Difference between revisions of "Talk:Pre-release protocol"
Line 3: | Line 3: | ||
− | According to the [https://github.com/sadimusi/mc3p/blob/master/mc3p/messages.py#L336 | + | According to the [https://github.com/sadimusi/mc3p/blob/master/mc3p/messages.py#L336 mc3p proxy server code], the last field of packet 0x36 (Block Action) is a short. I think the article is wrong. |
--[[User:DavidEGrayson|DavidEGrayson]] 10:57, 29 July 2012 (MST) | --[[User:DavidEGrayson|DavidEGrayson]] 10:57, 29 July 2012 (MST) |
Revision as of 17:57, 29 July 2012
In the discussion of the removal of the Chunk Allocation Packet, the article says: "To deallocate memory you need to send a MapChunk packet, which only contains biome data for plains(all biome bytes set to 1)." According to my testing of a custom client with a vanilla Minecraft 1.3 server, when a chunk goes out of range the server actually sends a Chunk Data (0x33) packet ground_up_continuous=true, primary_bit_map=0, add_bit_map=0, and real actual biome data (in my case it was mostly plains and some desert). This suggests that the sentence should be edited to: "To deallocate memory you need to send a Chunk Data (0x33) packet which only contains biome data.". If someone else can confirm this, please post here. I think this makes more sense, and it's easier to test for in the client because you don't have to look at the compressed data at all. --DavidEGrayson 23:33, 28 July 2012 (MST)
According to the mc3p proxy server code, the last field of packet 0x36 (Block Action) is a short. I think the article is wrong.
--DavidEGrayson 10:57, 29 July 2012 (MST)