02:11 < bridge> <-StormAx> any devs, i know furo gonna blame me, but how can i stop looping this mf? 02:11 < bridge> <-StormAx> https://cdn.discordapp.com/attachments/293493549758939136/1191911600546906312/image.png?ex=65a7292e&is=6594b42e&hm=abe60aa8892a3560a4ba77913913002a5c2b85a81fdaec45c9d23751dcb3d135& 02:11 < bridge> <-StormAx> ```cpp 02:11 < bridge> <-StormAx> 02:11 < bridge> <-StormAx> bool IsParsed = false; 02:11 < bridge> <-StormAx> const char *pname = Client()->PlayerName(); 02:11 < bridge> <-StormAx> if(!IsParsed) 02:11 < bridge> <-StormAx> { 02:11 < bridge> <-StormAx> 02:11 < bridge> <-StormAx> m_pClient->m_Stats.FetchPlayer(&s_StatsPlayer, pname); 02:11 < bridge> <-StormAx> 02:11 < bridge> <-StormAx> 02:11 < bridge> <-StormAx> 02:11 < bridge> <-StormAx> if(!s_StatsPlayer.m_pGetStatsDDStats) 02:11 < bridge> <-StormAx> return; 02:11 < bridge> <-StormAx> 02:11 < bridge> <-StormAx> if(s_StatsPlayer.m_pGetStatsDDStats->State() == s_StatsPlayer.m_pGetStatsDDStats->STATE_PENDING) 02:11 < bridge> <-StormAx> return; 02:11 < bridge> <-StormAx> 02:11 < bridge> <-StormAx> if(!s_StatsPlayer.StatsParsed && s_StatsPlayer.m_pGetStatsDDStats->State() == s_StatsPlayer.m_pGetStatsDDStats->STATE_DONE) 02:11 < bridge> <-StormAx> return m_pClient->m_Stats.ParseJSON(&s_StatsPlayer); IsParsed = true; 02:12 < bridge> <-StormAx> } 02:12 < bridge> <-StormAx> ``` 02:32 < bridge> you’re returning before you set IsParsed to true 02:43 < bridge> the `return m_pClient->m_Stats.ParseJSON(&s_StatsPlayer);` does what it should, return!, playing `isParsed = true` after that has no effect 02:43 < bridge> the `return m_pClient->m_Stats.ParseJSON(&s_StatsPlayer);` does what it should, return!, placing `isParsed = true` after that has no effect 02:44 < bridge> the `return m_pClient->m_Stats.ParseJSON(&s_StatsPlayer);` does what it should, return!, placing `isParsed = true` after that has no effect (yes Ewan you just said that but i know Stormi, he needs full explanations!) ❤️ 02:47 < bridge> the `return m_pClient->m_Stats.ParseJSON(&s_StatsPlayer);` does what it should, return!, placing `isParsed = true` after that has no effect 03:13 < bridge> this is not python, use {} for multi line statements after if 03:13 < bridge> btw, i don't see any loop in there 04:42 < bridge> pls use auto format 05:22 < bridge> how long until a ddnet server fix until its live in the servers? 05:28 < bridge> cause I still cant see finish messages 05:45 < bridge> how long until a ddnet server fix is live in the servers? 07:41 < bridge> There is no fixed protocol. Its whenever deen restarts the servers. You can speed it up by pinging deen to do so. Which finish messages do you mean by the way? For 0.7 (they for sure should be rolled out) or the new ones for 0.6 ddnet protocol? 07:53 < ChillerDragon> hmm fudgy so even the ddnet version works for me on ger and usa servers already so it should have worked for you too ._. 07:53 < ChillerDragon> https://zillyhuhn.com/cs/.1704264765.png 07:53 < ChillerDragon> my hax client just doesnt support the texture yet 07:56 < bridge> is this efficient ? located in ddrace.cpp 07:56 < bridge> ```cpp 07:56 < bridge> int m_TileIndex = GameServer()->Collision()->GetTileIndex(MapIndex); 07:56 < bridge> int m_TileFIndex = GameServer()->Collision()->GetFTileIndex(MapIndex); 07:56 < bridge> 07:56 < bridge> //Sensitivity 07:56 < bridge> int S1 = GameServer()->Collision()->GetPureMapIndex(vec2(pChr->GetPos().x + pChr->GetProximityRadius() / 3.f, pChr->GetPos().y - pChr->GetProximityRadius() / 3.f)); 07:56 < bridge> int S2 = GameServer()->Collision()->GetPureMapIndex(vec2(pChr->GetPos().x + pChr->GetProximityRadius() / 3.f, pChr->GetPos().y + pChr->GetProximityRadius() / 3.f)); 07:56 < bridge> int S3 = GameServer()->Collision()->GetPureMapIndex(vec2(pChr->GetPos().x - pChr->GetProximityRadius() / 3.f, pChr->GetPos().y - pChr->GetProximityRadius() / 3.f)); 07:56 < bridge> int S4 = GameServer()->Collision()->GetPureMapIndex(vec2(pChr->GetPos().x - pChr->GetProximityRadius() / 3.f, pChr->GetPos().y + pChr->GetProximityRadius() / 3.f)); 07:56 < bridge> int Tile1 = GameServer()->Collision()->GetTileIndex(S1); 07:56 < bridge> int Tile2 = GameServer()->Collision()->GetTileIndex(S2); 07:56 < bridge> int Tile3 = GameServer()->Collision()->GetTileIndex(S3); 07:56 < bridge> int Tile4 = GameServer()->Collision()->GetTileIndex(S4); 07:56 < bridge> int FTile1 = GameServer()->Collision()->GetFTileIndex(S1); 07:56 < bridge> int FTile2 = GameServer()->Collision()->GetFTileIndex(S2); 07:56 < bridge> int FTile3 = GameServer()->Collision()->GetFTileIndex(S3); 07:56 < bridge> int FTile4 = GameServer()->Collision()->GetFTileIndex(S4); 07:56 < bridge> ``` 07:56 < bridge> is this efficient ? located in ddrace.cpp 07:56 < bridge> ```cpp 07:56 < bridge> int m_TileIndex = GameServer()->Collision()->GetTileIndex(MapIndex); 07:56 < bridge> int m_TileFIndex = GameServer()->Collision()->GetFTileIndex(MapIndex); 07:57 < bridge> 07:57 < bridge> //Sensitivity 07:57 < bridge> int S1 = GameServer()->Collision()->GetPureMapIndex(vec2(pChr->GetPos().x + pChr->GetProximityRadius() / 3.f, pChr->GetPos().y - pChr->GetProximityRadius() / 3.f)); 07:57 < bridge> int S2 = GameServer()->Collision()->GetPureMapIndex(vec2(pChr->GetPos().x + pChr->GetProximityRadius() / 3.f, pChr->GetPos().y + pChr->GetProximityRadius() / 3.f)); 07:57 < bridge> int S3 = GameServer()->Collision()->GetPureMapIndex(vec2(pChr->GetPos().x - pChr->GetProximityRadius() / 3.f, pChr->GetPos().y - pChr->GetProximityRadius() / 3.f)); 07:57 < bridge> int S4 = GameServer()->Collision()->GetPureMapIndex(vec2(pChr->GetPos().x - pChr->GetProximityRadius() / 3.f, pChr->GetPos().y + pChr->GetProximityRadius() / 3.f)); 07:57 < bridge> int Tile1 = GameServer()->Collision()->GetTileIndex(S1); 07:57 < bridge> int Tile2 = GameServer()->Collision()->GetTileIndex(S2); 07:57 < bridge> int Tile3 = GameServer()->Collision()->GetTileIndex(S3); 07:57 < bridge> int Tile4 = GameServer()->Collision()->GetTileIndex(S4); 07:57 < bridge> int FTile1 = GameServer()->Collision()->GetFTileIndex(S1); 07:57 < bridge> int FTile2 = GameServer()->Collision()->GetFTileIndex(S2); 07:57 < bridge> int FTile3 = GameServer()->Collision()->GetFTileIndex(S3); 07:57 < bridge> int FTile4 = GameServer()->Collision()->GetFTileIndex(S4); 07:57 < bridge> ``` 07:57 < bridge> every time a player get contact with a tile it makes so many ints 07:58 < ws-client> making ints is cheap 07:58 < ws-client> its more the method call where the int value comes from but i think those are also only index lookups 07:58 < bridge> isn't it more efficient to make all of the ints in character.h and valuate them here ? 07:59 < ChillerDragon> they are on the stack either way 07:59 < ChillerDragon> its not allocating any memory 07:59 < ChillerDragon> on the heap 07:59 < bridge> hum 08:00 < ChillerDragon> So the GetPureMapIndex is basically taking a pos and then doing some simple math like multiply by width and height and offset etc. Thats a fast operation without any iteration. 08:00 < ChillerDragon> Then GetTileIndex() is just a array index ``return m_pTiles[Index].m_Index;`` 08:04 < bridge> Where the ints are stored has very little to do with the cost of that operation. "Creating an int" has no inherent runtime cost unless it's getting allocated dynamically 08:04 < bridge> i see 08:12 < bridge> what git commands you guys use in your every day ? 08:16 < bridge> what git commands you guys use in your every day of development ? 08:23 < ChillerDragon> @mods bro, checkout this user: piDOR on ddnet copy love box 08:23 < ChillerDragon> rus blocker #1 08:23 < chillerbot> https://github.com/ddnet/ddnet/issues/1 08:24 < ChillerDragon> nvm he left 08:24 < ChillerDragon> omg chilerbot u stoopid 08:25 < ChillerDragon> @mamad_melanin ``git add .