https://github.com/supercollider/supercollider/blob/master/SCClassLibrary/Common/Control/ServerStatus.sc#L140

                    this.updateRunningState(true);
                    this.changed(\counts);

It's the Server object, not the ServerStatusWatcher, that should be calling changed.

My refactored status watcher hasn't been merged, @crucialfelix wanted to be careful and incremental. I am now not sure, however, what is the best way to proceed.

Post a PR with a bug fix please.

That's the whole with huge unreadable PRs - even with spending hours with it something like that gets through. I'm not sure it it was something I missed or something in the original PR.

@crucialfelix I can try, but I don't think I can reimplement the whole recording pause-update-etc. thing in the IDE. It was already over my head, and I won't be able to do it all over. I understand that the PR was too big, but it just became that way and I had no way of simplifying it.

I guess better to just throw it away, even if it was several day's work.

Isn't the bug reported above just one small missing change ? It isn't getting the status.

The whole recording pause-update thing is separate. The rest of what you did is still waiting in that old PR to be cherry picked into a new smaller PR for each feature. If you (or me or somebody) starts from master now, that person can see what things are already merged in. I basically just did the first feature and then stopped.

I'll try to look at the issue above, but I'm pretty busy right now.

OK, yes I see.

Is this fixed? I can't reproduce on the current master.

the only thing that seems missing is that the recording little R field in the IDE GUI doesn't update when you start and stop recording with the s.makeGui server window.

I can reproduce. In particular, avg CPU, number of synths, etc. aren't displaying.

Ah ok, I can reproduce this, too. I didn't look at them, because I misunderstood the issue description.

Apologies for vagueness in the issue description. I had thought it would be clear from the screenshot that it should be inconsistent to have a server GUI showing "running" status while all the data fields are empty. Then again, I guess nobody is using the server GUI anymore so it may have been forgotten how it works.

No problem! Yes, I also think it would be better to make a new GUI once the refactoring is done. If it is OK even with the recording quirk, this would fix it.

@jamshark70 your change seems to fix it. Will submit a PR now.