WMS Time: postgis layers should not use date_trunc function

[assefa]

It seems that using date_trunc function could be slower that using simple comparison.

There was a thread on this in mapserver list:

 http://n2.nabble.com/WMS-TIME-parameter-slow-td4709981.html

[assefa]

the docs related to WMS time is: http://mapserver.org/ogc/wms_time.html

The reason date_trunc is used to accomodate requests where queries like time=2001-01-01 expects to return all values. I believe this assumption should still be respected for WMS.

Maybe It should be possible to not always use date_trunc and be able to use regular comparisons.

[pramsey]

Please see #3374