Opened 7 years ago
Closed 7 years ago
#4044 closed defect (fixed)
PostgreSQL 11 regression failures on topology upgrade caused by pg_proc catalog changes
Reported by: | robe | Owned by: | pramsey |
---|---|---|---|
Priority: | blocker | Milestone: | PostGIS 2.4.4 |
Component: | postgis | Version: | master |
Keywords: | Cc: |
Description (last modified by )
Sorry guys been distracted lately and seems debbie has been chocking on PostgreSQL 11 runs. I think it's changes upstream.
Has been happening since March 3rd.
One issue is topology https://debbie.postgis.net/job/PostGIS_Regress/8825/consoleFull
Upgrading PostGIS in 'postgis_reg' using: ALTER EXTENSION postgis UPDATE TO '2.5.0devnext' Upgrading PostGIS Topology in 'postgis_reg' using: ALTER EXTENSION postgis_topology UPDATE TO '2.5.0devnext' Died at ../../regress/run_test.pl line 1485. failed (Error encountered altering EXTENSION POSTGIS_TOPOLOGY: /var/lib/jenkins/workspace/postgis/tmp/2_5_pg11w64/regress_log) ----------------------------------------------------------------------------- sh: 1: createlang: not found CREATE EXTENSION CREATE EXTENSION ALTER EXTENSION ERROR: column proc.proisagg does not exist LINE 15: ...nd e.extname = $1 and c.relname = $2 AND COALESCE(proc.prois... ^ HINT: Perhaps you meant to reference the column "proc.prolang". QUERY: SELECT 'ALTER EXTENSION ' || e.extname || ' DROP ' || $3 || ' ' || COALESCE(proc.proname || '(' || oidvectortypes(proc.proargtypes) || ')',typ.typname, cd.relname, op.oprname, cs.typname || ' AS ' || ct.typname || ') ', opcname, opfname) || ';' AS remove_command FROM pg_depend As d INNER JOIN pg_extension As e ON d.refobjid = e.oid INNER JOIN pg_class As c ON c.oid = d.classid LEFT JOIN pg_proc AS proc ON proc.oid = d.objid LEFT JOIN pg_type AS typ ON typ.oid = d.objid LEFT JOIN pg_class As cd ON cd.oid = d.objid LEFT JOIN pg_operator As op ON op.oid = d.objid LEFT JOIN pg_cast AS ca ON ca.oid = d.objid LEFT JOIN pg_type AS cs ON ca.castsource = cs.oid LEFT JOIN pg_type AS ct ON ca.casttarget = ct.oid LEFT JOIN pg_opclass As oc ON oc.oid = d.objid LEFT JOIN pg_opfamily As ofa ON ofa.oid = d.objid WHERE d.deptype = 'e' and e.extname = $1 and c.relname = $2 AND COALESCE(proc.proisagg, false) = $4; CONTEXT: PL/pgSQL function postgis_extension_remove_objects(text,text) line 27 at FOR over EXECUTE statement ----------------------------------------------------------------------------- make[2]: *** [Makefile:79: check] Error 1 make[2]: Leaving directory '/var/lib/jenkins/workspace/postgis/branches/2.5/topology/test' make[1]: *** [Makefile:131: check] Error 2 make[1]: Leaving directory '/var/lib/jenkins/workspace/postgis/branches/2.5/topology' make: *** [GNUmakefile:16: check] Error 1 Build step 'Execute shell' marked build as failure An attempt to send an e-mail to empty list of recipients, ignored. Triggering a new build of PostgreSQL stop
Change History (10)
comment:1 by , 7 years ago
Description: | modified (diff) |
---|
comment:2 by , 7 years ago
Component: | postgis → topology |
---|---|
Owner: | changed from | to
comment:3 by , 7 years ago
comment:4 by , 7 years ago
proisagg is used in a query by our postgis_extension_helper: extensions/postgis_extension_helper.sql So this is not topology-specific but affects any need to update aggregates (remember the "brave enough" statement ? it's time to show that off ! )
comment:5 by , 7 years ago
Component: | topology → postgis |
---|---|
Owner: | changed from | to
Weird it only broke in topology wonder if that's the only thing we are testing.
strk you are brave enough I know you are or perhaps pramsey is braver.
I challenge you two and don't try to cop out with "Can't fight with a girl"
Anyway I took PG11 out of the tests until we fix it. I'm tied up with other things at moment so feel free to work on it if you have time. If you don't will have to wait till I'm less busy.
And if I do it, I'm not back-porting it to 2.4 pramsey you hear that - robe will not back-port. robe does not want people using 2.4 for PG 11 and this is as perfect an excuse as any. Sure they'll still be able to compile, but they can't upgrade
comment:6 by , 7 years ago
Milestone: | PostGIS 2.5.0 → PostGIS 2.4.4 |
---|
I guess since this would prevent upgrading of postgis_topology I should backport fix.
Updating for regular postgis though works fine without fixing. Seems the regular postgis is not calling this function for some reason
comment:8 by , 7 years ago
I've turned back on debbie's testing of 11 and will backport to 2.4.4 if she's happy with the change.
comment:9 by , 7 years ago
Summary: | PostgreSQL 11 regression failures on topology since March 3 → PostgreSQL 11 regression failures on topology upgrade caused by pg_proc catalog changes |
---|
found the culprit, and the dev run missed it because it only fails on upgrade.
https://git.postgresql.org/gitweb/?p=postgresql.git;a=commit;h=fd1a421fe66173fb9b85d3fe150afde8e812cbe4