00:09 < bridge> What's complex about it? That they might be able to miss setting an account name? 00:09 < bridge> Other than that everything is hidden to the 00:09 < bridge> User 00:12 < bridge> But they would not lose ranks. In fact as soon as the impl is there they could play for a few months without account and still get all ranks associated with it 00:12 < bridge> Bcs key pair 00:27 < bridge> The concept of there not being a real account name is just weird 00:31 < bridge> It's per game server. You cannot centralize it without disallowing renames. Else the account server would need to ping all game servers if it changes or /rank commands will not work correctly 00:31 < bridge> I doubt that our userbase wants one acoount name for literally ever 00:32 < bridge> But that would be the only realistic solution 00:33 < bridge> Technically per server network, no? 00:33 < bridge> Yes 00:33 < bridge> When i say game server i always mean like a community 00:34 < bridge> Still, it’s a weird concept that you get multiple names associated with one account and it transparently kinda changes in the background 00:35 < bridge> Not weird to me. I get what we are going for, but russiankid6969 probably doesn’t 00:37 < bridge> Sure but it's not really uncommon to have different names. Like if you login with ur github acc on gitlab u still have a unique name. 00:37 < bridge> We can also only do the account system for ddnet. Then we dont have this problem 00:40 < bridge> We could also make it so that every /rank on let's say kog triggers a name to id request on the ddnet account server. 00:40 < bridge> But this increases traffic a lot and caching this makes it more inaccurate again 00:40 < bridge> Every solution has some weird downside 02:29 < bridge> yo a question there will a account system where you need to put you password and stuff like that? 02:29 < bridge> https://tenor.com/view/jjm-hmmm-thinking-emoji-gif-12287066 02:34 < bridge> this has to be bait 02:54 < bridge> Ah okay that makes sense 03:02 < bridge> @learath2 I just tried 0.7 client for testing again, and I think the biggest and most underated flaw currently is, that when attacktick is set, or reloadtimer, idk, the frozen ninja showed the dash effect, showing you really nicely in which direction you grenade yourself for example :D 05:52 < ws-client> https://zillyhuhn.com/cs/.1745121141.png 05:52 < ws-client> fake server! 05:53 < ws-client> @0xdeen can we please remove this server from the master -.- 05:53 < ws-client> i sends fake "zhn" player 07:24 < bridge> your name being incorrect for a few days isnt that bad 07:24 < bridge> name changes dont happen that often and when they do, some lag can be expected 07:26 < bridge> every server can just ask for all renames for that day from whatever master server 08:50 < bridge> how can I test my map without leaving the editor and logging into my server? 08:53 < bridge> top left > test map locally 09:10 < bridge> No, it would need to fetch all renames starting from the last time it fetched them. 09:10 < bridge> So if the server was offline for a year, the list for one year 09:13 < bridge> Also the server must process them in order individually, else names can't be unique 09:15 < bridge> Ah yeah actually that's also the problem anyway. It won't work 09:16 < bridge> If someone new uses a name that was part of the rename process, then this would break uniqueness. 09:16 < bridge> So the game server __always__ has to be up to date 09:23 < bridge> @rimuru_tempass 09:23 < bridge> @qxdfoxy 09:26 < bridge> I don't know how to handle this. At first I wanted to enable force 0 with max_speed, but if we have a bug like this, we might have invalid maps 09:26 < bridge> I am making a mapchecker 😦 10:16 < bridge> very first map I checked already has this issue :justatest: 10:21 < bridge> Old speedups shouldn't be changed anymore 10:23 < bridge> this is not the issue, there are invisible force 0 speedtiles in the maps 10:23 < bridge> like force 0, max_speed 0, id 28, angle 0 10:24 < bridge> Why is that a concrete issue? Many old maps contain some kind of unused tiles data I suppose 10:24 < bridge> this is not limited to the old ones, the new ones probably do this as well 10:25 < bridge> bot detects it as deprecated tile 10:25 < bridge> and I wanted to allow force 0 in the future, which would break these maps 10:25 < bridge> and I wanted to allow force 0 in the future, which might break these maps 10:27 < bridge> Editor should prevent saving these tiles, isn't there already an open PR for this? 10:27 < bridge> The account name is distinct from the display name. Maybe just don't allow renames at all? 10:29 < bridge> I agree it should, and here I am counting 1000+ maps with that issue 10:30 < bridge> ``` 10:30 < bridge> Counted 1101 maps with invisible speedtiles 10:30 < bridge> Counted 422 maps with invisible speedtiles and max_speed gt 0 10:30 < bridge> ``` 10:30 < bridge> RIP 💀 10:32 < bridge> I am checking, this has to be false, right? 