Find unapplied changes and ensure we get updates from the backend about those.

[rtibbles]

Summary

Description of the change(s) you made

• Ensures that tasks that are long running in their application and emit their own changes during their running (like copying, publishing, syncing etc.) get properly finalized and marked as succeeded
• Does this by looking for changes that are marked as synced but with no errors nor disallowed, which means they have been sent to the backend and are awaiting application
• If this value is less than the currently registered max rev for the channel, it sends that instead.

Manual verification steps performed

1. Do a long running copy.
2. Ensure that the sync returns allowed, then that it returns the in progress tracking, and then that the change event is included in success (and that the channel rev sent to the backend is the server_rev for the copy change until it has succeeded

References

Fixes#3753

[marcellamaki]

I started by confirming that I could reproduce the original bug locally - which I could. After running the devserver with these changes, I cannot reproduce it -- so looks good to me 🎉 Did a code walkthrough with @rtibbles and I think the changes make sense, and it does not seem to be breaking anything else with my initial QA pass. Will merge, and @radinamatic and @pcenov can confirm that what I see locally is also reflected in hotfixes 🤞

[pcenov]

Hi @rtibbles, now the originally reported issue is fixed as the contents of the imported folders are displayed.
The newly introduced issue however is that the Publish button remains disabled after both importing or uploading a resource. I tried refreshing the page or going back to the channels page but this results in a brief activation of the button and then it gets disabled again:

2022-10-26_16-49-47.mp4

And a screenshot with the console errors:

[rtibbles]

So, I think the specific issue you are seeing here is a result of residual changes left in the local database from earlier testing. I have replicated with the same residual data. Signing out and signing back in before doing any more follow up testing will be important!

It did however make me realize that there is a bug in this fix which I will submit a follow up PR for.

[pcenov]

The issue with the disabled Publish button was tested and observed with different users in several browsers in incognito mode with no preexisting database from earlier testing. The errors in the console however might be from residual changes left in the local database in one of the browsers...

[rtibbles]

OK, that makes sense, thanks!