Re: [PATCH] refactor ATExec{En,Dis}ableRowSecurity

2021-03-01 Thread Michael Paquier
On Mon, Mar 01, 2021 at 03:30:44PM +0900, Michael Paquier wrote: > 0001 is a clean simplification and a good catch, so I'll see about > applying it. 0002 just makes the code more confusing to the reader > IMO, and its interface could easily lead to unwanted errors. 0001 has been applied as of fab

Re: [PATCH] refactor ATExec{En,Dis}ableRowSecurity

2021-02-28 Thread Michael Paquier
On Sun, Feb 28, 2021 at 02:27:44PM -0800, Zhihong Yu wrote: > For 0002-Further-refactoring.patch, should there be assertion > inside ATExecSetRowSecurity() on the values for rls and force_rls ? > There could be 3 possible values: -1, 0 and 1. 0001 is a clean simplification and a good catch, so I'l

Re: [PATCH] refactor ATExec{En,Dis}ableRowSecurity

2021-02-28 Thread Zhihong Yu
Hi, For 0002-Further-refactoring.patch, should there be assertion inside ATExecSetRowSecurity() on the values for rls and force_rls ? There could be 3 possible values: -1, 0 and 1. Cheers On Sun, Feb 28, 2021 at 1:19 PM Justin Pryzby wrote: > tablecmds.c is 17k lines long, this makes it ~30 lin