13:24 <unique_id> Since no one has said anything in 24hours+ according to the log: Currently working on reliable in-order messaging over UDP for gamedev. Thought I'd give it a go myself to learn how it works rather than relying on a blackbox I didn't understand.

13:26 <unique_id> Zig is really nice for this

13:42 <bheads> Sounds fun, I put zig on hold until the syntax issues have been resolved. I have one project that needs to be updated already

13:44 <ltr_> some kind kind of related, i was reading about udp protocols, and revisiting UDT, i found it an amazing protocol, it uses my bandwidth at full from anywhere

13:48 <bodie_> unique_id, that sounds like a blast. is your source open?

13:49 <bodie_> s/open/public/

13:51 <bheads> Whats your plan for dealing with lost messages?

13:52 <unique_id> bodie_: I haven't open sourced anything at the moment. bheads: I simply haven't updated to that commit since I found it extremely likely that the syntax change would be altered.

14:02 <unique_id> bheads: I'm figuring that out at the moment. What I know is that each packet has a sequence number and lost packets are not sent with the same sequence number. Sequence numbers just keep incrementing until they wrap around. This way acking can be a single sequence number + a bit field of the 32 previous sequence numbers. So you duplicate acks like crazy but in an incredibly compact way. I'm learning this from Glenn Fiedler's

14:02 <unique_id> articles. Reliability is built on top of this

14:04 <unique_id> So each bit indicates whether a package has been received, for something like 32 total packets.

14:04 <unique_id> packet*

14:05 <unique_id> I'm never good at explaining things, simple or not :)

18:19 <MajorLag> andrewrk, or anyone else who knows: I want to see how zig will pack a struct on a big-endian arch. Whats the best way to do that considering there are no big-endian archs or OSs supported by std? I've tried to emit asm for MIPS but it doesn't seem to be doing anything.

19:07 <wink_> MajorLag: There is some big endian support in llvm for aarch64 maybe that's available on zig: https://www.google.com/search?q=llvm+big+endian+aarch64 via some switches???