Conversation with hectorofchad on 1/27/2010 5:11:18 PM:
(5:11:18 PM) hectorofchad:
Someone should also designate a new patching format because UPS is clunky for something so revolutionary.

Zahlman and Hextator to the rescue~

(Or just Hextator if Zahlman acts like he has more important things to do (he doesn't))
(5:11:35 PM) zahlman@gmail.com:
"something so revolutionary"?
(5:11:44 PM) hectorofchad:
the original thread that it was posted in
(5:11:51 PM) hectorofchad:
had everyone going "omg this is way better" and shit
(5:11:56 PM) hectorofchad:
of course that was in like 2001
(5:12:00 PM) hectorofchad:
but it was a big deal then
(5:12:13 PM) hectorofchad:
and then for some reason we were so blinded by all that glory that we didn't notice it has terrible flaws
(5:12:25 PM) hectorofchad:
ever notice how much more sensible it is to stack IPS patches than UPS patches
(5:12:29 PM) zahlman@gmail.com:
you could make it like... basically a script for a command-line FEditor
(5:12:46 PM) hectorofchad:
UPS patches force you to disable checksum checking if you're using more than one
(5:12:49 PM) hectorofchad:
that's Bad
(5:12:51 PM) zahlman@gmail.com:
yeah
(5:13:00 PM) zahlman@gmail.com:
in a lot of cases patches just don't stack regardless though
(5:13:01 PM) hectorofchad:
basically my spec is just
(5:13:04 PM) hectorofchad:
"UPS
(5:13:13 PM) hectorofchad:
with the checksum applied to sections instead of the entire file"
(5:13:37 PM) hectorofchad:
or hell
(5:13:40 PM) hectorofchad:
a single checksum
(5:13:45 PM) hectorofchad:
but for only the relevant bytes
(5:13:54 PM) hectorofchad:
this way
(5:13:57 PM) hectorofchad:
you can stack the patches
(5:14:01 PM) hectorofchad:
and still undo them
(5:14:10 PM) hectorofchad:
as well as check if they conflict
(5:15:12 PM) zahlman@gmail.com:
hmm
(5:15:23 PM) hectorofchad:
and of course
(5:15:31 PM) hectorofchad:
it would be able to patch files that are over 16 megabytes
(5:15:33 PM) hectorofchad:
as usual
(5:15:47 PM) hectorofchad:
if we wanted, we could be cocky/stupid
(5:15:54 PM) hectorofchad:
and store addresses as variable length ints
(5:16:02 PM) hectorofchad:
then it'd be future proof!
(5:17:45 PM) hectorofchad:
and we could provide provisions for extendability for things like patch file encryption and compression and anything else we can think of pertaining to mangling the data
(5:17:55 PM) zahlman@gmail.com:
hmm.
(6:04:17 PM) hectorofchad:
let me know if you have any outstanding ideas on the matter