1.8.9 Porting #681
Comments
|
Just curious here, what exactly is the nightmare with rendering? |
|
There are some long winded explanations around but basically it requires a complete rewrite of all the mod's rendering code. The player really won't notice other than the time it takes to get the mod ported, but the developer will pretty much look like this to get it done:
|
|
Checked out how bad Numina will be with 1.8.8, so far got it down to 45 errors. |
|
Cool, let's prioritize that for now then. I'll have to fiddle with the build scripts to get a proper separation between 1.7.10 and 1.8.8 builds. |
|
It's the render code that will be the toughest to deal with. I looked at a short block tutorial for 1.8.8 and almost immediately lost my faith in humanity. |
|
Yes, I remember begging Grum not to do this. |
|
Just a thought, but how is the player going to charge the armor in 1.8.x? On a side note, LOL!!
I was expecting that, just not worded quite like that. |
|
I'm sure there is some kind of power API in 1.8+. Regarding rendering, we will have to circumvent the new rendering system rather than using it directly, since it is limited to cubic shapes - pretty much just like before. Does the updated Forge have working wavefront obj support? |
|
If i remember correctly, in Forge 1.8.9/1.8.8 there is a new rendering system, Made by @RainWarrior. |
|
Might be a good opportunity to go power-free instead. Limit modules by power generation of a generator module, instead of batteries. |
|
Yes, it appear according to one of the issues on Forge's github. As far as power goes, an API is one thing, a way to generate it is another. |
|
Alternatively just keep it as-is and let people charge using MPSA generators. |
|
They are very slow to charge. |
|
ok. We could also finally implement a 'storage' module with a UI that contains several slots which are unlockable by different modules:
This was a planned feature from the start but I got frustrated with minecraft's UI system (obviously) and never got around to it. |
|
Sounds interesting. Looks like this is going to be quite an overhaul. By the looks of it, none of the modules with external dependencies will carry over for awhile. |
|
Progress so far, LOL!!! : https://gist.github.com/lehjr/1709dfde96643ee8ff2d That's Numina and MPS in the same workspace. |
|
Update directly to 1.8.9 if possible |
|
Actually, I'm currently considering falling back to 1.8 and starting from there. 1.8.9 isn't really an option at this time. With a new version every day, it's too unstable to target, I'm not even sure about 1.8.8 at the moment since several dependencies including NEI are missing. |
|
I think you can use JEI, every 1.8.9 mod uses It as far as I know, like blood magic, tech reborn and others. I read that is a nei replacement with an easy-to-use API. Maybe @mezz can clarify. |
|
1.8.9 should be the current target for mods doing updates. By the time you finish, 1.8 will be completely dead, especially now that Thaumcraft has updated to 1.8.9. |
|
Well, at least the MCP mappings for 1.8.9 haven't changed for a few days. Wait, "This set of MCP mappings was designed for MC 1.8.8. Use at your own peril." Um... |
|
Yes, a few other sources have confirmed that JEI+1.8.9 is the standard to shoot for if we can get this done in any reasonable amount of time. Though there is still much uncertainty about power systems. |
|
I don't think it's impossible. I wish my skill level and experience were a little more fitting for the task, but I'll learn as I go, as I always do. My biggest setback is a lack of working examples and documentation, making it hard to figure out how to go from "what was" to "what is". Some of it is cleaner with the use of pos instead of x, y, and z, but some things are quite intimidating. And ffs, it's like someone is getting a commission for every json file needed. Right now, the only working example I have to work from is this: It does have a couple of things though, like OBJ rendering. Edit: for power I'm using the RF API copied from buildcraft's experimental builds. |
|
This is a great read: |
|
So it looks like Buildcraft has an Alpha build for 1.8.9. That should be a good start for getting power. For 1.8.9 I think it would be best to add "optional recipes" so that the default recipes are overwritten on update. That will address the insanity of having to constantly address recipe issues have already been fixed. |
|
Sort of have something rendering, too bad it's stuck to the player: |
|
ahahaha adorable :> |
|
Sadly, I'm still looking at what to use for rendering the TinkerTable. Since IBSR's were removed in 1.8 I used a TESR just to see it working in the current workspace. I'm not entirely sure if I can avoid using one, but I'm looking at alternatives. |
|
The armor models will be the next challenge. I'm reasonably confident that will be easier. On a side note, the 1.8.9 mods seed to have more than doubled in the past week. |
|
