Pedro Côrte-Real

What @klauspost describes there is indeed happening but the output seems to be completely empty so it's not the same issue.

@klauspost any opinion on this? getShortArray() and getIntArray() make bounds checking non-obvious so I'd rather just move everything into getInt() and getShort() as described. Any objection to that?

@axxel I haven't worked on this in a while but it's still a simplification opportunity in the code. Of the sketch I describe above only the first point is implemented....

@axxel not really, if you're interested go ahead

Do you know of any cases where we definitely do _not_ want to add zero_is_bad? Because if this is consistent across the format making it the default shouldn't really break...

I've just ran a full regression test on all the RW2 images I have, forcing zero_is_bad to true. They all worked fine. dcraw seems to agree as it uses zero_is_bad...

One last doubt @klauspost, currently the zero_is_bad code only works for the Rw2 encoding when it should actually apply to all of them and can even be used for other...

@LibRaw there's only one piece of code to do the interpolation itself but the zero_is_bad code is the one in the RW2 decoder that actually detects the pixels based on...

> but only RW2 has zero is bad In reality, all of Panasonic/Leica has it (.RW2 or .RAW) and apparently the CHDK cameras also have it. > That way we...

That's actually not the case. The files really are big endian or little endian. On little-endian architectures the bits are swapped on big endian they are not. It's the nature...