GIFtoIFF Version 2.0 - Fixed Bugs 0) Stack size should be set to 8000 to 16000 with larger files. This hasn't been fixed, just mentioned. 1) Unlace would correctly deallocate memory upon a failure but would not deallocate memory upon a success. This has been fixed. 2) I ran across a couple GIF files which would bomb out of GIFtoTMP. I found the problem and fixed it. Several other GIF readers also had trouble with these files. GIFtoTMP is a little more forgiving now. If you see <13?> while running GIFtoTMP then you know you would've been hosed. 3) Due to memory limitations, TMPtoIFF and Unlace would run out of memory on my 1Meg system for 640x400 files. I modified both programs to use less memory by a factor of two. This greatly speeded up Unlace in the process. (I'm also getting more memory). 4) TMPtoIFF would run into problems if the user asked for more colors than the original image had. I didn't run across this until later because all the images I tested had 256 colors. Now TMPtoIFF will change the final color number to that of the image if HAM or EHB are NOT selected. Otherwise it will politely abort. 5) This set of files had code that exited if the image size was greater than 640. This was in case of corrupt files and my brain damage. This limit has now been removed. 6) TMPtoIFF was not correctly padding the lines of the IFF files to 16 bit boundarys. This has been fixed. It manifested itself as garbage at the bottom and sides of images whose widths weren't a multiple of 16. 7) CAMG and the x,y aspects weren't being set up quite correctly. They probably still aren't, but unless something breaks I won't bother fixing it. 8) TMPtoIFF should run much faster. By modifying the algorithms I was able to do a lot of the work up front, which greatly speeded up HAM conversion. Optimizing the source code helped as well. I now print out every 16th line number instead of each one. 9) In cases where the final color number is equal to the image color number, TMPtoIFF doesn't bother calculating Histograms and Color Substituiton tables. It just goes right into the output routine. Mark Podlipec 27-7 Yorkshire Terrace Shrewsbury MA 01545