Thread: Geometry regression tests (was Re: [PATCHES] Alter domain)

Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Tom Lane wrote:
>> That's a pain.  Is there no way for config.guess to tell the difference
>> between your system and the -STABLE versions?

> As I remember, the issue is that the only info is in a system header
> file.

This is a bit of a kluge, but what about switching geometry over to the
style Peter set up for locale differences?  Instead of calling out
platform-by-platform expected files, we could arrange it so that
pg_regress will accept a match on-the-fly to either of two (or more,
but I think two will be enough now) geometry.out files.  Essentially
we'd be saying that we don't really care whether specific platforms
show positive or negative zeroes in that test.

            regards, tom lane

Tom Lane wrote:
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > Tom Lane wrote:
> >> That's a pain.  Is there no way for config.guess to tell the difference
> >> between your system and the -STABLE versions?
>
> > As I remember, the issue is that the only info is in a system header
> > file.
>
> This is a bit of a kluge, but what about switching geometry over to the
> style Peter set up for locale differences?  Instead of calling out
> platform-by-platform expected files, we could arrange it so that
> pg_regress will accept a match on-the-fly to either of two (or more,
> but I think two will be enough now) geometry.out files.  Essentially
> we'd be saying that we don't really care whether specific platforms
> show positive or negative zeroes in that test.

I have to say I like this approach.  ALl our regression files are valid
in some way.

--
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073