`` ``git commit`` ``git push`` ``git commit --amend`` ``git push -f`` and a few more that i frequently use but those are my daily musts xd 08:27 < ChillerDragon> oh and omg ``git status`` 08:27 < ChillerDragon> and ``git diff`` and ``git diff --cached`` ``git log`` i use those bithses daily for sure 08:28 < bridge> `fetch`, `push`, `rebase`, `reset`, `status`, `stash` 08:28 < bridge> Also `diff` sometimes, and very rarely `cherry-pick` 08:28 < ChillerDragon> oh true ``git pull`` 08:28 < ChillerDragon> also a classic 08:29 < bridge> git help 08:29 < ChillerDragon> xd 08:29 < bridge> :KEKSad: 08:29 < ChillerDragon> i also reset, stash, cherry-pick, rebase, checkout quite often but its not guranteed to happen every day 08:29 < ChillerDragon> ``git add . && git reset --hard`` is bae 08:31 < bridge> thank you ❤️ 08:33 < ChillerDragon> ayo lerato lately i have been using cherry-pick to replace stash 08:33 < ChillerDragon> my stash stacks became so bloated idk wat is wat xd 08:34 < ChillerDragon> so i commited on branches instead and then applied only the changes without the commit with ``git cherry-pick -n commithash`` 08:36 < ChillerDragon> https://zillyhuhn.com/cs/.1704267381.png 08:36 < ChillerDragon> but im clearly git nob 08:36 < ChillerDragon> i need to create 24 backup branches to feel safe xd 08:43 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1192010376167051304/GC2Qu8jWcAAi4YW.png?ex=65a7852c&is=6595102c&hm=2cc1b4bce70c71575ca9e4f34d02715a36d582008e3e0897540031c9370130c7& 09:41 < bridge> damn, looks like i really am a nerd 09:41 < bridge> aueigorifugu good morning 🌞 clan 09:49 < bridge> no u 10:08 < bridge> What's html doing there :justatest: 10:10 < ChillerDragon> HtMl iS nOt a ProGAMING lanGuage 10:11 < bridge> `git reset --hard HEAD` 10:11 < ChillerDragon> heinrich will totally respond with "--hard is dangerous" when he reads the backlog in 6 hours axaxaxax 10:11 < bridge> I always use hard 😏 10:12 < ChillerDragon> i like it hard too 10:12 < ChillerDragon> 😏 10:12 < ChillerDragon> hard and with force 10:12 < bridge> Oh, and also `git push origin master --force` 10:12 < ChillerDragon> yes 10:12 < bridge> hell yeah 10:12 < ChillerDragon> classic 10:13 < ChillerDragon> we are vibin 10:13 < ChillerDragon> also call it master instead of main because edgy 10:13 < bridge> Some of my repos have master 10:13 < bridge> Some main 10:13 < ChillerDragon> i set my github default to master xd 10:13 < ChillerDragon> and git too 10:14 < ChillerDragon> masterrace 10:14 < ChillerDragon> ok enough irc stalling 10:14 < bridge> Did u try something with passwords? 10:14 < ChillerDragon> no time for such trols 10:14 < ChillerDragon> i just starred 10:14 < ChillerDragon> i will watch the repo too 10:14 < ChillerDragon> in case u accidentally push secrets i want an email 10:15 < ChillerDragon> also have auto mirroring active 10:15 < ChillerDragon> in case you need an backup ;) 10:15 < bridge> milkeeycat@gmail.com, here u go 10:15 < ChillerDragon> leak 10:15 < ChillerDragon> enjoy spam from irc log scrapers 10:15 < ChillerDragon> axaxaxax 10:15 < ChillerDragon> ok fr now i close irc ffs 10:49 < bridge> Hi 10:49 < bridge> I also use git hard reset. Easiest, and yet I never fckd it up xd 10:51 < bridge> Chiller leak 10:51 < bridge> git stash list 11:01 < bridge> git commit -m "FUCK" 11:02 < bridge> `git commit -m "FUCK"` 11:23 < ChillerDragon> https://zillyhuhn.com/cs/.1704277415.png 11:23 < ChillerDragon> idk wat any of those are 11:24 < ChillerDragon> @blaiszephyr beauty of bash is that everything is a string so you can just do ``git commit -m FUCK`` 11:28 < ChillerDragon> omagawd snap deltas so complicated 11:32 < bridge> At least i think i understood the purpose of a alt snap now. There is the raw snap the client stores for future deltas. And then there is the alt snap which the client might modify due to invalid items. 11:40 < bridge> Ah, that's the alt for xd 11:49 < bridge> I think 11:50 < bridge> Zwelf probably knows everything about them ^^ 11:50 < bridge> He mentioned the first part once 11:53 < bridge> BTW does the first part even 100% work? 11:53 < bridge> I guess one constraint for it to work is that there is a single ordered data stream for snapshots and acks 11:53 < bridge> Right? 11:53 < bridge> Else the server and client cannot really proof in finite time if they actually received a snapshot for Delta 11:55 < bridge> Tw network code is a mystery to me.. I don't understand why it works xd 11:55 < bridge> But I sometimes think if it's broken. Which is only noticeable when u have huge lags 11:55 < bridge> But I sometimes wonder if it's broken. Which is only noticeable when u have huge lags 12:01 < bridge> @zwelf2 light us up: 12:01 < bridge> 12:01 < bridge> Assuming there is only one "raw"-snapshot that is used for delta comparision. 12:01 < bridge> 12:01 < bridge> How can the client and server be sure they talk about the same raw snapshot? 12:01 < bridge> 12:01 < bridge> - if raw-snap is #1 and snap #2 comes in, and the client tells the server it uses snap #2 now, if that packet is lost, server still assumes it must be #1. even if the snap-ack is resent. during that time how does the client proceed? 12:01 < bridge> 12:01 < bridge> it can't wait for the server to also ack the snap-ack. then we have a infinite cycle. 12:01 < bridge> there are so many questions in my head. if tw network code is so failsafe & well written gg 12:05 < bridge> i also wonder how much traffic is saved by the delta comparision. 12:05 < bridge> based on that i'd decide if it's even worth it. 12:05 < bridge> Also i wonder if e.g. other delta algos would suite tw better 12:05 < bridge> so many questions 12:06 < bridge> @zwelf2 light us up: 12:06 < bridge> 12:06 < bridge> Assuming there is only one "raw"-snapshot that is used for delta comparision. 12:06 < bridge> 12:06 < bridge> How can the client and server be sure they talk about the same raw snapshot? 12:06 < bridge> 12:06 < bridge> - if raw-snap is #1 and snap #2 comes in, and the client tells the server it uses snap #2 now, if that packet is lost, server still assumes it must be #1. even if the snap-ack is resent. during that time how does the client proceed? 12:06 < bridge> 12:06 < bridge> it can't wait for the server to also ack the snap-ack. then we have a infinite cycle.=> it must use multiple raw snaps to be correct (?) 12:06 < bridge> chillerdragon: when implement "edit" for irc 12:24 < bridge> @ryozuki please add proper static asserts to rust 🥹 12:26 < bridge> Macros have no kind of type information and checking e.g. a const of a trait can't be done at compile time directly. Currently the only way I know is to let usize overflow in a const expression xD 12:52 < bridge> xd 13:00 < bridge> `std::unique_ptr test = std::make_unique(10);`. deleting is handled automatically 13:03 < bridge> there are multiple raw snaps stored on the client 13:04 < bridge> @[matrix]: @heinrich5991 @murpi @deen @snailx3 @learath2 @robyt3 @pipou_tw Hello i would like to report the moderator ReD he has banned me ip (213.31.170.238) for until january 10th for ban evasion when im wasnt banned and i was 1 hour into a race with my fat friend Deamons 13:04 < bridge> don't mass-ping 13:04 < bridge> thanks 13:04 < bridge> what why are you deleting my support messages 13:05 < bridge> stop 13:05 < bridge> the range "last snapshot that was used as a delta" to "newest snapshot that I received" is stored by the client 13:05 < bridge> "newest snapshot that I received" is sent to the server 13:05 < bridge> im trying to report abusing staff who has banned me 13:05 < bridge> kinda the wrong channel 13:06 < bridge> but you also mass-pinged 13:06 < bridge> what channel do i speak to admins on 13:06 < bridge> developer has all the admins 13:06 < bridge> send me an email: heinrich5991@gmail.com 13:07 < bridge> Thanks for your ticket, we will fix this awful injustice 13:07 < bridge> i dont know how to send email 13:07 < bridge> i use this chat 13:07 < bridge> you could also use discord, but you might be banned here 13:07 < bridge> not sure 13:08 < bridge> yes the discord owner "jao" does not like me 13:08 < bridge> banned for complaining? :justatest: 13:08 < bridge> he ban me forever for no reaosn 13:08 < bridge> send me an email. thanks 13:08 < bridge> but my current problem is ReD has banned me for evasion until 10th january when i was 1 hour into map and i wasnt evading 13:09 < bridge> he has banned my ip 213.31.170.238 13:09 < bridge> kindly unban me and remove ReDs account 13:10 < bridge> OK then it makes sense 13:11 < bridge> (look for `PurgeUntil` in src/engine/client/client.cpp) 13:12 < bridge> And can u also rate the effective traffic saving? 