Random Chitchat 2012-2016

[xoft]

Frankly, I have no idea what tolua.setpeer() does, it is not exported by MCS, but rather by the Tolua++ glue code generator itself.

You cannot read other plugins' memory at all.
1. Each plugin runs in a separate Lua context, there's no way to leave a context in a plugin; the cPlugin:Call() function actually copies stuff from one context to the other in order to make the call.
2. Even if you did have a way of reading other plugins' memory, you couldn't use it because potentially the plugins may run in parallel - for example one plugin serving a webadmin page while the other is ticking the world. You know what happens in thread-unsafe accesses, right?

---

Just wondering, but what is currently stopping us from, creating a plugin that handles end portals + nether portals ?

End portals should be okay. Nether portals are currently probably not possible for one simple reason - they require the player to be in some kind of "limbo" while the world around them gets generated and *then* the portal gets placed. MCS doesn't yet have the concept of this limbo, players in MCS need to always have specific coordinates in a world and then all the physics from the surroundings is applied to them.

Another sub-reason is that plugins cannot force MCS to keep chunks in memory so that they can look for a suitable portal placement. A special class, cChunkStay, is used in MCS code, but it's not available in the API (yet?)

[tonibm19]

I'm amazed of what can a 128mb server do. MCPC (bukkit + forge) with MoCreatures, 2 players, and no lag.

[bearbin]

I assume your tickrate will be terrible and you are swapping, but cool

[tonibm19]

19-20 tps, swapping of course.

[tonibm19]

20 tps, 3 mods, java 8.

[xoft]

We seem to have broken OhLoh - it currently lists languages used in MCServer include "DOS batch script" as having -22 lines. How can we have a negative number of lines?
https://www.ohloh.net/p/mc-server/analys...es_summary

[bearbin]

Derp. What is ohloh btw?

[ThuGie]

Btw xoft, is it not possible to give the server a option to choose zlib compression level, aka 1 = fast compress high bandwidth, 9 = slow compress, low bandwidth ?
Or even compression strategy ?
compress2 does now give you access to it but altering zlib code itself its quite easy to change it.
I wanted to mess with it see the big differences as on mc classic RLE strategy gives a huge gain in speed and lower bandwidth as well.
Even if current map format is really different, compress speed and map size might still have some gains, also would defining FASTEST give any improvement ?

Some sample of a guy using FASTEST
http://mail.madler.net/pipermail/zlib-de...02674.html

[xoft]

What is ohloh btw?

It mostly calculates statistics related to the code size, languages etc. Also has kind of a "ranking" system for all the developers - when you create an account there, it calculates all the code you've written and how the community sees you (you can give people "kudos"). I enrolled MCS mainly to see the code statistics.

Btw xoft, is it not possible to give the server a option to choose zlib compression [...]

Interesting thought, I've never considered it, but it could be done, theoretically. Both for the network protocol and for the world storage.

The network data is compressed in the ChunkDataSerializer: $/src/Protocol/ChunkDataSerializer.* and the level parameter is directly changeable there compile-time, if you want to experiment.

The world storage uses StringCompression for the actual compressions, you might need to add a parameter to the CompressString() function to allow setting the compression level.

[ThuGie]

How about the strategy, and FASTEST ?
Btw where its located etc i already know, been messing with it, but compiling takes so long..