03:22 <+bridge> [ddnet] <Patiga> @Not Keks wanted to figure out how the camera width/height is calculated from the aspect ratio. is `CRenderTools::CalcScreenParams` the correct function, and `WMax`/32 and `HMax`32 are the amounts of tiles horizontally/vertically?
03:24 <+bridge> [ddnet] <Patiga> if so, what is `Amount` supposed to be and what is `f` with that weird calculation
08:14 <+bridge> [ddnet] <Jupstar> Amount is clear I guess. You want resolution independent scaling.
08:14 <+bridge> [ddnet] <Jupstar> The backend's screen mapping will scale down all variables so we have normalized values across all resolutions, so that you end up in a NDC.
08:14 <+bridge> [ddnet] <Jupstar> Basically how pretty much all projection matrices would also do it.
08:14 <+bridge> [ddnet] <Jupstar> Now sqrt(Amount) is the length of one side of the new resolution. I guess this is clear too?
08:14 <+bridge> [ddnet] <Jupstar> Pretty basic geometric math.
08:14 <+bridge> [ddnet] <Jupstar> 
08:14 <+bridge> [ddnet] <Jupstar> Now lets do some math. Imagine we subsitute the `Amount` at the very last step, even if it breaks math.
08:14 <+bridge> [ddnet] <Jupstar> But we want to transform into a different resolution anyway.. in the resolution of Amount:
08:14 <+bridge> [ddnet] <Jupstar> ```
08:14 <+bridge> [ddnet] <Jupstar> x * y = x * y | / y
08:15 <+bridge> [ddnet] <Jupstar> (x * y) / y = x | * x
08:15 <+bridge> [ddnet] <Jupstar> (x * y) * (x / y) = x^2 | sqrt
08:15 <+bridge> [ddnet] <Jupstar> sqrt(x * y) * sqrt(x / y) = x
08:15 <+bridge> [ddnet] <Jupstar> ```
08:15 <+bridge> [ddnet] <Jupstar> 
08:15 <+bridge> [ddnet] <Jupstar> At this point it's clear where our journey ends xd
08:15 <+bridge> [ddnet] <Jupstar> x / y is the Aspect ratio.. The function takes the Aspect ratio as parameter
08:15 <+bridge> [ddnet] <Jupstar> so our new x is:
08:15 <+bridge> [ddnet] <Jupstar> 
08:15 <+bridge> [ddnet] <Jupstar> ```
08:15 <+bridge> [ddnet] <Jupstar> x = sqrt(Amount) * sqrt(AspectRatio) 
08:15 <+bridge> [ddnet] <Jupstar> and y is
08:15 <+bridge> [ddnet] <Jupstar> y = sqrt(Amount) / sqrt(AspectRatio)
08:15 <+bridge> [ddnet] <Jupstar> now look at the original formular
08:15 <+bridge> [ddnet] <Jupstar> 
08:15 <+bridge> [ddnet] <Jupstar> float f = sqrtf(Amount) / sqrtf(Aspect);
08:15 <+bridge> [ddnet] <Jupstar> *w = f * Aspect;
08:15 <+bridge> [ddnet] <Jupstar> *h = f;
08:15 <+bridge> [ddnet] <Jupstar> 
08:15 <+bridge> [ddnet] <Jupstar> its the exact same. f is basically = y
08:15 <+bridge> [ddnet] <Jupstar> and we go from y to x by multipling AspectRatio
08:26 <+bridge> [ddnet] <chillerdragon> There is Patigas twmap which can turn maps into json and back. Do you can track json files in git.
08:26 <+bridge> [ddnet] <chillerdragon> (@cauldron)
08:31 <+bridge> [ddnet] <Jupstar> Maybe a tip for imagination:
08:31 <+bridge> [ddnet] <Jupstar> You want make the new resolution (Amount) fit into our aspect ration
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> So side_length * sqrt(Aspect)
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> and sqrt bcs the AspectRatio is on the "full" old resolution. But you only do the ratio multiplikation on one side of the new resolution not on the full length.
