00:24 <+bridge_> <reitw> nop, but I can tell you to not use awx lmao
01:07 <+bridge_> <soulyvevo> nice license type https://github.com/ddnet/ServerStatus
01:07 <+bridge_> <soulyvevo> https://cdn.discordapp.com/attachments/293493549758939136/1515855151053930516/image.png?ex=6a30859f&is=6a2f341f&hm=3aad6786d2d8c5c876e8193685abcdafd14d123beee328d2cae298b433e8d9aa&
01:45 <+bridge_> <hooliga> bro
01:45 <+bridge_> <hooliga> https://cdn.discordapp.com/attachments/293493549758939136/1515864843599478967/1.jpg?ex=6a308ea6&is=6a2f3d26&hm=e63f53ec9994b8e8d9d597d2a4993d56ffebd9b7fe20bfca1387b35d39cae506&
01:45 <+bridge_> <hooliga> https://cdn.discordapp.com/attachments/293493549758939136/1515864843985616957/2.jpg?ex=6a308ea6&is=6a2f3d26&hm=7a290eb8432ae2a402979e57283bd9dbdf8dbe5c64b19bd0fbba6d2de7fd871a&
01:45 <+bridge_> <hooliga> https://cdn.discordapp.com/attachments/293493549758939136/1515864844388008148/3.jpg?ex=6a308ea6&is=6a2f3d26&hm=d1ba684b9d4fd9c0a04b406334567f861d012a9a1c8d936404a72e0580b4c01d&
01:45 <+bridge_> <hooliga> https://cdn.discordapp.com/attachments/293493549758939136/1515864844774150344/4.jpg?ex=6a308ea6&is=6a2f3d26&hm=783a692b2485a70ad5981f96a6fb2f3432fe60470dee5cba774dd0fae59f87a4&
09:12 <+bridge_> <chillerdragon> so annoying that i can never run clang-tidy locally to test if the CI would pass because my arch clang-tidy version finds a GAZZILION errors always
09:15 <+bridge_> <milkeeycat> I use https://github.com/muttleyxd/clang-tools-static-binaries for clang-format, you can download clang-tidy from there as well
09:26 <+bridge_> <Fussel> Such "useless parenthesis" can be even more than just useless, but dangerous :happy:
09:30 <+bridge_> <chillerdragon> @milkeeycat  sounds annoying to setup but ye i should probably do that
09:30 <+bridge_> <chillerdragon> lets spam a few prs for now to reduce the noise :D
09:38 <+bridge_> <chillerdragon> most annoying one is the warning in color.h which i am not too sure about what to do and it throws every time color.h is included which is everywhere
09:38 <+bridge_> <chillerdragon> https://paste.zillyhuhn.com/kH
09:38 <+bridge_> <chillerdragon> https://github.com/ddnet/ddnet/blob/36336392ea7635cb35530b7388525347cbc8ccf4/src/base/color.h#L160
09:38 <+bridge_> <chillerdragon> https://github.com/ddnet/ddnet/blob/36336392ea7635cb35530b7388525347cbc8ccf4/src/base/color.h#L96
09:39 <+bridge_> <chillerdragon> fakin template magic
09:47 <+bridge_> <Assa> is this maybe missing the "override" keyword? Otherwise this is actually shadowing
09:48 <+bridge_> <Fussel> 160 shadows 96, w/o any apparent reason. Just delete 160
09:49 <+bridge_> <Assa> wait this is just a wrapper which does nothing
09:49 <+bridge_> <Fussel> override requires the function member to be virtual
09:49 <+bridge_> <Assa> you really could just delete the Pack function here right?
09:49 <+bridge_> <Fussel> override requires the function member to be virtual (and therefore not constexpr :happy:)
09:49 <+bridge_> <Assa> it just calls the base function anyways
09:50 <+bridge_> <Assa> and should be visible
09:52 <+bridge_> <Fussel> Yep.
09:55 <+bridge_> <Fussel> The whole class is a bit strange …
09:57 <+bridge_> <Fussel> chillerdragn if you want to flex around there you can replace this whole union stuff by std::variant, don't use `union`
09:57 <+bridge_> <Fussel> chillerdragon if you want to flex around there you can replace this whole union stuff by std::variant, don't use `union`
09:58 <+bridge_> <Assa> I am sure this is very old, where std::variant wasn't a thing
09:58 <+bridge_> <Assa> I remember reading the color class 15 years ago already
09:59 <+bridge_> <Fussel> and this whole pseudo-CRBP exists for completely no reason, classes can have function member templates w/o being class templates
09:59 <+bridge_> <Assa> I see we often use the same colors at multiple spaces, like menu color or others - do you think it would make sense to add a list of color constants?
