00:38 < bridge> <Assa> ❤️
01:34 < bridge> <jimpanse16> 14.5.1 tried to join any ddnet servers, same pic. No OS Changes or Graphics changes done. Mapres folder contains 50 elements, so i guess that folder's fine too. You guys got any idea what changed?
01:34 < bridge> <jimpanse16> https://cdn.discordapp.com/attachments/293493549758939136/748267089399250964/unknown.png?ex=68322208&is=6830d088&hm=8465d22ba3112b1eac41dd301846ee97870f3add3e7ba58bf4c182721c161ed5&
01:38 < bridge> <kebscs> https://github.com/ddnet/ddnet/pull/10000
01:38 < bridge> <kebscs> bump
01:57 < bridge> <Ewan> 10k wtff
10:03 < bridge> <sans._.> Is there any plans to support ddnet on xbox/ps or switch?
10:07 < bridge> <Jupstar ✪> No
10:10 < bridge> <pioooooo> I wonder how did you decide to launch on steam
10:14 < bridge> <Jupstar ✪> I mean steam is the logic choice for computer games?
10:14 < bridge> <Jupstar ✪> Tbh pretty much all other platforms are rather hostile for our project
10:28 < bridge> <Assa> all hail gaben, our lord and savior
10:29 < bridge> <Assa> tbh, for a company having a game selling platform, steam does a lot right at least compared to ... (checks notes) ... all others
10:33 < bridge> <Jupstar ✪> On Epic store you'd get ddnet for free
10:33 < bridge> <Jupstar ✪> gog is actually best, bcs no drm xD
10:33 < bridge> <Jupstar ✪> and i can use heroic games launcher, which doesnt take years to start
10:34 < bridge> <pioooooo> What about Itch.io
10:34 < bridge> <Jupstar ✪> who uses that xd
10:34 < bridge> <pioooooo> Many puzzle games release on that((
10:35 < bridge> <Jupstar ✪> i mean we could always add other stores, but steam defs helped us most yet for visibility
10:35 < bridge> <Jupstar ✪> only problem is confining their TOS.
10:35 < bridge> <Jupstar ✪> 
10:35 < bridge> <Jupstar ✪> for itch.io it's probs ok.. for like nintendo switch it's probs impossible xD
10:35 < bridge> <Jupstar ✪> i doubt they like free open source games
10:56 < bridge> <learath2> Ew free
12:24 < bridge> <0xfaulty> Is anyone do something like wireshark plugin to parse teeworlds/ddnet protocol? And if not, is it realistic to write one or are there complications? I want to find reason why one of the packets during connection not reached client in some cases
12:26 < bridge> <milkeeycat> https://github.com/heinrich5991/libtw2/tree/master/wireshark-dissector
12:29 < bridge> <0xfaulty> ty
13:31 < ws-client> <ChillerDragon> @Jupstar ✪ ids for accounts in db or username as foreign key? whats your take on that?
13:32 < bridge> <qxdfoxy> is there any way to check how much cl_prediction_margin is set to for clients?
13:36 < bridge> <kebscs> no, prediction is client-sided
13:36 < bridge> <qxdfoxy> sad
13:47 < ws-client> <ChillerDragon> having a way for the server to dump client settings would be insane but also really cool
13:48 < bridge> <qxdfoxy> im trying to send offset quad positions based on players latency so that moving quads seem to be in the correct place for that client
13:48 < bridge> <qxdfoxy> prediction margin could mess this up
14:03 < ws-client> <ChillerDragon> @qxdfoxy just block ddnet clients on your server then you know that they dont have a prediction margin, ez
14:05 < bridge> <Jupstar ✪> chiller what?
14:06 < bridge> <Jupstar ✪> I'd always use an increment primary key.
14:06 < bridge> <Jupstar ✪> 
14:06 < bridge> <Jupstar ✪> You can still add unique names regardless of IDs
14:06 < bridge> <Jupstar ✪> But names potentially change, IDs not. So I prefer having IDs for everything
14:10 < bridge> <qxdfoxy> doing it server side instead of client side so it *should* work for everyone
14:10 < bridge> <qxdfoxy> ill just send a broadcast to tell em nvm
14:26 < bridge> <qxdfoxy> im doing it server side instead of client side so it *should* work for everyone
14:40 < ws-client> <ChillerDragon> Hmm @Jupstar ✪ i got inspired by ddnet table which does not have a id
14:40 < ws-client> <ChillerDragon> i used to also always have ids
14:40 < ws-client> <ChillerDragon> but if the username is already unique it feels redundant to also have a unique id
14:41 < ws-client> <ChillerDragon> also i keep hearing auto increment bad
14:41 < ws-client> <ChillerDragon> yes if the username is used as a foreign key then changing it will be a mess. I wonder if I want to support changing username anyways.
14:43 < ws-client> <ChillerDragon> i cant really decide. One the one hand using strings is bloat and i feel like its much more storage and cpu waste than an integer id.
