On Mon, Mar 18, 2024 at 10:50:25AM +0500, Andrey M. Borodin wrote: > Maybe consider function injection_points_attach_local(‘point name’) > instead of static switch? > Or even injection_points_attach_global(‘point name’), while function > injection_points_attach(‘point name’) will be global? This would > favour writing concurrent test by default…
The point is to limit accidents like the one of this thread. So, for cases already in the tree, not giving the point name in the SQL function would be simple enough. What you are suggesting can be simply done, as well, though I'd rather wait for a reason to justify doing so. -- Michael
signature.asc
Description: PGP signature