#ChangeBrazil
- Joined
- Nov 4, 2012
- Messages
- 928
- Reaction score
- 545
Why don't you give a sh*t about the opinion of 2 people out of 16?The fun part of this crap, i have posted a topic about a season 5.4 GMO and made a poll:
But the n00bs that voted no, never come back to post a valid reason to that.
And i doubt that someone wants to do the things for the community would be encouraged because of this n00bs.
Again, forget it. Is a waste of time.
Why don't you give a sh*t about the opinion of 2 people out of 16?
Anyways, a bit offtopic: You rely on sources of MuEmu, but don't release the sources of your own changes? Of course, you can do that, but then you don't have to wonder why this community is like it is... My opinion: I as a developer don't see much value for the community in contributing to a closed source project.
i can try help to decompile a main....
wow, i think you decompiled function make effect for dl cape.can you share it?Few years ago I decompiled like 70% of S6 main as IGCN team member. I had literally everything I wanted from this game, file formats, whole render (maps, objects, effects) and internal structs (items, game objects like players and monsters). Parts of this decompilation were included in IGCN S6 Dll with custom monsters and other stuff (we called it silver edition). I decided to drop this project, and to move everything to external project where I tried to rewrite MU engine (basing on this decompilation). While it was very successful I realized that writing engine itself is just waste of time. Please don't repeat my mistakes, because this engine is literally black hole.
Later I moved whole project to Unity engine. Today after few more years, it's still not ready. This is not something you can code it one one or two years. And larger team won't make it any better, maybe more chaotic. After all this years I'm sure that community will never finish such complex task. Actually publishing SRC would kill that project. Look at other MU projects. Most of them are successful because they are one-man projects, or are made by very hermetic teams like IGCN. Even Deathway 0.99.60 decompilation (which is probably the best thing that happened to this community like 10 years ago?) it wasn't large community project, it was mostly two guys with some help (correct me if I'm wrong, it was long time ^^).
I can share few screens from our work:
You must be registered to see links
You must be registered to see linksYou must be registered to see links
While It's playable, it's not even close to be finished.
Last time we have converted all files to .fbx (autodesk), since it was more Unity friendly and it is much more easy to manage...How did you manage to read the mu client resources? Did you convert all, or did you write code to read them in as-is?
I made this decompilation about 6 years ago. I don't even have this IDA project or even CPP version. All stuff i have right now is entirely rewritten to C# and unity shaders. But I can give you some insights of how mu handles these effects.wow, i think you decompiled function make effect for dl cape.can you share it?
no effect when add custom cape
I read them as-is. I tried with conversion, but it's black hole. You will end up with hundreds of issues every time you need newer Data folder from client. Writing code to load BMDs and map files is one time job. Also I can easily create Unity assets with this code, but I think it's worse than dynamic loading. The hardest thing is animation loading. Creating animation clips from code in unity is really painful, since unity lack documentation in this area.@supron: Thanks for your insights. I also think decompiling the main.exe and rebuilding it's engine is a huge task. Writing an own engine nowadays is always a bad idea, since there are a half dozen good ones out there. How did you manage to read the mu client resources? Did you convert all, or did you write code to read them in as-is?
You need public repository for such project, but you can't create it. You need stable platform to handle multiple branches, pull requests and good issue tracking.I understand why a larger team could make it worse... I see it at work (~10m LOC, ~30 Devs) every dayHowever, this problems are more organizational... I think it's totally possible to work on such a project with more people. It's all about splitting the work in as many unrelated parts and work items as possible, so more teams/people can work on it independently. It doesn't matter for the success if it's open source or not... the fact that closed source projects had more success here, is based on the fact that there are not many serious open source projects here. And that's because most developers here want to make money instead of helping the community
Actually classic VBO layout is very easy to reconstruct from BMD file, and can be done in real time. You have to do it anyway if you want to convert BMD->FBX.Also if you want to use "as it is", you will need to rearrange all the data on the memory, since you can't use VBO because WZ interleaved all data information.
Actually classic VBO layout is very easy to reconstruct from BMD file, and can be done in real time. You have to do it anyway if you want to convert BMD->FBX.
We are loading files to structs in multiple background threads, and creating unity Mesh class in main unity thread. The slowest part is AnimationClip class in unity. You can't create it in background thread, and it's a bit slow (especially SetCurve method). But it's still way faster than webzen's client.
OP would be a dream trying to unified all of mu developers..its a impossible task....but I think that exist one thing that would be helpfull for mu community and its:
Fix the limit 25fps of the main and get at least 40fps playable without lag spike...if all of the team made a reward for a person that do that..for example each 10 dollars o whatever...maybe someone can achieve that.
We as player dont need a complete remake of the client...we only need play a Castle Siege...a arkawar with more than 100 player without lag spikes or playing at 10 fps
As far as I know, MU is software rendered. Is there a way to decompile the rendering pipeline and try to change some drawing calls to OpenGL or DirectX?
You must be registered to see links
Our emulator is optimized to work with stable 80 fps. On simple scenes like this (few NPCs and effects) I can have even more (up to 200-250 fps). It's Intel HD 530 card
Also this is almost "final" version. We won't add any resource expensive features from this point.
You must be registered to see links
Our emulator is optimized to work with stable 80 fps. On simple scenes like this (few NPCs and effects) I can have even more (up to 200-250 fps). It's Intel HD 530 card
Also this is almost "final" version. We won't add any resource expensive features from this point.