On Sat, Jan 18, 2025 at 08:44:00PM +0100, Mats Kindahl wrote: > For PostgreSQL 16, Peter extended the palloc()/pg_malloc() interface in > commit 2016055a92f to provide more type-safety, but these functions are not > widely used. This semantic patch captures and replaces all uses of palloc() > where palloc_array() or palloc_object() could be used instead. It > deliberately does not touch cases where it is not clear that the > replacement can be done.
I am not sure how much a dependency to coccicheck would cost (usually such changes should require a case-by-case analysis rather than a blind automation), but palloc_array() and palloc_object() are available down to v13. Based on this argument, it would be tempting to apply this rule across the stable branches to reduce conflict churn. However this is an improvement in readability, like the talloc() things as Peter has mentioned, hence it should be a HEAD-only thing. I do like the idea of forcing more the object-palloc style on HEAD in the tree in some areas of the code, even if it would come with some backpatching cost for existing code. Thoughts? Perhaps this has been discussed previously? -- Michael
signature.asc
Description: PGP signature
