00:04 <+bridge> I could perhaps be convinced to drop ogl1 if we keep server support for the last version that supports ogl1 indefinitely 00:10 <+bridge> I think Jupstar isn't against properly engineered backcompat 00:10 <+bridge> but what we have isn't properly organized 00:12 <+bridge> it's all over the codebase 00:12 <+bridge> instead of inside a backcompat module 00:18 <+bridge> @gerdoe 00:19 <+bridge> oops, unintended ping 00:22 <+bridge> it's on my to do list, but not very high. quic/accounts are definitely higher 00:23 <+bridge> How would a backcompat module work if the code is scattered everywhere as it is. 00:27 <+bridge> it would take the network traffic and translate it to old network traffic 00:27 <+bridge> I built a 0.6-0.5 using that pattern 00:27 <+bridge> the code almost entirely resided in one module 00:29 <+bridge> 0.5 😯 00:29 <+bridge> Anyone uses that? 00:29 <+bridge> past tense 00:30 <+bridge> ? 00:30 <+bridge> he did that most likely 10+ years ago, I'd guess 00:30 <+bridge> ah I guess maybe 00:31 <+bridge> https://github.com/heinrich5991/teeworlds/blob/eedc8065b0340ad5659fab2b91b09b8f94354336/src/proxy/proxy/0.5-0.6/proxy_cl05_srv06.cpp 00:31 <+bridge> 10 years ago seems correct 00:32 <+bridge> I should gamble 00:32 <+bridge> The massive debt of backcompatability already in the codebase makes me uncomfortable often but it's not my issue I guess. 00:33 <+bridge> Massive is a little overstatement 00:33 <+bridge> I don't think it's a debt of backcompat, I think it's a debt of spaghetti code ^^ 00:34 <+bridge> <0xdeen> Anyone want to submit https://github.com/ddnet/ddnet/pull/8538 ? It currently breaks nightly build 00:35 <+bridge> Well I don't think this is in small part due to back compatability requirements. Huge refactorings need to preserve old functionality 00:36 <+bridge> Also the nature of open source dev where most people don't want to take long term responsibility for large sections of the code 00:36 <+bridge> done 00:44 <+bridge> I actually don't know of any other multiplayer games that keep compatability for mismatched server/client versions. Does anyone know of any? 00:44 <+bridge> 00:44 <+bridge> I guess it is very nice that you can have a 4 year old client and be confident you can play on any server in the list. 01:25 <+bridge> @ryozuki I just stumbled upon this very old commit of mine: https://github.com/heinrich5991/libtw2/commit/76355d4a34870b336d3adb48cf61689a5281d504 01:26 <+bridge> `IpAddr` would only become stable in rust 1.7.0 😄 01:26 <+bridge> lol 03:36 <+bridge> The minecraft community is big on version translation. But I guess that is because the actual official game it self just breaks versions all the time xd 03:37 <+bridge> @totar: it’s just sad that latest ddnet does not even have the full list :P 03:39 <+bridge> Minecraft is quite interesting but that's a unique situation where only the server is controller by the developers so they can't control the client at all. 03:40 <+bridge> In a way ddnet is the same but we still add plenty of client features that depend on server functionality so it's not really the same 03:42 <+bridge> Was that proxy ever in use? How does it work? Does every user need one to connect to a server? Or can a server Hoster slap it in front of a server? If every user needs it it’s a bit annoying. And I guess servers have no full proxy support so they would think all players joined through the proxy are coming from one ip address. 