00:00 < bridge> But yeah, zed is the way - the first editor that handled DDNet-rs without issues :kek: 00:01 < bridge> chillerdragon: wow actually based and not insane 00:01 < bridge> arrow keys are the right way 00:02 < bridge> first time building it froze my laptop 00:02 < bridge> ran out of ram 00:02 < bridge> Yeah that's rust for you 00:02 < bridge> At least the binaries are smol :justatest: 00:03 < bridge> are you sure about that 00:03 < bridge> hello world is 10x larger than c 00:03 < bridge> i think im wrong, it was 40x 00:04 < bridge> depends on c compiler optimizations and other smart stuff 00:04 < bridge> > You get a compiler warning and a failing CI already. Thats why switch is so op. But could also add an assert. Do you think it is needed? 00:04 < bridge> @chillerdragon why do you get a compiler warning? At least the two maintainers agree with me here 😄 00:05 < bridge> wgpu apps are at least 15 mb 00:05 < bridge> if the compiler warning is about unreachable code, try to use switch default here 00:05 < bridge> If you switch on an enum and don’t cover all the compiler throws a warning. 00:05 < bridge> Rust optimizes for speed and adds a ton of symbols and whatnot 00:05 < bridge> 00:05 < bridge> Use strip and build in release 00:05 < bridge> The CI runs with warnings as errors 00:05 < bridge> yes and i like it 00:05 < bridge> So in ddnet this change would be caught in the pr that adds a new enum value 00:05 < bridge> The assert would be redundant 00:06 < bridge> ping robyt or lerath then 00:06 < bridge> Yes but what do you say? 00:06 < bridge> It’s your comment 00:06 < bridge> I'd add an assert, so you get an error while testing locally already 00:06 < bridge> not everyone build with -Wall and fail-on-warning 00:08 < bridge> I am okay with CI driven development but sure I can add an assert if you still need it knowing the CI already catches it 00:13 < bridge> I am not so okay with ci driven development, it's really slow and also expensive in terms of resources used 00:13 < bridge> I try to do as much locally as I can, but clang-tidy is still my enemy 00:15 < bridge> also I am not a maintainer, I am just trying to help. If they disagree with me they'll tell us 😄 00:17 < bridge> GitHub neutered all free actions runners lately 00:17 < bridge> does ddnet not pay for github actions at all? 00:17 < bridge> nope 00:17 < bridge> also from what money? 00:17 < chillerbot1> You can see donors and how to donate here https://ddnet.org/funding/ 00:18 < bridge> xd 00:18 < bridge> good bot 😄 00:18 < bridge> :3 00:18 < bridge> i think there was a limit on the free plan 00:18 < bridge> OUR money 00:19 < bridge> we can take some money from all the ddnet mod salaries 00:19 < bridge> that all goes to actual servers 00:19 < bridge> you can't take from zero 00:19 < bridge> any free infrastructure is a must for a donation-driven project like ddnet 00:20 < bridge> servers come and go all the time due to unstable and sometimes inadequate funding 00:20 < bridge> I think its more like regional sanctions 00:21 < bridge> deen is literally unable to pay for the servers 00:21 < bridge> it's both, it's a free to play game guys 00:21 < bridge> and this only works because people including deen but also a lot of others are spending money and free time on it 00:21 < bridge> Still the free runners do not have enough ram to build ddnet++ .\_. So annoying 00:21 < bridge> @blaiszephyr sux 00:21 < bridge> If we wanted more donations we can just put the donation progress bar on the main menu until its filled 00:22 < bridge> wikipedia style 00:22 < bridge> I dont think we need a beg banner 00:24 < bridge> i dont think it will help 00:24 < bridge> its on the home page already 00:24 < bridge> Im pretty sure it would help 00:25 < bridge> I think most people aren't aware theres a donation goal in the first place 00:25 < bridge> are there random server messages that say something about funding? 00:26 < bridge> We need to bring the fun into funding 00:26 < bridge> Skins! 