Hard Light Productions Forums
Community Projects => The FreeSpace Upgrade Project => Topic started by: captain-custard on June 20, 2009, 02:34:14 am
-
just a question
last night it was mentioned in IRC that there is a problem with the mvp complete tbls , and that it should not be run in multiplayer in combination with individual mvps .......
can this be explained , and if it is true why we have been testing our way through multi and no one informed us of this clash ....
thanks
-
No, I said that one of the problems about it that will cause for it to no longer be offered is to avoid this being a problem:
WARNING: Never use conflictive tbms inside the same priority location, (vp or /data/tables folder). For example, do not put aaa-shp.tbm and bbb-shp.tbm inside the same vp if they have conflictive entries. Due technical design there is no easy way to know their loading order. For the final user, finally applied entries will seem randomly selected. Alphabetical sorting is not applied here. Only vps are alphabetically ordered when loaded; individual files, (like tbms), aren't, it doesn't matter if they are obtained either from hard disk folders or from inside a vp archive.
The tbm's in MV_Complete would have to be merged, which would break their capability to play with non MV_Complete, or require the ability to multi-validate both regular packages as well as MV_Complete.
And while, as a unified large file, it does not pose any problems due to the size of MV_Complete, it poses a problem by having multiple tbm's in a single VP that modify the same table entries. And at some point, the filesize WILL get large enough that it could cause a problem.
And since validation has not been actually enforced apparently, and won't be until the exe and packages are final, and since MV_Complete will not be offered as a final solution, you've been testing multiplayer to test multiplayer.
As to why the tbm's were NOT merged to begin with for MV_Complete:
A: I completely spaced the warning on modular tables
B: Ran out of time to do so in an effort to make sure we got the release out when we did
C: made the simple mistake of allowing both A and B to be ignored in favor of wanting to put some thing cool out there for everybody.
We're still hashing things out, even now, for what it means to be a team and how to approach things. I don't know if the FSU was ever previously a team in the past, but when I came on board it was essentially DaBrain, in the midst of everything else, packaging and bringing us Teh Cool. Now we have a few more people, most (if not all) also engaged in other projects. Lots of new content on the horizon by community members (Just look at all the ship WIPs for this year alone!) and lot's of enhancements in the engine.
I appreciate the questions, and I apologize if I wasn't able to better clarify it on IRC. If you feel I still haven't clarified things well enough, let me know.
And a question of my own to the coders: How well do we know that validation actually is working, if it hasn't actually been implemented yet? I provided table hashes on the release of the MVP's, so "hacked tables" messages with a clean install shouldn't be happening, but it apparently has/is,
-
thanks zacam ..... lots