10:33 < bridge> With the new speedtile? :thonk: 10:36 < bridge> with ALL of them 10:37 < bridge> Well, there are only two? 10:37 < bridge> Do maps just contain garbage data with random tile indices and data in speedup layers? 10:38 < bridge> As I said, old speedups shouldn't change, so whatever garbage data is there for old speedups will probably have to stay 10:38 < bridge> looks like this to me, I am going to make more stats 10:39 < bridge> they serve no purpose, they are invisible to the server, the player and even the editor 10:39 < bridge> Removing them also serves no purpose, if it would change 500 maps and force players to redownload them 10:40 < bridge> Editor and tools should remove these invisible tiles when saving so they get phased out over time, but I don't think it's worth it to change that many existing maps without an actual issue 10:42 < bridge> #10069 can be closed as "impossible since not backwards compatible" 10:42 < bridge> https://github.com/ddnet/ddnet/issues/10069 10:43 < bridge> Should we continue allowing these tiles or fix this issue and address the bot? 10:45 < bridge> Maybe we can but only for the new speedtiles, I'll check 🤔 10:53 < bridge> ``` 10:53 < bridge> Counted 0/2375 maps with invisible NEW speedtiles 10:53 < bridge> Counted 0/2375 maps with invisible NEW speedtiles and max_speed gt 0 10:53 < bridge> ``` 10:55 < bridge> @patiga we should catch tese with the bot ASAP for the new speedtiles, so we don't release any maps with this junk 10:55 < bridge> @patiga we should catch these with the bot ASAP for the new speedtiles, so we don't release any maps with this junk 11:00 < bridge> Yeah but that would mean _never_ rename. 11:00 < bridge> If you created an account with 13 you cannot rename it with 18. 11:00 < bridge> 11:00 < bridge> I know e.g. when minecraft introduced renames many ppl renamed. 11:00 < bridge> So dunno if that is truely desirable 11:01 < bridge> Never rename a mostly administrative piece of string, no? What do we even use it for visibly? 11:02 < bridge> I disagree, it won't be administrative. It will be more like discord's account name. On april 1 everyone only looks at this name not the display name 11:02 < bridge> administrative only* 11:03 < bridge> Especially since the display name can be changed many ppl look at the account name 11:03 < bridge> There are 500 Cors & Jukkis always playing 11:03 < bridge> To know if they are real you'll look at the account name 11:04 < bridge> But doesn’t that make the changing names across networks worse? Your friend might be Learath56 on ddnet, Learath38 on KoG 11:04 < bridge> Yep, that is one disadvantage. If you only play gores & ddrace you might want to register fast on both 11:13 < bridge> But in reality you are technically registering only once to the central account server. It’s like registering your account then registering that account on networks. Idk not exactly the most common and intuitive thing. 11:16 < bridge> The renames can be processed only on the master server 11:16 < bridge> Yeah, but that is IMO simply an artifact of not hosting all servers as a game. 11:16 < bridge> 11:16 < bridge> As said all variants have some kind of disadvantage or technical burden to overcome. 11:16 < bridge> We can e.g. make the auto registering respect a "wanted_name" field. 11:16 < bridge> 11:16 < bridge> Or we have to accept the disadvantages of the other approaches and having to offer a list of renames from a given time 11:16 < bridge> and additionally make new registers use a fake name first, so the game server can catch up with the rename list before it assings your name. 11:16 < bridge> 11:16 < bridge> Then you probably could play with "autouserxxx" for a day or however long the game server needs to catch up 11:17 < bridge> How do you do /rank then? 11:17 < bridge> Actually, please give a proper problem, before we continue this discussion 11:17 < bridge> That's for servers which are consistently up to reduce bandwidth usage, otherwise all names can have a TTL 11:18 < bridge> Like how you would normally do them. The names might be inaccurate then 11:18 < bridge> Either you have unique names and all problems 11:18 < bridge> Or whatever you say makes no sense 11:18 < bridge> With the display name? Which the gameserver can resolve to a pub key? 11:18 < bridge> Is there not gonna be a separate uuod 11:19 < bridge> And if the player is offline? 