10:01 <+bridge_> <Fussel> akshually this whole thingy doesn't do even what it's meant to do …
10:09 <+bridge_> <Fussel> `requires(!std::is_same_v<DerivedT, OtherDerivedT>)` can't be very old :>
10:10 <+bridge_> <Fussel> and I didn't blame the original author, but we have better alternatives now
10:10 <+bridge_> <Assa> doesn't mean it didn't change with time :p
10:11 <+bridge_> <Assa> and if you want to rewrite it using std::variant, go ahead - I'd just make sure that this is already available for the minimal gcc version we support
10:12 <+bridge_> <Assa> nah this came with cpp17 so go ahead 😄
10:17 <+bridge_> <Assa> just write me if you do, otherwise I might suddenly write motivation again xD
10:17 <+bridge_> <Assa> just write me if you do, otherwise I might suddenly find motivation again xD
10:28 <+bridge_> <Fussel> I'm not even sure if it's a good idea to solve this via inheritance, traits would be another approach
11:30 <+bridge_> <Fussel> https://cdn.discordapp.com/attachments/293493549758939136/1516012086751268914/image.png?ex=6a3117c7&is=6a2fc647&hm=bf49405786e79a5e2adf234d607e48ac1b0b0dce93e8e50c548cd68c8bec4a0e&
11:30 <+bridge_> <Fussel> erm …
11:31 <+bridge_> <Fussel> shouldn't it be more like this?
11:31 <+bridge_> <Fussel> ~~~cpp
11:31 <+bridge_> <Fussel> template<typename H, typename ... T> 
11:32 <+bridge_> <Fussel> constexpr H minimum(H const& h, T... t)
11:32 <+bridge_> <Fussel> {
11:32 <+bridge_> <Fussel>     if constexpr (sizeof... (t) > 0)
11:32 <+bridge_> <Fussel>         return std::min(h, minimum(t...));
11:32 <+bridge_> <Fussel>     else
11:32 <+bridge_> <Fussel>         return h;
11:32 <+bridge_> <Fussel> }~~~
11:32 <+bridge_> <Fussel> shouldn't it be more like this?
11:32 <+bridge_> <Fussel> ```cpp
11:32 <+bridge_> <Fussel> template<typename H, typename ... T> 
11:32 <+bridge_> <Fussel> constexpr H minimum(H const& h, T... t)
11:32 <+bridge_> <Fussel> {
11:32 <+bridge_> <Fussel>     if constexpr (sizeof... (t) > 0)
11:32 <+bridge_> <Fussel>         return std::min(h, minimum(t...));
11:32 <+bridge_> <Fussel>     else
11:32 <+bridge_> <Fussel>         return h;
11:32 <+bridge_> <Fussel> }```
11:46 <+bridge_> <Assa> I guess more than 3 arguments is not that wanted, doing this recursive might also not be the best decision
11:47 <+bridge_> <Assa> well maybe this doesn't matter since it's constexpr
11:50 <+bridge_> <Fussel> Code duplication is hardly ever wanted at all
11:50 <+bridge_> <heinrich5991> `constexpr` only means "it's possible to evaluate this at compile time, there are no guarantees as to whether this is done or not"
11:51 <+bridge_> <Fussel> … and it's tail recursion, it would be curious if the compiler doesn't unfold it
11:53 <+bridge_> <Fussel> + the current solution does call by value, not good
11:54 <+bridge_> <heinrich5991> the current thing looks just fine for the things it's likely used with
11:54 <+bridge_> <heinrich5991> ints and floats
11:55 <+bridge_> <heinrich5991> we're not trying to be a general purpose template library, we do what we need for our purposes
11:55 <+bridge_> <teero777> why would you use a reference on an integer? xd
11:56 <+bridge_> <aegisub> me when I don't write 200 lines of template metaprogramming to take the maximum of three integers
11:56 <+bridge_> <Fussel> 1. It copies, calls by reference are easier to optimize out
11:56 <+bridge_> <Fussel> 2. a reference is (if at all) 8 bytes, float and double are larger
11:57 <+bridge_> <Assa> same reasoning, there is no guarantee
11:57 <+bridge_> <teero777> a reference is 8 bytes on 64bit machines, a float is 4bytes
11:57 <+bridge_> <teero777> idk what you are yapping
11:57 <+bridge_> <teero777> the compiler is not stupid enough to not recognize a simple function like that
11:58 <+bridge_> <Assa> yes, hopefully
12:02 <+bridge_> <Fussel> Now you have your answer ↑
14:32 <+bridge_> <heinrich5991> `ChillerDragon:pr_nullptrnullabnullnotNULLornul`
14:32 <+bridge_> <heinrich5991> you ok? 😛
14:33 <+bridge_> <chillerdragon> 🤪
