Document changes in large-object privilege checking.
Commit 5ecc0d73 removed the hard-wired superuser checks in lo_import and lo_export in favor of protecting them with SQL permissions, but failed to adjust the documentation to match. Fix that, and add a <caution> paragraph pointing out the nontrivial security hazards involved with actually granting such permissions. (It's still better than ALLOW_DANGEROUS_LO_FUNCTIONS, though.) Also, commit ae20b23a caused large object read/write privilege to be checked during lo_open() rather than in the actual read or write calls. Document that. Discussion: https://postgr.es/m/CAB7nPqRHmNOYbETnc_2EjsuzSM00Z+BWKv9sy6tnvSd5gWT_JA@mail.gmail.com
Showing
Please register or sign in to comment