Partyka Robert <bobson@saturn.alpha.pl> writes:
> #create user bobson with password '1' nocreatedb nocreateuser;
> CREATE
> #create table a (a int4);
> CREATE
> #revoke all on a from public;
> CHANGE
> and now from user bobson after conecting to test database:
> #insert into a values ('1');
> INSERT 19104 1
> hmmm... looks like bug. Or I miss something?
Oops. Strange though, this looks like it must be a very long-standing
bug: aclinsert3 thinks it can delete any zero-permissions item from an
ACL array, whereas aclcheck has a hard-wired assumption that the world
item is always there. Could we have missed this for this long?
regards, tom lane