Re: Remove redundant code in pl_exec.c
Japin Li writes: > On Fri, 09 Sep 2022 at 23:34, Tom Lane wrote: >> I don't like this particularly --- it puts way too much premium on >> the happenstance that the MakeExpandedObjectReadOnly call is the >> very last step in the callback function. If that needed to change, >> we'd have a mess. > Sorry, I don't get your mind. Could you explain it more? Thanks in advance! This refactoring cannot support the situation where there is more code to execute after MakeExpandedObjectReadOnly. regards, tom lane
Re: Remove redundant code in pl_exec.c
On Fri, 09 Sep 2022 at 23:34, Tom Lane wrote: > Japin Li writes: >> IMO, we can invoke plpgsql_param_eval_generic in >> plpgsql_param_eval_generic_ro >> to avoid the redundant. > > I don't like this particularly --- it puts way too much premium on > the happenstance that the MakeExpandedObjectReadOnly call is the > very last step in the callback function. If that needed to change, > we'd have a mess. > Sorry, I don't get your mind. Could you explain it more? Thanks in advance! -- Regrads, Japin Li. ChengDu WenWu Information Technology Co.,Ltd.
Re: Remove redundant code in pl_exec.c
Japin Li writes: > I found there are some redundant code in pl_exec.c, > plpgsql_param_eval_generic_ro is same as plpgsql_param_eval_generic > except it invokes MakeExpandedObjectReadOnly. Which is exactly why it's NOT redundant. > IMO, we can invoke plpgsql_param_eval_generic in plpgsql_param_eval_generic_ro > to avoid the redundant. I don't like this particularly --- it puts way too much premium on the happenstance that the MakeExpandedObjectReadOnly call is the very last step in the callback function. If that needed to change, we'd have a mess. regards, tom lane
