00:00 < bridge> dp looks weird 00:00 < bridge> maybe dd xd 00:00 < bridge> oh no diddy 00:00 < bridge> ddnetnet_Token 00:00 < bridge> because ddnet network 00:01 < bridge> chatgpt suggested ddp_ 00:01 < bridge> i think its nice 00:01 < bridge> `dummydragraceneteworkprotocol_SecurtyToken t = 2` 00:02 < bridge> lmfao and u made t just one letter 00:02 < bridge> ddp looks like ddpp which I use to refer to ddnet++ xd 00:02 < bridge> @kollpotato: xd 00:03 < bridge> `ddnetp_` `ddproto_` `ddnet_proto_` 00:03 < bridge> Maybe just [ 00:03 < bridge> A 00:03 < bridge> \_ 00:03 < bridge> wat 00:03 < bridge> `[A_` 00:03 < bridge> \_Token 00:04 < bridge> Or are there other solutions than prefixes and long names? 00:04 < bridge> namespaces? 00:04 < bridge> Symbol link giga mangling? 00:04 < bridge> xd too much for such small lib 00:05 < bridge> Shadowing 00:05 < bridge> Dependcy injection 00:05 < bridge> Recursion 00:05 < bridge> Idk 00:05 < bridge> Polymorphism? 00:05 < bridge> hi random word generator 00:05 < bridge> Reverse symbol lookup proxy encapsulation 00:06 < bridge> xD 00:06 < bridge> just use char[4] 00:06 < bridge> lol ok 00:06 < bridge> What about function names tho 00:06 < bridge> char[4]_decode_packet() 00:07 < bridge> uhhh 00:07 < bridge> Closures? 00:07 < bridge> Should all enums be defines instead? Then they can be undefed by users in case of conflicts 00:07 < bridge> game engine? use ansii escapes 00:07 < bridge> 00:07 < bridge> https://github.com/sollybunny/minesweeper 00:08 < bridge> do i see 8 space indentation 00:08 < bridge> Xd here we go again 00:08 < bridge> nah those are tabs, based and good 00:09 < bridge> Wait till you see that one project that uses 8 tab indentation 00:09 < bridge> lmfao what 00:09 < bridge> i wanna start a debate on tabs vs spaces 00:09 < bridge> but no 00:09 < bridge> im better than that 00:09 < bridge> i know spaces are better 00:09 < bridge> :kek: 00:09 < bridge> We just had that discussion 00:09 < bridge> ik 00:09 < bridge> i wasnt a part of it 00:10 < bridge> And I was banned for weeks because I questioned main instead of master xd 00:10 < bridge> oh ddnet.org is down 00:10 < bridge> i thought just the forum 00:10 < bridge> banned from what?? 00:10 < bridge> This channel 00:10 < bridge> @0xdeen: sos website down 00:11 < bridge> lmaooo 00:11 < bridge> but how can they ban irc 00:11 < bridge> btw libtw2 doc still works https://ddnet.org/libtw2-doc/ 00:11 < bridge> bcus its rust 00:11 < bridge> rust never breaks 00:11 < bridge> You can mute in irc and ban 00:11 < bridge> I'm still a grumpy master user because I thought the reasoning was stupid and lame 00:12 < bridge> wow irc more advanced than i thought 00:12 < bridge> wow irc is more advanced than i thought 00:12 < bridge> where did master come from 00:12 < bridge> main makes sense 00:12 < bridge> because its main 00:12 < bridge> Idc if master or main as long as it’s never mixed 00:13 < bridge> And since master ain’t changing there shouldn’t be any new main introduced either 00:13 < bridge> I have no issue with main, I think it's a more sensible name than master tbh 00:14 < bridge> git is linux propaganda to inject spyware 00:14 < bridge> Should propose to kog to rename main maps to master maps 00:14 < bridge> XDDDD 00:14 < bridge> novice maps should be init maps 00:14 < bridge> it's master because 00:14 < bridge> > The original of a document or of a recording. 