2004.09.24 12:36 "[Tiff] Possible (tiny) bug in LZW code", by Joris Van Damme
In tiff_lzw.c, inside LZWPreDecode, there is zeroing of memory of the code table, passed the initial entries. A comment describes the reason why:
* Zero entries that are not yet filled in. We do
* this to guard against bogus input data that causes
* us to index into undefined entries. If you can
* come up with a way to safely bounds-check input codes
* while decoding then you can remove this operation.
If this is so, then shouldn't the same reasoning apply after receiving a CODE_CLEAR and resetting the table, in LZWDecode as well as in LZWDecodeCompat? Or am I misunderstanding things?
Joris Van Damme
Download your free TIFF tag viewer for windows here: