Exiv2

Thanks for the report and patch.
Could you attach a sample image as well?
(I downloaded one from dpreview.com but that works fine and my internet connection is currently so ridiculously slow that it doesn't make sense to keep trying)

I'd love to, but the smallest image I could find is still
5.5 MB, which appears to be too large to handle for the
bugtracker.

You can grab the image from:

http://mhx.dyndns.org/pics/dsc00025.jpg

Also, this issue makes batch processing files incredibly slow,
as the code usually jumps past the SOS marker after the SOF
segment and gets lost in the compressed image data. So it
scans the whole file instead of exiting just after discovery
of the SOS marker.

mhx@r2d2 ~/data/images/2008-06-11 $ time ~/src/c/test/exiv2-0.17/src/exiv2 *.jpg >/dev/null 2>&1

real 0m21.170s
user 0m0.990s
sys 0m0.680s

mhx@r2d2 ~/data/images/2008-06-11 $ time ~/src/c/test/exiv2-0.17-mhx/src/exiv2 *.jpg >/dev/null 2>&1

real 0m3.176s
user 0m0.330s
sys 0m0.050s

I did flush the filesystem cache before each run, the timings
are reproducible. So that's almost a slowdown by a factor of 7,
which makes a huge difference in applications that always scan
the whole image folder.

r1503

This fix will be part of 0.17.1 to be released within the next few days.
I usually mention all contributors by name in the ChangeLog (w/o email address). Please let me know how I should refer to you in the ChangeLog.

Thanks,
Andreas