00:14 < bridge> is one of the definitions of the word 00:14 < bridge> Good idea 00:15 < bridge> alright makes sense 00:15 < bridge> and because master sounds cool too 00:15 < bridge> I just think the reason for switching from master to main is super far fetched and stupid and nobody is actually offended 00:15 < bridge> it saves 2 bytes 00:15 < bridge> :gigachad: 00:16 < bridge> I am too white old straight man to comment on if it’s offensive or not. I don’t think its hard to change to main so if someone is offended we can do it 00:16 < bridge> I mean, people were offended by the argument on both sides, but nobody was actually offended in the way they claim it's offensive 00:16 < bridge> chiller have u read teewars 0.1 code 00:17 < bridge> its great 00:17 < bridge> No 00:17 < bridge> everything was so simple back then 00:17 < bridge> no varints 00:17 < bridge> inputs in snapshots 00:17 < bridge> epic 00:17 < bridge> ```c 00:18 < bridge> struct snapshot 00:18 < bridge> { 00:18 < bridge> int num_items; 00:18 < bridge> int offsets[1]; 00:18 < bridge> 00:18 < bridge> struct item 00:18 < bridge> { 00:18 < bridge> int type_and_id; 00:18 < bridge> char data[1]; 00:18 < bridge> 00:18 < bridge> int type() { return type_and_id>>16; } 00:18 < bridge> int id() { return type_and_id&(0xffff); } 00:18 < bridge> }; 00:18 < bridge> 00:18 < bridge> char *data_start() { return (char *)&offsets[num_items]; } 00:18 < bridge> item *get_item(int index) { return (item *)(data_start() + offsets[index]); }; 00:18 < bridge> }; 00:18 < bridge> ``` 00:18 < bridge> entire snapshot structure 00:18 < bridge> Nested structs wtf 00:18 < bridge> ddnet also does that 00:18 < bridge> but with classes 00:18 < bridge> Ah true this is C++ already 00:19 < bridge> ```c 00:19 < bridge> // these are used to prepend data in the packet 00:19 < bridge> // used for debugging so we have checks that we 00:19 < bridge> // pack and unpack the same way 00:19 < bridge> enum 00:19 < bridge> { 00:19 < bridge> DEBUG_TYPE_SHIFT=24, 00:19 < bridge> DEBUG_SIZE_MASK=0xffff, 00:19 < bridge> DEBUG_TYPE_INT=0x1, 00:19 < bridge> DEBUG_TYPE_STR=0x2, 00:19 < bridge> DEBUG_TYPE_RAW=0x3, 00:19 < bridge> }; 00:19 < bridge> ``` 00:19 < bridge> this is funny for me 00:19 < bridge> debug mode for protocol 00:19 < bridge> Xd 00:20 < bridge> Hardware failure? 00:20 < bridge> Rip ddnet points 00:20 < bridge> its just the website down 00:20 < bridge> not the db right 00:20 < bridge> Isn’t it all ger1? 00:21 < bridge> Or whatever it was 00:21 < bridge> I thought there is one beefy main vps 00:21 < bridge> master\* 00:21 < bridge> ger1 is a lie 00:21 < bridge> its actually ger10 00:21 < bridge> then ger3 00:22 < bridge> there are players 00:22 < bridge> a lot of linear ones 00:22 < bridge> Blame them 00:22 < bridge> Ok gn 00:23 < bridge> gn 00:24 < bridge> here? 00:25 < bridge> yes 00:25 < bridge> this channel 00:29 < bridge> gg 00:30 < bridge> o whoa 00:30 < bridge> oop nvm 00:52 < bridge> #developer's thoughts on peter thiel 00:53 < bridge> i want to know 01:02 < bridge> all developers sleep 01:07 < bridge> no thoughts 01:08 < bridge> just eepy 04:28 < bridge> hi chiller, how to create map background? for my map 04:28 < bridge> hi ewan 04:28 < bridge> hello 07:39 < bridge> chillerdragon: Ye, we need some prefix for functions and types 07:47 < bridge> @blackkkent: in the editor 07:48 < bridge> @milkeeycat: so name collision is a real problem and prefix the only option? I wasn’t sure. If so which prefix .\_. 07:49 < bridge> What about ddnet\_ 07:49 < bridge> could be, we already prefixed `Packet` type with `DDNet` 07:50 < bridge> So adapt the prefix casing to the name? 07:50 < bridge> no, I think functions should have `ddnet_` prefix 07:50 < bridge> ddnet\_decode\_packet() and struct DDNetPacket 07:50 < bridge> ? 07:50 < bridge> lgtm 07:51 < bridge> But EVERYWHERE? 