Thread: make check and make installcheck

Hi

I was playing with a snapshot and doing some regression tests just in
case when i discovered that if i use a locale (fr_FR for instance), some
tests fails because the output doesn't not look like the default one
(without locales).
I believe the failure of these tests are not really bad as it's just
things like :


*** 246,256 ****
  SELECT '' AS to_char_13, to_char(q2, 'L9999999999999999.000')  FROM
INT8_TBL;
   to_char_13 |        to_char
  ------------+------------------------
!             |                456.000
!             |   4567890123456789.000
!             |                123.000
!             |   4567890123456789.000
!             |  -4567890123456789.000
  (5 rows)

  SELECT '' AS to_char_14, to_char(q2, 'FM9999999999999999.999') FROM
INT8_TBL;
--- 246,256 ----
  SELECT '' AS to_char_13, to_char(q2, 'L9999999999999999.000')  FROM
INT8_TBL;
   to_char_13 |        to_char
  ------------+------------------------
!             | F              456.000
!             | F 4567890123456789.000
!             | F              123.000
!             | F 4567890123456789.000
!             | F-4567890123456789.000
  (5 rows)

  SELECT '' AS to_char_14, to_char(q2, 'FM9999999999999999.999') FROM
INT8_TBL;

======================================================================

does someone think that someone should do something just not to scare
people ? :)

--
Mathieu Arnold

Mathieu Arnold writes:

> I was playing with a snapshot and doing some regression tests just in
> case when i discovered that if i use a locale (fr_FR for instance), some
> tests fails because the output doesn't not look like the default one
> (without locales).

> does someone think that someone should do something just not to scare
> people ? :)

If you run against a temporary postmaster (make check) then this is
guarded against, but other than that there's not much to be done about it
other than perhaps adding a note in the documentation.

--
Peter Eisentraut      peter_e@gmx.net       http://yi.org/peter-e/

Peter Eisentraut wrote:
>
> Mathieu Arnold writes:
>
> > I was playing with a snapshot and doing some regression tests just in
> > case when i discovered that if i use a locale (fr_FR for instance), some
> > tests fails because the output doesn't not look like the default one
> > (without locales).
>
> > does someone think that someone should do something just not to scare
> > people ? :)
>
> If you run against a temporary postmaster (make check) then this is
> guarded against, but other than that there's not much to be done about it
> other than perhaps adding a note in the documentation.

well, (make check) works just fine, that was the aim of my email and the
"not to scare people" just someone could add it to the doc

--
Mathieu Arnold