too.

(And wow, f-strings are also really cool! Thanks for introducing me to all this awesome stuff!)

Great - let's make it 3.7+ then!!

Totally agree - dataclasses and f-strings are really awesome!! Especially f-strings are very addictive!

Indeed I thought about maybe using it for contacts, or even just the url collection we have to talk to the api - it would provide nice dot-notation:

urls.ws or urls.contacts

closed

Indeed I thought about maybe using it for contacts, or even just the url collection we have to talk to the api - it would provide nice dot-notation:

urls.ws or urls.contacts

Let's do this and make this wrapper library awesome