Command line evolution

[Mark]

This has been a long time coming! I'm excited to finally be working on the command line situation more formally. Attached you'll find a description of the current state of affairs, and the work I plan on doing in this area in the short term.

Happy to receive questions, comments and suggestions on this. And if you've made a command line application on Urbit before, let me know about your experience!

[Matilde Park]

“More formally” — as in, you’re doing terminal (RIP) userspace development full-time?

Curious about the webdojo changes. I’ve had them brought up to me before; you would open two different connections — a Dojo session and a websocket to the runtime? Or would we just pipe one to the screen and send the Dojo session sole-actions without rendering its response?

> Web dojo currently pretends to be a sole client. This works for rendering
> output, but makes input handling awkward (turning input into sole-actions,
> instead of passing it on as-is), and restricts it to a single application.

It’s a port of Anton’s former implementation, which itself let you use :talk, so I’m curious why you say that we lost that. I mean, I just didn’t implement that (...mine lets you paste tho), but I figured it was a front-end specific change we could’ve integrated. This implies it’s back-end not possible?

Honestly I’m excited to see the fruits of your terminal work; the “power user vim experience” for Urbit is quite overdue and I think in particular it’s the inability to edit Arvo source that feels like a gap.

—
~haddef-sigwen
https://urbit.org

> On Sep 18, 2020, at 4:51 PM, Mark <ma...@tlon.io> wrote:
>
> This has been a long time coming! I'm excited to finally be working on the command line situation more formally. Attached you'll find a description of the current state of affairs, and the work I plan on doing in this area in the short term.
>
> Happy to receive questions, comments and suggestions on this. And if you've made a command line application on Urbit before, let me know about your experience!
>

> <command-line-evolution.txt>
>
> —
> ~palfun-foslup
> https://urbit.org
>

[Mark]

> “More formally” — as in, you’re doing terminal (RIP) userspace development full-time?

I don't think current circumstances (or ~poldec) will let me do 100% full-time, but it's a management-approved part of my stack now, as opposed to something I hack on during weekends.

>> Web dojo currently pretends to be a sole client.

>

> It’s a port of Anton’s former implementation, which itself let you use :talk, so I’m curious why you say that we lost that. I mean, I just didn’t implement that (...mine lets you paste tho), but I figured it was a front-end specific change we could’ve integrated. This implies it’s back-end not possible?

Oh, I didn't mean to imply that. You're right that with the current frontend stack, you could add support for chat-cli. But what if we could reduce the amount of code in the frontend, while getting all of the functionality in drum (app linking/switching, ctrl-key shortcuts, etc) for free?

Right now the frontend reimplements a lot of drum logic, around mirroring sole state, tracking command history etc. If we connect lower down the stack, all of that stuff happens above us, and we only have to concern ourselves with primitive input handling & rendering. The result is increased parity with the terminal experience at lower cost.

Thinking this over again, maybe we want to connect to drum instead of dill? It speaks dill- belts and blits... Might be even easier than hooking up to dill itself.

—
~palfun-foslup
https://urbit.org