13:14 < bridge> the traffic savings by using deltas instead of sending the whole snap? 13:15 < bridge> Yes 13:16 < bridge> It kinda must depent on the ping and other factors 13:17 < bridge> it means we can just dump rarely changing data into the snap 13:17 < bridge> e.g. all the player data and metadata about the game 13:17 < bridge> and it only consumes traffic at the start of the session and then only when it changes 13:18 < bridge> vs sending game/player data every tick 13:18 < bridge> Yes but how effective is it actually 13:19 < bridge> Player data could be moved out, so id not count that 13:19 < bridge> player data could be moved out, but then you're implementing your own diffing algorithm for player data 13:19 < bridge> which the snap already is 13:20 < bridge> OK but I'm interested how effective it is for the more often changing stuff 13:21 < bridge> I don't know. I'd need to modify the libtw2 code a little to sample it, I guess 13:22 < bridge> I was always surprised that standing still on man server still creates significant enough traffic 13:22 < bridge> Lan 13:34 < bridge> Windows not saving the command history between sessions is absolutely infuriating infuriating 13:34 < bridge> Windows not saving the command history between sessions is absolutely infuriating 13:36 < bridge> Windows's cmd line tool not saving the command history between sessions is absolutely infuriating 13:48 < bridge> https://www.vim.org/vim-9.1-released.php 13:48 < bridge> > This release is dedicated to Bram Moolenaar, Vims lead developer for more than 30 years, who passed away half a year ago. The Vim project wouldn't exist without his work!. 14:06 < bridge> https://github.com/kyren/piccolo 14:08 < bridge> Do u remember her? I told you about here before xdd 14:08 < bridge> She also maintainer of the hashlink create I use 14:09 < bridge> ah xd 14:10 < bridge> https://github.com/kyren/turbulence 14:10 < bridge> Lead Dev of starbound or smth 14:11 < bridge> epyc 14:12 < bridge> https://github.com/kyren/webrtc-unreliable 14:12 < bridge> sounds interesting for ddnet 14:18 < bridge> Go implement 14:19 < bridge> easier to implement after https://github.com/ddnet/ddnet/pull/6961 14:20 < bridge> because transports become abstracted with it 14:20 < bridge> my next plan after quic would be steam networking, so we could try getting ddos-proof servers with steam 14:25 < bridge> Sounds interesting 14:25 < bridge> You should join team zwelf & patiga 14:56 < bridge> Yeah got it, thanks :) 14:56 < bridge> Also read that smart pointers inside std containers (like vectors) are not good practice - and that's what I need :/ 14:56 < bridge> But works fine deleting in the base class! 14:56 < bridge> 14:56 < bridge> However, just wanted to ask another question: 14:56 < bridge> ```cpp 14:56 < bridge> struct Base { 14:56 < bridge> virtual int number() = 0; 14:56 < bridge> virtual ~Base() = default; 14:56 < bridge> }; 14:56 < bridge> 14:56 < bridge> struct Derived : public Base { 14:56 < bridge> int number() override { 14:56 < bridge> return 10; 14:56 < bridge> } 14:56 < bridge> }; 14:56 < bridge> 14:56 < bridge> int main() { 14:56 < bridge> vector x; 14:56 < bridge> x.emplace_back(new Derived); 14:56 < bridge> 14:56 < bridge> cout << x.size() << " -> " << x[0] << "(" << x[0]->number() << ")" << endl; 14:56 < bridge> 14:56 < bridge> for (const auto &item: x) { 14:56 < bridge> delete item; 14:56 < bridge> } 14:56 < bridge> } 14:56 < bridge> ``` 14:57 < bridge> 14:57 < bridge> There are more structs like `Derived` which inherit `Base`. What would be the best way to clone `vector x`? I need the `Base *` values inside also to be deep copied. Played around with copy/constructors but that didn't work out since the generic is a pointer I think. 14:57 < bridge> Do I need to implement a `virtual clone()` method manually in the base class and overwrite it in the subs, then call them in a loop to fill a new vector? 14:57 < bridge> where did you read that smart pointers inside std containers aren't good practice? 