Add some notes about why "ALTER TYPE enum DROP VALUE" is hard.

In hopes of putting these where any would-be implementer is sure to
find them, make a placeholder grammar production for ALTER DROP VALUE
and put them there.  This is really just a docs patch, though.

Vik Fearing, with a bit more wordsmithing by me

Discussion: https://postgr.es/m/9fffd149-da0f-0c9c-6745-731fb688642a@postgresfriends.org

1 files changed, 27 insertions, 0 deletions

diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y
index fec9b2604fe..e56cbe77cb7 100644
--- a/src/backend/parser/gram.y
+++ b/src/backend/parser/gram.y
@@ -6412,6 +6412,33 @@ AlterEnumStmt:
 				n->skipIfNewValExists = false;
 				$$ = (Node *) n;
 			}
+		 | ALTER TYPE_P any_name DROP VALUE_P Sconst
+			{
+				/*
+				 * The following problems must be solved before this can be
+				 * implemented:
+				 *
+				 * - There must be no instance of the target value in
+				 *   any table.
+				 *
+				 * - The value must not appear in any catalog metadata,
+				 *   such as stored view expressions or column defaults.
+				 *
+				 * - The value must not appear in any non-leaf page of a
+				 *   btree (and similar issues with other index types).
+				 *   This is problematic because a value could persist
+				 *   there long after it's gone from user-visible data.
+				 *
+				 * - Concurrent sessions must not be able to insert the
+				 *   value while the preceding conditions are being checked.
+				 *
+				 * - Possibly more...
+				 */
+				ereport(ERROR,
+						(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+						 errmsg("dropping an enum value is not implemented"),
+						 parser_errposition(@4)));
+			}
 		 ;
 
 opt_if_not_exists: IF_P NOT EXISTS              { $$ = true; }