@Tyriar Thats interesting - since I removed the internal worker I dont need to patch eslint anymore. Before with the worker for image decoding, eslint always would run into OOM w'o patching, no clue why. Could it be, that eslint expands the whole project tree for different lib settings, thus doubling the mem load for a separate webworker.d.ts setup?

@Tyriar Read for review. The code is the same as the last patch release (0.4.2), except the license entries (all adjusted).

@Tyriar Not sure whats going on with azure atm, both macos and windows runner give me different timeout issues. Tried to mitigate them somewhat, not sure if they are all gone now...

@Tyriar Those timeout issues with heavy load CI runners are really nasty. Its a bit tricky to wait for the state change in the remote browser engine. Tried to replace several waits with pollFor, sadly idk how to express "ensure that value xy does not change" with it. We kinda lack an event from the render logic telling us, that all pending write buffer data was finally applied to the output states. How do you work around that issue in the canvas and webgl renderer tests?

Jesus, the failing tests are really annoying, all versions I tested above pass locally just fine, grrr. 🤯

@jerch good to merge when you're ready as this is all pretty separated from the core code. We have a few more changes from the old repo to include as well is all.

@Tyriar I will go though your comments in jerch/xterm-addon-image#55 and transfer the missing changes here as well before merging.