14:57 < bridge> maybe its more that u should simply give ownership to the vec? 14:57 < bridge> but idk 14:57 < bridge> I read they could leak when `emplace_back` fails for example 14:58 < bridge> in that case, your example would also leak, I think 14:58 < bridge> https://stackoverflow.com/questions/3283778/why-can-i-not-push-back-a-unique-ptr-into-a-vector 14:58 < bridge> Comments in accepted answer 14:58 < bridge> the comments in the accepted answer tell you to use `std::unique_ptr` over raw pointers, I think 14:59 < bridge> because raw pointers might leak 14:59 < bridge> wait I had another source 15:00 < bridge> bonus question: no, C++ will stay relevant for ddnet. barely anything is rust yet 15:01 < bridge> yes, `virtual std::unique_ptr clone() const` is the only way I can think of 15:01 < bridge> I want to know why you need polymorphism though @stnwtr 15:01 < bridge> Hm, I can’t imagine unique_ptrs leaking on throw. If emplace fails it probably just results in the old unique_ptr keeping the ownership and if that happens to be a temporary it would then get destroyed properly 15:02 < bridge> I am searching for the other posts I read ... maybe I misunderstood because coming from Java, C++ is super hard haha 15:02 < bridge> dont try to do java in c++ 15:02 < bridge> i am trying 15:03 < bridge> well, I had 2 years rust in between ... and I am "unlearning" the java way 15:03 < bridge> a common problem is not adapting to the language idioms and trying to force one from another lang 15:03 < bridge> its a common problem in rust learners too :d 15:04 < bridge> It’s a massive problem in Rust 15:04 < bridge> Because my teacher wants me to do so 15:04 < bridge> Having a base class building and some subclasses like "SolarPowerPlant", "GiantHouse" and so on 15:04 < bridge> which consists of Materials (baseclass) like Wood, Metal, ... 15:04 < bridge> 15:04 < bridge> And I need vectors (or arrays) of the base* 15:04 < bridge> but well c++ kinda has classes like java 15:04 < bridge> but with pointers xD 15:05 < bridge> Because Rust is just absurd compared to other languages. No other language really has such strictly enforced ownership hierarchies and lifetime restrictions 15:05 < bridge> ah typical teacher teacing OOP 15:05 < bridge> blegh 15:05 < bridge> why ppl teach OOP and not useful stuff 15:06 < bridge> How would you do this without oop? 15:06 < bridge> Enums would be nice imo, but well, school 15:06 < bridge> @stnwtr btw did ur teacher ever say the phrase "composition over inheritance"? 15:06 < bridge> Not yet 15:06 < bridge> And no 15:06 < bridge> ask him 15:06 < bridge> about it 15:07 < bridge> Almost forgot to thank you :) 15:07 < bridge> Well, oop is useful when you have shared behaviour. What sort of behaviour would a `Wood` class even have? 15:07 < bridge> I should use `unique_ptr` everywhere I can? 15:07 < bridge> > To favor composition over inheritance is a design principle that gives the design higher flexibility. It is more natural to build business-domain classes out of various components than trying to find commonality between them and creating a family tree. For example, an accelerator pedal and a steering wheel share very few common traits, yet both are vital components in a car. What they can do and how they can be used to benefit the car are easily 15:07 < bridge> > 15:08 < bridge> > Initial design is simplified by identifying system object behaviors in separate interfaces instead of creating a hierarchical relationship to distribute behaviors among business-domain classes via inheritance. This approach more easily accommodates future requirements changes that would otherwise require a complete restructuring of business-domain classes in the inheritance model. Additionally, it avoids problems often associated with relatively 15:08 < bridge> Nothing, thats so stupid. Just a single variable, the price. 15:08 < bridge> In part 1 of the exercise we used an enum tho. 15:08 < bridge> Now we need a class and are not even allowed to use a static variable for the price 15:08 < bridge> School is trolling us, right? 15:08 < bridge> Smart pointers are indeed the way in modern C++. You want to think in terms of ownerships 15:09 < bridge> gimme that sweet static void abstract MainBuilderProducer 15:09 < bridge> gimme that sweet static void abstract MainBuilderProducerFactory 15:09 < bridge> in java its where u can manage to put 100 prefixes before the function name 15:09 < bridge> So I should avoid raw pointers as good as possible 15:09 < bridge> public final static const void urmom 15:10 < bridge> (sry i forgot a lot of java kek) 15:10 < bridge> good! 15:10 < bridge> u know 15:10 < bridge> the next java is called dart 15:10 < bridge> im using it for flutter apps 15:10 < bridge> But then `Wood` being the baseclass of `WoodHouse` wouldn’t really make much sense, why would you ever want `WoodHouse` to inherit the price of wood? So that’s kinda why this is not the greatest use of OOP 15:10 < bridge> https://dart.dev/ 15:10 < bridge> i though its c# :justatest: 15:10 < bridge> Its not, sorry 15:11 < bridge> no 15:11 < bridge> dart is like go for making apps 15:11 < bridge> stupidly simple and focused on being good at making apps 15:11 < bridge> Baseclasses are `Material` (for Wood, Metal, Plastic, ...) and `Building` for a few more where Building has a `vector` 15:11 < bridge> yes 15:11 < bridge> Baseclasses are `Material` (for Wood, Metal, Plastic, ...) and `Building` for a few more (like SolarPowerPlant) where Building has a `vector` 15:11 < bridge> only if u need a pointer 15:11 < bridge> shouldnt u favor references 15:12 < bridge> if you need ownership 15:12 < bridge> yes 15:12 < bridge> i would even favor std::option> 15:12 < bridge> but thats me being rusty 15:12 < bridge> the OOP examples by your teacher seem sketchy. just follow the class and forget the OOP practices later 15:12 < bridge> altho idk if option is zero cost 15:12 < bridge> I used option, expected and tuple for a few reasonable things, my teacher called it overengineered :( 15:12 < bridge> I see, well that sounds saner, anyway it’s just a conceptual exercise 15:13 < bridge> it's not zero-cost 15:13 < bridge> Tell him you find his solution underengineered 15:13 < bridge> not even for a reference? 15:13 < bridge> rip c++ 15:13 < bridge> is `unique_ptr` zero-cost or are there some bytes overhead? 15:13 < bridge> for shared_ptr there is a counter 15:13 < bridge> but unique_ptr? 15:13 < bridge> it's zero-cost in terms of space 15:14 < bridge> Alllmost zero runtime cost too 15:14 < bridge> ye 15:14 < bridge> and a huge +++ in terms of memory management 15:14 < bridge> https://www.gamingonlinux.com/2024/01/linux-hits-nearly-4-desktop-user-share-on-statcounter/ 15:14 < bridge> YEAR OF LINUX DESKTOP!!! 15:15 < bridge> unless you are coding for a very very very weak mcu the runtime cost won’t ever be something you think about 15:15 < bridge> tldr; 15:15 < bridge> a lot of things i am learning in software engineering classes are outdated and bad practice? 15:15 < bridge> probably 15:16 < bridge> unclear. software engineering is probably never outdated. coding classes more so 15:16 < bridge> unclear. software engineering is probably never outdated. classes where you learn to code more so 15:16 < bridge> I found courses range from wildly outdated to slightly outdated and the practices range from very bad to meh 15:17 < bridge> We'll do server side web services using Java Server Pages next term .... 15:17 < bridge> Oh yeah, the actual meta software engineering part is never really out of date 15:17 < bridge> Now that’s one course I’d just delete from my brain after passing 15:18 < bridge> done already 15:18 < bridge> r u doing uni? 15:18 < bridge> sadly, yes 15:18 < bridge> Unless you aspire to work at airline IT. They love their java legacy stuff 15:18 < bridge> prof about 600 years old 15:18 < bridge> in my uni mfs teach nothing :justatest: 15:18 < bridge> in my uni mfs teach us nothing :justatest: 15:18 < bridge> i did no uni :justatest: 15:19 < bridge> good! 15:19 < bridge> if i wasnt in uni id probably go on the war :justatest: 15:19 < bridge> The most out of date course I had was computer architecture, but I do kinda understand that one. I doubt you can teach anything about any modern architecture to any bachelor student 15:20 < bridge> I liked architecture a lot. Learning about MOSFET, MUX, Latches, Boolean Algebra and all that stuff. That was fun 15:20 < bridge> is it just me or it looks like cum(new wallpapers btw) :greenthing: 15:20 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1192110162622812250/image.png?ex=65a7e21b&is=65956d1b&hm=794bd2331701ca05b67de25c47f188371351043c49df615acd01b8e88ec3ddd4& 15:20 < bridge> btw von neumann is a genius among geniuses 15:20 < bridge> the guy did everything 15:20 < bridge> And the