Conversation
|
minor conflict to be cleaned up and I'll get this merged |
pipermerriam
left a comment
There was a problem hiding this comment.
I think the querying in this strategy is going to need a bit more machinery (most of which I'm inclined to have tested to be sure we're actually doing things correctly).
First, I worry about race conditions in the "queue". If I'm reading the code correctly, I think that this strategy will continue to select the same items if the rate the queue is being processed is slower than the rate this loop runs, which I believe results in the queue getting filled up with multiple copies of the same item.
Second, I am unsure if this is going to do the correct thing when a piece of content has multiple audits, some of which are failing and some of which are succeeding. Are we sure that it will only pick the ones who's most recent audit has failed? The code currently looks like it will filter out all successful audits, and then select the most recent failed one... which doesn't seem correct.
Agree with this observation. It will need to be addressed
Agree. Understanding how to effectively execute a subquery for the solution using sea-orm remains elusive to me. |
Did you find this bit in the docs? |
I did. I just haven't been able to put it together to do the sequence of things required for the strategy |
Issue addressed
Proposed changes
Adds the
SelectionStrategy::FailedAdditional info