00:26 < bridge> add accounts 00:26 < bridge> so it will work properly 00:27 < bridge> Personally I'd also handle things a bit different 00:27 < bridge> like give donators a role in the discord and a donator emoji 00:28 < bridge> oh it is, but when you got billions of them, then it's billions times free which accounts to 100 watts 00:28 < bridge> Every donated euro should be 1 point! 00:28 < bridge> lmfao 00:28 < bridge> pay to win ddnet 00:28 < bridge> 1 ddcoin 00:28 < bridge> Why finish Kobra when you can donate for points instead 00:28 < bridge> Yes pay2win 00:28 < bridge> no 00:29 < bridge> skin2win > pay2win 00:29 < bridge> alternate map parts that are blocked by a donation wall 00:29 < bridge> lmao 00:29 < bridge> I'd open a skin shop before even considering pay2win 00:29 < bridge> must donate at least $50 for this skip 00:29 < bridge> charge 20€ to get unbanned 00:29 < bridge> @kollpotato: Genius 00:29 < bridge> Especially with the random ip range vpn bans 00:30 < bridge> many small russian minecraft servers run from those donations 00:30 < bridge> so it works 00:30 < bridge> Unban donations? 00:30 < bridge> It’s a thing? Xd 00:30 < bridge> yes, but they charge much less 00:30 < bridge> nah, they use noise and production process variation to create a stream of random bits which is then decimated 00:30 < bridge> Nice 00:31 < bridge> Can I buy a deen? 00:33 < bridge> where a deen 00:33 < bridge> xd 01:03 < bridge> you could have a donation progress bar as a channel 06:26 < bridge> I can't run what I want to run by typing !multeasymap in my Discord bot. It will list the players but I don't know if the site is the wrong API. 06:59 < bridge> https://blog.nginx.org/blog/native-support-for-acme-protocol 07:24 < bridge> im moving my main linux partition from a 500gb nvme to a 2tb nvme 07:24 < bridge> and in doing so ill migrate from ext4 to xfs 07:24 < bridge> wish me luck 07:42 < bridge> awesome 07:42 < bridge> did u get one of those gen 5 ssds? and ur on the latest ryzen right? it's gonna be so fast 07:43 < bridge> i'm partial to btrfs because it's really rounded itself out over the past few years. it has pretty much every feature, usually with the better implementation, it's COW, somewhat fast 07:43 < bridge> and best of all, the WinBTRFS driver is very very good, has a more extensive config menu than even NTFS itself gives on Windows and is officially signed & easy to install 07:44 < bridge> whereas accessing ext4 on Windows without the use of WSL2 is very difficult. ext2fsd is really terrible 07:44 < bridge> it's very nice for dualbooters 08:09 < bridge> ask the master server 08:09 < bridge> 08:09 < bridge> i think u can ask the server directly aswell but idk how 08:16 < bridge> im not using btrfs, im using xfs, which is the new recommended modern fs and really well tested 08:16 < bridge> gentoo recommends it on the install book now before ext4 08:16 < bridge> it has cow too and reflinks 08:16 < bridge> https://wiki.gentoo.org/wiki/XFS 08:17 < bridge> i can read, im partial to btrfs for the reasons listed 08:17 < bridge> i think the new nvme is faster yeah, from 3500 of the old to 5000 or 7000 i forgot 08:17 < bridge> the old one i bought when nvme where new in market 08:18 < bridge> ah not gen 5 08:18 < bridge> rip 08:19 < bridge> CT2000O3OSSD8 08:20 < bridge> CT2000P3PSSD8 * 08:20 < bridge> https://www.crucial.com/ssd/p3-plus/ct2000p3pssd8%20CT2000O3OSSD8 08:20 < bridge> i think 08:20 < bridge> SSD Series P3 Plus 08:20 < bridge> Interface NVMe (PCIe Gen 4 x4) 08:20 < bridge> Total Density 2TB 08:20 < bridge> Kit Quantity 1 08:20 < bridge> Form Factor M.2 (2280) 08:20 < bridge> SSD Endurance TBW 440TB 08:20 < bridge> Speed & timing 08:20 < bridge> Sequential Write 4,200 MB/s 08:20 < bridge> Sequential Read 5,000 MB/s 08:21 < bridge> https://github.com/TheChymera/mkstage4 08:21 < bridge> im using cpstage 08:21 < bridge> im on a live usb and mounted the 2 disks 08:21 < bridge> + new boot partitions 09:03 < bridge> Sounds great, lmk how it works! 