Hi all,

I did some testing of the last commits, and found that 8d17a compiles on my machine, but I'm not able to execute the file. The program crashes immediately after starting. e9106c6 perfectly works.

TMO and preview panel: If I understood the preview panel correctly, when you click on a small preview image you get it as a bigger image, and the selection of the TMO parameter is changed to the selected algorithm. So far so good. When you now click on "Tonemapping" leaving the default values (meaning small default resolution) you should see the same image as the image from the preview panel (ignoring the different resolution). Moreover when you try to do TM with the full resolution, you should
again see the same image. All these steps should not apply to Fattal, since it is the only size dependent algorithm.

Now my findings (ignoring Fattal):
With L 2.1.0: The assumptions above apply to all TMO, except Reinhard05. The result image after TM with default resolution gives a nearly white image. The problem here, I think, is because of some default values, which changed ages ago.

With L trunk: The problem of Reinhard05 is still present. And moreover Mantiuk06 gives different result when doing TM on full resolution and with default resolution. At the higher resolution, the image becomes somehow darker. I found that this difference was introduced before c75e3.
Can someone have a deeper look at this, please?

Libraw: On the windows, as doing regular library updates, I wanted to use libraw 0.14.3 instead of the 0.13.8. Does this cause some problems on unix?