[raster] regress failing on PostGIS 2.2 on winnied and debbie for regress_index

[robe]

As of r12813

This is on:

PostgreSQL 9.3.4, compiled by Visual C++ build 1600, 64-bit
  Postgis 2.2.0dev
 - r12814
 - 2014-07-21 17:00:00
  scripts 2.2.0dev r12814
  GEOS: 3.5.0dev-CAPI-1.9.0 r3989
  PROJ: Rel. 4.8.0, 6 March 2012
  SFCGAL: 1.0.4

 regress_index .. failed (diff expected obtained: /projects/postgis/tmp/2.2.0dev_pg9.3_geos3.5.0dev_gdal1.11.0w64/test_28_diff)
-----------------------------------------------------------------------------
--- regress_index_expected	2014-07-21 12:57:55 -0400
+++ /projects/postgis/tmp/2.2.0dev_pg9.3_geos3.5.0dev_gdal1.11.0w64/test_28_out	2014-07-21 13:06:49 -0400

Details of diff are:

--- regress_index_expected	2014-07-21 12:57:55 -0400
+++ /projects/postgis/tmp/2.2.0dev_pg9.3_geos3.5.0dev_gdal1.11.0w64/test_28_out	2014-07-21 13:06:49 -0400
@@ -5,7 +5,7 @@
 3618|POINT(130.447205 131.655289)
 7245|POINT(128.10466 130.94133)
 3+=5:true
-924+=60:true
+924+=60:false:72
 12621+=500:true
 50000+=600:true
 expr|3+=5:true

[robe]

Failure on debbie too for some test, though different error: Testing:

  /var/lib/jenkins/workspace/postgis/branches/2.2/regress/00-regress-install/share/contrib/postgis/postgis_comments.sql
PostgreSQL 9.2.4 on x86_64-unknown-linux-gnu, compiled by gcc (Debian 4.4.5-8) 4.4.5, 64-bit
  Postgis 2.2.0dev - r12814 - 2014-07-21 16:42:23
  scripts 2.2.0dev r12814
  GEOS: 3.4.3dev-CAPI-1.8.3 r3982
  PROJ: Rel. 4.7.1, 23 September 2009

regress_index .. failed (diff expected obtained: /var/lib/jenkins/workspace/postgis/tmp/2_2_pg9.2w64/test_28_diff)
-----------------------------------------------------------------------------
--- regress_index_expected	2014-07-21 09:41:09.000000000 -0700
+++ /var/lib/jenkins/workspace/postgis/tmp/2_2_pg9.2w64/test_28_out	2014-07-21 09:43:47.000000000 -0700
@@ -4,11 +4,5 @@
 2594|POINT(130.504303 126.53112)
 3618|POINT(130.447205 131.655289)
 7245|POINT(128.10466 130.94133)
-3+=5:true
-924+=60:true
-12621+=500:true
-50000+=600:true
-expr|3+=5:true
-expr|924+=60:true
-expr|12621+=500:true
-expr|50000+=600:true
+ERROR:  unsupported XML feature
+ERROR:  unsupported XML feature

[robe]

Actually I just realized for debbie the unsupported XML feature is because I didn't compile PostgreSQL with XML support. Debbie is testing the EDB packaged one which does have XML support. I guess I can recompile with XML support. Then again I'm not sure if all folks just compiling PostGIS will compile against a PostgreSQL with xml support.

[strk]

The failure +924+=60:false:72 is because ANALYZE takes a random sample and is evidently taking one with which we get a bigger error (72) than tolerated (60). I guess the best would be to increase the STATISTICS target for the test table, what do you think ?

As for XML I guess the code could be made to parse the text EXPLAIN ANALYZE output, but let's use a different ticket for that, ok ?

[strk]

With r12815 I set the stats target to its limit (10k). That's 1/5 of the total data.

[strk]

And r12816 drops dependency of the testcase on XML. Marking as fixed, please reopen if you get more issues.