This patch implements native support for JPEG images in POV-Ray. By native support, I mean that POV-Ray can output directly to a JPEG file, and it can also input JPEG files as image maps. Although there is a loss of quality when going from 24 bit lossless (like TGA or PNG) to lossy JPEG, most people convert their output files to JPEG anyways, to save space.
The jpeg patch works exactly like the support for built-in file formats of POV-Ray. For command line JPEG output, use +fj as an option. For using a JPEG image as an image map, use the "jpeg" keyword like you would use "gif" or "tga". Sorry, but bump maps, texture maps, heightfields, etc are not supported. This is not a difficult addition, but I don't see a need for it. If there is a demand for this, I may add it in as time permits.
The changes in the POV code are very minor. The file jpegpv13.tar.gz contains context diffs, applied using the GNU patch utility, which should be on all unix systems, but if not, you can get it here too.
This patch also requires libjpeg, a totally free jpeg compression/decompression library available via ftp. I have used libjpeg v5a.