misos1

It would be nicer to preserve original bit depth and color type as close to original as possible.

Also this message looks strange: `InvalidBitDepth(nr) => write!(fmt, "Invalid dispose operation {}.", nr)` Should not it be something about invalid bit depth? There is already something about "dispose op": `InvalidDisposeOp(nr)...

Ok, but who would expect such difference between local and global version of event?

Maybe it is not so bad to have some separate event fired on failed attempt "failed_attempt". But I suppose "failed" everywhere else means "definitely failed". Or at least I hope...

You meant reject as it is returning promise. There will still be at least 5 seconds waiting in Job#finished when happens that job is removed during waiting after was isFinished...

Another good reason to have global:removed is to delete associated resources together with job. I assume it is not very good to store large files in job data. Jobs with...

@manast I suppose in bull 4.0 Scripts.isFinished throws if job is already removed. But consider these 2 cases: **Removal of delayed job:** 1. Add job with delay lets say 2...

Seems good now.

Feel free to copy over my issue. Just `fail_with_message` feels like it should be public so you can override a message like you could do for example with `assert!`. But...

Without Config::cargo_metadata() is this package not so usable: https://github.com/alexcrichton/pkg-config-rs/issues/69