14:34 <+bridge_> <chillerdragon> its replacing 0 with nullptr in antibot\_null.cpp so that branch name is pretty descriptive if you ask me
14:36 <+bridge_> <chillerdragon> i have been using funny branch names for a while since nobody seems to be offended by it so its my canvas where i do let my creativity flurish
14:36 <+bridge_> <chillerdragon> i dont think flurish is a word
14:37 <+bridge_> <milkeeycat> on this beautiful day I'd like to share with you my favorite chiller's commit name https://github.com/ChillerDragon/json/commit/64762085555c6cd58e070f0856de478211360b92
14:37 <+bridge_> <chillerdragon> xd
14:38 <+bridge_> <chillerdragon> not sure if assa noticed the branch name he merged here https://github.com/ddnet/ddnet/pull/12293
15:03 <+bridge_> <Assa> I ignore branch names as long as ppl ignore mine
15:04 <+bridge_> <Assa> I wonder if wuthering is a reference about a game or about fish
15:51 <+bridge_> <jxsl13> probably: https://www.imdb.com/title/tt32897959/
15:55 <+bridge_> <patrickmancini499> https://cdn.discordapp.com/attachments/293493549758939136/1516078719716950188/image.jpg?ex=6a3155d6&is=6a300456&hm=289a5c861595a26a9c1cbd5c9d5b585a5e3e21bb402505703edb767e3d2e807d&
15:55 <+bridge_> <patrickmancini499> https://cdn.discordapp.com/attachments/293493549758939136/1516078720140710139/image.jpg?ex=6a3155d6&is=6a300456&hm=452c55b58abad3a3f69d638e41c6818071038141b1ca22d9b06dc068f9836fc3&
15:58 <+bridge_> <chairn> more like the book though 🙂
16:20 <+bridge_> <jxsl13> preferrably provide an econ alternative that is not based on telnet but encrypted :'/
16:21 <+bridge_> <jxsl13> or encrypted telnet 👀
16:25 <+bridge_> <melon> `shell: "echo '{{ command }}' > {{ deploy_path }}/servers/{{ server.id }}.fifo"` 🥲
16:38 <+bridge_> <jxsl13> that's no better than current econ. you need to be connected to the server and run on localhost mire or less.
16:38 <+bridge_> <jxsl13> more*
16:39 <+bridge_> <jxsl13> maybe with some sh shenanigans feasible
16:39 <+bridge_> <jxsl13> ssh
16:40 <+bridge_> <jxsl13> ssh + log parsing (where econ is actually in its current form the better solution)
16:45 <+bridge_> <melon> econ over TLS or similar would be really neat indeed, but i'm unsure how applicable it is
16:46 <+bridge_> <melon> maybe its something that could be done with QUIC over UDP
17:35 <+bridge_> <learath2> the problem with econ isn't the connection method, sure it can be improved but that's not what makes it pretty much useless
17:35 <+bridge_> <learath2> it's the completely unstructured and uncorrelated input and output
17:36 <+bridge_> <learath2> you send a command, you have almost no way to figure out what the result of the command is reliably
17:41 <+bridge_> <robyt3> Create another FIFO/Named pipe in the opposite direction where the server writes the command output to? :justatest:
17:52 <+bridge_> <jxsl13> econ with extra steps 👀
17:53 <+bridge_> <jxsl13> :kek:
17:54 <+bridge_> <jxsl13> but via ssh it'd at least be secure without ddnet devs maintaining it
17:55 <+bridge_> <Fussel> how about a more general approach: let the client use an arbitrary tag (let's say 5 letters of their choice) and every response to this request gets prefixed with this tag?
17:56 <+bridge_> <Fussel> like `tagged_status 4711`
17:56 <+bridge_> <Fussel> ```4711 blupp
17:56 <+bridge_> <Fussel> 4711 blah
17:56 <+bridge_> <Fussel> 4711 foo```
17:56 <+bridge_> <Fussel> I can remember it working well with /who requests on IRC back in these days
17:56 <+bridge_> <robyt3> FIFO over SSH is easier than implementing SSH in the server
17:56 <+bridge_> <jxsl13> that's what i am saying
17:57 <+bridge_> <jxsl13> you use fifo files over ssh without having to maintain the ssh implementation
17:57 <+bridge_> <jxsl13> which would be neat
17:58 <+bridge_> <robyt3> Yeah, that's also what I was thinking. Associate the command/session with a UUID, then write output for the specific command/session to a file/socket/FIFO/pipe and include the UUID to identify the session.
