On Wed, Mar 2, 2016 at 5:39 AM, Alexander Farber <alexander.far...@gmail.com > wrote:
> Thanks Vitaly, but instead of inverting the IF-condition I would prefer to > make my function more robust, since it is kind of security-related and I > might forget about the special IF-condition later when using it elsewhere... > > Merlin's point about inlining and SQL language functions not withstanding: You should consider writing a variation of the check_user function that returns VOID or raises an exception and use is like an assertion. I would consider raising an exception if in_auth is NULL as I'd potentially consider such a situation to represent mis-usage of the function which should gets it own error instead of simply indicating that the validation failed. David J.
