Have you all seen this announcement today about another cloud implementation of VS Code?
Admittedly, it seems further along than Theia. I’ve just started looking through the repo but Theia’s progress could be pushed forwards by months or a year if some parts of the codebase were merged. Thankfully Coder open-sourced their efforts.
For any Theia devs, what downsides do you see to their implementation vs Theia in it’s current state?
[original thread by Josh Bradley]
They did not implemen it, but patch VS Code source base to make it run in the browser somehow. I don’t have an idea how i would extend it if i want to change somehing. Apply my patch on the top of their patch? And then hope that next VS Code version does not break all of us.
my testing confirmed that i couldn’t debug with it in it’s current state (known issue). I’m sure they are working on that. but it is going to be an interesting maintenance dance for them.
they also expose internal APIs of VS Code which are really subject to move, we have troubles with several APIs from Monaco and it takes a week to catch up, don’t know what they are hoping for
from featur perspective, i don’t know how they can ensure quality after each migration, basically they need to retest the whole VS Code, since code not their or study each commit of VS Code (which is monkey job)
then some VS Cod extensions don’t run in browser, like a new cool preview access native browser, they will need to reimplement it, Theia already has own implemented
I’ve gone yesterday through vscode.d.ts for Theia and coverage is quite decent, very good job from RedHat. Something work bogusly, but we need just more testing. And then aligning UI a bit to look like VS Code, implementing built-in commands and theming will be big pieces. But it is not years of work and after that we will be compatible based on stable APIs which does not move so fast as VS Code code base.
I would observe how code-server develops but I have really doubts who is having an advantage here.