17:59 <+bridge_> <robyt3> That's already how DDNet works
17:59 <+bridge_> <jxsl13> 👀 output as well
18:00 <+bridge_> <jxsl13> hmmm are there any example docs?
18:00 <+bridge_> <robyt3> ddnet-scripts is the public part
18:01 <+ChillerDragon> just bake ssh server into ddnet server
18:01 <+bridge_> <robyt3> The wrapper script would have to pass that back over SSH too when the output can be identified
18:06 <+bridge_> <Fussel> With "more general" I meant using it for rcon too. This would allow the client to issue multiple concurrent requests instead of artificially serializing them. And most of the multiline response commands don't have a clear end of reply marker
18:06 <+bridge_> <Fussel> With "more general" I meant using it for rcon too. This would allow the client to issue multiple concurrent requests instead of artificially serializing them. And most of the multiline response commands don't have a clear-end-of-reply marker
18:07 <+bridge_> <Fussel> rn I'm parsing the response based on a line pattern, this has limitations :happy:
18:49 <+bridge_> <mpft> a more general way of sending secure traffic could also be helpful if it's extended to chat (kog /login comes to mind, maybe rcon credential transport also fall into this)
18:50 <+bridge_> <louis> you could already do this if you echo smth before your commands i guess
18:51 <+bridge_> <louis> structured logging is much needed though
18:53 <+bridge_> <Fussel> I'm using a modified client, I grab the stuff from rcon itself hooked into `CGameClient::OnRconLine`
18:53 <+bridge_> <melon> tbh fifo over ssh is the easiest we could implement without much hussle :D
19:35 <+bridge_> <fokkonaut> @heinrich5991 LOTM
19:36 <+bridge_> <robyt3> Would be neat to have separate netmessages for rendering structured commands and output in the console. Quick sketch with paint.
19:36 <+bridge_> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1516134219183816794/image.png?ex=6a318986&is=6a303806&hm=c6ce58ab0185cf577b29ea49ca35aa3c1028c1585ddfd6401602e898cb89ddf0&
19:36 <+bridge_> <Fussel> like a separate NET_MSG with further subdivision?
19:37 <+bridge_> <Fussel> I thought about something like this for the account system :happy:
19:37 <+bridge_> <fokkonaut> shouldnt this area collapse up maybe? useful for f1 too
19:37 <+bridge_> <fokkonaut> https://cdn.discordapp.com/attachments/293493549758939136/1516134628967059517/image.png?ex=6a3189e7&is=6a303867&hm=6f0d5e23bb83adac9d489b93d2eee4ba5aff406a565a432919d01a8789b0523b&
19:38 <+bridge_> <robyt3> Like the server tells the client that a command was executed. This includes the meta data like key, command and channel (rcon, econ, FIFO, etc.) in different fields. Each command group has some UUID so the server can add messages specifically to each block. Could also indicate that the block is fully received.
19:39 <+bridge_> <robyt3> Yeah, I also have this on my TODO list for a console rework so the completion options work like in the editor.
19:39 <+bridge_> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1516135158418374777/image.png?ex=6a318a66&is=6a3038e6&hm=b639271b1baeec69ea4db804af23ee72b4dc5c07e1e996ffa3ce23d369cef402&
19:40 <+bridge_> <fokkonaut> hmm
19:41 <+bridge_> <Fussel> Sounds like it's almost too late to shift the config variables and commands to json :s
19:54 <+bridge_> <learath2> Would be nice to have something like json that the client can render however it likes if it knows it
21:08 <+bridge_> <robyt3> That could be added on top of this relatively easily. Add a `format` property to the netmessage to differentiate between `plain` and `json`.  With `json` format, the client parses the output as json, with the json data expected to contain a more detailed `"format"` attribute that specifies the exact format (e.g. table of bans, antibot specific output etc.). If the command output exceeds the size of one netmessage, the output is continued in separat
21:15 <+bridge_> <rei_ayanami_tee> hey
21:16 <+bridge_> <rei_ayanami_tee> i need help
21:17 <+bridge_> <rei_ayanami_tee> i cant join some fng servers with normal github build of 19.8.3 and 19.8.2
21:17 <+bridge_> <rei_ayanami_tee> what should i do
21:21 <+bridge_> <melon> @_noby :angy:
21:21 <+bridge_> <bamcane_tee237> why can't you join
21:22 <+bridge_> <bamcane_tee237> since your version?
