Hi all is there some agreement on this behave of ASSERT statement?
I would to assign last patch to next commitfest. Possible changes: * I would to simplify a behave of evaluating of message expression - probably I disallow NULL there. * GUC enable_asserts will be supported * a assert exception should not be handled by PLpgSQL handler -- like CANCEL exception Regards Pavel 2014-12-14 19:54 GMT+01:00 Pavel Stehule <pavel.steh...@gmail.com>: > > > 2014-12-14 18:57 GMT+01:00 Alvaro Herrera <alvhe...@2ndquadrant.com>: >> >> Pavel Stehule wrote: >> > Hi >> > >> > any comments to last proposal and patch? >> >> WTH is that pstrdup("hint is null") thing? Debugging leftover? >> > > No, only not well error message. I propose a syntax ASSERT > boolean_expression [, text expression ] -- text expression is >>hint<< for > assertion debugging > > This text expression should not be null when it is used. I am not sure, > what is well behave - so when assertions fails and text expression is null, > then I use message "hint is null" as hint. > > Regards > > Pavel > > >> >> -- >> Álvaro Herrera http://www.2ndQuadrant.com/ >> PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services >> >