[patch] Bug in pg_dump/pg_restore using --no-publication
Hi,
Attached is a patch that fixes a bug in pg_dump since 10.0 and
reproducible in master. When using option --no-publication : ALTER
PUBLICATION orders are still present in the dump.
Steps to reproduce:
postgres=# CREATE DATABASE test;
CREATE DATABASE
postgres=# \c test
You are now connected to database "test" as user "postgres".
test=# create table t1 (id integer);
CREATE TABLE
test=# CREATE PUBLICATION p1 FOR TABLE t1;
CREATE PUBLICATION
test=# CREATE PUBLICATION p_all FOR ALL TABLES;
CREATE PUBLICATION
test=# CREATE PUBLICATION p_insert_only FOR TABLE t1 WITH (publish =
'insert');
CREATE PUBLICATION
test=# \q
$ pg_dump -p 5400 --no-publication test | grep PUBLICATION
-- Name: p1 t1; Type: PUBLICATION TABLE; Schema: public; Owner:
ALTER PUBLICATION p1 ADD TABLE ONLY public.t1;
-- Name: p_insert_only t1; Type: PUBLICATION TABLE; Schema: public; Owner:
ALTER PUBLICATION p_insert_only ADD TABLE ONLY public.t1;
pg_restore suffers the same problem:
pg_restore --no-publication test.dump | grep PUBLICATION
-- Name: p1 t1; Type: PUBLICATION TABLE; Schema: public; Owner:
ALTER PUBLICATION p1 ADD TABLE ONLY public.t1;
-- Name: p_insert_only t1; Type: PUBLICATION TABLE; Schema: public; Owner:
ALTER PUBLICATION p_insert_only ADD TABLE ONLY public.t1;
pg_restore --no-publication test.dump -l test.dump| grep PUBLICATION
2230; 6106 16389 PUBLICATION TABLE public p1 t1
2231; 6106 16392 PUBLICATION TABLE public p_insert_only t1
Should I add it to current commitfest ?
--
Gilles Darold
Consultant PostgreSQL
http://dalibo.com - http://dalibo.org
Attachments:
patch_pg_dump_option_no_publication-v1.difftext/x-patch; name=patch_pg_dump_option_no_publication-v1.diffDownload
diff --git a/src/bin/pg_dump/pg_backup_archiver.c b/src/bin/pg_dump/pg_backup_archiver.c
index 7d1d439ba2..2e86664c17 100644
--- a/src/bin/pg_dump/pg_backup_archiver.c
+++ b/src/bin/pg_dump/pg_backup_archiver.c
@@ -2920,7 +2920,8 @@ _tocEntryRequired(TocEntry *te, teSection curSection, ArchiveHandle *AH)
return 0;
/* If it's a publication, maybe ignore it */
- if (ropt->no_publications && strcmp(te->desc, "PUBLICATION") == 0)
+ if (ropt->no_publications && (strcmp(te->desc, "PUBLICATION") == 0
+ || strcmp(te->desc, "PUBLICATION TABLE") == 0))
return 0;
/* If it's a security label, maybe ignore it */
diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c
index 0687a81914..c8d01ed4a4 100644
--- a/src/bin/pg_dump/pg_dump.c
+++ b/src/bin/pg_dump/pg_dump.c
@@ -3907,6 +3907,7 @@ getPublicationTables(Archive *fout, TableInfo tblinfo[], int numTables)
PQExpBuffer query;
PGresult *res;
PublicationRelInfo *pubrinfo;
+ DumpOptions *dopt = fout->dopt;
int i_tableoid;
int i_oid;
int i_pubname;
@@ -3914,7 +3915,7 @@ getPublicationTables(Archive *fout, TableInfo tblinfo[], int numTables)
j,
ntups;
- if (fout->remoteVersion < 100000)
+ if (dopt->no_publications || fout->remoteVersion < 100000)
return;
query = createPQExpBuffer();
On Fri, Sep 21, 2018 at 05:44:02PM +0200, Gilles Darold wrote:
Attached is a patch that fixes a bug in pg_dump since 10.0 and
reproducible in master. When using option --no-publication : ALTER
PUBLICATION orders are still present in the dump.
Thanks for the report, the patch, and the test case, Gilles!
pg_restore --no-publication test.dump -l test.dump| grep PUBLICATION
2230; 6106 16389 PUBLICATION TABLE public p1 t1
2231; 6106 16392 PUBLICATION TABLE public p_insert_only t1
This command does not actually work ;)
Should I add it to current commitfest ?
No need to. I have committed your patch down to v10 after making sure
that we are not missing any other spots, and testing each branch
manually.
--
Michael