cDc

This is a ex code for Windows, maybe it helps: ``` #ifdef _MSC_VER CoInitialize(NULL); LPITEMIDLIST pidlFolder(ILCreateFromPath(Util::getPath(fileName))); LPITEMIDLIST pidlFile(ILCreateFromPath(fileName)); LPCITEMIDLIST pidlItems[1] = { pidlFile }; SHOpenFolderAndSelectItems(pidlFolder, 1, pidlItems, 0); ILFree(pidlFile); ILFree(pidlFolder);...

Pls find them bellow; you only need to add a way to read the PLY file, I am not familiar enough with igl to do that quick. Thx! [globe.zip](https://github.com/libigl/libigl/files/3888493/globe.zip) PS:...

Commenting out that assert, fails later in `circulation` in what looks like an infinite loop that keeps growing the memory through: `N.push_back(fi);`

For reading the EXIF info [easyexif](https://github.com/mayanklahiri/easyexif) library can be used (stand alone), or [TinyEXIF](https://github.com/cdcseacave/TinyEXIF) library, an extension supporting also parsing the XMP tags.

Try my lib then, it was completely rewritten. And if you still find bugs, I will fix them as soon as reported.

Indeed, all the above comments were for the case you want to actually implement this _issue_.

your change looks good, but I am not convinced it is following the standard where it is placed; more exact, the format already supports little/big endian, as you can see...

most of the values should be set to an invalid default value, so you can check if it was read or not; do you have something in particular you need...

one way is to modify the code to not set it to 0 by default, but a value that can not exist

yes, the patchmatch algorithm greatly benefits from a GPU implementation and for that `EstimateDepthMapTmp()` should be ported; however the code can not just be directly ported to CUDA, the propagation...