07:51 < bridge> for every public symbol 07:51 < bridge> DDNET\_ERR\_NONE gg 07:52 < bridge> hmm, Idk about enums, but probably yes :\ 07:52 < bridge> ERR\_NONE is super conflict candidate 07:52 < bridge> check any sdk enum value xd 07:52 < bridge> how do they name it 07:52 < bridge> sdl* 07:52 < bridge> SDL\_ does this 07:52 < bridge> Also they use defines a lot o think 07:53 < bridge> And curses doesn’t do that too much which broke my project 07:53 < bridge> xd 07:54 < bridge> chillerdragon: `read` is `read` in 3 forms btw https://www.usingenglish.com/reference/irregular-verbs/read.html, only pronunciation changes, in present tense is read but in others it's red 07:55 < bridge> WTF 07:59 < bridge> Gumo ^.^ 08:13 < bridge> how to change color bg? 08:56 < ws-client> @blackkkent green dots 08:57 < bridge> @ryozuki can we reopen #9991 ? As you said in you want to reopen this if this feature is wanted. We now have an issue about it and yet another discussion in #mapping 08:57 < bridge> https://github.com/ddnet/ddnet/pull/9991 09:02 < ws-client> @milkeeycat i just tried making the ``buf`` const for normal packet decoding. And I had to bubble down a bunch of consts to get it to compile. Now it compiles without warnings but i think its fake. 09:02 < ws-client> we write to the buffer in str_sanitize 09:03 < ws-client> is that bad? now users always have to pass writeable memory to decode hm 09:03 < ws-client> ah no its not user passed its internally malloced xd not like we talked about this yesterday all day 09:04 < ws-client> at least when they decode a full packet 09:04 < ws-client> i begin to agree with heinrich. this fake const stuff is super weird. Why is it possible to just ignore const xd when fix C 09:05 < bridge> wdym by just ignore const? 09:10 < bridge> interesting, pushing to a branch of a closed PR doesn't update it anymore 09:12 < bridge> idk how to from mobile, ask a other or ill do when im home 09:12 < bridge> no worries, we are not sitting on a bomb that explodes, if you do not reopen it asap 💥 09:20 < ws-client> @milkeeycat the ``const uint8_t *buf`` gets passed and casted around until it becomes a ``char *`` and is being written to 09:20 < ws-client> so the user can pass in a pointer thinking its data will not be changed only to find out it will actually be written to 09:21 < bridge> welp, then it's not const xd 09:21 < ws-client> ye its fake const 09:21 < ws-client> the user only sees const on the surface 09:22 < ws-client> ``size_t fetch_chunks(const uint8_t *buf, size_t len, PacketHeader *header, OnChunk callback, void *ctx, Error *err) {`` 09:22 < ws-client> this is what the user would see and can call passing in a const buf only to find out it replaced some bytes with the ascii space 09:23 < ws-client> seems like a bug in C to me 09:23 < ws-client> i mean casting around void * to allow polymorphism type shit is cool and all but casting away const should be a syntax error 09:58 < bridge> Btw wouldn't it be useful for the bridge to make nicks **bold**? 10:26 < bridge> cool :happy: 10:29 < bridge> Sounds like a good idea for discord. IRC has no bold 11:38 < bridge> @essigautomat 11:38 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1394251650029588570/image.png?ex=68762189&is=6874d009&hm=c54836cf27d55cb5fe1a2e3128494c32dfd91ef4bc6144a8ee910787b27a9293& 11:38 < bridge> i cant 11:48 < ws-client> @milkeeycat OGMMGMGMG i had bug in the test!!B!BDDD 11:48 < ws-client> just debugged the code for like the entire morning 11:48 < ws-client> but code was fine 11:48 < ws-client> az89u0doa0üdpka 11:48 < ws-client> d 11:48 < ws-client> i need test test driven development where tests have tests 12:03 < bridge> ❓ 12:09 < bridge> But what about backwards compatibility? 