09:03 < bridge> 09:03 < bridge> I want to get a bigger drive for my stuff as well, I also need a better PSU 09:03 < bridge> 09:03 < bridge> Any recommendations? My GPU sucks my current one dry 09:09 < bridge> im gooing full new 09:09 < bridge> im using limine instead of grub 09:09 < bridge> fck grub from now on! 09:09 < bridge> :KEKW: 09:09 < bridge> https://wiki.gentoo.org/wiki/Limine 09:09 < bridge> New month new ryo 09:09 < bridge> i also plan to add limine for my kernel 09:09 < bridge> they have a modern boot protocol 09:10 < bridge> grub uses ancient multiboot 09:10 < bridge> (limine supports multiboot too) 09:10 < bridge> You really making me look into Gentoo again 09:10 < bridge> 09:10 < bridge> I just don't quite understand the gain in it in relation to arch based or Debian based distros 09:10 < bridge> I gave up on the installation after I think 2 hours xd 09:10 < bridge> full customization 09:10 < bridge> fir example limine supports bios, eufi64, uefi32, uefirv64 09:11 < bridge> on arch if u install limine 09:11 < bridge> u get all enabled 09:11 < bridge> i can configure it just for uefi64 09:11 < bridge> which is my needs 09:11 < bridge> and this with all packages 09:11 < bridge> also gentoo has imho more packages than arch 09:11 < bridge> gentoo repos are a better system 09:11 < bridge> arch aur is gentoo guru, but not only that, u can add more repos or making ur repos 09:12 < bridge> i have my own gentoo repo 09:12 < bridge> :poggers2: 09:12 < bridge> Can you link it, I wanna see how Gentoo handles that 09:13 < bridge> I like nix because I can just grep for a package in the nixpkgs repo, if it doesn't exist it's not packaged 09:13 < bridge> https://wiki.gentoo.org/wiki/Ebuild_repository 09:15 < bridge> :EssexWow: 09:15 < bridge> ryo we share similar specs iirc? 09:15 < bridge> Can you share you PSU? I don't know why I saved money on that 09:15 < chillerbot1> You can see donors and how to donate here https://ddnet.org/funding/ 09:15 < bridge> Chilleeeer :angy: 09:23 < ws-client> **** go donate melon :melon: 09:23 < ws-client> **** @Solly (bun bun) souly bun bun when progress #10308 09:23 < chillerbot1> https://github.com/ddnet/ddnet/issues/10308 09:25 < bridge> @chillerdragon I stopped cuz Iade it for something specific and then it stopped being in the direction of what I wanted. You can take it over but if you don't want to I can finish it when I'm free which is not for a week or so 09:25 < ws-client> **** you have 24h! 09:26 < ws-client> **** or i open 20 stale prs in tclient repo 09:26 < ws-client> **** nah sry i dont want to take it over right now too busy my self 09:30 < ws-client> **** @ryozuki when squash https://github.com/ddnet/ddnet/pull/9572 09:33 < ws-client> **** @learath2 when stop ignore heino https://github.com/ddnet/ddnet/pull/10063 09:45 < bridge> just made the map rendering standalone from the client, possibilites are endless :3 10:08 < bridge> hi 10:08 < bridge> from new system 10:10 < bridge> @chillerdragon wtf is an enum class? whats the difference xD 10:10 < bridge> im not making prs or reviewing 10:10 < bridge> u can take over the pr if u want 10:11 < bridge> okay I am unaware, didn't know D: 10:17 < bridge> @blaiszephyr hi from limine bootloaded os, using xfs on the main / drive 10:18 < bridge> Show hardware details 10:18 < bridge> *fetch 10:18 < bridge> i didnt change any pc components xd 10:19 < bridge> 🥱 10:19 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405465757747249212/image.png?ex=689eed7c&is=689d9bfc&hm=eb7a7eb122e64618eddbb586dfc22728e27d7ecd044eb1c2d4257f14fca64de8& 10:19 < bridge> Only changed furniture? 