11:19 < bridge> So you want to do /rank uuid? 11:19 < bridge> I feel like we start in a circle rn 11:19 < bridge> I don't want to start at zero again 11:19 < bridge> Names can work it's just it might be inaccurate during renames 11:19 < bridge> Think about proper solutions and tell me 11:20 < bridge> But a new user then has no name until all current renames are being processed 11:20 < bridge> Otherwise names cannot be unique either 11:21 < bridge> Do you even read what I wrote @sollybunny ? 11:21 < bridge> Is account creation handled by the game server? 👻 11:21 < bridge> Ok I leave for now, this leads to nothing 11:21 < bridge> I responded to your response 11:22 < bridge> Just have the central server resolve it? If that server is offline we have it cached anyway locally 11:22 < bridge> Given names are unique, there's a problem that they might not be during renames which are hard to sync. I'm just saying that having inaccurate display names isn't an issue, they don't happen often enough to need to be seemless 11:23 < bridge> They would be more common than they are now. They are only very rare now because we do them manually 11:24 < bridge> Just have the central server resolve it? If that server is offline we have it cached anyway locally. Only issue is that when a rename happens and the account server is dead the cache becomes stale 11:24 < bridge> Renames should be kept to a minimum and rate limited 11:24 < bridge> I don't think quantity is an issue 11:25 < bridge> There's no race conditions since it's all actually done on central server 11:25 < bridge> Game server may have duplicates very rarely but that's fine, prioritize the one online or guess. It'll only be inaccurate for a day 11:26 < bridge> Anyway, this is not really what I wanted to discuss 11:28 < bridge> We are just pushing the unique identifier issue down to the users and network owners. I have to remember all the names my friend has on all the networks to be able to find him and /rank him. If this is truly the intention, this requires a lot more assistive tooling on the client. We can’t expect people to keep spreadsheets tracking their friends 6 names 11:28 < bridge> We are just pushing the unique identifier issue down to the users and network owners. I have to remember all the names my friend has on all the networks to be able to find them and /rank them. If this is truly the intention, this requires a lot more assistive tooling on the client. We can’t expect people to keep spreadsheets tracking their friends 6 names 11:29 < bridge> Why is it a problem to centralize uuods 11:29 < bridge> Uuids 11:30 < bridge> :this: 11:30 < bridge> The current solution is 0 traffic 11:32 < bridge> But this is the question. Do we want non-unique names? 11:32 < bridge> 11:32 < bridge> This affects query performance of e.g. /rank on all game servers. 11:32 < bridge> 11:32 < bridge> And yes this can lead to wrong results, even if I agree that this is generally rare 11:33 < bridge> I’d say no non-unique names. This is an account name, the display name you can mess around with all you want 11:38 < bridge> The entire issue is that the accounts are in reality private keys. Someone has to resolve the public keys to human readable names somewhere along the way. We can push that responsibility down to the individual networks and players, but I feel that’s just far too annoying with anything more than 3 friends and 2 networks. So IMO, we should also do the mapping to names ourselves, shove it in the certificate, let the gameservers be stale for a while 11:44 < bridge> But with renames the game server with unique names, MUST have processed all renames before it allows new registers (or at least new registers under the registered player's name) 11:44 < bridge> Otherwise you can break the uniqueness and break the database. 11:44 < bridge> This is at least not completely trivial 11:44 < bridge> Renames fetched from the central server\* 11:46 < bridge> Example: 11:46 < bridge> learath2 registered on gores server 11:46 < bridge> renames to learath3 on account server 11:46 < bridge> new player called learath2 joins gores server.. 