This is what I was talking about with the wall of warnings: https://gist.github.com/lehjr/d1acec7f03bd1d3e62f6 |
|
decided to hold back on 1.9 at least until I can get 1.8.9 to a beta state. I still need to write a JEI recipe handler (in progress), fix the Lux Capacitor, and fix the power fist rendering (maybe use item entity?) first. I'm not entirely convinced the current OBJ model rendering will work for the armor since groups seem to be tied to blockstates. The JEI recipe handler is the priority at the moment, since it seems players want to be able to craft items. |
|
finally got enough quiet time to get the recipe handler working. :D |
|
I'm gonna push out the update for Numina with the recipe handler this weekend. In the mean time I want to see what I can do with the model rendering. The big thing is to resize the textures without hosing the UV mapping. Edit: looks like as long as the textures are re sized by stretching, the mapping should stay the same for the OBJ models. |
|
I'll be working on a 1.9.4 port in the next couple weeks. From a gameplay perspective, the biggest issue of the moment is the lack of high end power generators, making charging anything but the most basic power level nearly impossible. For 1.8.9, I'm using Advanced Generators and Pressure pipes. From a development standpoint, I'm going to need to fix a couple of things first, like the recipe handler which is missing the ability to place the items in the crafting interface as well as fixing the handling of recipes. There's also a weird behavior where if the first line of the recipe doesn't have all 3 placeholders used, the recipe displayed is rotated and invalid. The model issues will likely be addressed in 1.9.4. The power fist model needs to be fixed. I don't know if we'll ever get animated item models back again though. The other thing I'm not sure about is the customization for the armor models like we had in 1.7.10. I'm just not even sure how to implement it. Also, on the Lux Capacitor. I'm thinking about getting rid of the custom color for it and maybe just having a couple preset colors. In doing so, we can completely eliminate using a TE and a TESR, allowing these to be useful for base and tunnel lighting without really effecting performance. I'm also considering changing the recipe setup a bit so that there are both builtin/default recipes that will get overwritten on loading and optional additional custom recipes to help eliminate the bug reports for recipes that have already been fixed but don't get overwritten. |
Or you can have a config option to allow both variants. Technically I'm rendering (5x5x5 - 3x3x3) FastTESR models on one tank in my mod as well and it doesn't cause any visible lag. Still, I'd like to page @RainWarrior on this, since he's a bit more experienced with rendering and that stuff, iirc :P |
|
I did an experiment with the current setup and it took about 50 or so before I started seeing lag on my PC. A higher end video card would take more, but the point is better optimization. Currently the only thing the TESR is being used for is the colored lens, the model itself is a JSON model, but it doesn't always work. And quite frankly, when it comes to the rendering code, I'm not just horrible at it, I'm short-busing it all the way. |
|
Is the newest code on your Github as we speak? |
|
*affecting performance As far as limited colour options are concerned, if you decide to do that, I'd suggest going with the standard 16, possibly excluding black. |
|
Be warned, this code is such a mess that it may cause paralysis, temporary blindness, and insanity (no really, I'm not exaggerating): MPS code here: https://github.com/lehjr/MachineMusePowersuits/tree/1.8.9 |
|
If an issue of yours with 1.9.4 is power, there's EnderIO. I absolutely LOVE that mod, it has vibrant capacitors, and vibrant energy conduits transferring up to 20k RF/t. High end enough? By the way, EnderIO did some fancy stuff with rendering. If that helps in any way. |
|
I love EnderIO, but a fully charged power suit with maxed out power storage can hold millions of RF per piece and it would take several top end generators and days of charging. My last setup in 1.7.10 was using a Mekanism fusion reactor. My current setup in 1.8.9 is using Advanced Generators Steam Generator with bronze turbines and 2 heat exchangers with lava pumping from the nether and an EnderIO wireless charger. |
|
@lehjr maybe you can use the creative capacitor bank from EnderIO or debug item+MFSU from IC2 to charge items. |
|
@Xen1290 for power I was referring to actual gameplay as a regular player would play. Without the ability to charge the armor, it becomes useless rather quickly and likely would be unused, and not generating the feedback I need to find and fix issues. I just took a couple weeks off from working on the mod, to work with the mod, playing with it as an actual player to find and fix issues. The recipe handler was horrible and there were broken plating recipes that I didn't know about. I also found the tooltips for the power are missing. I fixed the recipes and the handler, but the broken tooltips are still on my todo list. The hoe module also seems a bit off (it's supposed to be able to do more than a 1x1 square). I'll probably fix those last 2 issues before moving to 1.9.4. |