14:44 < ws-client> <ChillerDragon> But on the other hand you might be able to reduce queries if you do not have to lookup username or id
14:53 < bridge> <Jupstar ✪> I can only tell you, your users probably want to rename
14:53 < bridge> <Jupstar ✪> Look at kog discord. many renames per month xd
14:53 < bridge> <Jupstar ✪> If they could do it manually/themself maybe even more
14:55 < bridge> <Jupstar ✪> But honestly, the ID can't hurt can it?
14:55 < bridge> <Jupstar ✪> In worst case never use it somewhere
14:55 < bridge> <Jupstar ✪> I dunno what's bad about auto increment, any example?
14:59 < ws-client> <ChillerDragon> do ppl on kog really change their account name or just the display name that shows up in the ranking?
15:00 < bridge> <Jupstar ✪> I think there are no account names on kog, so their display names which are their ranks
15:00 < ws-client> <ChillerDragon> idk if thats the same thing or not on kog actually but i have one username to login and then also the option to claim a display name
15:00 < ws-client> <ChillerDragon> i see
15:04 < bridge> <Jupstar ✪> also isn't the query perf better with integers over strings? xd
15:04 < bridge> <Jupstar ✪> In case you do joins
15:04 < ws-client> <ChillerDragon> yes thats one of the main reasons why i am thinking about adding an id
15:06 < bridge> <teero777> https://cdn.discordapp.com/attachments/293493549758939136/1375822232579997816/Screenshot_20250507-231030.png?ex=683315cd&is=6831c44d&hm=52027b42f81db801607b32a43844347cdf487ae2d3bb5b573cba34080975d75b&
15:06 < ws-client> <ChillerDragon> ?xd teero
15:06 < bridge> <meloƞ> based
15:07 < ws-client> <ChillerDragon> clang format joined the game
17:02 < bridge> <teero777> we need fall damage
17:08 < bridge> <chillerdragon> Fr
20:17 < bridge> <learath2> Mostly because the next id and prev id are predictable. So you can guess valid user ids and know the amount of users
20:20 < bridge> <learath2> Another big reason would be that they don't allow sharding. All db nodes need to coordinate to find the next id. UUIDs as an alternative can be generated independently
21:08 < bridge> <Jupstar ✪> yeah that is indeed a problem, but uuid will have some overhead again ig. so kinda a trade for what you want
21:10 < bridge> <learath2> Aye, and usually it doesn't end up really being 0 coordination since there is usually more than just the id that needs to be unique
23:38 < bridge> <meloƞ> #10147
23:38 < bridge> <meloƞ> 
23:38 < bridge> <meloƞ> could we get a second opinion on this please? - i dont like stale issues espacially with things that can and want to be used
23:38 < bridge> <DDNet> https://github.com/ddnet/ddnet/issues/10147
23:44 < bridge> <kebscs> If you don't like stale issues, you're in for a surprise in ddnet repi
23:44 < bridge> <meloƞ> oh really ._.
23:44 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1375952692257489007/image.png?ex=68338f4d&is=68323dcd&hm=7080feae7aabbaaea889f70e90c1d494f104da60c088a163ff6846aa752ae166&
23:48 < bridge> <learath2> I guess the question is whether this is "emulable" with just normal cp and whether that is so unacceptable
23:51 < bridge> <meloƞ> While yes, you could do that, it would require 2 cps minimum for each part, cannot be hidden and would clog up tight parts in entities where it's possible to drop into the same tele as the person that is getting dragged
23:51 < bridge> <louis> i think fixing current tp is probably a nicer solution
23:52 < bridge> <louis> invisible tiles is kinda a nono imo
23:53 < bridge> <meloƞ> @jsontw ^ go talk! (Afk 30 mins)
23:54 < bridge> <louis> from lynns comment:
23:54 < bridge> <louis> 
23:54 < bridge> <louis> > This doesn't cover the secondary purpose of this which is to force the order for retrying drag parts quickly. You wouldn't have to "fix" which side you're on, it would be consistent
23:54 < bridge> <louis> 
23:54 < bridge> <louis> this can probably be done with vertical teleTos and a no-bounce fix
23:54 < bridge> <louis> you either spawn on top or below dummy, each scenario you just tap [D] to get on the correct side of the dummy
23:55 < bridge> <jsontw> it will make every part look spammed with cp
23:55 < bridge> <jsontw> why do people not use it like that then?
23:55 < bridge> <kebscs> The point of entities is to see the map without editor, this defeats the purpose
23:56 < bridge> <kebscs> Also predicting would be issue if there's so many of them
23:56 < bridge> <jsontw> it doesn't do anything with the gameplay and doesn't have a hidden purpose behind it
23:56 < bridge> <jsontw> just qol
23:57 < bridge> <jsontw> this is a bad solution tho
23:58 < bridge> <jsontw> would rather just take the perfect spawn without having to press something every time
23:59 < bridge> <jsontw> that could work but idk if that is in favor of ddnet to change the current cto/to.
23:59 < bridge> <jsontw> also the 2nd purpose of the invisible cp would be a nice feature as well