We have some corrupted tiffs we use for code validation/testing/verification.
And low and behold once I updated from 3.5.6/3.5.7 era to 4.0.7 – these are now readable according to the code.
Found to be changes to TIFFReadEncodedStrip calling directly into TIFFReadRawStrip1() for uncompressed files of certain conditions.
Without the extra validations being done on the stripbytecounts – the file becomes valid – which is not good.
In our case the strip offsets/lengths were stomped to ‘0’ by something – so a bit of a nasty file.