Gez
Why don't I have a custom title by now?!
Posts: 9165
Registered: 07-07 |
esselfortium said:
Using an MD5 check to detect a specific wad and apply different behavior to that wad specifically for the sake of supporting its non-1.9-compatible behavior is not "for some reason", it's "because we fix it". :P
Again, irrelevant. What matters is that mods can be played without the user having to be forewarned about subtleties and intricacies. That the engine applies special compatibility settings or not doesn't change anything, what matters it that it works "right out of the box" or not.
Quasar said:
I would very much like to have Eternity auto-apply the appropriate compatibility settings rather than expecting users to understand and anticipate which are needed for various levels; however, there is no MD5 or SHA-1 database that corresponds to the settings which EE supports, so it would be a nearly impossible undertaking that would never be finished.
There are many other places where you can put checks to determine the nature of a mod and which compatibility settings should apply to it. For example, the presence or absence of a mapinfo lump, and the format of said lump. The format of a map, as well. For instance, you could have it so that any Doom-format map uses the old clipping code by default (unless compat-optioned in MAPINFO), while any UDMF map uses the new portal-friendly clipping code (again, unless compat-optioned in MAPINFO). If a map makes use of MBF features, then certain MBF features such as torque, friction and wind effects affecting monster movements are on by default; otherwise they're off. Etc. With such rough checks, you greatly reduce the number of potentially troublesome maps; which can then simply be added to the database as they are reported. Certainly Eternity wouldn't need to maintain a larger database than ZDoom, since the code has remained closer on most respects.
Quasar said:
BTW you shouldn't say "never" when it comes to MD5, as that hash was verified to have collisions a long time ago. Of course, the odds of two different WAD files having the same MD5 hash are astronomically poor ;)
The calculation isn't made for the file; it's made for the level itself -- just the part of the file that corresponds to the map's lumps. I don't know if this affects probability in any way; but in either case I think MD5 are reliable enough. Worst case, some compat settings are applied to a map that doesn't need them.
|
03-03-10 18:26 