10:20 < bridge> furniture and bootloader and went from a old nvme 512gb to a new with 2tb 10:20 < bridge> Nice 10:20 < bridge> i dont rly have any need to upgrade hardware further yet 10:20 < bridge> i have a rather powerfull system xd 10:21 < bridge> Already did 10:21 < bridge> Hot! 10:21 < bridge> Now leak your psu 10:21 < bridge> @chillerdragon do it i dare you (lmao which one am i replying too, youll never know) 10:22 < bridge> how can i know 10:23 < bridge> i think its not smth i can know from linux 10:23 < bridge> its been years so the cpu isnt new 10:23 < bridge> but i think its gold standard 10:23 < bridge> for psu, i think gold is better over more W 10:24 < bridge> cuz failure in psu can be bad 10:39 < bridge> one more guys 10:42 < bridge> Is this line wrong? I think switch 0 is special and shouldn'https://github.com/ddnet/ddnet/blob/d135c3f412e5c0dfd75522e442b31090acaf7481/src/game/server/gamecontext.cpp#L3160 10:42 < bridge> Is this line wrong? I think switch 0 is special and shouldn't be used? https://github.com/ddnet/ddnet/blob/d135c3f412e5c0dfd75522e442b31090acaf7481/src/game/server/gamecontext.cpp#L3160 10:46 < bridge> isn't switch 0 the switch used by the game/front layer? 10:46 < bridge> yeah it is confusing 10:53 < bridge> done spamming PRs for today :justatest: 10:53 < bridge> Get on your knees :kek: 10:53 < bridge> done spamming PRs for today :justatest: `sorry for the inconvenience` 10:54 < bridge> this one might be something for jupey next week, because 30 - 39 = -9 LOC right? :justatest: 10:54 < bridge> umm what does tthat mean 10:54 < bridge> you can place doors and lasers with the game and front layer 10:55 < bridge> they are still switchers :justatest: 10:55 < bridge> I wait for 24 core desktop cpu :brownbear: 10:55 < bridge> X3d 10:55 < bridge> so what does that do with switch 0 10:56 < bridge> i mean switch 0 is a dummy that has no use isn't it? 10:56 < bridge> oh I wouldn't trust in that, but honestly I don't know 10:59 < bridge> ah yes switch 0 10:59 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405475914367893554/switch-0-test_2025-08-14_10-58-51.mp4?ex=689ef6f2&is=689da572&hm=050cfebed546d60aa6985ae081af161f3c6f352e3aadee111cf6c168de959c45& 11:08 < bridge> hmm seems like it does work but visuals are all messed up 11:12 < bridge> :peepoFlower: 11:22 < bridge> Maybe buy multiple CPUs and build a supercomputer lol 11:29 < bridge> this is cursed 11:29 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405483458494468156/c3c683a4f08e4c4f.png?ex=689efdf9&is=689dac79&hm=6e9bf48cd910546fc60a0b8a96c41aa71f2ab8e9c6e167a5ae7be9a576066198& 11:30 < bridge> the penalty tile counts in the switch id 11:39 < bridge> I hate so much that it's `CQuad *m_pQuads` and not `CQuad* m_pQuads`, one is the name the other is a type, get your shit together C++ 11:43 < bridge> It's clang tidy style 11:43 < bridge> Not c++ 11:43 < bridge> You can set either 11:55 < bridge> here a puzzle for you: 11:55 < bridge> `int *a, *b, c;` 11:55 < bridge> 11:55 < bridge> What type do b and c have? 11:56 < bridge> Pointer and no pointer 11:56 < bridge> I have a rule, that if I don't immediately understand something it's probably bad code 11:57 < bridge> that is bad code, yes 😄 11:58 < bridge> `int x = x++ + ++x -1;` I had to do cursed stuff like this at uni. And always thought, why don't you just write good code instead? 12:01 < bridge> What was even the point of this code 12:02 < bridge> this was just an example for understanding pre and post increase operator 12:02 < bridge> like the guy who buys 217 water melons 12:33 < bridge> Who even made you write this in uni? This is clearly UB 12:34 < bridge> Left one is the sane one, as you pointed out in your follow up with the declaration puzzle 12:35 < bridge> I know, but for me the declaration puzzle is also wrong 😄 but even big programming languages struggle from backwards compat. 12:36 < bridge> Declaration follows use, it's just how declarations are in C++. It's not kept as a weird quirk, it's just how it's designed 12:37 < bridge> like ` , , ` and no pointers, because that would be illegal syntax would be cleaner imo 12:37 < bridge> You might find the way go does it more palatable I guess 12:37 < bridge> yes, coming from other programming languages 12:51 < bridge> `*` is a part of the type so why would it be grouped with the name? 12:59 < bridge> C and C++ declarations explicitly are defined to follow how you'd use the identifier (6.7.6p2 in C++20), the * is technically part of the declarator 12:59 < bridge> C and C++ declarations explicitly are defined to follow how you'd use the identifier (6.7.6p2 in C20), the * is technically part of the declarator 13:04 < ws-client> **** @ryozuki oof you also stepped down as a contributor? ._. 