Cold Blood

Members
  • Content count

    106
  • Joined

  • Last visited

Community Reputation

0 Neutral

About Cold Blood

  • Rank
    Advanced Member
  1. Actually I was thinking of adding dynamic parameters passed to the compound, like arg1="value1" arg2="value2" and so on. And in the compound you would have cond="arg1 == 1 and arg2 == 2". And is seems similar to the arr attribute. And it seems more flexible to me and easier to implement . Also I think I can get it to work with arrays too. So instead of arg3="value" you can have arg3="array[2]". This seems pretty easy to implement. All in all I want to code in a way that is compatible with the rest of the stuff, like nifskope and niflib and to make the nif.xml compatible with everything as much as possible. I think I know the problem with nifskope, you pass the arguments for each vertex and if you have 15 thousand vertices, it might take a while.
  2. I understand now, it all makes sense, why you need an array for passing arguments. I don't know why the code generated in niflib doesn't pass the arguments.
  3. This is the code from the nif xml that uses BSVertexData compound and sends a parameter to it. But as I see it, you can only send one parameter to it which is wrong. cond="(Data Size > 0) && ((VF7 & 64) == 0) && ((VF7 & 1) == 0) && ((VF6 == 176) || (VF6 == 16))" /> And this is the component being used: Catch all Vertex Data compound, SLOW I thin it's not just one flag, but multiple flags in those conditions which use more than one parameter.
  4. I am not too sure but I think figment also made some changes in the code that is regenerated from the xml so the changes would get lost. But I am not too sure, I just looked a bit into this. If he did that, then maybe there are some restrictions in the niflib code generation scripts, I do not know. I will post the xml parts with the issues here once I get back from work .
  5. There were also some invalid type declarations. I saw arg but was used in a different way for templates, this one is used to pass parameters and not template parameters which are a bit different. I am referring to passing some parameters from another xml block to a condition in a child xml block. I still need to study niflib a bit more. There also some changes between figment's nif xml and the latest xml in nifskope alphas. So I got the updated nif xml.
  6. I almost finished updating niflib to the latest nif.xml from the alpha versions of nifskope that support fallout 4. But I ran into a problem, the code generator for niflib doesn't support some new stuff from the updated nif.xml. I am talking about passing arguments to a nested compound type from the parent. This could be a problem and hard one. I can easily fix it by removing the passed arguments and adding the condition directly in the compound type. But I don't know if it's supported in the rest of the library, especially in nifskope. But this is similar to conditions that use user version and user version 2, so I guess I can use it. And there were a ton of bugs to fix. Enum values clashes, wrong characters and names in members. The good news is that I might be able to import segments and export segments. I don't know what they do but they seem important. I think the max exporter doesn't do that .
  7. Also, could this thread be renamed to Maya Fallout 4 support and put as a sticker on top? Can the older ones be removed since they are obsolete?
  8. I looked into half floats and found a library for them. I can store them internally as unsigned short and that will be the type in the nif.xml. I need to update the fallout 4 nif xml with an actual valid type for the half floats. I will just create get/set with float values and convert them internally to half floats I can't tell if I will be able to make this work with fallout 4 yet. I also don't have a clue about dismember partitions and how they work in fallout 4. There was some kind of submesh info node. I tried to update the maya nif plugin repository and got an error with niflib saying it's not a valid repository. I still don't know if niflib can be updated with the new fallout 4 xml...
  9. I know about those bugs, fixed them when I updated the plugin for maya 2016. Now I am looking into fallout 4 support. I figured out how to regenerate the niflib from the fallout4 nif xml. I know about the "special" half floats. There are a ton of libraries on the internet. I just need to store them as unsigned short and encapsulate them with get and set which accepts a normal float and converts them to the half float. But the weird bitagent doesn't make any sense to me . Don't know how to generate that. I am not too sure if I can get the plugin to work with fallout 4. I think the max one already works with fallout 4 meshes .
  10. Hello, saw your reply here right now. I have been busy. I made some new mods for skyrim and so far I have improved the plugin here and there to work properly for Skyrim. Nice to see that people are still active here .
  11. Hmm, I can get the xml but I need to run a tool on it to update the c++ code as far as I know. But I can't find anywhere how to run the tool and what tool.
  12. Is there an updated niflib available so I can make the maya exporter work with fallout 4 meshes? I can also contact the people for fallout 4 script extender for the ba2 extractor, I think I can use it for the maya plugin with slim chances. I think I will run into some big issues making it work but who knows. I will at least look into it.
  13. Hello, I developed a custom version of the maya plugin available here: http://www.nexusmods.com/skyrim/mods/65421/? That is if I don't get banned on nexus for some crazy reason in which case I will post it here . On seconds thoughts I will post it here..
  14. I think I am making a confusion here. Isn't the "push" operation how you put back your changes into the remote repository from the local repository? And doesn't "pull" just notify the main fork that some changes happened in a branch and it should be updated?
  15. My main headache is the fact that the plugin depends on the niflib library. I made some changes and since the current submodule is tied to the main niflib branch, I can't push my changes since I do not have admin rights for the main branch. I would need to change the local niflib submodule to point to my personal niflib branch. I am used to team foundation server and svn that integrated great with visual studio. Personally I hate the shell because it is a pain to use and really slow. At work I sometimes modify as much as 200 files at once, adding, deleting and modifying which is too much to manually track the changes with the git command line. At most I changed 600 files at once and I had to check each of them to see that they are modified properly .