RISC-V, but term ended too soon 15:20 < bridge> I’m a sucker for old computer architecture. Back then everything made so much sense 15:20 < bridge> > ohn von Neumann (/vɒn ˈnɔɪmən/ von NOY-mən; Hungarian: Neumann János Lajos [ˈnɒjmɒn ˈjaːnoʃ ˈlɒjoʃ]; December 28, 1903 – February 8, 1957) was a Hungarian-American mathematician, physicist, computer scientist, engineer and polymath. He had perhaps the widest coverage of any mathematician of his time,[9] integrating pure and applied sciences and making major contributions to many fields, including mathematics, physics, economics, 15:21 < bridge> > John von Neumann (/vɒn ˈnɔɪmən/ von NOY-mən; Hungarian: Neumann János Lajos [ˈnɒjmɒn ˈjaːnoʃ ˈlɒjoʃ]; December 28, 1903 – February 8, 1957) was a Hungarian-American mathematician, physicist, computer scientist, engineer and polymath. He had perhaps the widest coverage of any mathematician of his time,[9] integrating pure and applied sciences and making major contributions to many fields, including mathematics, physics, economics 15:21 < bridge> Pipelines were nice and short, caches MSI 15:21 < bridge> 3rd term out of 7 now and I want to give up because its either boring or too much 15:22 < bridge> @stnwtr i recommend this book, its post uni stuff tho https://en.algorithmica.org/hpc/ 15:22 < bridge> but its rly rly interesting 15:22 < bridge> Yeah, truly a genius 15:24 < bridge> Also got a few recommendations for books about building compilers? Need some for a presentation / paper I need to write 15:24 < bridge> Is it 4 terms a year? 15:24 < bridge> thanks 15:24 < bridge> 3 1/2 years 15:24 < bridge> 1 year = 2 terms 15:24 < bridge> The classic recommendation for compiler theory is the dragon book 15:24 < bridge> the book if for making truly perfomant code tho 15:24 < bridge> Yep, already got the dragon book 15:24 < bridge> But I don’t think it’s very beginner friendly 15:24 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1192111340593086514/dennard.png?ex=65a7e333&is=65956e33&hm=f2f262d9b4dd46c7175dfaac9248f810ce5390c5726705255eeb852c4835f6c3& 15:25 < bridge> I am not *that much* a beginner in compilers I think. Did the craftinginterpreters thing and some stuff. And I only need to write about 15 pages, that's done fast I hope 15:26 < bridge> Nice 15:26 < bridge> Ah, I was just about to recommend crafting interpreters 15:26 < bridge> https://os.phil-opp.com/ 15:27 < bridge> https://llvm.org/docs/tutorial/ :troller: 15:28 < bridge> troller? D: 15:28 < bridge> You don't recommend the llvm tutorial? 15:28 < bridge> oh i do 15:28 < bridge> but its not simple xd 15:28 < bridge> probably too high for me then 15:29 < bridge> A friend of mine recommended me “Appel. Modern Compiler implementation in C” but I never got around to reading it so not sure if good 15:29 < bridge> Might be worth looking into 15:29 < bridge> oh i didnt read this 15:29 < bridge> well i currently work making compilers :d 15:29 < bridge> Hope libgen has this books :hAA: 15:29 < bridge> but did u do a compiler or a interpreter 15:29 < bridge> It should, it’s a classic recommendation 15:30 < bridge> It was an interpreter for the so called "lox" language 15:30 < bridge> Lexing, Parsing and stuff is similar in building compilers/interpreters i think 15:31 < bridge> that part is similar in both yeah 15:31 < bridge> learn llvm 15:31 < bridge> then learn MLIR 15:31 < bridge> :BASED: 15:32 < bridge> :BASEDDEPT: 15:32 < bridge> ☕ 15:32 < bridge> @stnwtr did u do type analysis 15:32 < bridge> or is it dynamic 15:32 < bridge> mlir? something like llvm-ir? 15:32 < bridge> mlir is a new thing 15:33 < bridge> above llvm ir 15:33 < bridge> sounds nice 15:33 < bridge> my blog https://edgarluque.com/blog/mlir-with-rust/ 15:33 < bridge> ah nice, thanks 15:33 < bridge> dynamic as far as i remember? hmmm 15:33 < bridge> im not rly good blogger but hope its ok :d 15:33 < bridge> its fine :D 15:37 < bridge> And the reason I got into computer science is teeworlds. I started playing 7 or 8 years ago and since then always wanted to create a custom gamemode xD 15:37 < bridge> Now I am starting learning c++, woah xD 15:37 < bridge> Progress :stonks: 15:38 < bridge> kinda same 15:38 < bridge>