Em sex, 28 de jul de 2023 11:54 PM, Michael Paquier <michael@paquier.xyz> escreveu:
On Tue, Jul 25, 2023 at 04:45:22PM +0200, Daniel Gustafsson wrote: > Skimming the tree there doesn't seem to be any callers which aren't exiting or > ereporting on failure so the real-world impact seems low. That being said, > silencing static analyzers could be reason enough to delay allocation.
A different reason would be out-of-core code that uses rmtree() in a memory context where the leak would be an issue if facing a failure continuously? Delaying the allocation after the OPENDIR() seems like a good practice anyway.