Re: perlcritic and perltidy

Andrew,

* Andrew Dunstan (andrew.dunstan@2ndquadrant.com) wrote:
> On 05/08/2018 01:18 PM, Stephen Frost wrote:
> > * Andrew Dunstan (andrew.dunstan@2ndquadrant.com) wrote:
> >> On 05/08/2018 12:51 PM, Stephen Frost wrote:
> >>> * Andrew Dunstan (andrew.dunstan@2ndquadrant.com) wrote:
> >>> There's not much point adding the ',' unless you're also putting the
> >>> ');' on the next line, is there..?
> >> No, not really.
> >>
> >>> Or is that going to be handled in a follow-up patch?
> >> No, the current proposal is to keep the vertical tightness settings for
> >> parentheses, which is precisely this set of cases, because otherwise
> >> there are some ugly code efects (see Peter's email upthread)
> >>
> >> So I think we're all in agreement to fortget this trailing comma thing.
> > Well, agreed, for parentheses, but for curly-brace blocks, it'd be nice to
> > have them since those will end up on their own line, right?
> 
> Yes, but there isn't a perlcritic policy I can find that detects them
> reliably. If you know of one we can revisit it. Specifically, the one
> from the Pulp collection called RequireTrailingCommaAtNewline didn't
> work very well when I tried it.

Ok, perhaps we can't automate/enforce it, but if everyone is agreed on
it then we should at least consider it something of a policy and, as you
said up-thread, clean things up as we come to them.  I'd love to clean
up the pg_dump regression tests in such a way to make it simpler to work
with in the future, as long as we're agreed on it and we don't end up
getting complaints from perlcritic/perltiday or having them end up
being removed..

Thanks!

Stephen