Fetch can't be used to drain a queue which has workers

[chrisDeFouRire]

Hi! Thanks for your great job with pg-boss!

I've encountered a limitation:

• I have workers working on a queue which are stuck, for instance because an FTP (don't blame me 😸 ) upload breaks
• the queue is configured to retry jobs later, after start_after
• I know the FTP server will remain unavailable for a while
• so I'd like to drain the queue with a fetch, and collect all the jobs into an archive to send by email (again, don't blame me 😸)
• fetch doesn't return anything, because of the line below, which states you can't return jobs before their start_after... except in this case, that's exactly what I would want
• If I stop the workers, and wait until the start_after date, I can fetch the jobs... but it's not convenient at all
• if I let the workers work, they'll pick the job, retry it and fail it, preventing the fetch.

The line in question:
https://github.com/timgit/pg-boss/blob/97604249556276fdbe36d5d665f53e7644ed81ae/src/plans.js#L512C1-L513C1

What's your take on this case? I see why waiting until start_after makes sense, but also why not have an option to skip waiting?

[timgit]

Adding an option to fetch to ignore start_after looks like a good idea to me.