21:25 <+bridge_> <bamcane_tee237> :thonk:
21:26 <+bridge_> <bamcane_tee237> I remembered some useless anti-bot system would block some versions, not sure which fng server you join.
21:43 <+bridge_> <rei_ayanami_tee> idk it just crashes
21:44 <+CN-Bridge> <Bamcane>  I guess it's snapshot issue. D:
21:45 <+bridge_> <rei_ayanami_tee> isnt 19.8.2 alr released?
21:46 <+bridge_> <rei_ayanami_tee> https://cdn.discordapp.com/attachments/293493549758939136/1516167092636877011/image.png?ex=6a31a823&is=6a3056a3&hm=fb48055b48d99437880904e94ec7016a2641c6218dd97ffe3aa11c4e8a213a8f&
21:47 <+CN-Bridge> <Bamcane>  not sure if the issue it is
22:12 <+bridge_> <milkeeycat> @robyt3 I have a question about #7544, do you imagine all binds be layout aware or only some of them/maybe ones outside of the game itself(like only in the editor)
22:12 <+bridge_> <milkeeycat> @robyt3 I have a question about https://github.com/ddnet/ddnet/issues/7544, do you imagine all binds be layout aware or only some of them/maybe ones outside of the game itself(like only in the editor)
22:12 <+bridge_> <milkeeycat> @robyt3 I have a question about https://github.com/ddnet/ddnet/issues/7544, do you imagine all binds be layout aware or only some of them/maybe ones outside of the game itself(like only in the editor)
22:12 <+bridge_> <milkeeycat> @robyt3 I have a question about https://github.com/ddnet/ddnet/issues/7544, do you imagine all binds be layout aware or only some of them/maybe ones outside of the game itself(like only in the editor)
22:22 <+bridge_> <robyt3> I think we don't want to change how it works for binds by default. The only case where it matters currently are Ctrl+Z/Y hotkeys to undo/redo. Would be nice if the UI could be agnostic to the layout. https://github.com/MilkeeyCat/ddnet/commit/eb912a49188ea26121f889860bac0683451d409b seems to work for me to correctly use Ctrl+Z/Y with `inp_translated_keys 1` whereas the setting previous did nothing.
23:19 <+bridge_> <milkeeycat> I don't know if you investigated why current `inp_translated_keys 1` doesn't work, but here's my 2 cents about this.
23:19 <+bridge_> <milkeeycat> 
23:19 <+bridge_> <milkeeycat> SDL under the hood has [`keymap`](https://github.com/libsdl-org/SDL/blob/1414dbf29d272f15a755afc7eee8242ad89b7226/src/events/SDL_keyboard.c#L49) array, its size is `SDL_NUM_SCANCODES` and it stores `SDL_Keycode`s. In other words it's an array that allows to look up what keycode corresponds to what scancode and vice versa.
23:19 <+bridge_> <milkeeycat> 
23:19 <+bridge_> <milkeeycat> On QWERTY layout, `keycode[SDL_SCANCODE_Y]` is `y` and `keycode[SDL_SCANCODE_Z]` is `z`.
23:19 <+bridge_> <milkeeycat> On QWERTZ layout, `keycode[SDL_SCANCODE_Y]` is `z` and `keycode[SDL_SCANCODE_Z]` is `y`.
23:19 <+bridge_> <milkeeycat> 
23:19 <+bridge_> <milkeeycat> Current implementation of `inp_translated_keys` uses [`SDL_GetScancodeFromKey`](https://github.com/libsdl-org/SDL/blob/1414dbf29d272f15a755afc7eee8242ad89b7226/src/events/SDL_keyboard.c#L1163-L1175)a to get the scancode from the keycode.
23:19 <+bridge_> <milkeeycat> If you look at the code, you can see that it simply iterates over this `keymap` array and checks if `keymap[scancode]` equals to the passed keycode.
23:19 <+bridge_> <milkeeycat> 
23:20 <+bridge_> <milkeeycat> On QWERTY layout, `SDL_GetScancodeFromKey(SDLK_y)` will return `SDL_SCANCODE_Y` because `keymap[SDL_SCANCODE_Y]` is going to be the place that has `SDLK_y` as a value.
23:20 <+bridge_> <milkeeycat> On QWERTZ layout, `SDL_GetScancodeFromKey(SDLK_y)` will return `SDL_SCANCODE_Z` because `keymap[SDL_SCANCODE_Z]` is going to be the place that has `SDLK_y` as a value.
23:26 <+bridge_> <robyt3> When pressing `Y` on my QWERTZ keyboard I get `sym=121 scancode=29 scancode-from-key=29`. When pressing `Z` I get `sym=122 scancode=28 scancode-from-key=28`.