13:05 < bridge> Where the confusion comes from is contexts where you need a `type-name`, the * will appear as part of the type, but in a declaration expression there is no `type-name` 13:08 < bridge> (e.g. a `sizeof` expression, that's where you'll find a `type-name`) 13:08 < bridge> not fully as contrib but im not interested right now 13:09 < bridge> chiller did u miss the whole drama 14:55 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405535174678282362/photo_2025-03-02_15-46-31.jpg?ex=689f2e23&is=689ddca3&hm=c91534798365e99f112e9a27fef5140c5ade569233099242ddd569010cb20834& 14:57 < bridge> https://tenor.com/view/me-atrapaste-es-cine-its-cinema-cinema-esto-es-cine-gif-17729711691959966457 15:04 < bridge> my fixed async lock for use in my kernel executor 15:04 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405537533940731945/mutex.rs?ex=689f3055&is=689dded5&hm=f3a24dd3ca8e01a7135cbdfafdeacfb1499fc9fce82db9d3bc9a32d2a6fe6896& 15:04 < bridge> i wish i could share big snippet and it should be scrollable 15:04 < bridge> skill issue discord 15:05 < bridge> the most interesting part is implementing the future manually in AsyncMutexLockFuture 15:07 < bridge> basically an async mutex must keep a queue of wakers: 15:07 < bridge> - wakers are used to tell the executor, hey this future is ready, poll me please" and it basically adds the related future it to the ready queue of the executor. 15:07 < bridge> atomic to actually do the lock yada yada and drop is the important bit too: 15:07 < bridge> ```rust 15:07 < bridge> impl<'a, T: ?Sized> Drop for AsyncMutexGuard<'a, T> { 15:07 < bridge> fn drop(&mut self) { 15:07 < bridge> let next_waker = self.queue.pop(); 15:07 < bridge> self.lock.store(false, Ordering::Release); 15:07 < bridge> if let Some(next) = next_waker { 15:07 < bridge> next.wake(); 15:07 < bridge> } 15:07 < bridge> } 15:07 < bridge> } 15:07 < bridge> ``` 15:07 < bridge> which wakes the next waker in the queue, important bit is to get the next waker before unlocking or u have a race 15:07 < bridge> well not a real race but u would not get the real next waker 15:08 < bridge> i forgot if this is would be a race condition or data race 15:08 < bridge> i guess race condition which i think is not a memory safety issue 15:08 < bridge> yeah 15:09 < bridge> thanks for coming to my ted talk 15:09 < bridge> i wonder how a channel would look 15:24 < bridge> @kollpotato: i missed part of it because i went for dinner xd but i know ryo stepped down as maintainer i just expected him to finish his pr still .\_. 15:35 < bridge> @aoetw: I think it is pretty cool that you are active on GitHub btw. Much appreciated! 15:39 < bridge> but you can get tricked if you declare multiple variable like this with comma because the * applies only to the first variable 15:57 < bridge> chillerdragon: yeah but looks like edos is more important now 15:59 < bridge> why is there `#[inline(always)]` everywhere 15:59 < bridge> performance 16:00 < bridge> it basically forces the compiler to inline the fn 16:00 < bridge> i dont think its clearly defined 16:00 < bridge> i have read somewhere that the compiler can just ignore it 16:00 < bridge> and do what it wants 16:00 < bridge> it can ignore #inline 16:00 < bridge> but inline always is almost always inlined 16:00 < bridge> the edge cases are for example if u take the ptr of the function and call it 16:01 < bridge> i guess 16:01 < bridge> alright im gonna go spam inline always everywhere 16:01 < bridge> cant wait for your 5gb binary 16:02 < bridge> instructions are small af 16:02 < bridge> unless my function is bazillion lines of code 16:02 < bridge> generics can increase binary size too 16:03 < bridge> wait i'll ask AI 16:04 < bridge> ``` 16:04 < bridge> Ryozuki was correct that #[inline(always)] is much more forceful than #[inline], though even always can be ignored in extreme cases. 