In source/fitz/load-tiff.c Line:272 272 *dst++ = a << (8 - tiff->bitspersample); if tiff->bitspersample = 16, 8 - tiff->bitspersample will revert to a very large value 4294967288, which is not our expectation. I think it is an integer overflow issue.
Do you have a file that can be used to reproduce the issue?
Sorry, I just reviewed the source code. No POC now.
Fixed in commit b7892cdc7fae62aa57d63ae62144e1f11b5f9275 Author: Sebastian Rasmussen <sebras@gmail.com> Date: Mon Jun 10 13:58:28 2019 +0200 Bug 701176: Handle TIFFs with alpha that have more than 8 bits per sample.