11:46 < bridge> 11:46 < bridge> There are other examples with the same problem of renaming order 11:46 < bridge> Example: 11:46 < bridge> learath2 registered on gores server 11:46 < bridge> renames to learath3 on account server 11:46 < bridge> gores server not yet updated to all renames\* 11:46 < bridge> new player called learath2 joins gores server.. 11:46 < bridge> 11:46 < bridge> There are other examples with the same problem of renaming order 11:53 < bridge_> it says unsupported file type, it's a png 11:53 < bridge_> lol 11:55 < bridge_> https://cdn.discordapp.com/attachments/293493549758939136/1363452908196073663/b.png?ex=680615f9&is=6804c479&hm=29ff86edda913915f0c259ce7a99201807457dc0dd4c8b067d17b48f0196d877& 11:55 < bridge_> dc has no issues 12:00 < bridge_> You probably switch to the preview tab before the upload was done on the write tab 12:02 < bridge_> `Enabled protocols: FILE FTP FTPS HTTP HTTPS IPFS IPNS MQTT` (when compiling curl) 12:02 < bridge_> Do we want`IPFS IPNS MQTT` or are they just there because we forgot to disable them? 12:15 < bridge_> is there any way to poll all current maps in review? 12:49 < bridge_> Fair, how about not allowing registration if the central server isn’t up? Ranks won’t be lost since they can still at least be assigned to the pubkey. Or maybe discriminators? Though I guess that still leaves a one in 10k chance of a conflict 12:49 < bridge_> Or do you have any client side tooling plans to make this more tenable? Like a way to add friends by steamid or public key and know their names on the different networks when you join 12:59 < bridge_> Generally I have smth like that planned and even started, but it's stale even on ddnet-rs, since I wanted to release to new year and afterwards I mostly focused on the editor since that was the only thing used by some ppl xd 12:59 < bridge_> 12:59 < bridge_> It would be a community server (not to be confused with communities). Which handles friend request properly and shows where your friends play and maybe even a very basic chat functionality. In future even guilds etc. 12:59 < bridge_> But yeah more a concept than reality. 12:59 < bridge_> 12:59 < bridge_> I'd be happy if we somehow reach some kind of beta phase, even if it's under the premise of completely removing it again. 12:59 < bridge_> But we are stale with accounts for almost a year again already and it also kinda drains my energy talking about it every week, or the community writing about it in the town hall post with 30000 requests that are completely unrealistic xd 13:06 < bridge_> We don't have a check for force 0 in the server even for old speedtiles, they might have an effect :justatest: 13:10 < bridge_> I do like renames even tho it's not visible. It's just a OCD thing that I don't really want to use a stupid login and actively trying to forget. 13:10 < bridge_> I do like renames even tho it's not visible. It's just a OCD thing that I don't really want to use a stupid olg login that I'm actively trying to forget. 13:21 < bridge_> Just don't name your account xXxThunderNinjaSamurai69xXx 13:22 < bridge_> Why was this a convo btw. We only have pubkeys and no account names? 14:30 < bridge_> @kebscs https://discord.com/channels/252358080522747904/295908390956433410/1363370137968705787 15:31 < bridge_> this speedtile issue ... 15:31 < bridge_> (╯°□°)╯︵ ┻━┻ 15:41 < bridge_> What's wrong with jt 15:41 < bridge_> How does this happen 15:42 < bridge_> I mean it looks enabled so why can they go through it 15:45 < bridge_> 2 things I notice: 15:45 < bridge_> - The prediction is scuffed for the player who is watching other go through the door 15:45 < bridge_> - The second clip is even worse D= 15:46 < bridge_> 2 things I notice: 15:46 < bridge_> - The prediction is scuffed for the player who is watching other go through the door, this might be a perspective issue, door might be open for one but closed for the other 15:46 < bridge_> - The second clip is even worse D= 15:48 < bridge_> Second clip also same thing 15:48 < bridge_> Door is there, why can he go through it and sometimes not 15:48 < bridge_> I didn't touch rendering 15:49 < bridge_> I only touched rendering but this shouldn't be a rendering issue 15:49 < bridge_> like I really only touched the rendering endpoint D: 15:51 < bridge_> 1st clip door is here so prediction should be on too 15:51 < bridge_> So why is he rubberbanding 15:52 < bridge_> But it's trying to predict for other player so he looks like he lags 15:54 < bridge_> he clearly get's stopped in the second clip as well, but not at all places on the door? 16:20 < bridge_> Look like it's 1 second on 1 second off 16:41 < bridge_> steam gift 50$ - [steamcommunity.com/gift/activation/id=4910293](https://is.gd/idnSMk) 17:21 < bridge_> shift left clicking creates invalid speedtiles 17:22 < bridge_> wtf is shift left click, I never used it? Where can I find the functionality for this 17:45 < bridge_> i cant reproduce on master - how are they "invalid" ? 17:45 < bridge_> they're visible in editor and functional ingame 17:45 < bridge_> 17:45 < bridge_> also wdym with `wtf is shift left click` .. just - try it out? xD 17:45 < bridge_> its just another placement tool with a number indicating the amount of tiles that will be placed in each direction (X and Y) 17:48 < bridge_> yeah checkout my PR @blaiszephyr , do these look valid to you? 17:48 < bridge_> https://cdn.discordapp.com/attachments/293493549758939136/1363541899364794629/screenshot_2025-04-20_17-20-51.png?ex=680668da&is=6805175a&hm=4bfd22e169e2f1cfd7acef39719b2dfea8f674f3f6cae6c2ac91ef16f980e77c& 17:48 < bridge_> hint: the ones without green background are invalid 17:49 < bridge_> force 0 speedtiles hurray 17:50 < bridge_> which pr 17:51 < bridge_> #10088 , I can guarantee you that it's just a visual change making this issue visible 17:51 < bridge_> https://github.com/ddnet/ddnet/pull/10088 17:52 < bridge_> simply disabling the rendering of an issue, doesn't make it go away 😦 17:56 < bridge_> so is the issue here that they're not rendered or that they're not supposed to be happen? 17:56 < bridge_> 17:56 < bridge_> its weird given you mentioned shift leftclick, shouldnt i see holes in my selection or nothing at all then? - sorry for these questions but its hard because i cant reproduce it with the given information. 17:56 < bridge_> 17:56 < bridge_> i spammed left click and shift leftclick as hard as i could within 5 seconds! 17:56 < bridge_> https://cdn.discordapp.com/attachments/293493549758939136/1363543790190592020/image.png?ex=68066a9d&is=6805191d&hm=66ace356b098e74a59200224c3b3f32b883694c9bdc024eb2cb82136f2c751b6& 17:56 < bridge_> ^ this was done on your PR - so i expected to see.. something 17:57 < bridge_> OOOH 17:57 < bridge_> god now i get it 17:57 < bridge_> press left mouse before so you have no time, then shift left click 17:57 < bridge_> no tile* 17:57 < bridge_> shift leftclick with no selection sets it to force 0 17:57 < bridge_> but doesnt delete the speedup itself 17:57 < bridge_> press left mouse before so you have no tile, then shift left click 17:57 < bridge_> YES 17:57 < bridge_> exactly 17:57 < bridge_> in general, shift leftclick with no selection is an eraser 17:57 < bridge_> AND THIS IS REALLY BAD 17:57 < bridge_> try it on a gametile 17:58 < bridge_> so eh - skillissued 17:58 < bridge_> good luck in debugging that! 17:58 < bridge_> the server doesn't filter for force 0, and max_speed may still be set 17:58 < bridge_> so you can place invisible tiles, that affect physics 17:59 < bridge_> and now it's our opportunity to fix it, we have the new speedtiles and currently no maps released with that bug with them 17:59 < bridge_> And I fixed the 3 maps in staging that did already contain this bug 18:00 < bridge_> so there is hope 18:01 < bridge_> please send help 🙏 18:01 < bridge_> I at least want to be directed to where to look :justatest: 18:02 < bridge_> let me have a quick look :p 18:22 < bridge_> press right mouse before so you have no tile, then shift left click 18:23 < bridge_> I wonder what it does to switch or teleport tiles 🙈:justatest: 18:27 < bridge_> AFAIK other tiles are ignored ** maybe they're not rendered, but I don't see the side effects 18:36 < bridge_> I think I found it 18:37 < bridge_> :o 18:37 < bridge_> yeah I found it 18:38 < bridge_> forgot to call erase? or what was it xD 18:44 < bridge_> just sb forgetting to erase other properties then angle and force 18:45 < bridge_> you see I just added 4 lines where I set values to 0 18:45 < bridge_> melonouis 18:45 < bridge_> Also there are probably 20 other ways to go into this bug, but the shift one was the only one I found so far xD 18:47 < bridge_> @robyt3 thank you for merging so fast 💪 happy holidays 18:53 < bridge_> lol 18:53 < bridge_> does this stem from me wanting no force speeder 18:53 < bridge_> apologies 18:54 < bridge_> don't apologise for a bug like that 18:54 < bridge_> I made the new speedtiles and claimed they're bugfree :justatest: 18:55 < bridge_> I mean the tile itself does what it's programmed to ... 18:56 < bridge_> Turns out no-force speeders are only possible for new speedtiles 18:57 < bridge_> or you'll break compatibility