16:04 < bridge> So while Koll Potato isn't wrong about instruction size, their plan to "spam inline always everywhere" would likely hurt performance more than help it. The 5GB binary joke from melonFC isn't that far-fetched for a large codebase. 16:04 < bridge> ``` 16:04 < bridge> 16:04 < bridge> 💥 16:04 < bridge> wow it expects than my codebase is huge 16:04 < bridge> wow it expects that my codebase is huge 16:05 < bridge> Change the asset mapres or whatever it’s called to a transparent image? Like how it’s done in this video with skins: https://youtu.be/e4l1lZ5uw6M 16:05 < bridge> 14.5.1 tried to join any ddnet servers, same pic. No OS Changes or Graphics changes done. Mapres folder contains 50 elements, so i guess that folder's fine too. You guys got any idea what changed? 16:05 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/748267089399250964/unknown.png?ex=689ee5c8&is=689d9448&hm=2a64ab33d0f80e4e8bc32a98a9ab173df440f676f613612a97e25ae950dafe1e& 16:05 < bridge> im not ddnet or rust compiler 16:06 < bridge> i have to learn how to not quit a project after 500 lines of code 16:06 < bridge> just never exceed 500 loc 16:07 < bridge> oh right i can use semicolons and have multiple statements per line 16:09 < bridge> I don’t understand which block I should place in the editor for the spawn in tournament, and how to change the 1v1 arena position (X, Y) in DDNetPP? 16:09 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1380166894471872522/ZUV0koZ.png?ex=689f2d15&is=689ddb95&hm=2c2129d16ef9585e199b2778627dd78545ae7d92c2a56cca8c7c33e9d8f07797& 16:14 < bridge> here is a tipp: only do one project at a time and finish it before starting anything new, it helped me a lot 16:15 < bridge> yes, I have like 5 open PRs simultaniously, yes I don't follow my own advice and I am a hypocrite, doesn't mean what I say is not valid 😠 16:18 < bridge> yeah that might actually work 16:18 < bridge> i can start up to 5 projects per day 16:18 < bridge> and in the end have no progress at all 16:22 < bridge> Only useful on certain situations, as where basic atomic calculations do not benefit from it in a real environment. 16:22 < bridge> my kernel has 5800 lines 16:23 < bridge> And yeah, #[inline] only hints to the compiler, that it should be inlined, but the compiler could decide not to inline it to X amount of reasons 16:24 < bridge> Atomic calculations and even memory allocations do not need inlining. It might give an overhead instead 16:29 < bridge> So, easier to debug and benchmark ye? 16:29 < bridge> I recommend using Criterion, as in that measures the overall performance of your functions by continuously iterating it for, 5 seconds as default. 16:30 < bridge> This way, you essentially stress test the function, where #[inline] or #[inline(always)] would make a difference, if there is any. 17:04 < bridge> > Atomic calculations and even memory allocations do not need inlining. 17:04 < bridge> those concepts are not related 17:04 < bridge> inlining allows other optimizations to take place 17:05 < bridge> atomic are synchornization primitives, used in threading 17:05 < bridge> memory allocations are that just memory allocations 17:05 < bridge> in lining what would matter is how the values are "moved" between functions, aka the ptr u hold from a allocation, or the argument u pass (either in register or stack), when inlining, u simply dont pass arguments, because the function to be called body is inserted where the function is called directly 17:06 < bridge> and calls inherently have a small overhead 17:06 < bridge> plus if ur function is small but takes a giant sized parameter, inlining saves a lot 17:07 < bridge> rust abuses inlining heavily 17:07 < bridge> iterators are only perfomant thanks to inlining 17:07 < bridge> this is why rust is so slow in debug mode 17:07 < bridge> because the iterator api has a lot of calls inside 17:07 < bridge> and in debug no inlining is done 17:07 < bridge> after inlining the compiler can also do const folding and easier bounds checks 17:07 < bridge> (both help optimizing the iterators fort example) 18:05 < bridge> fun fact You can compress make most c/c++ codebases to only a few lines 18:40 < bridge> I