On Fri, Nov 24, 2017 at 3:08 PM, Erikjan Rijkers <er@xs4all.nl> wrote:
> (debian 8)
>
> make check fails:
>
> foreign_data ... ok
> window ... FAILED
> xmlmap ... ok
>
> The diff is:
>
> $ ( cd /var/data1/pg_stuff/pg_sandbox/pgsql.frame_range/src/test/regress &&
> cat regression.diffs )
> ***
> /var/data1/pg_stuff/pg_sandbox/pgsql.frame_range/src/test/regress/expected/window.out
> 2017-11-24 15:36:15.387573714 +0100
> ---
> /var/data1/pg_stuff/pg_sandbox/pgsql.frame_range/src/test/regress/results/window.out
> 2017-11-24 15:38:35.290553157 +0100
> ***************
> *** 1034,1043 ****
> (10 rows)
>
> SELECT pg_get_viewdef('v_window');
> ! pg_get_viewdef
> ! ----------------------------------------------
> ! SELECT i.i, +
> ! sum(i.i) OVER (ORDER BY i.i) AS sum_rows+
> FROM generate_series(1, 10) i(i);
> (1 row)
>
> --- 1034,1043 ----
> (10 rows)
>
> SELECT pg_get_viewdef('v_window');
> ! pg_get_viewdef
> !
> ---------------------------------------------------------------------------------------
> ! SELECT i.i,
> +
> ! sum(i.i) OVER (ORDER BY i.i ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING)
> AS sum_rows+
> FROM generate_series(1, 10) i(i);
> (1 row)
>
>
> This small hickup didn't prevent building an instance but obviously I
> haven't done any real tests yet.
>
I think something was committed recently that changed the spacing of
pg_get_viewdef on Windows. I had the same spacing as you until I
pulled this morning, so I updated my expected output but now it looks
like whatever's changed on Windows hasn't changed on Linux.....
I'll try and find what caused this change.