12:09 < bridge> i just reopened, not that i agree 12:09 < bridge> but it will just sit there forever 12:09 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1394259553792692367/image.png?ex=687628e5&is=6874d765&hm=8b969a1cc87432f52046aa6508f30040bf1fe76e777dab8daf0bd1ee7b595acd& 12:16 < bridge> You could just act like you didn't see 12:16 < bridge> He has the power to reopen too 12:16 < bridge> :kek: 13:09 < bridge> @kebscs https://github.com/ddnet/ddnet/issues/10508 13:09 < bridge> 13:09 < bridge> "they didnt work" - on your local machine? you're on windows iirc? 13:10 < bridge> @kebscs https://github.com/ddnet/ddnet/issues/10508 13:10 < bridge> 13:10 < bridge> "they didnt work" - on your local machine? you're on windows iirc? 13:10 < bridge> 13:10 < bridge> are you sure its not just a local missmatch of your ABI? (msvc vs gnu) 13:12 < bridge> .... microsoft moment 13:12 < bridge> (╯°□°)╯︵ ┻━┻ 13:24 < bridge> on the ci 13:24 < bridge> not on my machine 13:24 < bridge> i see - weird 13:25 < bridge> https://github.com/KebsCS/ddnet/actions/runs/16167086728/job/45631432472 13:25 < bridge> this were the errors 13:25 < bridge> ``` = note: Warning: corrupt .drectve at end of def file␍ 13:25 < bridge> Warning: .drectve `-exclude-symbols:_ZN4core10intrinsics19copy_nonoverlapping18precondition_check17h553e1dbcd4616456E ' unrecognized␍ 13:25 < bridge> Warning: .drectve `-exclude-symbols:"_ZN4core3fmt2rt38_$LT$impl$u20$core..fmt..Arguments$GT$16new_v1_formatted17h6c7da83799dfed05E" ' unrecognized␍ 13:25 < bridge> Warning: .drectve `-exclude-symbols:_ZN4core3fmt2rt9UnsafeArg3new17hd3bff8496e06fd94E ' unrecognized␍``` 13:25 < bridge> etc 13:29 < bridge> okay, linker error - yeah most likely because of an ABI missmatch 13:35 < bridge> if you know you can pr a fix 13:35 < bridge> i have no idea 13:36 < bridge> its an issue within the CI itself, i dont want to touch that 13:36 < bridge> its an issue within the CI itself then, i dont want to touch that 13:36 < bridge> its an issue within the CI itself then, i dont want to touch that.. :kek: 13:40 < ws-client> ddnet protocol so goofed ngl 13:40 < ws-client> did the game server just give me the same security token for multiple reconnects? 13:41 < ws-client> does that mean i can spoof my room mates? 13:42 < bridge> bro 40 year old still having room mates 13:46 < bridge> https://github.com/ddnet/ddnet-rs/actions/runs/16265488999/job/45920103707?pr=127 13:46 < bridge> 13:46 < bridge> :angy: any rust chad who can help me find out why the CI hates me 13:46 < bridge> i cant figure it out for the life of me 13:46 < bridge> master is fine 13:46 < bridge> 14:19 < bridge> What's the err msg 14:20 < bridge> ``` 14:20 < bridge> x86_64-pc-windows-gnu-0.4.0\lib'` 14:20 < bridge> error: failed to run custom build command for `ffmpeg-sys-next v7.1.3` 14:20 < bridge> 14:20 < bridge> Caused by: 14:20 < bridge> process didn't exit successfully: `D:\a\ddnet-rs\ddnet-rs\target\release\build\ffmpeg-sys-next-ae6bc9808fc18b8a\build-script-build` (exit code: 101) 14:20 < bridge> --- stdout 14:20 < bridge> cargo:rustc-link-search=native=D:\a\ddnet-rs\ddnet-rs\target\release\build\ffmpeg-sys-next-f44695e8ae3c4c34\out\dist\lib 14:20 < bridge> cargo:rustc-link-lib=static=avcodec 14:20 < bridge> cargo:rustc-link-lib=static=avdevice 14:20 < bridge> cargo:rustc-link-lib=static=avfilter 14:21 < bridge> cargo:rustc-link-lib=static=avformat 14:21 < bridge> cargo:rustc-link-lib=static=avutil 14:21 < bridge> cargo:rustc-link-lib=static=swresample 14:21 < bridge> cargo:rustc-link-lib=static=swscale 14:21 < bridge> cargo:rustc-link-arg=-Wl,--no-as-needed 14:21 < bridge> ``` 14:27 < bridge> Sounds unrelated to your pr 14:27 < bridge> ye.. 14:28 < bridge> Remind me in 2 weeks, then I'll look into it xd 14:28 < bridge> 2 weeks? oh lord! 14:28 < bridge> happy vacation 14:28 < bridge> Ty xd 14:28 < bridge> any gibberish speakers? what does upside down letter A mean? 14:28 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1394294564826779759/image.png?ex=68764980&is=6874f800&hm=f084ab290879cb1e44f94753398c71888e4a4d10649788c23dbf7c1f119983fa& 14:29 < bridge> for All in t 14:29 < bridge> ^ 14:30 < bridge> is there a name for this notation? 14:30 < bridge> u (is element of) precolored or (every t (is element of) Adjacent(v) and OK(t, u)) 14:30 < bridge> mathematical notation ig? 14:30 < bridge> universal quantifier 14:30 < bridge> or just the entire notation? 14:31 < bridge> If any of you want to give it a quick glance 14:31 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1394295122505765004/SecureBoot.pdf?ex=68764a05&is=6874f885&hm=e855d1cc70db88d07eb2e071c8063ac17f085989b932115fff27d0747a27d112& 14:31 < bridge> If any of you want to give it a quick glance it'd be appreciated 14:31 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1394295122505765004/SecureBoot.pdf?ex=68764a05&is=6874f885&hm=e855d1cc70db88d07eb2e071c8063ac17f085989b932115fff27d0747a27d112& 14:31 < bridge> Master thesis? 14:31 < bridge> Someone is going into optimization theory 14:32 < bridge> 14:32 < bridge> I want to add register coalescing :owo: and rewrite register allocator :pepeW: 14:32 < bridge> Nooo, just needed some example of academic writing so I wrote some stuff 14:33 < bridge> I believe this is an actual usecase where asking AI would help 14:33 < bridge> It lied to me so many times, I don't trust it anymore xd 14:34 < bridge> Well it kept lying to me all the time. I just used it to find typos for me instead 14:34 < bridge> XDD 14:34 < bridge> can we just convince AI to be honest about NOT knowing something 14:34 < bridge> You just have to pay 200$ for the premium ai, trust me bro 14:35 < bridge> It gives me an insane claim. I tell it it's insane, it says no it's real. I give it the UEFI spec and the UEFI PI spec, ask it to find me this insane quote. It tells me there is a related quote on page 2800 of a 2300 page document, in section 32.3.3. The related quote it tells me is not in the document and section 32.3.3 has nothing to do with it 14:35 < bridge> Anyone who thinks these things can think at any capacity is frankly insane 14:35 < bridge> guys do you not proof read what the ai says/does? like sometimes it's easier to go in reverse from a "solution" 14:36 < bridge> I use another ai to proof read ofc 14:36 < bridge> fr, ask claude 14:36 < bridge> run it through grok 14:36 < bridge> factcheck it by chatgpt 14:36 < bridge> apply fixes by cursor 14:36 < bridge> have memes created by phind 14:36 < bridge> I agree, totally even, still I find use of them 14:37 < bridge> I just use it for much much simpler things, finding typos, finding me sources to read 14:38 < bridge> Daily question: 14:38 < bridge> 14:38 < bridge> Imagine you have a function that takes 14 parameters, but has a short body, like creating a struct. 14:38 < bridge> 14:38 < bridge> Does forcing inline improve perfomance? 14:38 < bridge> Ppl that care about their solutions will defs notice that ai gets worse with increasing complexity 14:38 < bridge> Most likely 14:39 < bridge> It would be no worse, since the compiler would probably inline it even if you didn't ask it to 14:39 < bridge> maybe, if the compiler doesn't force it inline anyways 14:39 < bridge> what's the purpose of ddnet-rs? is it just a project for fun or what 14:39 < bridge> jupstar just dabbed on all of us with it and i like so it so i use it pretty often 14:39 < bridge> It's serious af 14:40 < bridge> jupstar just dabbed on all of us with it and i like it so i use it pretty often 14:40 < bridge> correct, preparing lot of parameters has overhead, and most compilers wont optimize a big function because one of the first hints they use is "will inlining this make the caller way bigger?" and is often true for thise, but it prevents looking at optimizations after inlining such big function, which may make the code smaller actually 14:40 < bridge> correct, preparing lot of parameters has overhead, and most compilers wont inline a big function because one of the first hints they use is "will inlining this make the caller way bigger?" and is often true for thise, but it prevents looking at optimizations after inlining such big function, which may make the code smaller actually 14:41 < bridge> The purpose was actually to write a ddnet clone without all the burden we have right now with backward compat, breaking forks etc. 14:41 < bridge> 14:41 < bridge> Technically up to date, good modding support etc 14:41 < bridge> i asked this cuz today i improved perf by 30% on a benchmark forcing inline of a stupid constructor taking 14 parameters 14:41 < bridge> maybe more like 40% 14:41 < bridge> Instead you wrote a teeworlds clone, with no compat, no forks, no mods 😛 14:41 < bridge> Oh wow, that thing must have been very hot path 14:41 < bridge> i fork, it has compat 14:41 < bridge> it is in the hot hot path 14:42 < bridge> and i'll make the best twinbop twerking mod you'll ever see! 14:42 < bridge> Is the benchmark realistic though? If it's calling the constructor a billion times a second but your actual program isn't then it's not an accurate measurement of improvement 14:42 < bridge> At least it was a huge bonus project for hiring interviews yet xd 14:43 < bridge> Yeah you know i just slammed 200k rust project xd 14:43 < bridge> :pog: landed a rust job yet? 14:43 < bridge> Yes :deen_star: 14:43 < bridge> make on stream 14:43 < bridge> LETSGOO 14:43 < bridge> i would say its like 20% realistic 14:43 < bridge> xd 14:43 < bridge> but its a requirement to improve this 14:43 < bridge> from bossman 14:44 < bridge> does that mean ddnet-rs will be worked on less? 14:44 < bridge> Yes xd 14:44 < bridge> i still think its worth it 14:44 < bridge> So overall it's still 20% of 40%, which gives an 8% real improvement, so still not bad 14:45 < bridge> classic rust rewrite 14:45 < bridge> arentall non comercial projects for fun? but actually ddnet-rs has more features than ddnet 14:45 < bridge> i'll try it when it finishes compiling 14:45 < bridge> Cya tomorrow 14:46 < bridge> why does this download more packages than npm? 14:46 < bridge> Depends on the project 14:47 < bridge> cargo is well known for downloading a billion things 14:47 < bridge> it's like a cargo ship, a million containers 14:50 < bridge> wow it actually looks kinda nice 14:55 < bridge> that's weird 14:55 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1394301193622261880/image.png?ex=68764fad&is=6874fe2d&hm=bb65d47b259e08c241ec2a701b7588d77d7939d07f214df17588f7b7ab3f791c& 14:57 < bridge> interesting port 15:04 < bridge> it's definitely not ready but still pretty cool 15:07 < bridge> i have no idea how to change my nick lol 15:13 < bridge> Yeah no idea why servers don't always respond, maybe something related to libtw2 15:13 < bridge> Should be rather similar to normal ddnet 15:34 < bridge> there are way too many acronyms, it's almost impossible to follow without a general scheme 15:35 < bridge> > For the purposes of this discussion we assume Verified mode is 15:35 < bridge> > used. The ACM now uses the OEM IBB key to verify the IBB. IBB 15:35 < bridge> > usually contains only the SEC and PEI parts of the UEFI firmware, 15:35 < bridge> > leaving the DXE to the OEM Boot Block (OBB). The OBB is not 15:35 < bridge> > verified by Boot Guard, that job is delegated to the platform vendor 15:35 < bridge> it's unbearable mate 😄 15:36 < bridge> Mh, I had all of them expanded but then it read weird to me 15:37 < bridge> Prefer a scheme which will be much more comprehensible 15:38 < bridge> Scheme as in another flowchart? 15:39 < bridge> for instance 15:40 < bridge> or whatever makes more sense, but a graphical representation would greatly help 15:40 < bridge>