figured the devs would be annoyed, by me spamming issues all the time lmao xD 18:41 < bridge> I wish I was able to help out, and not just report… 18:47 < bridge> Learn C++ and send PRs :greenthing: 18:53 < bridge> Forget C++ and hope that other developers will fix your issues 19:06 < bridge> https://media.discordapp.net/attachments/1403100340810940611/1404437664672907374/1.jpg?ex=689b3000&is=6899de80&hm=2e1f1d26248dd68076f2a2986a04c43104211a8db288fe3d993d842f1e1b310a&=&format=webp&width=1156&height=1542 https://media.discordapp.net/attachments/1403100340810940611/1404437665423818883/2.jpg?ex=689b3000&is=6899de80&hm=a076d97c5ef5e068a214d549b349d90e94d2a2a24a55173268b5a75be80b1aef&=&format=webp&width=1156&height=1542 https://media.di 19:15 < bridge> @aoetw: you can do surprisingly much without being a fluent C++ developer. Even testing the proposed code changes by running client and server is quite valuable. I would be happy to show you some stuff if you are curious. 19:16 < bridge> But yes even spamming issues is good. I mean we want to reduce the amount of open issues of course. But every valid bug report is very valuable. 19:18 < bridge> you can open as much issues as you want, as long as this are mistakes/bugs you want to have fixed and this isn't a duplicate of an existing issue 19:18 < bridge> Yes that would be really cool, I always download the latest nighty anyways!. 19:18 < bridge> Btw, when are you downloading Discord, so we can privately message each other? 19:18 < bridge> @aoetw: you can browse the other open issues. And can try to reproduce existing bug reports in the latest version. Comment your opinion on feature requests. And also download the client and server from pull requests and run them. See if you like the new changes and if they are bug free 19:18 < bridge> :p 19:18 < bridge> devs appreciate bug reports, specially if its not just "it doesnt work" 19:18 < bridge> Yes 19:19 < bridge> if u tell how to reproduce the issue its even better 19:19 < bridge> easy to test and iterate on the issue to fix it 19:19 < bridge> there wer bugreports where I thought as a dev "this person is full of shit" but they made a guide to reproduce and they were right 19:19 < bridge> @aoetw: nightly is the already merged stuff. You can test ahead of nightly as githubber. You can shape nightly! 19:20 < bridge> sometimes it's really hard to reproduce, like the editor history issues ... 19:20 < bridge> I see, I am very new to Github! 19:20 < bridge> Need a proper guide xD 19:20 < bridge> No need for DMs we can spam this channel @aoetw 19:20 < bridge> There is a green check mark in the pr where you can download the build 19:20 < bridge> Took me forever to figure out how to even download the nightly build. 19:20 < bridge> https://wiki.ddnet.org/wiki/Development 19:20 < bridge> Sadly on phone right now can’t give you good guide 19:21 < bridge> you can do this directly from the homepage or activate BETA in steam, both is fairly simple 19:21 < bridge> !reminder show aoe how to download pr binaries 19:21 < chillerbot1> Helo chillerdragon I will remind you in 86400000 time units about your matter again. 19:21 < bridge> if I was on ubuntu, I'd have a nightly crontab :3 19:21 < bridge> yeah just go on steam right lcick ddnet and betas 19:21 < bridge> then put nightly 19:21 < bridge> Ye I've used Steam in the past, but I've reasontly switched to Tclient because of the Fast input 19:21 < bridge> what is fast input 19:21 < bridge> maybe we should merge it 19:22 < bridge> remind @chillerdragon to improve remind function 19:22 < bridge> @ryozuki: lives under a rock 19:22 < bridge> [Fast input](https://github.com/ddnet/ddnet/issues/9845) 19:22 < bridge> 19:22 < bridge> With “fast inputs” enabled, the client predicts your movement instantly one tick ahead of time, even before the server confirms it. 19:22 < bridge> Fast inputs essentially removes the 20 millisecond delay and prevents you from over-shooting your movements. It makes the game feel more responsive and you'll be able to control your tee a lot better. 19:23 < bridge>