JunoLab / atom-ink

IDE toolkit for Atom
MIT License
228 stars 40 forks source link

Console Improvements #54

Closed pfitzseb closed 6 years ago

pfitzseb commented 8 years ago
MikeInnes commented 8 years ago

Do we have a reasonable path to implementing the un-docking stuff? It seems to me like this basically involves implementing the feature in general for Atom; we'd need to figure out how it should work for general panes, move over CSS and keybindings and so on. I have another potential solution for the julia-client issue but will mention it there.

MikeInnes commented 8 years ago

Still can't overwrite stuff, but still.

At the moment I basically see interpreting console control codes as a total non-goal, not least because the use cases (e.g. progress bars) are things we can do better graphically. Is that way off base?

pfitzseb commented 8 years ago

Well, I thought about opening a new Electron window and connecting that to the Julia server once we have that (and the server then "broadcasts" STDOUT to all clients or something like that) -- didn't think about the dependencies though...

About the console stuff: Yeah, I agree, we shouldn't jump through too many hoops. I'm mostly just annoyed with not being able to print pretty stuff to the console because I get funky linebreaks everywhere. On a related note: That hack in julia-client where you hide the require(Symbol) method overwrite doesn't work for me, because that line gets split into multiple cells on my machine :D

MikeInnes commented 8 years ago

Oh yeah, well handling STDOUT is yet another issue to think about...

So the newline stuff is still affecting you quite badly then? It's been working great for me, but yeah I guess we do need an approach for that.

pfitzseb commented 8 years ago

console Yup ;)

petertuuk commented 7 years ago

I'll bump the multi-window support topic. With Julia 0.6 and then 1.0 coming, updates to the UI could be well-timed for return on investment.

jmboehm commented 6 years ago

I would prefer the line in the console where one enters text to be 'separate' from the output (in other words, it should remain visible when I scroll up). That way one does not always need to scroll down to type a new command. Of course tastes may differ.