On Mon, Feb 03, 2025 at 09:30:33PM -0800, Jeff Davis wrote: > That sounds useful, but not necessarily required, for the HashAgg tests > I just posted[1].
I thought so based on what you have posted on the other thread, as you are relying on a three-step sequence to happen in a specific order, multiple times. A stack of repeated NOTICE messages is a bit noisy in the output, still it should be enough. > One extra benefit of supporting arguments is that it would be a more > flexible way to change the local state around the injection point. > Right now the only way is by using IS_INJECTION_POINT_ATTACHED(), which > doesn't permit callback-defined conditions, etc. Agreed. The line I'm drawing here (mentioned upthread as well) is that any changes done in the core backend for injection_point.c should have one or more use cases in the tree. > If you do add suppport for arguments, would it make sense to just have > all callback functions take a single "void *" argument, rather than > adding branches for the zero-argument and the one-argument case? Yep. Just passing down a full structure would be enough with a single argument. > +1 to the idea, but I'm fine waiting for additional use cases to get > the API right. Thanks for the input and the comments. -- Michael
signature.asc
Description: PGP signature