|
If you you are thinking of making your own armor charger, maybe an armor stand that you could right click and have it transfer armor and power tool onto it, maybe burn coal, or hook up your own generator to it. But also allow rf pipes to hook up from an external mod. |
|
No, I'm merely pointing out that the top tier power modules require more than simple generators and that without those, the mod receives less usage than needed in order to get the bugs worked out. There are MPS modules for charging, but they are quite underwhelming. If I were to add any further methods of charging, it would likely be through modules as adding external devices doesn't fit well into the theme of the mod. An armor stand that does charging would be better as a stand alone mod as there are other armors out there that use power. The problem with implementing high tier charging in MPS is balance and mod interaction. If you add an internal method of charging at the highest level possible, then those high end power mods really don't have as much draw to them. Not only that, but the armor itself could end up being a source of infinite power. It would really nice to implement the EV Capacitor, which would become the highest tier power system, but again balance and power issues. Charging is much more problematic than you would imagine. In fact, there is a partially implemented coal generator in the code, but in order to make it balanced, it needs to put out the same amount of charge as a stand-alone generator, which isn't much. Keep in mind though that you can also use charged devices as power sources by carrying them in your inventory, such as EnderIO capacitor banks and various devices from IC2. |
|
Oh sorry, now I understand what you mean |
|
I still don't get one thing - is enderIO's Vibrant tier too low for you? |
|
@lirannl there's no high tier generators to go with them. Right now, in my actual gaming setup, the Advanced Generator setup is producing over 10k RF/tick. EnderIO has no such power generation. Yes, you can use capacitor banks to store power, but you still need a way to charge them in a reasonable amount of time. |
|
Top tier MPS armor piece stores 50 million RF for each piece. So the question becomes how do you charge 250 Million RF in a reasonable amount of time. Edit: I mean without using top tier power generation mods. |
|
I think when you have "Top tier MPS armor", you also have "top tier power generation" O_o //EDIT: Also, one Endergenic Generator setup from RFTools can yield to ~25k RF/t or even more. //EDIT 2: Taking 250m RF into account for the full armor, + 25k RF/t from the generator: So around 8 minutes. |
|
Not without relying on mods that may not be out. With "Top tier MPS armor" you have top tier power consumption and storage capability, but very low tier power generation capabilities in the mod itself. |
|
@Lordmau5 right, but again, that would be considered top tier generation rather than something like EnderIO's Stirling Generator at around 20-60 RF/t where it would not be considered a reasonable about of time. |
|
@lehjr |
Sure, at a max of 60 RF/tick you'll need at least 250 of them to make it worth while.
The power output wasn't enormous. Fuel consumption scaled with the output and at times became impractical. Even if they come back, it's not likely they'll change the way you charge the armor.
Such as? EnderIO has no way of generating a lot of RF since it doesn't need it. I'm still not seeing all that many other options yet. Maybe Mekanism if/when it's done for 1.9.4.
or maybe I do some crazy stuff and charge the armor directly with mana :P (probably not, but it might be an option)
I actually do HAVE to worry about such things as being able to charge the armor. This isn't really a situation where "if you build it, they will come". Things have to be done in a practical manner or there really isn't much point. Anyway, it's possible that I could implement the EV capacitor which would become the new top tier power storage and then re-scale the lower tiers while not really raising the ceiling much higher. I don't know yet. It may or may not really be worth while to implement. |
1.7.10 has been mostly busy work for awhile now, for mods, for Forge, and for players stuck on the sidelines waiting for updates. Aside from the nightmare of the rendering, there are also several dependencies that may not be available. Maybe it's time to start working on the port and disable the modules that have dependencies that can't be met until they become available. I just hope the performance is better than 1.7.10.
The text was updated successfully, but these errors were encountered: