DDraceNetwork - developer

[ddnet/ddnet] Pull request opened: #1344 CMakeLists.txt: Replace PythonInterp with Python

FindPythonInterp is deprecated since CMake version 3.12. Use Python to automatically detect Python 3 or 2.

Ooh, could go into ISUPPORT

btw, since i updated to 11.4.4 the client has these freezes, but since the update also older clients have it (different config_directories)

@noby how do i access a gameclient.h var from client.cpp?

10:29

GameClient()-> wont let me do it

is it private var

no

10:32

or @noby, how can i access a gameclient.h var in client.cpp or a client.h var in gamneclient.cpp?

10:32

not owrking xd

10:39

@Ryozuki help

10:39

xd

sry im doing otherstuff rn

10:39

dealing with huffman

xd

fokkonaut

[ddnet/ddnet] Pull request opened: #1345 add DDRace Score server message

the server admin can edit sv_ddrace_score in his autoexec file. if its 1, it allows the client to use ddrace scoreboard, if its 0 it disallows it. this would work nicely with IsDDNet by timakro, to have a server with a custom gametype, ddrace features, but also the normal score (not ddrace score with times). it can also be used to run mutliple gametypes on a single server without having weird time scores (need more editing serverside ofc).

why tw uses huffman with a hardcoded frequency table? wouldnt it be better to create the freq table based on the packet and put it on the header or somewhere

11:10

or maybe i didnt get how huffman works..

11:10

but i think what i say makes sense

11:10

you also get more compressed data i think

11:11

-test-

[quakenet] <Ryozuki> test

because that's table based on already computed character frequency for accumulated possible tw messages if you will be recompute new tabel for each packet that's will be more slow and packets size will be bigger than now

11:54

if i right understand y

hmm not sure

Also, sending the table with the packet is going to be useful only for large packets (since otherwise you won't decrease the size)

yeah i thought that too

and there is a limit on UDP packet size, anyway

and why use huffman and not another algorithm?

11:56

i never heard of it until tw xD

11:56

tho im not much into these things anyway

tw use small packets more often, 95-99% packets small (small for any compressing advantages with recomputed included table), can y give an example of often used big packets? (edited)

bors[bot]

[ddnet:master] 2 new commits

c1523d9 CMakeLists.txt: Replace PythonInterp with Python - yangfl 865678a Merge #1344 - bors[bot]

11:58

bors[bot]

[ddnet/ddnet] Pull request closed: #1344 CMakeLists.txt: Replace PythonInterp with Python

@Ryozuki Huffman is fast and simple, should be the default choice for something like this. What else would you use?

you mean

13:11

I joined the server that full players (edited)

13:15

also, people playing with testing

will 0.7 bring us normal (and awesome) official client updating like in ddnet?

yes

0.7

xD

I ask because somebody write on forum that's 0.7 so fresh for using and it's not rly good idea release it as solution for current connection without handshake problem, but if there are updating system be available, I think official (maybe marked as beta) must me released as soon as it possible

@deen i was just curious if there was an alternative

@Ryozuki https://en.wikipedia.org/wiki/Arithmetic_coding

Arithmetic coding

Arithmetic coding is a form of entropy encoding used in lossless data compression. Normally, a string of characters such as the words "hello there" is represented using a fixed number of bits per character, as in the ASCII code. When a string is converted to arithmetic encodi...

heinrich5991

[ddnet/ddnet] Issue opened: #1346 Test on older CMake versions

#1344 bumped the minimum required CMake version without us noticing, that shouldn't happen. I think our current minimum version was somewhere around 2.8.12.

am i dumb or this doesnt make sense?

14:29

CNode m_aNodes[HUFFMAN_MAX_NODES];

14:30

isnt this undefined behaviour or smth like that, it increades m_Numnodes when its already HUFFMAN_MAX_SYMBOLS

14:30

am lost

14:30

m_aNodes[m_NumNodes] shouldn't even work?

HUFFMAN_MAX_NODES != HUFFMAN_MAX_SYMBOLS?

@timakro