model-writtenrun 7079 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Tender Wang proposed standardizing tuple validity checks in tablecmds.c on HeapTupleIsValid(), attaching a patch limited to that file #1.
- Heikki Linnakangas preferred explicit NULL checks because the macro misleadingly suggests additional validity tests, and floated replacing its uses while retaining it for compatibility #2.
- Tom Lane warned that broader macro removal would create code churn and backpatching difficulties #3; Heikki nevertheless considered PointerIsValid() pointless while recognizing more value in OidIsValid() #4.
- Michael Paquier favored leaving existing styles alone to avoid backpatching pain #5, and Tender accepted that reasoning without submitting another patch #6.
- Peter Eisentraut supported removing redundant validity macros #7 and closed the supplied discussion by arguing that they obscure ordinary C checks, citing stable-branch history to suggest limited backpatching impact #8.
- Peter later committed PointerIsValid() removal as a5b35fcedb from the matching patchset #600776 in thread #52301, retaining HeapTupleIsValid() to reduce churn; this thread contributed earlier design discussion, while Tender's standardization patch remained uncommitted.