WCS scaling operation dosen't work with georeferenced coverage

[gxinghua]

All the scaling operations: scalefactor, scalesize, scaleextend and scaleaxis can not correctly response with georeferenced coverage.

For example, to the system_test coverage irr_cube_2, the WCS request: "service=WCS&version=2.0.1&request=GetCoverage&coverageId=irr_cube_2&scalefactor=2.0", petascope give this response:

<ows:Exception exceptionCode="InvalidSubsetting">
  <ows:ExceptionText>
Error converting WCPS query to rasql query: E axis: subset (0.0:31.0) is out of bounds.
</ows:ExceptionText>
</ows:Exception>
</ows:ExceptionReport>

and in catalina.out:

TRACE [18:54:08] CrsUtil@800: Checking order, format and bounds of axis E ...
 ERROR [18:54:08] CrsUtil@848: Subset [0.0, 31.0] is out of coverage bounds [75042.72735943, 705042.72735943]
 ERROR [18:54:08] DimensionIntervalElement@249: Error while transforming geo-coordinates to pixel coordinates. The metadata is probably not valid.
 ERROR [18:54:08] Wcs2Servlet@261: Error stack trace:
InvalidSubsetting: Error converting WCPS query to rasql query: E axis: subset (0.0:31.0) is out of bounds.

Similar result repeat when using other scaling options, no matter the given parameters of scalesize or scaleextent are in grid coordinate or in georeferenced coordinate.

[Piero Campalani]

Thanks for creating the ticket Xinghua: yes as discussed this is a critical bug in our WCS Scaling Extension. The systemtest which were added for securing it unluckily I see they target only IndexCRS coverages.

I can look at this not before some 10 days: if somebody will want to take it beforehand, just reassign it to him/herself.

[Piero Campalani]

Fixed in changeset:81d7fca (new WCS systemtests n.5[5678] added).

Xinghua, please verify this then either close the ticket or let me know if some of issue persists.

[gxinghua]

Still has some problems about factor scaling. Since a new ticket #788 has been created, this one should be close.

[Piero Campalani]

Fix complete actually with changeset:5d16411 to fix a bad BigDecimal handling.