Hi folks,
We've discovered (or possibly rediscovered?) a corruption when restoring large
objects.
If 'standard_conforming_strings = on' is set in our DB (which is required for
our app) then the piped restore method (e.g. pg_restore -O backup.dat | psql)
results in the large objects being corrupted.
However, when we use the DB-direct restoration method
(e.g. pg_restore -O -d dbname backup.dat) it works just fine.
If I take the output of the non-DB pg_restore command and edit it to set the
standard_conforming_strings off just prior to the large object creation and
then back on again immediately after the last one is created, it works.
It appears that the escaping of the strings used in the lowrite() functions
is not happening properly.
Is this something that is known and should just be avoided? Or is it something
that needs reporting?
TIA.
Bosco.