I like Python, but the using the REPL w/ Python's indentation is not fun for me.

I'd venture a guess that anyone advocating REPL-based development in Python is pitching IPython/Jupyter. IPython's REPL supports multi-line editing, trims common leading whitespace from pasted input, among a whole host of other features. The standard REPL is definitely painful for all but the most basic tasks.

That's why you write code in your editor, and just ask your editor to send code to the REPL.

Why not just run the file? To be more specific, I get a lot of value out of Ruby, Elixir, and JavaScript REPLs because I can type right into them to try stuff out in one long, ridiculous command. If I have to type in a file, I'm just gonna run the file & print the result.

>Why not just run the file?

It's faster. That's the point of using a REPL.

>To be more specific, I get a lot of value out of Ruby, Elixir, and JavaScript REPLs because I can type right into them to try stuff out in one long, ridiculous command.

You can type long, ridiculous commands in a file too. Just clean it up as you go along.

I like to use my editor to send inputs to the REPL instead of typing directly into the REPL because different REPLs provide different editing experiences. Using my editor means all the editing features I like to use are always available.

Oh dear lord yes. Autoreload is my saviour in many respects.

I suppose that I come at this from R, where REPL (sortof) driven programming/analysis is a core feature, and until I discovered %autoreload I was feeling very left out in Python.

I guess I sortof understand why it is that way, but it definitely struck me as weird that one wouldn't allow code to be reloaded into a process buffer.

Serious question: Is there a way to use a Python REPL with multiple long-running asyncio tasks, some of which need a turn to run every X seconds?