Hi,
On Wed, Jun 16, 2021, at 12:37, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > On 2021-06-15 22:44:29 -0400, Tom Lane wrote:
> >> + memset(&popt, 0, sizeof(popt));
> >> + popt.header = true;
> >> + popt.align = true;
> >> + popt.fieldSep = "|";
> >> + PQprint(stdout, res, &popt);
>
> > Is there an argument for not aligning because that can make diffs larger
> > than the actual data changes? E.g. one row being longer will cause all
> > rows in the result set to be shown as differing because of the added
> > padding? This has been a problem in the normal regression tests, where
> > we solved it by locally disabling alignment. It might be unproblematic
> > for isolationtester, because we don't often have large result sets...
>
> I tried it that way first, and didn't much like the look of it.
>
> I think the result sets in the isolation tests don't have a big
> problem here: as you say, they aren't very large, and in most of them
> the column widths are fairly uniform anyway.
Cool. Just wanted to be sure we considered it.
Andres