Modding extras by CA (unsupported)

**SigniferOne** · 07-07-2005, 17:49

Trying to test now!

EDIT: Jerome, you'll have to give us more tips on what to do to test this out: should the four animation pack files be deleted? Should descr_skeleton.txt be deleted? Am I supposed to see anything happen when running the game with -util:animdb enabled? Etc.

**Lord Adherbal** · 07-07-2005, 18:31

my guess is that the four anim files need to be removed, and that those are regenerated by using that parameter. Have you tried that dsyrow1 ?

**JeromeGrasdyke** · 07-07-2005, 18:50

Originally Posted by [cF]Adherbal

my guess is that the four anim files need to be removed, and that those are regenerated by using that parameter. Have you tried that dsyrow1 ?

That's pretty much correct. What should happen is that the game will parse descr_skeletons.txt, load the .evt and .cas files and generate a new batch of animation pack files. Your best bet for checking wether the process has run correctly is probably just to sort your directory view by date, any files that were just created should be easy to recognise.

**SigniferOne** · 07-07-2005, 19:55

Okay, here's what I did: I unpacked the animation packs using Verc's XIDX program, then I extracted the EVT files into the same folder, and then removed the four animation files. When I ran the game, it complained that it couldn't find the very first file in descr_skeleton.txt. Then I realized that I extracted the animation files in their native directory, i.e. into data/animations/data/animations, so then I simply moved the entire folder structure into the same folder as where the four anim pack files used to be. Then when I ran the game, it crashed without a comment, like it used to in our previous tests. I did add the -util:animdb to the command line, but have not noticed any different behavior because of it. The game still crashed without saying anything, as before. That's why I asked if there was something else I should have done, to bring out the new -util behavior. Maybee someone else can confirm this while Jerome is still in office.

**SigniferOne** · 07-07-2005, 22:37

Ok here's some new interesting behavior. I decided to play the game, so I restored all folders to their original state, and when I ran the game it again said that I was missing the first animation referenced in descr_skeleton.txt. But when I removed the -util:animdb from the shortcut, the game ran fine. This tells me that the game IS reading descr_skeleton.txt and is trying to parse it, even though the four animation pack files are still present.

**JeromeGrasdyke** · 07-08-2005, 11:28

What the code does is this:

* if you run with -util:animdb on the command line, it parses descr_skeleton.txt and tries to build the animdb files from it, regardless of other factors;
* if you run without the command line option, it looks for the packed files, and if it finds them it does not use descr_skeletons.txt as all the information is inside the packs;
* if it fails to load the packs, it falls back on loading descr_skeletons.txt, but does not regenerate the packs -- renaming pack.dat, pack.idx, skeletons.dat and skeletons.idx inside data/animations/ should cause it to take this code path;

In order to load from descr_skeletons.txt you need to have a complete set of all the files it references in the correct places - that includes all the .evt files in the zip file that I supplied some time ago, plus the .cas files generated by Vercingetorix's unpacking tool. If it can't find any of the files, the code should warn you, providing that you have -show_err on the command line. In general there should be matched pairs of .evt and .cas files.

This process works correctly when I try it with a dev build in release mode, so in theory it should work for you, but to the best of my knowledge no modder has yet managed to get it to complete... Exactly why is not really clear at this point.

**Lord Adherbal** · 07-08-2005, 11:52

I gave it a try, and like dsyrow it crashed with no error (-show_err parameter is there).

I will try it with an unmodded RTW version later today, just to be sure the new skeletons I'm using in descr_models_battle.txt (which are not defined in descr_skeleton) dont cause the problem - although I would expect it to return some kind of error message then.