Luanti logo

IRC log for #minetest-dev, 2023-02-14

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Enable nick filtering
Time Nick Message
15 more elements. Show/hide.
11:11 Zughy[m] #12315 can be merged rubenwardy
11:11 pgimeno https://github.com/minetest/minetest/pull/12315 -- Add world-independent storage directory for mods by rubenwardy
16:34 sfan5 rubenwardy: did you test #13205
16:34 pgimeno https://github.com/minetest/minetest/pull/13205 -- Android: Increase minimum SDK version to 21 by srifqi
3 more elements. Show/hide.
18:41 caffblks Anyone know the maximum amount of data that may be sent between client and server in a single UDP request?
18:42 caffblks I'm building a proxy for minetest and want to ensure that I don't have buffer overflows or other issues
18:43 sfan5 the hard limit is 2^16 but that obviously doesn't fit in a single packet
18:43 sfan5 (not one sent over the internet that is)
18:45 caffblks hard limit for UDP datagram is 65507 over IPv4, no?
18:46 rubenwardy the limit depends on the network, it's called the MTU
18:47 caffblks I get that.
18:48 caffblks I'm not concerned about MTU, I'm concerned about maximum payload
18:48 caffblks (sans fragmentation)
18:49 caffblks For some reason, my proxy handles the authentication just fine, but whatever happens next before it starts downloading media, that's where it's crashing.
18:58 caffblks Actually, maximum UDP datagram is 65535 - (60 + 8) = 65467, but from what I can infer from my Googlefu, most people stick to a maximum packet size somewhere between 512 and 8192 bytes.
18:58 caffblks I currently have a max buffer of 4096, and I'm wondering if that's the issue.
19:07 rubenwardy are you working at the UDP or Minetest protocol layer
19:11 caffblks UDP
19:12 caffblks It's just a basic UDP proxy, essentially.
19:29 caffblks The recvfrom() function shall return the length of the message written to the buffer pointed to by the buffer argument. For message-based sockets, such as [RS] [Option Start]  SOCK_RAW, [Option End] SOCK_DGRAM, and SOCK_SEQPACKET, the entire message shall be read in a single operation. If a message is too long to fit in the supplied buffer, and MSG_PEEK is not set in the flags argument, the excess bytes shall be discarded.
19:30 caffblks I think that's the problem (see: excess bytes discarded)
19:32 caffblks I'm so new to c++ I guess there are just things I don't understand yet. I figure if I have my buffers too large, I may add additional latency / overhead as I'm sending excess bytes both directions. That's why I wanted to know what the maximum expected size might be.
19:36 sfan5 if you proxy the minetest packets transparently it will work
19:37 sfan5 that it doesn't work is an indicator that you have something wrong, consider a tool like wireshark to see what is actually being sent
19:37 caffblks Indeed, good idea!
19:44 kilbith https://github.com/minetest/minetest/pull/13206
4 more elements. Show/hide.

| Channels | #minetest-dev index | Today | | Google Search | Plaintext