From cc6974df16e0dfb106c4caefd9d42d332a04247d Mon Sep 17 00:00:00 2001 From: Tom Lane Date: Tue, 13 Jun 2023 15:58:37 -0400 Subject: Correctly update hasSubLinks while mutating a rule action. rewriteRuleAction neglected to check for SubLink nodes in the securityQuals of range table entries. This could lead to failing to convert such a SubLink to a SubPlan, resulting in assertion crashes or weird errors later in planning. In passing, fix some poor coding in rewriteTargetView: we should not pass the source parsetree's hasSubLinks field to ReplaceVarsFromTargetList's outer_hasSubLinks. ReplaceVarsFromTargetList knows enough to ignore that when a Query node is passed, but it's still confusing and bad precedent: if we did try to update that flag we'd be updating a stale copy of the parsetree. Per bug #17972 from Alexander Lakhin. This has been broken since we added RangeTblEntry.securityQuals (although the presented test case only fails back to 215b43cdc), so back-patch all the way. Discussion: https://postgr.es/m/17972-f422c094237847d0@postgresql.org --- src/backend/rewrite/rewriteHandler.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'src/backend/rewrite/rewriteHandler.c') diff --git a/src/backend/rewrite/rewriteHandler.c b/src/backend/rewrite/rewriteHandler.c index 5086b90df5f..cc00ee47776 100644 --- a/src/backend/rewrite/rewriteHandler.c +++ b/src/backend/rewrite/rewriteHandler.c @@ -477,6 +477,8 @@ rewriteRuleAction(Query *parsetree, /* other RTE types don't contain bare expressions */ break; } + sub_action->hasSubLinks |= + checkExprHasSubLink((Node *) rte->securityQuals); if (sub_action->hasSubLinks) break; /* no need to keep scanning rtable */ } @@ -3348,7 +3350,7 @@ rewriteTargetView(Query *parsetree, Relation view) view_targetlist, REPLACEVARS_REPORT_ERROR, 0, - &parsetree->hasSubLinks); + NULL); /* * Update all other RTI references in the query that point to the view -- cgit v1.2.3