Thread: pgsql: Avoid using platform-dependent floats in test case.
Avoid using platform-dependent floats in test case. The number of decimals printed for floats varied in this test case, as noted by several buildfarm members. There's nothing special about floats and arrays in the code being tested, so replace the floats with numerics to make the output platform-independent. Branch ------ master Details ------- http://git.postgresql.org/pg/commitdiff/73c8e8506cd1933ccf5c5d61088ca171a5f414c0 Modified Files -------------- src/pl/plpython/expected/plpython_types.out | 19 ++++--------------- src/pl/plpython/expected/plpython_types_3.out | 19 ++++--------------- src/pl/plpython/sql/plpython_types.sql | 11 ++--------- 3 files changed, 10 insertions(+), 39 deletions(-)
Heikki Linnakangas <heikki.linnakangas@iki.fi> writes:
> Avoid using platform-dependent floats in test case.
> The number of decimals printed for floats varied in this test case, as
> noted by several buildfarm members. There's nothing special about floats
> and arrays in the code being tested, so replace the floats with numerics to
> make the output platform-independent.
Numerics aren't entirely platform-independent in Python, either,
cf prairiedog. Is it okay to use text?
regards, tom lane
On 10/26/2016 04:42 PM, Tom Lane wrote: > Heikki Linnakangas <heikki.linnakangas@iki.fi> writes: >> Avoid using platform-dependent floats in test case. >> The number of decimals printed for floats varied in this test case, as >> noted by several buildfarm members. There's nothing special about floats >> and arrays in the code being tested, so replace the floats with numerics to >> make the output platform-independent. > > Numerics aren't entirely platform-independent in Python, either, > cf prairiedog. Is it okay to use text? Oh, thanks for prodding, I didn't notice it was still failing, I thought it just hadn't run since that fix. There is actually a similar test for text already. I don't really have any reason to believe that this would work differently with different datatypes, adding tests with different datatypes was perhaps overkill in the first place. I'm going to just remove the offending case. - Heikki