Re: Segfault when restoring -Fd dump on current HEAD

On Mon, Feb 25, 2019 at 11:01:05AM +0100, Dmitry Dolgov wrote:
> > On Mon, Feb 25, 2019 at 8:45 AM hubert depesz lubaczewski <depesz(at)depesz(dot)com> wrote:
> >
> > I did upgrade of my test pg. Part of this is pg_dump -Fd of each
> > database, then upgrade binaries, then initdb, and pg_restore.
> >
> > But - I can't restore any database that has any data - I get segfaults.
>
> Thank for reporting. Unfortunately, I can't reproduce this issue on the master
> (for me it's currently bc09d5e4cc) with the dump you've provided - do I need to
> do something more than just pg_restore to trigger it?

What's crashing for me is restoring the (12dev) dumpfile using v11 psql:

[pryzbyj(at)database tmp]$ pg_restore backup-20190225074600.10361-db-depesz.dump >/dev/null
Segmentation fault (core dumped)
[pryzbyj(at)database tmp]$ pg_restore -V
pg_restore (PostgreSQL) 11.2

I would restore dump into v12dev and re-dump and compare dump output, except it
seems like pg_restore -d no longer restores into a database ??

[pryzbyj(at)database tmp]$ PGHOST=/tmp PGPORT=5678 PATH=~/src/postgresql.bin/bin pg_restore backup-20190225074600.10361-db-depesz.dump -d postgres |wc -l
44

Justin