Re: Assorted improvements in pg_dump

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>

To: Andres Freund <andres@anarazel.de>

Cc: Hans Buschmann <buschmann@nidsa.net>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>

Date: 2021-10-26T22:31:05Z

Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

pg_dump: avoid unsafe function calls in getPolicies().
- b7333e826955 11.19 landed
- a5b26aaafe4f 13.10 landed
- 1ed6f1b9116c 12.14 landed
- 03ac48549438 14.7 landed
- 3e6e86abca01 15.0 landed
Postpone calls of unsafe server-side functions in pg_dump.
- e46e986baef0 13.10 landed
- b1f106420b1a 11.19 landed
- 55f30e6c7640 14.7 landed
- 344b7849200f 12.14 landed
- e3fcbbd623b9 15.0 landed
Account for TOAST data while scheduling parallel dumps.
- 65aaed22a849 15.0 landed
Use PREPARE/EXECUTE for repetitive per-object queries in pg_dump.
- be85727a3df7 15.0 landed
Avoid per-object queries in performance-critical paths in pg_dump.
- 9895961529ef 15.0 landed
Rethink pg_dump's handling of object ACLs.
- 0c9d84427f44 15.0 landed
Refactor pg_dump's tracking of object components to be dumped.
- 5209c0ba0bfd 15.0 landed
pg_dump: fix mis-dumping of non-global default privileges.
- 2acc84c6fd29 15.0 cited

Attachments

0001-fix-component-masking-2.patch (text/x-diff) patch 0001
0002-rethink-ACL-handling-2.patch (text/x-diff) patch 0002
0003-move-unsafe-function-calls-2.patch (text/x-diff) patch 0003
0004-bulk-queries-2.patch (text/x-diff) patch 0004
0005-fix-toast-sched-2.patch (text/x-diff) patch 0005
0006-prepare-repeated-queries-3.patch (text/x-diff) patch 0006

Andres Freund <andres@anarazel.de> writes:
> I guess we could move the prepared-statement handling into a query execution
> helper. That could then use a hashtable or something similar to check if a
> certain prepared statement already exists. That'd then centrally be extensible
> to deal with multiple connects etc.

That seems like more mechanism than is warranted.  I tried it with a
simple array of booleans, and that seems like not too much of a mess;
see revised 0006 attached.

(0001-0005 are the same as before; including them just to satisfy
the cfbot.)

			regards, tom lane