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> E.g. lets say you take 1920x1080
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> thats the full mass.. EVERYTHING XD
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> if you only want the amount of everything that was created by the width
08:31 <+bridge> [ddnet] <Jupstar> u basicall do
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> (1920×1080) × 16/9
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> now you have a new mass, the mass of the width
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> do now get the side length of that new mass u simply put a sqrt around it
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> sqrt((1920×1080) × 16/9)
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> and tada. You are back to 1920
08:31 <+bridge> [ddnet] <Jupstar> 
08:31 <+bridge> [ddnet] <Jupstar> See hard to explain but kinda makes sense if u imagine it somehow
08:32 <+bridge> [ddnet] <Jupstar> thus the name Aspect `Ratio`
08:32 <+bridge> [ddnet] <Jupstar> It is the Ratio between the two sides
08:32 <+bridge> [ddnet] <Jupstar> xd
08:32 <+bridge> [ddnet] <Jupstar> ok i stop now
10:53 <+bridge> [ddnet] <cauldron> well you could even track binary files with git, couldn't you? i was thinking of managing soundpacks for ddnet with git, to just checkout another branch and have another soundpack
10:56 <+bridge> [ddnet] <cauldron> but json converter could make building a ddnet map history feature "natively" without git easy
10:57 <+bridge> [ddnet] <Jupstar> we had an undo manager before just saving the map, its simply very slow
10:57 <+bridge> [ddnet] <Jupstar> calculating the diff is not much faster
10:58 <+bridge> [ddnet] <Jupstar> would be cleaner to seperate all changes on a map from the editor and do undo events there
10:59 <+bridge> [ddnet] <cauldron> what do you mean very slow
10:59 <+bridge> [ddnet] <cauldron> even if its super slow i would use it
11:00 <+bridge> [ddnet] <Jupstar> as said there was one it was made async bcs it was too slow probs
11:00 <+bridge> [ddnet] <cauldron> just saving would be slow, right?
11:01 <+bridge> [ddnet] <cauldron> anyway, i am not into this. its too big for me. soundpack or color feature keep me busy xD
11:02 <+bridge> [ddnet] <cauldron> but is the web collabo editor built from scratch?
11:02 <+bridge> [ddnet] <cauldron> maybe it could have it
11:02 <+bridge> [ddnet] <Jupstar> i cannot say smth about the actual performance, but e.g. adding a layer somewhere on a 4kx4k map would require quite a lot of processing
11:02 <+bridge> [ddnet] <Jupstar> be it diff or just saving
11:03 <+bridge> [ddnet] <Jupstar> bcs it would shift all bytes after the layer at least
11:03 <+bridge> [ddnet] <cauldron> why?? after all its just the tiles mapped to locations or not?
11:03 <+bridge> [ddnet] <cauldron> i mean does the res play a role?
11:04 <+bridge> [ddnet] <cauldron> ah you mean 4k tiles
11:04 <+bridge> [ddnet] <cauldron> still
11:04 <+bridge> [ddnet] <Jupstar> see
11:04 <+bridge> [ddnet] <Jupstar> 
11:04 <+bridge> [ddnet] <Jupstar> you either have an undo manager that knows about the file format (then you can also just make it very efficient, bcs all actions you do in the editor can be cleanly seperated)
11:04 <+bridge> [ddnet] <Jupstar> or you have an undo manager doing smth like git: binary comparision... that can result in a full check in worst case
11:04 <+bridge> [ddnet] <cauldron> ok
11:05 <+bridge> [ddnet] <Jupstar> for git its obviously not annoying, bcs it does that exactly when u expect it
11:05 <+bridge> [ddnet] <Jupstar> but a weird lag everytime u add new tiles or stuiff like that can be really annoying
11:05 <+bridge> [ddnet] <cauldron> yes it has to be seamless
11:05 <+bridge> [ddnet] <cauldron> off
12:04 <+bridge> [ddnet] <Patiga> thank you a lot, your explanation was very good :)
13:21 <+bridge> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/982605095093030952/unknown.png
13:22 <+bridge> [ddnet] <cauldron> hah
13:31 <+bridge> [ddnet] <jao> lmao
15:08 <+bridge> [ddnet] <ReiTW> only retards uses -fr
15:09 <+bridge> [ddnet] <c0d3d3v> I have never build a good undo manager, but I would think if every possible action in the editor has a corresponding inverse action, you could just put the actions on a stack and revert them on a undo command one by one.
