types conversion mistakes in cli's server

Hi There are many loosely conversions in the CLI web server. Mostly from (u)int64 to size_t (on 32bit platforms) or in64t to char, or the other way 'round. What was the goal to support 64bit data range for content length (or read data) in 32 bit environments? If LFS support was the goal, it could be possible to make it server large files but the way it is now won't work. Any objection if I change it to use exclusively size_t to avoid any possible issues? Cheers, -- Pierre @pierrejoye | http://blog.thepimp.net | http://www.libgd.org