MVT simplification before gridding

[pramsey]

To generate somewhat more pretty outputs, and smaller ones, apply a repeated point filter and simplification prior to the affine transform and gridding steps. These also reduce number of vertices prior to the (optional) clipping stage, which might be a good thing.

[pramsey]

Done at r16096

[robe]

Is this an enhancement? I see no NEWS item for it.

[robe]

Not so fast. I see all bots that test MVT are broken presumably because of this.

Winnie's last run seems to be similar to debbie ​https://debbie.postgis.net/job/PostGIS_Regress/7775/consoleFull and dronie ​https://drone.osgeo.org/postgis/postgis/665

+++ /var/lib/jenkins/workspace/postgis/tmp/2_5_pg9.6w64/test_120_out	2017-11-09 08:51:53.456606198 +0000
@@ -4,16 +4,15 @@
 PG4|MULTIPOLYGON(((5 4096,10 4096,10 4091,5 4096)),((0 4096,0 4101,5 4096,0 4096)))
 PG5|
 PG6|POLYGON((600 594,894 2704,2791 594,600 594))
-PG7|POLYGON((1251 1904,1252 1905,1253 1906,1253 1905,1252 1904,1251 1904))
+PG7|POLYGON((1251 1904,1253 1906,1253 1905,1252 1904,1251 1904))
 PG8|MULTIPOLYGON(((5 4096,10 4096,10 4091,5 4096)),((0 4096,0 4101,5 4096,0 4096)))
 PG9|POLYGON((0 0,0 4096,4096 4096,4096 0,0 0))
 PG10|
 TG1|GiEKBHRlc3QSDBICAAAYASIECTLePxoCYzEiAigBKIAgeAI=
 TG2|GiMKBHRlc3QSDhICAAAYASIGETTcPwECGgJjMSICKAEogCB4Ag==
 TG3|GiYKBHRlc3QSERICAAAYAiIJCQCAQArQD88PGgJjMSICKAEogCB4Ag==
-TG4|GioKBHRlc3QSFRICAAAYAiINCQCAQBLoB+cH6AfnBxoCYzEiAigBKIAgeAI=
-TG5|GjgKBHRlc3QSIxICAAAYAiIbCQT8PxLoB+cH6AfnBwnRD9IPEugH5wfoB+cHGgJjMSICKAEogCB4
-Ag==
+TG4|GiYKBHRlc3QSERICAAAYAiIJCQCAQArQD88PGgJjMSICKAEogCB4Ag==
+TG5|GjAKBHRlc3QSGxICAAAYAiITCQT8PwrQD88PCdEP0g8K0A/PDxoCYzEiAigBKIAgeAI=
 TG6|GjIKBHRlc3QSHRICAAAYAyIVCRTYPxoyFBRFOwoPCQoUEh4JCR4PGgJjMSICKAEogCB4Ag==
 TG7|Gj0KBHRlc3QSKBICAAAYAyIgCRTEPyIAKCgKHh0xHQ8JABQSFAoTCg8JADsSMh4JEw8aAmMxIgIo
 ASiAIHgC
@@ -41,4 +40,4 @@
 TU2
 ERROR:  pgis_asmvt_transfn: parameter row cannot be other than a rowtype
 TU3|
-#3922|323
+#3922|91
-----------------------------------------------------------------------------

[pramsey]

Must have had a brainfart; there are changes to regress/ that I forgot to check in. I will add those and a NEWS item as soon as SVN is back up.

[pramsey]

In 16104:

Regression test changes and NEWS item (References #3925)