This issue (causing "Stack Overflow" in formerly correctly written
applications) seems like it will be important for some applications.
It certainly struck GraphicsMagick:
https://gitlab.com/libtiff/libtiff/-/issues/575
Yes, that's definitely annoying. The whole TIFFGetField/TIFFSetField API
is hard to use safely. In an ideal world, we'd have a set of different
signatures like TIFFGetFieldAsInt32(TIFF*, ttag_t tag, int32_t*
retValue), etc. but I've the feeling that would be a significant work
(especially if we want to be tolerant on using such a function to get
uint8/int8/uint16/int16 values as well)
The annoying aspect of #575 is that it breaks in a patch release. One
possibility would be to revert the tags added by the DNG update
(possibly just #if 0 / #endif the declaration of the new tags) for
4.5.1, and re-enable it for 4.6.0.
@Su, would that be OK for you ?
The suggested work-around caused the program to not work with older
libtiff, and even the most recent release. I just put a fix into
GraphicsMagick to solve that.
Can you point to the commit with your fix ? That could potentially be
interesting for others
Even
--
http://www.spatialys.com
My software is free, but my time generally not.
_______________________________________________
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff
