00:05 < bridge> they should open a ticket, but the PR robyt link was only integrated into yesterdays (mondays) nightly 00:33 < bridge> Ono crashes 00:33 < bridge> Where are the backtraces! 00:34 < bridge> @learath2: weren’t you working on automated crash reports? 03:42 < bridge> Today i joined a server and developed my way through LearnToPlay 09:42 < bridge> The hook collision line bug count is now officially at 8, making them more broken then the old speedtiles 💀 09:56 < bridge> Learath don't answer, he's trolling 09:56 < bridge> obv bait, no triangulation 09:56 < bridge> I’m working on 80 different things including moving across europe and doing a full time masters program. Soooo one day we might get automated crash reports 10:05 < ws-client> **** @Assa do we have an ddnet issue for https://github.com/teeworlds/teeworlds/issues/2977 ? 10:18 < bridge> I was writing one rn, you were faster 10:19 < ws-client> **** blazingly issuer 10:19 < ws-client> **** maybe you should hide your DDNET backtrace from the TEEWORLDS issue xd 10:19 < bridge> I already know where this is crashing btw 10:19 < ws-client> **** me too 10:19 < ws-client> **** its in the trace xd 10:19 < bridge> OnMapLoad it reads nullptr quads 10:19 < ws-client> **** `pSrc=0x0` 10:19 < ws-client> **** `pSrc[SrcIndex].m_Skip` 10:20 < bridge> it clearly says DDNet.exe :justatest: 10:20 < ws-client> **** yes 10:20 < ws-client> **** still troll 10:21 < bridge> I am troll, didn't even notice that I was writing in teeworlds issues xD 10:21 < ws-client> **** no you didnt ofc 10:22 < ws-client> **** but now that you know i think its not nice leaving other projects traces there 10:22 < bridge> also I think I found out how to crash ddnet as well with a bug since 18.4 10:22 < bridge> do you know how I see the screen values? 10:22 < ws-client> **** imo it is rude to report bugs from forks upstream 10:23 < ws-client> **** screen values? 10:23 < bridge> Yes, if you manage to line up ScreenX0 perfectly, the client should crash in rendering xD 10:23 < bridge> https://github.com/ddnet/ddnet/issues/7922#issuecomment-3322893594 10:24 < ws-client> **** u mean if they are in some dbg hud? 10:24 < ws-client> **** idk 10:24 < ws-client> **** i dont do gfx 10:45 < bridge> time to crash 18.4 11:01 < bridge> Crashed ✅ I can now crash 18.0 and know, that the bug is not present in the new version. I can also create a pull request removing this debug assertion and doing a better validity check 11:01 < bridge> That’s illegal, don’t be crashing clients 11:01 < bridge> I am crashing clients in order to crash less clients 😠 11:58 < bridge> Not crashing != not having a bug tho. 11:58 < bridge> 11:58 < bridge> Hiding bugs usually just leads to more bugs 12:00 < bridge> Yes, but I understand how this happened and tried to explain it. Ofc you can continue hunting ghosts 12:00 < bridge> I see no explaination at all 12:01 < bridge> > because X0 was able to be bigger then XR in principle 12:01 < bridge> Click on the issue, there is a bigger explanation 12:01 < bridge> That is not an explaination of the bug 12:01 < bridge> That is the explaination of the crash 12:02 < bridge> Guess I'll spin up the debugger and reproduce again 12:05 < bridge> The bug is, that X0 >= X1 and the layer is still beeing drawn, or do you want to know how X0 and X1 are beeing set? 12:06 < bridge> @jupeyy_keks btw the reported crash was on normal zoom and I can crash it on normal zoom 12:23 < bridge> OK then tell what leads to this value 12:23 < bridge> A bug in our projection? 12:24 < bridge> Especially on normal zoom 12:26 < bridge> The camera goes out of bounds? 12:26 < bridge> That is the issue right? 12:26 < bridge> Like out of map bounds or smth 12:27 < bridge> The crash itself is here `unsigned int NumVertices = ((Visuals.m_pTilesOfLayer[y * pTileLayer->m_Width + XR].IndexBufferByteOffset() - Visuals.m_pTilesOfLayer[y * pTileLayer->m_Width + X0].IndexBufferByteOffset()) / sizeof(unsigned int)) + (Visuals.m_pTilesOfLayer[y * pTileLayer->m_Width + XR].DoDraw() ? 6lu : 0lu); 12:27 < bridge> ` 12:27 < bridge> 12:27 < bridge> If X0 > XR, then NumVertices underflows 12:27 < bridge> link to old code: https://github.com/ddnet/ddnet/blob/36825f51715569c77598a206dbaf0b3212982184/src/game/client/components/maplayers.cpp#L905 12:28 < bridge> XR is X1 -1 12:29 < bridge> I find it really hard to reproduce right now, the debugger lags on the map border there 12:29 < bridge> it's a bug on the map border , because this clips X0 and X1 value 12:31 < bridge> it would help if I'd link the right old code from 18.0 https://github.com/ddnet/ddnet/blob/6888f75fac87096e153f61f2a5db5903450c5927/src/game/client/components/maplayers.cpp#L1004 12:31 < bridge> here is the bug: https://github.com/ddnet/ddnet/blob/6888f75fac87096e153f61f2a5db5903450c5927/src/game/client/components/maplayers.cpp#L1017 12:31 < bridge> Should have been X1 >= m_Width 12:31 < bridge> 12:31 < bridge> Any my PR didn't even fix it, it was fixed between 18.0 and 18.4 12:33 < bridge> but your pr moves random stuff 12:33 < bridge> "random" 12:33 < bridge> Ifc if you magically fixed it. Just explain wtf the bug is 12:34 < bridge> How hard can it be to reproduce it 12:34 < bridge> you need to be at the map border and produce exact x1 and x0 values 12:34 < bridge> how hard can this be? 12:35 < bridge> Yes that would be the correct answer 12:35 < bridge> It was fixed between 18.0.3 and 18.1, I want to pinpoint the commit now 12:38 < bridge> ~It was fixed between 18.1 and 18.2, I want to pinpoint the commit now~ 12:38 < bridge> ~~It was fixed between 18.1 and 18.2, I want to pinpoint the commit now~~ 12:40 < bridge> 18.3 to 18.4 12:47 < bridge> @essigautomat so why not simply check if x0 >= x1? 12:48 < bridge> in the continue 12:48 < bridge> the loop has nothing to do with x0 and x1, it iterates over rows 12:49 < bridge> and it skipped the row if x0 > x1 previously 12:49 < bridge> dunno what you refactored 12:50 < bridge> but the > simply gets >= 12:50 < bridge> it skips every row if X0 > x1 previously, you can move this out of the loop 12:50 < bridge> Anyway, without the assert it sounds to me that you could make zoom = -1 and crash client 12:51 < bridge> and then there is the X0 <= XR check in my PR, which prevents this 12:51 < bridge> Good, but you ignore the case 12:51 < bridge> previously if x0 really was bigger it crashed 12:51 < bridge> wow if you haven't seen https://discord.com/channels/252358080522747904/295908390956433410/1419937088232620094 yet, its worth it :heartw: 12:51 < bridge> my favorite ddnet animation as of yet by far 12:51 < bridge> that is not the same 12:52 < bridge> I don't think a crash is more desirable, I can add the debug message for the else case 12:53 < bridge> i just told you a case 12:53 < bridge> which could fuck up some stuff, if you dont prove otherwise 12:53 < bridge> anyway, @essigautomat now neutral again. this bug only happens in 18.0-18.4? 12:53 < bridge> not before not after? 12:54 < bridge> I only managed to crash it in 18.0, the issue was only about 18.0, as far as I can tell it was there until 18.3 12:59 < bridge> do the screen values swap with zoom = -1? 13:00 < bridge> because the PR i provided shouldn't crash, it should just render nothing 13:00 < bridge> because the PR i provided shouldn't crash, it should just render nothing, if this is the case 13:04 < bridge> This one? https://m.bilibili.com/video/BV1oHp2zPEPt 13:06 < bridge> 18.0 code: 13:06 < bridge> ``` 13:06 < bridge> if(X0 > X1) 13:06 < bridge> continue; 13:06 < bridge> int XR = X1 - 1; 13:06 < bridge> ``` 13:06 < bridge> 13:06 < bridge> `If X0 == X1, XR < X0` which underflows NumVertices 13:07 < bridge> 13:07 < bridge> How can this happen? At this point idk 13:07 < bridge> Solution: Check if XR >= X0, what my PR does 13:07 < bridge> yea, sry chiller ^^ 13:07 < bridge> yea it’s cute :3 13:30 < bridge> But before 18 it did not happen, right? So there was a check previously preventing it. 13:30 < bridge> 13:30 < bridge> But anyway, if you tested with negative value fine, just write it in your PR 13:32 < ws-client> **** wtf is `git replace` ?? and why did it not create conflicts???? 13:33 < ws-client> **** i just swapped out a old commit in the history not affecting any shas on top of it. Not creating conflicts even if the other commits were based on the now outdated changes?????? 13:33 < ws-client> **** it gets even weirder 13:33 < ws-client> **** the code still compiles 13:33 < ws-client> **** ????? 13:34 < MMC> https://desuarchive.org/qa/thread/3570832/#3571069 gem 13:35 < bridge> git replace me 13:35 < bridge> Do i want to click a random link by a random person? 13:35 < ws-client> **** looks like 4chan 13:35 < bridge> chiller you go first 13:35 < ws-client> **** pretty boring 13:35 < ws-client> **** and non tw related 13:35 < bridge> delete 13:36 < ws-client> **** git replace is giga troll ngl 13:36 < ws-client> **** it creates just a fake history 13:36 < ws-client> **** not doing any actual changes to the code 13:36 < ws-client> **** well that also sounds op to bypass reviews xd 13:36 < ws-client> **** but not useful for me who actually needs to replace 13:37 < bridge> wtf 13:37 < bridge> git replace is super epyc 13:37 < ws-client> **** u know it? 13:37 < bridge> didn't know it's so cool 13:37 < bridge> No, I just read about what it can do 13:37 < ws-client> **** u understand it? 13:37 < ws-client> **** for me it did random garbage 13:37 < bridge> Ok but export & importing kinda sucks 13:37 < bridge> Would be cool if it just works in a repo 13:38 < bridge> You can fix old commits without changing history kinda 13:38 < bridge> But only to a certain limit 13:38 < bridge> I don't get why this should crash with a negative zoom value, except if you think that this should invert the screen. If you set `zoom -29` you zoom out of the map. Can you specify what you mean? 13:38 < bridge> I expect that 13:38 < ws-client> **** @Jupstar ✪ but i want to change history.. i guess i need rebase then 13:38 < ws-client> **** i wonder if i can cherry pick during an interactive rebase xd 13:39 < ws-client> **** or just swap out the sha in the rebase editor 13:39 < ws-client> **** that might actually work 13:39 < bridge> You want to truely change history? 13:39 < ws-client> **** yes 13:39 < ws-client> **** its pr 13:39 < bridge> `git rebease --interactive HEAD~4` 13:39 < bridge> Best tool 13:39 < bridge> of git 13:39 < ws-client> **** i have two branches 13:39 < ws-client> **** branch a and b 13:39 < ws-client> **** b is based on a 13:39 < ws-client> **** a got force pushed 13:39 < ws-client> **** b has commits on top 13:40 < ws-client> **** now i need to get b updated with the new commits from a 13:40 < bridge> git rebase interactive you can just change commit hashes 13:40 < ws-client> **** interesting 13:40 < ws-client> **** i will try that 13:40 < ws-client> **** fakin git replace sidequest xd 13:40 < bridge> gl 13:40 < ws-client> **** i wonder if i need to delete .git because i dont understand git replace xd 13:40 < bridge> Honestly just start from scratch 13:41 < bridge> git init --bare 13:41 < ws-client> **** fr 13:41 < bridge> `git checkout b`, `git rebase -i a`, then discard commits from a in the interactive rebase log to keep only original commits from b 13:41 < bridge> sick 13:41 < ws-client> **** o 13:41 < bridge> Chiller just use llm no joke 13:41 < ws-client> **** i dont trust it 13:42 < bridge> I bet these questions got answered 200000 times 13:42 < ws-client> **** last time i asked llm with git question it couldnt help me 13:42 < bridge> And they are accurate enough 13:42 < ws-client> **** i lost my merge head 13:42 < bridge> lmao 13:42 < ws-client> **** i still dont know how to get it back 13:42 < ws-client> **** its so easy to lose it 13:42 < ws-client> **** and llm rly didnt get me 13:42 < bridge> I dunno, once it's pushed to github it will stay 13:42 < bridge> You can restore anything 13:42 < ws-client> **** no 13:42 < ws-client> **** imagine dis 13:42 < bridge> Even if you force push 13:42 < ws-client> **** u have conflicts 13:43 < bridge> So just randomly push your stuff for backup 13:43 < ws-client> **** then you solve them 13:43 < ws-client> **** then you do `git merge --continue` ye? 13:43 < ws-client> **** or rebase --continue 13:43 < bridge> I never merge 13:43 < bridge> I always rebase 13:43 < ws-client> **** same thing 13:43 < ws-client> **** but u wanna do rebase --continue rite? 13:43 < bridge> yes 13:43 < ws-client> **** now 13:43 < bridge> At some point 13:43 < ws-client> **** the big danger is 13:43 < ws-client> **** that it is quite easy to break `git rebase --continue` 13:44 < ws-client> **** if u ran the wrong command 13:44 < bridge> in worst case you do git rebase --abort 13:44 < bridge> Dunno wtf you done 13:44 < ws-client> **** you cant 13:44 < ws-client> **** you cant abort 13:44 < ws-client> **** or continue 13:44 < ws-client> **** if you lost the merge head thing 13:44 < bridge> wtf? XD 13:44 < bridge> Show me your prompt 13:44 < ws-client> **** it says THERE IS NO REBASE HAPPENING 13:44 < bridge> No way it suggested to fuck up that badly 13:44 < ws-client> **** happens quite often 13:44 < ws-client> **** for example if you do `git rebase --committer-date-is-author-date` 13:45 < bridge> What are you using, Cellegen AI? 13:45 < ws-client> **** i dont use ai 13:45 < ws-client> **** i can hallucinate on my own 13:46 < bridge> if you commited stuff you can restore, they have a hash somewhere 13:46 < bridge> How hard can you fuck up 13:46 < ws-client> **** thing is 13:46 < bridge> Only stupid you can do is force reset without ever having commited 13:46 < ws-client> **** some commands like --abort or --continue only work while rebasing or merging 13:46 < bridge> So always type git stash before resetting 13:46 < ws-client> **** and that depends on some file being there 13:46 < ws-client> **** and that file is easy to lose 13:47 < bridge> No matter if you use the stash ever again 13:47 < ws-client> **** then i am done solving all conflicts 13:47 < ws-client> **** and cant finish 13:47 < ws-client> **** i could only create a new commit 13:47 < ws-client> **** but not finish the rebase or merge 13:47 < bridge> Ok 13:48 < ws-client> **** maybe i suck at pomting idk 13:48 < bridge> Nah 13:48 < bridge> I simply don't believe you 13:48 < bridge> Ez 13:48 < ws-client> **** i dont think its that of an edge case given how often it happens to me 13:48 < ws-client> **** omg 13:48 < ws-client> **** i will show u sec 13:48 < ws-client> **** need to create conflict 13:48 < bridge> lies 13:48 < bridge> did noone ever notice that? 13:48 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420014040033267712/492792492-29ddc46c-f4c0-42d9-9133-29e6982c61cd.png?ex=68d3daa1&is=68d28921&hm=30d8026a3c133d51ccc7d20a92f384731df541c56a2c6c53da5aa34fdbc3ef09& 13:49 < bridge> Wasn't that on purpose 13:49 < bridge> Bcs normal entities have "cfrom" 13:49 < bridge> text 13:49 < bridge> or smth similar 13:49 < bridge> whats cfrom 13:49 < bridge> Cant boot up the game rn 13:49 < bridge> Just update your screenshots with default entities in the PR 13:49 < bridge> That is what matters more than your custom ones 13:50 < bridge> these are the default comfort 13:50 < bridge> Bro 13:50 < bridge> cl_entities "default" 13:50 < ws-client> **** ok i struggle to lose merge head aaa 13:50 < ws-client> **** omg wat i did all the last times 13:51 < ws-client> **** A 13:51 < ws-client> **** GOT IT 13:51 < ws-client> **** STASH 13:51 < bridge> I'll read in an hour, chiller. 13:51 < bridge> 13:51 < ws-client> **** adoihaihod 13:51 < bridge> Gl breaking your stuff 13:51 < ws-client> **** dont @Jupstar ✪ 13:51 < ws-client> **** i have it 13:51 < ws-client> **** `git merge conflictbranch` 13:51 < ws-client> **** solve all conflicts 13:51 < ws-client> **** `git add . && git stash && git stash pop` 13:51 < ws-client> **** `git merge --continue` 13:52 < ws-client> **** `fatal: There is no merge in progress (MERGE_HEAD missing).` 13:53 < bridge> Am Arsch 13:53 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420015306910339173/image.png?ex=68d3dbcf&is=68d28a4f&hm=e55190672feceab6d17eb1ba03f51dff9537a1f7b08daaec77782bc9c44067c3& 13:53 < ws-client> **** fr 13:53 < bridge> not that bad 13:54 < bridge> client's default is comfort for long time now 13:54 < bridge> Idc anyway, just update pr for whoever cares 13:54 < bridge> Wtf really? 13:55 < bridge> yea 13:55 < bridge> aaaaaaaaa this assertion issue makes me crazy 13:55 < bridge> why did 18.0 crash for me this one time exactly there? 13:56 < bridge> why can't I reproduce it 13:56 < bridge> all bs 13:56 < bridge> cl_assets_entities default value is default but im pretty sure it was comfort for me when i installed game 14:00 < bridge> or maybe is not, on phone i have default 14:00 < bridge> that's in general an issue if you have text written over something and text in an image itself 14:00 < bridge> wtf 14:01 < bridge> should be comfort though, since its shipped with the client and looks way more modern 14:01 < bridge> like you will unbreak sbs entities and break others instead 14:01 < bridge> yea thats why i dont like the old entities, text on texture is bad 14:02 < Jfif> Hi 14:03 < bridge> hi 14:03 < Jfif> Test 14:06 < bridge> @jupeyy_keks can we merge https://github.com/ddnet/ddnet/pull/10936 ? I am pretty sure you and I can count to 1 14:08 < bridge> And together, we're 2 ❤️ :owo: 14:15 < bridge> @kebscs since you tested the hook collision line, did you test different tuning values? 14:17 < bridge> because of tune hook_length 9999 and hook_fire_speed 1 :omo: 14:17 < bridge> Ok 14:18 < bridge> ah right 14:18 < bridge> i found 1 now with tune 14:18 < bridge> And i thought chiller is our new merge bot 14:18 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420021468649099294/image.png?ex=68d3e18c&is=68d2900c&hm=5b62b6e12f47e476e8f96b17f7d74badf8f70f8772653ba604070357d90e7302& 14:18 < bridge> this should extend to cursor, not player hit pos 14:18 < bridge> chiller is skillchecked for anything related to gfx 🙁 14:18 < bridge> tune hook_fire_speed 1; tune hook_length 600 14:19 < bridge> ah nvm 14:19 < bridge> idk whats going on 14:19 < bridge> wrong actually 💀 this is how it should look like 14:19 < bridge> yeah 14:20 < bridge> well idk why this works like that 14:20 < bridge> but ig works fine 14:21 < bridge> length with no tunes or basic tunes is good 14:21 < bridge> and with tele outs 14:21 < bridge> so better than before 14:21 < bridge> explanation: 14:21 < bridge> The hook takes longer than 5 seconds to hit the player, so the simulation ends and we hit the case I added for unknown hook states. Unknown hook is blue and for exceeding simulations the length is only an approximation 14:22 < bridge> same happens with 2 teleouts, hook will color blue 14:22 < bridge> yea but idk why it doesnt launches towards cursor 14:22 < bridge> jank physics 14:22 < bridge> exactly 14:22 < bridge> xD 14:22 < bridge> this is also why I give up on small hook lengths and hook_fire_speed <= 0 14:22 < bridge> I don't even want to know what a negative hook fire speed does 14:23 < bridge> wait 1 more thing 14:23 < bridge> line isnt attached to player when you hold hammer 14:23 < bridge> do same thing as with tele 14:23 < bridge> same thing as with the teleouts, do you wish the line to be visually connected (even if it doesn't color and doesn't hook there) 14:24 < bridge> alright 14:28 < bridge> @kebscs you can't do that, that would break the jank physics, Id need to add an extra line element for it 14:29 < bridge> add the short line to vLineSegments 14:34 < bridge> this is actually really useful, you can predict where a hook goes into a wall: 14:34 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420025458438963231/screenshot_2025-09-23_14-33-38.png?ex=68d3e544&is=68d293c4&hm=51a09b57c4555a8306de954f24f8abf77e26b9cda35988b007178dca37b8f7de& 14:36 < bridge> jank physics will have a kink of course 14:36 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1420026154869456906/screenshot_2025-09-23_14-35-53.png?ex=68d3e5ea&is=68d2946a&hm=ffeb2a56dc984ec3deb42ab5fc2078137d98a0a70447c2241a04279da86a04cf& 14:46 < bridge> ok now its perfect 14:47 < bridge> it still suffers from floating point precision issues, but compared to before, I would call it a day 14:47 < bridge> What is this even 😄 14:47 < bridge> the hook collision line rework I am working on 14:48 < bridge> and before you say the second image is wrong, it isn't 14:49 < bridge> Is that really how far the hook makes it before getting to the hook from tele? 14:49 < bridge> no, the second image is from a simulation time exceeded state with jank physics 14:50 < bridge> and the hook actually follows that line 14:51 < bridge> I don't get how we can have an indeterminate state for a finite process 14:51 < bridge> What does it mean for the "simulation time" to be exceeded? 14:51 < bridge> with the right physics, I can make your hook take a million ticks in the future and bring your PC into it's knees 14:52 < bridge> Do our tune bounds even allow that? Wouldn't using the hook in such a state also do that? 14:53 < bridge> the hook line simulates the actual hook up to 5 seconds in the future with my PR. Currently it's unlimited and I can bring my PC to lags with it 14:53 < bridge> no, the hook is calculated tick by tick, the hookline is calculated all in one frame 14:54 < bridge> the tune bounds allow that 14:54 < bridge> Ok, I see the issue. Why not simulate the hook line tick by tick too? 14:54 < bridge> Maybe do it 10 ticks at a time to simulate the "instant" effect 14:54 < bridge> this would need a full restrusture but would be possible 14:55 < bridge> also you would need to hold still in order to get this effect, what if the player moves in the next tick? 14:55 < bridge> Guess it's not worth it 14:55 < bridge> i like current code, maps should use these tunes 14:55 < bridge> I limited it to 250 ticks, e.g. 5 seconds 14:55 < bridge> i like current code, maps shouldnt use these tunes 14:56 < bridge> because I assume that if the hook takes longer, the hook line loses it's value 14:56 < bridge> Well move = restart simulation is probably the sanest idea if you want to give that a try 14:56 < bridge> For sane tunes it should look pretty much the same if you simulate like 50 ticks at a time or so 14:57 < bridge> But yeah very doubtful if the amount of reworking needed is worth the trouble 14:58 < bridge> also this wouldn't be stateless, I'd need to track this for every player of the possible 64 14:58 < bridge> Another good point though we probably already store enough state to derive your newly required state 14:59 < bridge> No, we don't 14:59 < bridge> So yeah, more state needed 14:59 < bridge> you'd at least need to store enough information to know where to continue the simulation 14:59 < bridge> Which can be derived from the tick the simulation started at, no? 15:00 < bridge> Anyway, we don't store that currently, so yeah, new state 15:01 < bridge> storing a tick is not enough, if you mean that <.< 15:01 < bridge> but yeah 15:01 < bridge> because we accumulate quantization errors 15:04 < bridge> so we'd at least store the position, the hook tick, and the direction 15:04 < bridge> 5s is probably good enough, it's just very arbitrary. So I'd worry there might be a combination of tunes and amount of players where it does still lag 15:04 < bridge> oh and the player target position to detect changes 15:05 < bridge> I agree, would you prefer a config variable for this? 15:06 < bridge> If this is about hook prediction you can curve the hook depending on ur and that's enough for me 15:06 < bridge> Ur velocity* 15:06 < bridge> and we don't do that lol 15:07 < bridge> Hm, not a horrible idea, but it's really hard to get config options in nowadays 15:07 < bridge> do we actually oO 15:08 < bridge> Idk what he means by curve, but technically if you are going fast enough you can get your hook inside a gap and by the time it hooks onto something you will be well past the gap. Soo, very technically the hook will curve 15:09 < bridge> Or more correctly the path of the hook will appear curved from the perspective of the tee 15:09 < bridge> the line shows the path the hook head takes, not what the hook animation should look like 15:10 < bridge> Yeah, it's not something you need to take into account at all 15:10 < bridge> because the chain will just connect the head position to the player, you can easily see this with telehooks 15:19 < Unidentified> Why there's so much sexual garbage on russian servers 15:19 < Unidentified> Do mods even exist here? 15:24 < Unidentified> . 15:24 < Unidentified> . 15:26 < bridge> . 15:52 < bridge> i mean #✉-create-a-ticket might be a better place to report such stuff 16:05 < Brap> IN LOVING MEMORY OF AOE 16:05 < Brap> We gather with heavy hearts to remember Aoe, Valid Trans Woman who lived simply but left a lasting mark on those who truly saw her. Short in stature and thin in frame, she resided in a modest hut, lived on her own terms, and walked through life with a spirit that was raw, real, and entirely her. Of Lesothese descent, Aoe cherished her roots and found deep joy in the bold, comforting flavors of Leso cuisine. 16:05 < Brap> Meals with her were more than sustenance-they were expressions of love, memory, and identity. Food was one of the many ways she kept her culture close, and she shared it generously. She was a person of contradictions-gentle but blunt, quiet but unforgettable. 16:06 < Brap> She bore a scent that many found hard to ignore. Earthy, unapologetic, and part of the unfiltered truth. She carried in every aspect of his life. Aoe did not pretend. She was who she was. Tragically, Aoe became heckin' infamous. We do not pretend to understand the weight she carried, nor the pain that brought her to that moment. But we do know this: she mattered. Her life, in all its complexity, was worthy of love, compassion, an 16:06 < Brap> Let us not define her by how she left us, but remember her for how she lived-with honesty, depth, cultural pride, and a stubborn refusal to be anyone but herself. May we carry forward her memory not only with sorrow, but with tenderness and truth. 16:06 < Brap> Rest peacefully, Aoe. You were seen. You are missed. You are loved. 16:08 < bridge> why can't I block this spam app 16:09 < Brap> OMG DA FUGGIN SPAMMERINOS RUINERINO OUR DISCORD SERVERINO 16:19 < bridge> Uhh 16:26 < bridge> mom, multischeduling is hard af 16:26 < bridge> i refactored my scheduler cuz i had bugs 16:27 < bridge> now i got a better scheduler, but i want to use oneshot timer in lapic, which means the timer is not periodic and i must reset the counter when i reschedule for example 16:27 < bridge> the problem is depending on ur logic u can not set the timer again and the current cpu goes idle, this is how u save power when nothing to do 16:27 < bridge> and u usually wake up with an external interrupt 16:27 < bridge> but i have a. bug where some schedulers go idle forever when they shouldnt xd 16:28 < bridge> rip 16:28 < bridge> damn atomics 16:46 < bridge>