AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2004.10.08 16:40 "[Tiff] TIFFReadEncodedStrip - unforgiving", by Chris Losinger
2004.10.09 08:14 "Re: [Tiff] TIFFReadEncodedStrip - unforgiving", by Andrey Kiselev
2004.10.09 08:26 "Re: [Tiff] TIFFReadEncodedStrip - unforgiving", by Andrey Kiselev

2004.10.09 08:26 "Re: [Tiff] TIFFReadEncodedStrip - unforgiving", by Andrey Kiselev

On Sat, Oct 09, 2004 at 12:14:04PM +0400, Andrey Kiselev wrote:

v3.6.1 is refusing to read some G3-compressed images in my test suite that have premature EOLs.

the problem seems to be that in 3.5.7, TIFFReadEncodedStrip treated the return value from the tif->tif_decodestrip function as a boolean (fail on 0, pass otherwise). and, the return code from fax decoder was -1. so, it always passed - and this was a good thing, since apparently many fax encoders create TIFFs with premature EOLs.

but now, in 3.6.1, TIFFReadEncodedStrip tests for return > 0, which means it errors-out on fax images (and, presumably other types, too) that it previously passed.

Could you share the samples with me (Bugzilla is preferred, but if the images should not be disclosed you can send them directly to me). I'm prefer to fix that problem in the codec itself. If the error is not fatal the decoding routine should not return the bad status.

Oh, just found the bug reprot with a sample. Please, skip my message.

Andrey

--
Andrey V. Kiselev

Home phone: +7 812 5274898 ICQ# 26871517