15:09 <+bridge> [ddnet] <Jupstar> yes
15:26 <+bridge> [ddnet] <cauldron> btw colemak user
15:26 <+bridge> [ddnet] <cauldron> _ 😎
16:38 <+bridge> [ddnet] <Chairn> I think i reached a new low with windows
16:38 <+bridge> [ddnet] <Chairn> a simple copy paste literally crashed all opened applications simulatenously
16:43 <+bridge> [ddnet] <Wohoo> I think save full state is easier than create inverse action to recover state
16:44 <+bridge> [ddnet] <cauldron> are you using windows as main os or just for gaming / compiling?
16:44 <+bridge> [ddnet] <Chairn> windows on my pc, linux for work and dev
16:45 <+bridge> [ddnet] <Chairn> i sometimes dev on windows like right now
16:45 <+bridge> [ddnet] <Chairn> requires mome memory if you have to save few MB every action
16:48 <+bridge> [ddnet] <cauldron> aha
16:48 <+bridge> [ddnet] <k2d222> I planned on doing that for the web editor although I'm not there yet
16:49 <+bridge> [ddnet] <k2d222> plus multi editor brings a lot of additional pain like concurrent edits
16:49 <+bridge> [ddnet] <cauldron> are you writing it from scratch? adding feature after feature
16:50 <+bridge> [ddnet] <cauldron> is it like etherpad if you know it?
16:50 <+bridge> [ddnet] <k2d222> yea no choice really bc there is a lot of specifics with the way you handle multiple users
16:50 <+bridge> [ddnet] <k2d222> huh no
16:51 <+bridge> [ddnet] <cauldron> etherpad is multiple user online text editor
16:51 <+bridge> [ddnet] <k2d222> oh like google docs
16:51 <+bridge> [ddnet] <k2d222> I went multiple times already to try and understand how they handle undo in case of conflicts
16:52 <+bridge> [ddnet] <cauldron> yeah but open source i think
16:52 <+bridge> [ddnet] <k2d222> I'll have a look then thx
16:53 <+bridge> [ddnet] <cauldron> theres also ethercalc
16:53 <+bridge> [ddnet] <cauldron> and more ethers idk
17:39 <+bridge> [ddnet] <Nagi01 {LAN}> true story
17:40 <+bridge> [ddnet] <ReiTW> Nagi best javascript developer :sue:
17:42 <+bridge> [ddnet] <Nagi01 {LAN}> js is for kiddos
17:43 <+bridge> [ddnet] <Nagi01 {LAN}> real chads use rust
17:43 <+bridge> [ddnet] <Nagi01 {LAN}> :greenthing:
17:44 <+bridge> [ddnet] <Chairn> @heinrich5991 is there a difference between
17:44 <+bridge> [ddnet] <Chairn> ```c++
17:44 <+bridge> [ddnet] <Chairn> std::atomic<uint64_t> *m_pTextureMemoryUsage
17:44 <+bridge> [ddnet] <Chairn> ```
17:44 <+bridge> [ddnet] <Chairn> and 
17:44 <+bridge> [ddnet] <Chairn> ```c++
17:44 <+bridge> [ddnet] <Chairn> std::atomic<uint64_t*> m_pTextureMemoryUsage
17:44 <+bridge> [ddnet] <Chairn> ```
17:44 <+bridge> [ddnet] <Chairn> and 
17:44 <+bridge> [ddnet] <Chairn> ```c++
17:44 <+bridge> [ddnet] <Chairn> std::atomic<uint64_t*> m_pTextureMemoryUsage;
17:44 <+bridge> [ddnet] <Chairn> ```
17:44 <+bridge> [ddnet] <Chairn> @heinrich5991 is there a difference between
17:45 <+bridge> [ddnet] <Chairn> ```c++
17:45 <+bridge> [ddnet] <Chairn> std::atomic<uint64_t> *m_pTextureMemoryUsage;
17:45 <+bridge> [ddnet] <Chairn> ```
17:45 <+bridge> [ddnet] <Jupstar> yes
17:45 <+bridge> [ddnet] <Chairn> which is?
