Movatterモバイル変換


[0]ホーム

URL:


Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Sign up
Appearance settings

Commitfe9b7b2

Browse files
committed
Fix plpgsql to re-look-up composite type names at need.
Commit4b93f57 rearranged things in plpgsql to make it cope better withcomposite types changing underneath it intra-session. However, I failed toconsider the case of a composite type being dropped and recreated entirely.In my defense, the previous coding didn't consider that possibility at alleither --- but it would accidentally work so long as you didn't change thetype's field list, because the built-at-compile-time list of componentvariables would then still match the type's new definition. The newcoding, however, occasionally tries to re-look-up the type by OID, andthen fails to find the dropped type.To fix this, we need to save the TypeName struct, and then redo the typeOID lookup from that. Of course that's expensive, so we don't want to doit every time we need the type OID. This can be fixed in the same way that4b93f57 dealt with changes to composite types' definitions: keep an eyeon the type's typcache entry to see if its tupledesc has been invalidated.(Perhaps, at some point, this mechanism should be generalized so it canwork for non-composite types too; but for now, plpgsql only tries tocope with intra-session redefinitions of composites.)I'm slightly hesitant to back-patch this into v11, because it changesthe contents of struct PLpgSQL_type as well as the signature ofplpgsql_build_datatype(), so in principle it could break code that ispoking into the innards of plpgsql. However, the only popular extensionof that ilk is pldebugger, and it doesn't seem to be affected. Sincethis is a regression for people who were relying on the old behavior,it seems worth taking the small risk of causing compatibility issues.Per bug #15913 from Daniel Fiori. Back-patch to v11 where4b93f57came in.Discussion:https://postgr.es/m/15913-a7e112e16dedcffc@postgresql.org
1 parentbb5ae8f commitfe9b7b2

File tree

7 files changed

+265
-37
lines changed

7 files changed

+265
-37
lines changed

‎src/backend/utils/cache/typcache.c

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -2116,6 +2116,16 @@ TypeCacheRelCallback(Datum arg, Oid relid)
21162116
if (--typentry->tupDesc->tdrefcount==0)
21172117
FreeTupleDesc(typentry->tupDesc);
21182118
typentry->tupDesc=NULL;
2119+
2120+
/*
2121+
* Also clear tupDesc_identifier, so that anything watching
2122+
* that will realize that the tupdesc has possibly changed.
2123+
* (Alternatively, we could specify that to detect possible
2124+
* tupdesc change, one must check for tupDesc != NULL as well
2125+
* as tupDesc_identifier being the same as what was previously
2126+
* seen. That seems error-prone.)
2127+
*/
2128+
typentry->tupDesc_identifier=0;
21192129
}
21202130

21212131
/* Reset equality/comparison/hashing validity information */

‎src/pl/plpgsql/src/expected/plpgsql_record.out

Lines changed: 29 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -447,6 +447,35 @@ alter table mutable drop column f3;
447447
select getf3(null::mutable); -- fails again
448448
ERROR: record "x" has no field "f3"
449449
\set SHOW_CONTEXT errors
450+
-- check behavior with creating/dropping a named rowtype
451+
set check_function_bodies = off; -- else reference to nonexistent type fails
452+
create function sillyaddtwo(int) returns int language plpgsql as
453+
$$ declare r mutable2; begin r.f1 := $1; return r.f1 + 2; end $$;
454+
reset check_function_bodies;
455+
select sillyaddtwo(42); -- fail
456+
ERROR: type "mutable2" does not exist
457+
LINE 1: declare r mutable2; begin r.f1 := $1; return r.f1 + 2; end
458+
^
459+
QUERY: declare r mutable2; begin r.f1 := $1; return r.f1 + 2; end
460+
CONTEXT: compilation of PL/pgSQL function "sillyaddtwo" near line 1
461+
create table mutable2(f1 int, f2 text);
462+
select sillyaddtwo(42);
463+
sillyaddtwo
464+
-------------
465+
44
466+
(1 row)
467+
468+
drop table mutable2;
469+
select sillyaddtwo(42); -- fail
470+
ERROR: type "mutable2" does not exist
471+
CONTEXT: PL/pgSQL function sillyaddtwo(integer) line 1 at assignment
472+
create table mutable2(f0 text, f1 int, f2 text);
473+
select sillyaddtwo(42);
474+
sillyaddtwo
475+
-------------
476+
44
477+
(1 row)
478+
450479
-- check access to system columns in a record variable
451480
create function sillytrig() returns trigger language plpgsql as
452481
$$begin

0 commit comments

Comments
 (0)

[8]ページ先頭

©2009-2025 Movatter.jp