axxel

Here is how I interpret the situation: 1. whether or not the encoder is given info about the encoding of the data, it may use any combination of the 4...

In my opinion that is not crazy at all. Those modes are means to compress data (there is no concept of string vs bytes at that level). They have nothing...

The code says otherwise: https://github.com/darktable-org/rawspeed/blob/develop/RawSpeed/RawDecoder.cpp#L579-L583 I was just about to suggest to implement just that when I remembered having seen it before. If there is some bug it should be...

Yes? Please have a closer look at what the code does: if `applyCrop` is false, then the 'cropped' image is simply the 'complete'/'uncropped' image. And since the CFA data is...

> I'd prefer to avoid OpenMP for library code. pthreads/etc has been around for a while, and is widely supported. I can not say the sample about OpenMP. clang has...

> Does that sound contradicting even now? No :). > True, that was a bit of FUD and "so i heard this somewhere". Right away i have no further proofs....

> One argument against those would be that currently pthread is used. Using both at the same time will be confusing, and a questions like "which one to use in...

VisualStudio 2010 has only _very_ patchy support for c++11. The decision to use the many improvements provided by c++11 (apart from threads) has been made and long been discussed (back...

Of course. Sorry for my 'bite reflex' ;). Then you will indeed miss out on new rawspeed improvements until you drop support for those two compilers. We (or I) don't...

To come back to the original question that @LebedevRI asked: you have your own build system based on qmake and therefore had no use for the VC solution and project...