On Fri, Apr 30, 2021 at 7:07 AM Justin Pryzby <pryzby@telsasoft.com> wrote:
>
> On Thu, Apr 29, 2021 at 02:39:42PM -0400, Alvaro Herrera wrote:
> > On 2021-Apr-29, Tom Lane wrote:
> > > Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> > > > I'd do it like this. Note I removed an if/else block in addition to
> > > > your changes.
> > >
> > > > I couldn't convince myself that this is worth pushing though; either we
> > > > push it to all branches (which seems unwarranted) or we create
> > > > back-patching hazards.
> > >
> > > Yeah ... an advantage of the if/else coding is that it'd likely be
> > > simple to extend to cover additional statement types, should we ever
> > > wish to do that. The rendering you have here is nice and compact,
> > > but it would not scale up well.
> >
> > That makes sense. But that part is not in Amul's patch -- he was only
> > on about removing the is_foreign_table Boolean. If I remove the if/else
> > block change, does the rest of the patch looks something we'd want to
> > have? I kinda agree that the redundant variable is "ugly". Is it worth
> > removing? My hunch is no.
>
> Getting rid of a redundant, boolean variable is good not because it's more
> efficient but because it's one fewer LOC to read and maintain (and an
> opportunity for inconsistency, I suppose).
Yes.
> Also, this is a roundabout and too-verbose way to invert a boolean:
> | transformCheckConstraints(&cxt, !is_foreign_table ? true : false);
I agree to remove only the redundant variable, is_foreign_table but
not the if else block as Tom said: it's not scalable. We don't need to
back patch this change.
With Regards,
Bharath Rupireddy.
EnterpriseDB: http://www.enterprisedb.com