17:45 <+bridge> [ddnet] <Jupstar> one is a pointer of atomic
17:46 <+bridge> [ddnet] <Jupstar> the other an atomic ptr
17:46 <+bridge> [ddnet] <Chairn> thanks...
17:46 <+bridge> [ddnet] <Jupstar> np
17:46 <+bridge> [ddnet] <Jupstar> u cannot just change it
17:47 <+bridge> [ddnet] <Chairn> because i see both in the code, so i wonder if it's intentional
17:47 <+bridge> [ddnet] <Jupstar> where is the 2nd?
17:48 <+bridge> [ddnet] <Chairn> there ```c++
17:48 <+bridge> [ddnet] <Chairn> std::atomic<ILogger *> m_pLogger;
17:48 <+bridge> [ddnet] <Chairn> ```
17:48 <+bridge> [ddnet] <Jupstar> there its probs intendet
17:48 <+bridge> [ddnet] <Chairn> mostly in the new log code
17:49 <+bridge> [ddnet] <Chairn> i was mostly wondering if i should init `std::atomic` variables
17:57 <+bridge> [ddnet] <Swarfey> only french use -fr
17:59 <+bridge> [ddnet] <ReiTW> I'm french I use -rf
18:06 <+bridge> [ddnet] <Chairn> im french and i use -rf
18:56 <+bridge> [ddnet] <cauldron> im german and i use -rf
18:57 <+bridge> [ddnet] <cauldron> ls -alh
18:58 <+bridge> [ddnet] <Jupstar> real pros use GUIs
18:59 <+bridge> [ddnet] <cauldron> eaglemode nice fs "gui"
19:00 <+bridge> [ddnet] <cauldron> http://eaglemode.sourceforge.net/
19:04 <+bridge> [ddnet] <cauldron> https://eaglemode.sourceforge.net/screenshots/em_b02.jpg
19:05 <+bridge> [ddnet] <cauldron> http://eaglemode.sourceforge.net/screenshots/em_b02.html
19:05 <+bridge> [ddnet] <cauldron> http://eaglemode.sourceforge.net/screenshots/em_b02.jpeg
19:37 <+bridge> [ddnet] <Chairn> heretic !!
21:27 <+bridge> [ddnet] <cauldron> nice
21:27 <+bridge> [ddnet] <cauldron> 😅
23:23 <+bridge> [ddnet] <cauldron> i switched to windows after like 9 years of arch, cause
23:25 <+bridge> [ddnet] <cauldron> challenge to use opensource on windows, like alacritty, nvim, co
23:25 <+bridge> [ddnet] <cauldron> # outcoming
23:26 <+bridge> [ddnet] <cauldron> i mean i was editing /etc/netctl profiles for every new wifi i was connecting
23:28 <+bridge> [ddnet] <cauldron> not working for me. rather use windows than spend a weekend to make blietooth headset work on arch.
23:28 <+bridge> [ddnet] <cauldron> sorry opensource, ill try
23:30 <+bridge> [ddnet] <cauldron> choosing react over vuejs
23:30 <+bridge> [ddnet] <cauldron> im outta here
23:34 <+bridge> [ddnet] <cauldron> 🙈
23:36 <+bridge> [ddnet] <Wohoo> @Ryozuki
23:36 <+bridge> [ddnet] <Wohoo> What about testing this tool with DDnet, to have full dump execution?
23:36 <+bridge> [ddnet] <Wohoo> https://rr-project.org/
23:36 <+bridge> [ddnet] <Wohoo> It's on HN front page.
23:36 <+bridge> [ddnet] <Wohoo> I've never tested it, it should probably dump network packets too, since it is mainly used with Firefox.
23:36 <+bridge> [ddnet] <Wohoo> Reverse execution feels like magic to me
23:49 <+bridge> [ddnet] <deen> Those are projects are very cool, but slow down execution by a lot
23:50 <+bridge> [ddnet] <deen> So it's something you can use in your central tests sometimes to try and catch hard to reproduce bugs