[docs] convert WCS doc and add examples

[jmckenna@…]

http://mapserver.gis.umn.edu/docs/howto/wcs_server should be converted to
restructured text, and formatted similar to wms/wfs server docs.

Also need to add note on SERVICE=wcs request:

"Note that when a request happens, it is passed through WMS, WFS, and WCS in
MapServer (in that order) until one of the services respond to it."

[jmckenna@…]

reassigning

[jmckenna]

I have updated the wcs_server doc (converted to reStructuredText, added examples, and tested with ms5.0.2 and 5.1-dev). It is committed in SVN /trunk/docs/.

While updating, I noticed 2 issues, which if you say are valid I can open separate tickets for:

1) the original wcs_server doc mentioned that the layer metadata "timeperiod" exists, however I could not get it to work (appear in a DescribeCoverage request that is), and in fact I see that mapwcs.c states "TODO: add TimePeriod" on line#996. Therefore I removed timeperiod from the latest doc.

2) I just could not get the layer metadata rangeset_axes (and the associated {rangeset_axis}_ metadata) to work (I got internal server errors on DescribeCoverage requests). However I was only testing with the gmap bathymetry.tif, and I was trying bogus values for that rangeset metadata...which could be the problem.

[tomkralidis]

timeperiod: See ​http://trac.osgeo.org/mapserver/browser/trunk/mapserver/mapwcs.c#L990 for use of timeperiod metadata. I think we need this clarified for this and #2487 (Steve explained this initially but I'm still fuzzy).

rangeset_axes: some clarification/explanation would be valued here as well

[jmckenna]

timeperiod: yes but if you look below line990 to 996 it states that timeperiod is not done, which would explain why I could never get timeperiod to display in a DescribeCoverage request. 990 seems to check if timeperiod or timeposition layer metadata exist, if so then create a temporalDomain tag, then create a timeposition tag, but the section where a timeperiod tag would be created is missing. Am i wrong?

[tomkralidis]

True. It's only checked and never output. Hence need for more clarification.

[sdlime]

When the WCS support was written we didn't have any temporal operator support so we couldn't leverage a time period even if provided. That probably isn't the case any more but this isn't the bug to track it with. So, from a documentation standpoint, as you suggest, comment out any mention of timeperiod metadata until the support is added. A WCS timeperiod bug could turn into a documentation bug at that point.

As for rangesets I've got to go back to my original demos. Those are used to describe image "bands" correct? I've got a demo at the DNR that allowed users to choose different bands that might be helpful here. I just need to find it and dust it off.

Steve