I found that it worked in GeoServer 2.6.5 which is using GeoTools 12.5.
It's not working to well using openlayers (which might be expected in this
version?), however using Normalize in other wms request formats, PNG, JPEG
etc. are working as expected. From what I recall it was failing all the
way back to 2.7 when you were helping me look into it.
On Thu, Nov 9, 2017 at 12:55 PM, Devon Tucker <dtuc...@boundlessgeo.com>
wrote:
> John, which version of GeoServer does this work in?
>
> I looked into this a bit as well. I *believe* that the root issue is an
> ArrayIndexOutOfBounds exception inside the IHSColorSpace conversion, which
> occurs using either JAI or JAIEXT.
>
> https://gist.github.com/dvntucker/b5198a7f7143c234a0613e03bb2970d3
>
> I get the same exception in the JAIEXT version of IHSColorSpace as well.
> This class is reading is reading past the raster bounds, which leads me to
> believe there is something wrong with the Image Layout when doing the IHS
> conversion.
>
> One thing I did notice is that ContrastEnhancementNode is re-using the
> same Hints over and over again, calling ImageWorker.setHints before every
> operation:
>
> https://github.com/geotools/geotools/blob/master/modules/
> library/render/src/main/java/org/geotools/renderer/lite/gridcoverage2d/
> ContrastEnhancementNode.java
>
> I'm not totally sure this is the correct way to use ImageWorker? As I
> understand it from looking at the code, ImageWorker updates its own Hints
> after each call in order to track the updated image layout. I _think_
> calling setRenderingHints each time might be causing issues? However, when
> I tried to remove all these calls it didn't seem to fix the issue, so I
> don't know for sure.
>
> It's also possible that the issue is simply a bug in the IHSColorSpace
> conversion. Possibly a simple off by one error? I didn't look too closely
> at the code to discern whether there might be an issue here. The bug is
> replicated in both the Java and EXT versions, so I assumed not, but it's
> worth looking at.
>
> On Thu, Nov 9, 2017 at 8:04 AM, John Schulz <jsch...@boundlessgeo.com>
> wrote:
>
>> Hi Daniele, and all,
>>
>> I'm looking into an issue involving a contrast enhancement bug for 3 and
>> 4 band rasters found here: https://osgeo-org.atlassian.ne
>> t/browse/GEOT-5860
>>
>> In tracking down the bug it appears to be a regression. I think that the
>> issue stems from the IHS colorspace conversion section of the
>> ContrastEnhancementNode, where the parameters of the image and the raster,
>> at the very least, appear to get out of synch.
>>
>> I'm just looking to start a conversation as I get more familiar with
>> issue and hoping to get some insight into when this regression came along.
>>
>> Cheers,
>>
>> John
>>
>> ------------------------------------------------------------
>> ------------------
>> Check out the vibrant tech community on one of the world's most
>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>> _______________________________________________
>> GeoTools-Devel mailing list
>> GeoTools-Devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/geotools-devel
>>
>>
>
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
GeoTools-Devel mailing list
GeoTools-Devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel
