00:02 < bridge> Seems fine, but it's a bit random when exactly nightly is build anyway 00:03 < bridge> nightly or master, a recentish commit 00:03 < bridge> idk when the reserve was added 00:04 < bridge> I think "nightly" is simply the latest commit on master when deen starts the build 00:04 < bridge> I can tell exactly, it was a week ago 00:04 < bridge> ghasp! 00:04 < bridge> You can see at which times the debugging symbols were updated on https://ddnet.org/downloads/symbols/ 00:04 < bridge> the idiot who wrote #10940 must have a fried brain 00:04 < bridge> https://github.com/ddnet/ddnet/pull/10940 00:07 < bridge> Thanks for fixing set\_team\_ddr I just had to use it this week! But for me it worked c: 00:11 < bridge> because you didnt start a run or were in practice 00:11 < bridge> i have `sv_practice_by_default` on my test server so it didnt work at all 00:12 < bridge> Rip you 00:12 < bridge> @blaiszephyr: you can’t tell me the Teeworlds pr was done by a human 00:13 < bridge> Probably some vibe dude in told some Agent: create a GitHub account and make it do open source 00:13 < bridge> Probably some vibe dude told some Agent: create a GitHub account and make it do open source 00:54 < bridge> is it still refactor if all whats done is 5 lines being moved 00:54 < bridge> and moved into basically same space, as there are 0 checks 01:00 < bridge> no flame, i think youre much better than moving few lines around 👍 01:03 < bridge> prefacing flame doesnt reduce flame 01:03 < bridge> prefacing with "no flame" doesnt reduce flame 01:03 < bridge> it was worse before, its better now, its a pr 01:04 < bridge> i dont distinguish between "clean" and "refactor" like i used to, i just use "refactor" 01:05 < bridge> just jokingly, that this is such a small pr 01:05 < bridge> 👀 01:13 < bridge> (atleast to me "no flame" doesnt sound very jokey, /j would probably be more clear) 01:13 < bridge> (should i, is the small effort of derterming whether something is a cleanup or refactor worth the small clarity for the reviewer) 01:14 < bridge> no :kek: 01:14 < bridge> i mean youre way better dev, can do more than just shift 5 lines 01:15 < bridge> it was in the process of doing 01:15 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420549146523340850/image.png?ex=68d5ccfd&is=68d47b7d&hm=685d5e5178ff346c7cddf42bf9f5bcdec3be81cfe587791b9c16b7b0b370f00f& 01:15 < bridge> i saw bad code so i fixed it 01:16 < bridge> I'm currently trying to create some new classes to interact with database, 01:16 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420549425306140813/lightquery.cpp?ex=68d5cd3f&is=68d47bbf&hm=3dcae3e1c41b95ea31917e9b3e66c06952d7918d2d37b163efb365629385ed27& 01:16 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420549425604071615/lightquery.h?ex=68d5cd3f&is=68d47bbf&hm=ff676ab8a0feacf7b14456d76c582df04e755978316c72d210770fd59a4453a5& 01:18 < bridge> The score.h is kind let me confused because it seems to put many stuff together 01:19 < bridge> And it requires some search to the code to find where is the logic handling the result of querying (finally I found CPlayer::ProcessScoreResult executed in CPlayer::Tick) 01:22 < bridge> There would be a manager class but I haven't written yet 01:24 < bridge> There is a obsolete version of kind "manager" class, I decided not to use it because its code was not good 01:24 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420551533023596634/image.png?ex=68d5cf36&is=68d47db6&hm=898ced929e1f5fe678196bcc43ba2b681ebbe9a3c5aa80ae1a065b13af280278& 01:27 < bridge> So my target is let database function can be "registered" in the initializing of server or gamecontext like console commands doing, and also have QueXXX conmmands like ConXXX do 01:32 < bridge> The CMessageBuilder would have more function to let it be more powerful like directly create send chat or set motd instructions in a few lines because it is already have the capability to parse format string and fill the value directly from the Result 05:02 < bridge> Yeah idk. 05:06 < bridge> it makes me sad. 05:06 < bridge> The constant "is this AI?" in the back of my head 06:14 < bridge> add player_flag and dummy_flag to f1 06:14 < bridge> to bind full profile change 06:14 < bridge> :justatest: 06:42 < bridge> morning 07:00 < bridge> I'm AI 08:24 < bridge> What are you working on btw? I mean what is it for? @thrava are you planning to pr that into ddnet? 08:27 < bridge> seems like i got a working scheduler again 08:27 < bridge> also maybe im moving out to a nearby city thats not barcelona 08:27 < bridge> cheaper and bigger flat 08:27 < bridge> :owo: 08:41 < bridge> Hi 08:41 < bridge> Hola? 08:41 < bridge> MERHABA 08:42 < bridge> 🙂 08:42 < bridge> Sprich deutsch 08:42 < bridge> como estas? 08:43 < bridge> bien 08:44 < bridge> kein deutsch sprechen 08:44 < bridge> :brownbear: 09:32 < ws-client1> **** xd 10:22 < bridge> I don't, I just wan't to have some method to auto update map info in motd when changing map 10:22 < bridge> I don't know, I just wan't to have some method to auto update map info in motd when changing map 10:23 < bridge> I just dislike the interaction method of database in score.h 10:24 < bridge> So I decide to rewrite a new one 10:24 < bridge> With mostly dependency injection 10:33 < bridge> It's powerful for queries that needn't return every value in absolutely one-time, because in this case, it can be collected in multiple queries 10:36 < bridge> is it beneficial to break things up 10:38 < bridge> for short time queries, it almost make no effects 10:39 < bridge> but for long time queries like queries related to race.sql, It porbably(I haven't test) take more time 10:40 < bridge> But it can be solved by adding link-list logic to the Result and Request struct I think 10:42 < bridge> If really added, there is no need to allocate memory chunk to store continuous values 12:31 < ws-client1> **** ddnet is deprecated on brew <:tee_thinking:478629518358085653> teeworlds is not 12:32 < ws-client1> **** https://github.com/Homebrew/homebrew-cask/blob/cacb4996f58b5323e97ecd6abec4da6dfb4e45b6/Casks/d/ddnet.rb#L15 14:07 < ws-client1> **** where do i write documentation comments? 14:07 < ws-client1> **** in the interface or in the implementation class? 14:16 < bridge> This is fucking huge, @robyt3 take a :deen_star: 14:26 < bridge> @chillerdragon what are you documenting? 14:26 < bridge> it depends 14:26 < bridge> i would go with the most general, if it applies to interface and implementation class (not implementation.cpp) then do interface 14:36 < ws-client1> **** just a general question i wonder about it every time 14:37 < ws-client1> **** i also am not sure which is supported best by editors 14:38 < ws-client1> **** maybe oop was a mistake and we should go back to C 14:38 < ws-client1> **** if it is non obvious where to search for documentation 14:38 < ws-client1> **** watafak is polymorpic documentation even supposed to do 14:39 < bridge> @chillerdragon idk what that means 14:39 < bridge> why is polymorphic documentation any different 14:39 < bridge> you put comments where they are relevant 14:39 < bridge> 👀 14:39 < bridge> for manual searching yeah its a bit confoozing 14:41 < ws-client1> **** if i ask my editor for docs which should it even show 14:46 < ws-client1> **** am i just a hater or have the integration become more annoying since the python rewrite? 14:46 < ws-client1> **** `FileNotFoundError: [Errno 2] No such file or directory: './integration_open_editor_h91bi5r5/client0.fifo' ` what does that even mean 14:49 < bridge> Document the interface. The implementation also inherits the documentation usually. 14:52 < bridge> Either the server failed to create the FIFO file, or maybe it failed before that 14:52 < bridge> Either the client failed to create the FIFO file, or maybe it failed before that 15:08 < bridge> @blazulite https://github.com/ddnet/ddnet/issues/10326 would be a good opportunity to help in ddnet if you want 🙂 15:08 < ws-client1> **** yea it looks like it crashed in the CI the error was more useful than locally 15:09 < ws-client1> **** i know basically nothing about gfx and i do not feel like looking into it we need someone help for all the assa and solly prs 15:09 < ws-client1> **** when send new maintainer 15:09 < ws-client1> **** i am more comfortable in the breaking phyiscs and back compat department 15:09 < bridge> you can just read netcode, this is where I am mostly skillchecked 15:10 < ws-client1> **** ye netcode i can also merge 15:10 < ws-client1> **** we need more gfx mergers 15:10 < bridge> because I don't get the "ddnet protocol" if you can call that 15:11 < bridge> I am also not perfect for gfx, as soon as it hits backend/shader expert level, I also start heavy sweating 15:11 < bridge> the next obvious shader would be a tee-render-shader btw, and this is not simple at all 15:12 < ws-client1> **** i want moving grass! 15:12 < bridge> and it wouldn't even guarantee a performance increase, because we are already using quad containers in the backend 15:12 < bridge> we could also do that, similar to the water shader I already made (but different) :justatest: 15:14 < bridge> omg you could actually do that automatically by adding a shared external white grass for everyone and detecting that one and animating it 15:15 < bridge> > Gotta find this soreu dude. Or also recreate that one from scratch :/ 15:15 < bridge> @chillerdragon wat? 15:17 < bridge> am i also missing smth 15:17 < bridge> do we need a creator of the image to copy paste grenade launcher into a png file? 15:17 < ws-client1> **** we should discuss this on gh if its not clear^^ 15:18 < ws-client1> **** but what i meant is we should stop editing png files and move to svg 15:18 < bridge> you can directly convert every weapon in the svg to a 64*64 tile 15:18 < ws-client1> **** soreu might have on otherwise i would propose to recreate it 15:18 < ws-client1> **** @soreu 15:19 < bridge> omg soreu is a person, I thought you _uwuiefied_ the word `sorry` 15:19 < bridge> isnt that the point of ddnet-data-svg 15:19 < ws-client1> **** it is 15:19 < ws-client1> **** but it does not have the file we need yet 15:19 < ws-client1> **** so its either find or recreate 15:19 < bridge> yeah so just paste it into png 15:19 < bridge> dont need to recreate it as svg for that issue 15:19 < bridge> yes ! 15:20 < ws-client1> **** hrmpf 15:20 < bridge> recreating all of the entities in svg and then exporting pngs is another issue 15:20 < ws-client1> **** ok fair i gues 15:20 < ws-client1> **** gimp it in then 15:21 < bridge> and we'd actually need sb to maintain it, a lot of tiles could need some overhaul 15:21 < ws-client1> **** but before big overhaul we should move to svg :p 15:21 < bridge> do you need help to split up the svg into weapon svgs? 15:22 < bridge> guess i am going to do that now for ddnet-data-svg 15:23 < ws-client1> **** who is "you" 15:23 < bridge> you 15:23 < bridge> :kek: 15:23 < ws-client1> **** im not splittin shit 15:24 < bridge> baba is you 15:24 < ws-client1> **** if you want weapons i already built them from hand here lel https://github.com/TeeworldsDB/images/tree/48ddb0b2b60b44e6a8d44ce2999daf71445e7afa/realistic/transparent/game 15:24 < bridge> I mean you as everyone and I mean kebs, since he also wants to work on this issue 15:24 < ws-client1> **** i do not want to work on the issue. 15:24 < bridge> are these rebuilds or the original? 15:24 < bridge> i dont 15:24 < ws-client1> **** rebuilds by me 15:25 < bridge> but if noone wants to i can fire up paint for 5 min 15:25 < ws-client1> **** that was before originals dropped 15:25 < ws-client1> **** i am not sure if we even got originals 15:25 < bridge> why did you rebuild them? You can extract them all from the game.svg 🙈 15:25 < bridge> and now I mean YOU CHILLERDRAGON 15:25 < ws-client1> **** when zatline randomly dropped some weird ass work in progress 0.8 svgs he found on his drive 15:26 < ws-client1> **** @Assa you know that the game.svg is a newer rebuild than mine right? 15:26 < bridge> I'll split the svg and give blazulite a chance to contribute 🙂 15:26 < bridge> svg :pepeW: 15:26 < bridge> yeah read that after I wrote my comment 15:27 < ws-client1> **** @Assa depends on how you define original 15:27 < ws-client1> **** i think teeworlds game was never published. I recreated in 2021 and the bois did it in 2022 again xd https://github.com/ddnet/ddnet-data-svg/pull/32 15:29 < bridge> hmm the ship is sinking 15:29 < bridge> Yup 15:29 < bridge> Where a Landil 15:30 < bridge> We need a Landil 15:30 < bridge> so theseus created a new one, I get it 15:30 < bridge> It became somewhat official for ddnet tho 15:30 < bridge> So ddnet is now proper temu fake of Teeworlds 15:30 < bridge> With recreated assets 15:30 < bridge> can you guarantee that your recreation or the one of the 3 ohters is 100% accurate? Which is better? 15:31 < bridge> It’s not accurate 15:31 < bridge> Neither accurate to DDNet nor Teeworlds 15:31 < bridge> Go with the ones from the ddnet repo 15:31 < bridge> Cut them out 15:32 < bridge> This svg was already used to generate the png we ship I think 15:32 < ws-client1> **** so its already tw fake xd 15:32 < ws-client1> **** and ddnet official 15:32 < ws-client1> **** even if it wasnt done yet we plan to do so eventually 15:32 < ws-client1> **** hmm latest commit 3 years ago https://github.com/ddnet/ddnet/blob/24fd04b7bc9dcea9f92624ca455eccf19d4f7ac6/data/game.png 15:33 < ws-client1> **** i think we actually still use tw one 15:34 < bridge> Hi, I have an idea about how to add voice chat to the game. 15:34 < ws-client1> **** how? 15:36 < bridge> There are many apps and plugins for voice chat. It would be great to talk not in Discord or other apps, but in the game itself. 15:36 < ws-client1> **** oh so your idea is just "voice chat" not "how to do it" 15:36 < ws-client1> **** iirc that has already been rejected 15:36 < ws-client1> **** i would also find it useful tbh 15:36 < bridge> proximity voice chat 15:37 < bridge> ddnet-rs :gigachad: 15:37 < ws-client1> **** https://github.com/ddnet/ddnet/issues/6547 15:37 < ws-client1> **** @Teero fr 15:39 < bridge> oh okay 15:41 < ws-client1> **** nah bro what does it want from me https://github.com/ddnet/ddnet/actions/runs/18008439189/job/51234313018?pr=10964 15:43 < bridge> flaky test 15:43 < bridge> failing on my pr too 15:45 < bridge> btw 15:45 < bridge> 0xDEADBEEF is a bad address to debug 15:45 < bridge> its not even 8 byte aligned 15:46 < bridge> in rust if u try to deref it on debug build it panics cuz it catches ur trying to deref an an unaligned ptr 15:46 < bridge> also its a addr that its not that hard it may have data 15:46 < bridge> > Frankly, 0xdeadbeef is a rather bad value for a pointer poison: It doesn't have enough leading zeros to be caught in unmapped area used to catch null pointer access (it has a high change of being a valid kernel address on linux if you have about 700mb of ram). And there are not enough zeros at the end to still recognize it if the access is to a offset within the array/structure (as it usually happens). 15:47 < bridge> so as of today im anti deadbeef 15:47 < bridge> alivecow 15:48 < bridge> 0xw00000DEAD000000000 15:48 < bridge> 0x00000DEAD000000000 15:48 < bridge> :topkekMate: 15:48 < bridge> @chillerdragon since bot is not connected to ddnet-data-svg: https://github.com/ddnet/ddnet-data-svg/pull/66 15:49 < bridge> the unaligned access only is ub if the underlying data requires a a non 1 alignment like u64 15:49 < bridge> you can argue for and against the usefullness of this PR 🤷‍♂️ 15:49 < bridge> ``` 15:49 < bridge> thread 'main' panicked at src/main.rs:7:9: 15:49 < bridge> misaligned pointer dereference: address must be a multiple of 0x8 but is 0xdeadbeef 15:49 < bridge> note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace 15:49 < bridge> thread caused non-unwinding panic. aborting. 15:49 < bridge> ``` 15:49 < bridge> missing hammer 15:50 < bridge> there is no hammer pickup :justatest: 15:50 < ws-client1> **** @kebscs well.... i disabled rcon commands so the prs test would get stuck on the client running `shutdown` fair i guess xd 15:50 < bridge> the pr says `weapons` :think_bot: 15:51 < ws-client1> **** its just the output always looks the same 15:51 < ws-client1> **** i wish it would be more specific about what timeouted 15:51 < bridge> it contains `weapons` (pl) :justatest: 15:51 < bridge> it does not contain `pickup` 15:53 < ws-client1> **** @Assa wdym bot the irc github bot was shutdown before i was born xd 15:53 < ws-client1> **** i get emails for that repo 15:56 < bridge> emails, that is soo oldschool 15:56 < bridge> (if emails wouldn't work suddenly, I guess the next world finance crisis would immediately start) 15:57 < bridge> wants you to go smoke 15:57 < bridge> i think 15:58 < bridge> i honestly think python errors are way harder to read than c++ or c errors 15:59 < bridge> actually if ur using deadbeef maybe u want that 16:09 < ws-client1> **** another morning w out rcon prs merged 16:10 < ws-client1> **** another night i cry my self into sleep 16:10 < ws-client1> **** end of shift, chilor out 16:10 < bridge> cries about his rcon prs, while I have PRs that are longer open 16:14 < bridge> 99 open prs! 🔔 16:52 < bridge> :owo: 🤭 16:59 < bridge> We can try to overflow the counter 17:28 < bridge> f-droid contributor answered :poggers2: 17:29 < bridge> @essigautomat answer him 17:29 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420794387364974714/image.png?ex=68d6b163&is=68d55fe3&hm=69d14c8bcc906f36c19cb5ab75dadb7c88253ca1655acf7e8f2acc0a4799f900& 17:32 < bridge> @teero777 how do we build those libs? 17:32 < bridge> i don't know he would do it 17:32 < bridge> or they 17:32 < bridge> idk 17:33 < bridge> well we don't xd 17:33 < bridge> thats the whole point of having static libraries 17:43 < bridge> what? no? 17:43 < bridge> I believe I let Heinrich answer, I am not too familiar with building android 17:43 < bridge> ofc 17:43 < bridge> im stupid 17:43 < bridge> youre right 17:44 < bridge> for android yes 17:44 < bridge> nice scripts 17:44 < bridge> ~~https://github.com/ddnet/ddnet-scripts/blob/master/ddnet-lib-update.sh~~ 17:44 < bridge> ~~this is how they are built~~ 17:44 < bridge> https://github.com/ddnet/ddnet-libs~~ 17:45 < bridge> ~~https://github.com/ddnet/ddnet-libs~~ 17:55 < bridge> wtf is this mechanic 17:55 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420800946182557696/2025-09-25_17-54-51.mp4?ex=68d6b77e&is=68d565fe&hm=60f3c55f234f6bb79bdb32d3b9dcd701fe1726d9e3ff9e4c90449d7de6888748& 17:56 < bridge> Its used in the map "just deep fly" 17:57 < bridge> One of the final solo parts after dummy dies 17:58 < bridge> I needed to add this to the hook line as well btw xD 17:58 < bridge> okay im not supporting that shit in my physics 17:59 < bridge> wait until he learns about the same trick without teleouts 18:05 < bridge> With 0 teleouts hook is glitched 18:05 < bridge> Won't attach 18:06 < bridge> really 18:06 < bridge> Yea I think I tested that 18:07 < bridge> :monkaStop: 18:13 < bridge> do I need a new hookline color for state _fucked_? 18:15 < bridge> Who doesn't know about the 4 hook line states: 18:15 < bridge> - hooking 18:15 < bridge> - not hooking 18:15 < bridge> - unknown 18:15 < bridge> - _fucked_ 19:39 < bridge> and you only want to know where the end of the hook is, and if it's hooking something or not 19:39 < bridge> and we have basically yes/no/maybe/bugged 19:40 < bridge> mega bool 20:29 < bridge> sorry i wa dealing with moving house stuff 20:30 < bridge> samply is a perf wrapper 20:30 < bridge> it formats the data for u into firefox profiler automatically 20:30 < bridge> firefox profiler is the best ui ever 20:30 < bridge> i already showed a profile of ddnet in this channel 20:30 < bridge> okay but i cant see the assembly xd 20:30 < bridge> like what if i have 5 operations in a single line 20:30 < bridge> i wouldnt know which one is slow 20:31 < bridge> thats like pretty important 20:31 < bridge> also i think the firefox profile ui is horrible 20:33 < bridge> like this is so ugly 20:33 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420840705911292068/image.png?ex=68d6dc86&is=68d58b06&hm=3ed8c2136c59f38cdf5deb47fe413d067cd10991295a6cd66bd0b6d6b7b09034& 20:35 < bridge> in perf you can instantly see whats wrong and cycle through the hot instructions 20:35 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420841102608437369/image.png?ex=68d6dce4&is=68d58b64&hm=a13c2a3088fd45463bb6bbbd84ad10f9c7eb0167588bc3779f56c2ecae8a7e5f& 20:36 < bridge> there must be a reason you think the firefox profiler is good that i haven't found yet 20:43 < bridge> https://github.com/firefox-devtools/profiler/issues/4520 20:43 < bridge> 20:43 < bridge> the issue says it supports native assemblyfor C/C++/Rust but it doesn't appear so 20:43 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420843251824590889/image.png?ex=68d6dee5&is=68d58d65&hm=aa15711d90f1f20d7dbfcc50183426bf90b7ac240fe9013fc602008273c14429& 21:19 < bridge> did u leave samply server open? 21:20 < bridge> well if perf works better use it, if im real i barely used perf cuz i didnt learn the cli tools to use it xd 21:20 < bridge> samply is rly simple, u wrap the run command and u get a shareable profile 21:20 < bridge> however i think u can see the asm 21:22 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420853046346518578/image.png?ex=68d6e804&is=68d59684&hm=31b399162d5ad354dc8d228eb37bb1aa843e00fca11a2f2ceb6701257196d0d2& 21:22 < bridge> ah yea mb the asm works 21:23 < bridge> is there no combined view? 21:24 < bridge> no idea 21:24 < bridge> samply can record for perf too btw 21:24 < bridge> iirc 21:24 < bridge> i guess its also more crossplatform 21:25 < bridge> it works on linux macos winshit 21:25 < bridge> @teero777 how do u use perf to view the source like that btw? 21:27 < bridge> idk seems simple enough xd 21:27 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420854284966170737/2025-09-25_21-26-05.mp4?ex=68d6e92b&is=68d597ab&hm=ba15e3a1d0aa1b9093a285f3ff99150a60e0cbc1d269f86c97f88f1805f088e9& 21:33 < bridge> you can do a lot with perf actually 21:33 < bridge> i mean if i wanted a flame graph i could just tell perf to make one 21:47 < bridge> nah this doesnt look easier than firefox graph at all 21:50 < bridge> how is blinding your eyes, moving your mouse, double clicking some elements, having 0 visual feedback and no combined asm+code view easier xd 21:51 < bridge> also my mouse wheel is broken and the firefox dragger thing is so small for big code files that you can't see it. so scrolling to some part of the code is horrible 21:52 < bridge> TheFatRat - Warbringer 21:54 < bridge> i have checked the firefox profiler once so cant say much 21:54 < bridge> but anything easier than arrow keys in console 21:54 < bridge> its actually not though. its a different song that they used the same section in 21:57 < bridge> True, arrow keys are awful 21:57 < bridge> it's gotta be hjkl 21:58 < bridge> i lied. it is the song 22:04 < bridge> huh, can you send link? I like that song a lot but I'd wish for a metal variant tbh xD 22:04 < bridge> i said i lied :( 22:04 < bridge> you failed me brother 🙁 22:15 < bridge> idk I kinda agree with teero, I used the vs profiler a lot but seeing the assembly inline with the code in perf lets you spot a lot more optimizations that I didn't think about otherwise. 22:15 < bridge> the interface is also not bad 22:15 < bridge> you can just around well 22:15 < bridge> you can jump around well 22:16 < bridge> lack of flamegraph is the worst part 22:19 < bridge> now I feel blind without assembly lol 23:32 < bridge> i just discovered there is literally an "export as flamegraph" button lmao 23:34 < bridge> you just go to run scripts for all samples and then press flamegraph xD 23:34 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420886080940412939/image.png?ex=68d706c8&is=68d5b548&hm=734c829a34b7cdfd19971acc844f9f406d15b3dfdf53a43b47ba92ef685f0290& 23:34 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420886081884389406/image.png?ex=68d706c8&is=68d5b548&hm=66bcb4c829bb059a6acaaa3b987011f5f5415a370cc3bed63119010e216c5f0f& 23:34 < bridge> Does it work 23:35 < bridge> it should, but i dont have the program installed 23:35 < bridge> it needs some sort of template 23:35 < bridge> the package is not in the any arch repo so id have to compile it manually 23:39 < bridge> Every time I've tried a tool thats supposed to give perf flamegraph it didnt work for some reason 23:39 < bridge> or there was some catch that isnt acceptable 23:43 < bridge> man modern cpus have to many instructions its probably impossible to properly write assembly 23:43 < bridge> man modern cpus have so many instructions its probably impossible to properly write assembly 23:54 < bridge> The instructions are just wrappers for microcode which has far fewer instructions, using a highly specific instruction isnt guaranteed to give you any real benefit. 23:55 < bridge> uh maybe i did something wrong 23:55 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420891572081528963/image.png?ex=68d70be5&is=68d5ba65&hm=05ed4662c8b9b16ffbd77695bf940122ad88a2bd7ff10378277f7cb459004f0b& 23:57 < bridge> "Properly" by hand is veeery hard nowadays. But the manuals technically have all the information you'll need