On 7 April 2017 at 10:31, Andres Freund <andres@anarazel.de> wrote:
> Hi,
>
> I personally, and I know of a bunch of other regular contributors, find
> context diffs very hard to read. Besides general dislike, for things
> like regression test output context diffs are just not well suited.
> E.g. in
> https://buildfarm.postgresql.org/cgi-bin/show_stage_log.pl?nm=prairiedog&dt=2017-04-06%2021%3A10%3A56&stg=check
> the salient point (ERROR: 50 is outside the valid range for parameter "effective_io_concurrency" (0 .. 0))
> is 130 lines into the diff, whereas it's right at the start in a unified diff.
> Issues with one error that causes a lot of followup output changes are
> really common in our regression suite.
>
> I personally have PG_REGRESS_DIFF_OPTS set to -dU10, but that doesn't
> help much analyzing buildfarm output.
>
> Therefore I propose changing the defaults in pg_regress.c.
You mean people actually use those diffs? I've never done anything
apart from using <my favourite text comparison tool>. That way I can
reject and accept the changes as I wish, just by kicking the results
over to the expected results, or not, if there's a genuine mistake.
-- David Rowley http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services