- 2020.08.06 14:55 "Re: [Tiff] Unable to read a large TIFF File (single image, one IFD, Strip image of size 2.6 GB) on Unix using TIFFClientOpen API", by Upanita Goswami
- 2020.08.06 19:43 "Re: [Tiff] Unable to read a large TIFF File (single image, one IFD, Strip image of size 2.6 GB) on Unix using TIFFClientOpen API", by Bob Friesenhahn
-
2020.08.07 04:13 "Re: [Tiff] Unable to read a large TIFF File (single image, one IFD, Strip image of size 2.6 GB) on Unix using TIFFClientOpen API", by Adam
Richter
-
2020.08.10 10:00 "Re: [Tiff] Unable to read a large TIFF File (single image, one IFD, Strip image of size 2.6 GB) on Unix using TIFFClientOpen API", by Adam
Richter
- 2020.08.10 11:11 "Re: [Tiff] Unable to read a large TIFF File (single image, one IFD, Strip image of size 2.6 GB) on Unix using TIFFClientOpen API", by Upanita Goswami
-
2020.08.10 10:00 "Re: [Tiff] Unable to read a large TIFF File (single image, one IFD, Strip image of size 2.6 GB) on Unix using TIFFClientOpen API", by Adam
Richter
2020.08.13 11:44 "Re: [Tiff] Unable to read a large TIFF File (single image, one IFD, Strip image of size 2.6 GB) on Unix using TIFFClientOpen API", by Adam Richter
Hi, Upanita.
When I mentioned that that number in your message was 0x7ffff000 hexadecimal and that I took this as an indication that there was likely an integer overflow, I was referring mostly to the possibility of an overflow in libtiff, not your code. I should also mention that 4096 from overflowing, and that 4096 is exactly the value of IO_CACHE_PAGE_SIZE in _TIFFPartialReadStripArray, so I think I really did find a bug that is causing problems. However, I could be wrong, and also, if this really is a bug that was not detected previously, it means that this execution path has probably not been used very often with sizes greater than 2**31 bytes, so I would guess that the chance that there might be another bug like this further along that execution path is higher than I would normally expect.
Anyhow, I look forward to your test results or test case when your teams gets a chance to provide either one of these. If that provides infeasible, please let me know, as I may try harder to generate my own test case.
Thanks for reporting this problem, and thanks in advance for any further updates.
Adam