Opened 5 years ago
Closed 4 years ago
#4533 closed defect (fixed)
3.0.0rc1 chaikin test failure on Ubuntu bionic/ppc64el
Reported by: | myon | Owned by: | komzpa |
---|---|---|---|
Priority: | low | Milestone: | PostGIS 3.1.0 |
Component: | postgis | Version: | master |
Keywords: | Cc: |
Description
On Ubuntu bionic, ppc64el only, there's a regression diff:
13:47:27 PostgreSQL 11.5 (Ubuntu 11.5-3.pgdg18.04+1) on powerpc64le-unknown-linux-gnu, compiled by gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0, 64-bit 13:47:27 Postgis 3.0.0rc1 - r17886 - 2019-10-08 09:53:04 13:47:27 scripts 3.0.0rc1 r17886 13:47:27 GEOS: 3.7.1-CAPI-1.11.1 27a5e771 13:47:27 PROJ: Rel. 4.9.3, 15 August 2016 13:47:46 chaikin .. failed (diff expected obtained: /tmp/pgis_reg/test_35_diff) 13:47:46 ----------------------------------------------------------------------------- 13:47:46 --- chaikin_expected 2019-10-08 06:18:28.000000000 +0000 13:47:46 +++ /tmp/pgis_reg/test_35_out 2019-10-08 11:47:46.687220275 +0000 13:47:46 @@ -4,4 +4,4 @@ 13:47:46 4|LINESTRING(0 0,4.5 4.5,6 7,6 9,4.5 11.5,0 16) 13:47:46 5|POINT(0 0) 13:47:46 6|GEOMETRYCOLLECTION(POINT(1 1),LINESTRING(1 1,1 2.125,1 2.75,1 3.25,1 3.875,1 5),POLYGON((5 5,5 7.8125,5.3125 9.0625,5.9375 9.6875,6.875 10,8.125 10,9.0625 9.6875,9.6875 9.0625,10 8.125,10 6.875,9.6875 5.9375,9.0625 5.3125,7.8125 5,5 5),(6 6,6 6.5625,6.0625 6.8125,6.1875 6.9375,6.375 7,6.625 7,6.8125 6.9375,6.9375 6.8125,7 6.625,7 6.375,6.9375 6.1875,6.8125 6.0625,6.5625 6,6 6))) 13:47:46 -7|POLYGON((2 2,6 6,6 10,2 14,0 12,0 4,2 2))|BOX(0 2,6 14) 13:47:46 +7|POLYGON((2 2,6 6,6 10,2 14,0 12,0 4,2 2))|BOX(-0 2,6 14) 13:47:46 -----------------------------------------------------------------------------
Other dists (except xenial, see #4532) and architectures are fine.
Change History (7)
comment:1 by , 5 years ago
Milestone: | PostGIS 3.0.0 → PostGIS 3.0.1 |
---|---|
Priority: | medium → low |
comment:2 by , 5 years ago
Owner: | changed from | to
---|
comment:4 by , 5 years ago
Since lwprint_double already handles the close to zero values correctly, we could fix this issue by using it in BOX2D_out and take out sprintf
. Not only it'd fix this bug but make the output consistent with ST_AsText and so on.
Same for gbox_to_string.
comment:5 by , 5 years ago
Milestone: | PostGIS 3.0.1 → PostGIS 3.1.0 |
---|
too low to care. Let's punt for now. Maybe we can revisit in 3.0.2 or not until 3.1
Note:
See TracTickets
for help on using tickets.
Zero vs negative zero on low usage platform. Could you just comment the test out in the packaging process?