Bug in crop mode?

[YourMJK]

If I understand the code right then the same i, which is actually the index of the wds.windows being iterated, is also used here to index pcs.compositionObject:

https://github.com/MonoS/SupMover/blob/9646a10a0f7aac351f9f18fad99f86e469ce7bf5/main.cpp#L973-L980

I don't know much about the PGS spec but this can't be right?

This should only work if a Display Set contains exactly one composition object/image per window (spec says can also be two) and their order is the same (not sure if that's guaranteed). Besides that, my example .sup files have Display Sets with pcs.numCompositionObject == 0 but wds.numberOfWindows > 0, so uninitialized t_compositionObjects are being manipulated in that case.

I supposed it should be something like this instead:

if (corrVer != 0 || corrHor != 0) {
    for (int j = 0; j < pcs.numCompositionObject; j++) {
        if (pcs.compositionObject[j].windowID != wds.windows[i].windowID) continue;
        pcs.compositionObject[j].objectVerPos -= corrVer;
        pcs.compositionObject[j].objectHorPos -= corrHor;
    }
    fixPCS = true;
}

[MonoS]

I think you are completely right, I've probably missed the windowID field when implementing the corrVer/Hor handling and worked fine as it is on my test case.

Please make a pull request if you want, I'll merge it immediately :)