Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Quote:
Originally Posted by
azaidi
What's up with discouraging other developers?
I'm not doing this for making my own hotel, I'm doing it for the coding experience, what you should do it for anyway. If you want to run a succesful hotel use Phoenix Cracked, because this will never beat it. But I'm still going to work on this untill it becomes usefull and code CMS'es on this.
Because Sulake told him to aha :P
OnTopic: Anyway nice, looked through this good stuff keep it up :P
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Quote:
Originally Posted by
FullmetalPride
Pretty awful for someone to say. Like your emu was finished. Sure Sulake forbid you but work under a Latino alias or something. They won't know the difference.
Oh god, you just ruined his opportunity to do that.
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Quote:
Originally Posted by
Quackster
Or you could just use a complete/near complete server instead. :laugh:
Haha haha like what near complete server? :laugh:
Butterfly is just crap to use as a base though it is stable its still crap.
Quote:
Originally Posted by
Aquel
Oh god, you just ruined his opportunity to do that.
Quacky wont do that since he listens to people that tell him what to do / not to do.
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Quote:
Originally Posted by
tdid
Haha haha like what near complete server? :laugh:
Butterfly is just crap to use as a base though it is stable its still crap.
Quacky wont do that since he listens to people that tell him what to do / not to do.
Butterfly aint crap..
Far from that, original uberEmu is crap! (the leaked released one.. sorry.)
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Update: Starlight 0.3.rar
Can someone find out why the pathfinding is so bugged? (It works but you walk like an autistic dog) Also please give me some feedback about the coding, I can't do anything with the usual "not useful" or "nice".
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Quote:
Originally Posted by
tdid
Haha haha like what near complete server? :laugh:
Butterfly is just crap to use as a base though it is stable its still crap.
Quacky wont do that since he listens to people that tell him what to do / not to do.
I wasn't talking about Butterfly. I was talking about Holograph, Thor. They're pretty decent.
Quote:
Originally Posted by
azaidi
What's up with discouraging other developers?
I'm not doing this for making my own hotel, I'm doing it for the coding experience.
If you're doing it for your own coding experience, I don't see why it has to be released. It doesn't have sufficient features to run a hotel so therefore it's useless.
Quote:
Originally Posted by
azaidi
Update:
Starlight 0.3.rar
Can someone find out why the pathfinding is so bugged? (It works but you walk like an autistic dog) Also please give me some feedback about the coding, I can't do anything with the usual "not useful" or "nice".
There are reasons for your multiple issues with your pathfinder. Firstly let me point them out to you;
- You create a new thread each time the user walks (very unstable). It should be per room instead.
- This method call
Should be 500 milliseconds instead, this processes the thread with less speed so it's like normal walking. If you want to see how a pathfinder is properly implemented, look at Room.java from the Sierra leak. The syntax is almost the same so you should be able to understand it.
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Quote:
Originally Posted by
Quackster
I wasn't talking about Butterfly. I was talking about Holograph, Thor. They're pretty decent.
If you're doing it for your own coding experience, I don't see why it has to be released. It doesn't have sufficient features to run a hotel so therefore it's useless.
There are reasons for your multiple issues with your pathfinder. Firstly let me point them out to you;
- You create a new thread each time the user walks (very unstable). It should be per room instead.
- This method call
Should be 500 milliseconds instead, this processes the thread with less speed so it's like normal walking. If you want to see how a pathfinder is properly implemented, look at Room.java from the Sierra leak. The syntax is almost the same so you should be able to understand it.
Sorry to not update the thread every 2 seconds, I already fixed pathfinding yesterday. I know about the roomhandling, I'm gonna recode it after this works better and I cleaned up some code so recoding is easier. Also, I released it here because I want feedback from some coders. If you'd like to come on Skype to give me a lot of feedback about my coding I'll close this thread if you want. I just want to improve my coding by listening to other people's advice.
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Add some try { } catch or it will hang.
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Ok, kids, let me explain how this stuff works. First of all, adding a try-catch block around ANY code would NEVER make it hang less.
Having one thread per room is shooting yourself in your foot. It would cause X = amount of loaded rooms to be added to your CPU to be cycled. It scales incredibly poorly and literary eats up system resources.
The way to go is pooling, as I did in Butterfly. It would probably be a better idea to use a timer or something like that, but the way I did it in bfly was good enough. The difference between this and the other approach is that the OS handles how many threads should be running, not the amount of rooms that is loaded. That way it is more scalable and the calls are asynchronous.
Edit:
Also, here is a way better socket which I used in bfly: https://github.com/martinmine/Connectivity It basically talks directly to the Winsock API so you don't need all the overhead caused by the generalized socket class in .NET
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Quote:
Originally Posted by
maritnmine
Ok, kids, let me explain how this stuff works. First of all, adding a try-catch block around ANY code would NEVER make it hang less.
Having one thread per room is shooting yourself in your foot. It would cause X = amount of loaded rooms to be added to your CPU to be cycled. It scales incredibly poorly and literary eats up system resources.
The way to go is pooling, as I did in Butterfly. It would probably be a better idea to use a timer or something like that, but the way I did it in bfly was good enough. The difference between this and the other approach is that the OS handles how many threads should be running, not the amount of rooms that is loaded. That way it is more scalable and the calls are asynchronous.
Edit:
Also, here is a way better socket which I used in bfly:
https://github.com/martinmine/Connectivity It basically talks directly to the Winsock API so you don't need all the overhead caused by the generalized socket class in .NET
Are 500ms timers for the pathfinders, and then one thread cycling through all rooms better?
Re: [Starlight] SocketAsyncEventArgs Sockets 0.1 - BETA
Quote:
Originally Posted by
azaidi
Are 500ms timers for the pathfinders, and then one thread cycling through all rooms better?
500ms is standard. Having one thread which calls the cycle function on ALL the rooms is just as bad as having one thread per room. If one thread lags, all the other rooms would lag as well. Therefore, thread-pooling.