03:53 <+bridge> What was the accident xd 04:02 <+bridge> Amazing thanks 04:08 <+bridge> We test the cuirect and stuff like that logic already in unit tests right? So what’s missing is probably generating a screenshot and doing a image diff. 04:11 <+bridge> What why? ssh is great? I don’t get it 04:12 <+bridge> https://cdn.discordapp.com/attachments/293493549758939136/1257519189594341376/Screenshot_20240701-163218.png?ex=6684b382&is=66836202&hm=ea286d6b3d752da868b625d7b939ab6ab29c13878b22976038662702901de527& 04:17 <+bridge> Oh that’s good to know! Thanks for sharing. I never had the problem that the WiFi website does not pop up but it might happen some day. Makes a lot of sense never thought about it. 04:23 <+bridge> Xd 04:26 <+bridge> ?xd 04:31 <+bridge> Yikes 04:32 <+bridge> Debian BTW 08:01 <+bridge> Debian FTW 08:23 <+bridge> Debian was my first big love 08:23 <+bridge> But its to old 🥲 08:23 <+bridge> But its too old 🥲 08:43 <+bridge> morning 08:43 <+bridge> gentoo ftw 08:49 <+bridge> mornin! 08:49 <+bridge> I Use arch btw 08:52 <+bridge> https://ersei.net/en/blog/fuse-root 08:58 <+bridge> @scrumplex are u around for some nixos troubleshooting 08:58 <+bridge> im a noob 09:13 <+bridge> I know the basics - and scrumpi taught me sum. Maybe i can Help 09:16 <+bridge> On phone atm but here's the gist 09:16 <+bridge> I have a flake that I wanna use as my system config https://github.com/ewancg/shit 09:16 <+bridge> Now put /boot on cloud too. Write an efi driver for cloud fs 09:16 <+bridge> in nixos dir 09:17 <+bridge> ``` 09:17 <+bridge> ewan@machine ~/shit (main)> nixos-rebuild --flake ./nixos#machine build 09:17 <+bridge> building the system configuration... 09:17 <+bridge> error: 09:17 <+bridge> … while calling the 'head' builtin 09:17 <+bridge> 09:17 <+bridge> at /nix/store/g83ji892byw4wd81rmb14v69yrhq408s-source/lib/attrsets.nix:1575:11: 09:17 <+bridge> 09:17 <+bridge> 1574| || pred here (elemAt values 1) (head values) then 09:17 <+bridge> 1575| head values 09:17 <+bridge> | ^ 09:17 <+bridge> 1576| else 09:17 <+bridge> 09:17 <+bridge> … while evaluating the attribute 'value' 09:17 <+bridge> 09:17 <+bridge> at /nix/store/g83ji892byw4wd81rmb14v69yrhq408s-source/lib/modules.nix:809:9: 09:17 <+bridge> 09:17 <+bridge> 808| in warnDeprecation opt // 09:17 <+bridge> 809| { value = builtins.addErrorContext "while evaluating the option `${showOption loc}':" value; 09:17 <+bridge> | ^ 09:17 <+bridge> 810| inherit (res.defsFinal') highestPrio; 09:17 <+bridge> 09:17 <+bridge> (stack trace truncated; use '--show-trace' to show the full trace) 09:17 <+bridge> 09:17 <+bridge> error: getting status of '/nix/store/pz24qq8rhnc918bg0n31xpli9fs3vz82-source/.config': No such file or directory 09:17 <+bridge> ``` 09:17 <+bridge> It's evil 09:18 <+bridge> Which File exactly ? 09:21 <+bridge> Looks Like a Syntax Error or Typo in one of your modules 09:21 <+bridge> Are you correctly importing it ? 09:25 <+bridge> Ah wait, i dont think you can invoke configuration.nix INSIDE a Home Module 09:28 <+bridge> On phone myself so hard to Doublecheck lmao 09:36 <+bridge> https://codeberg.org/melon/.dotfiles/src/branch/main/hosts/melon/flake.nix 09:36 <+bridge> 09:36 <+bridge> Thats how i do it, have to keep on working 09:36 <+bridge> 09:36 <+bridge> @mpft :pepeangryping: 09:37 <+bridge> https://codeberg.org/melon/.dotfiles/src/branch/main/hosts/melon/flake.nix 09:37 <+bridge> 09:37 <+bridge> Thats how i do it, have to continue working 09:37 <+bridge> 09:37 <+bridge> @mpft :pepeangryping: 09:37 <+bridge> ya i mean 09:37 <+bridge> im not sure it's an issue w my flake necessarily 09:37 <+bridge> 09:38 <+bridge> ppl seem to have a similar issue with obrcene fixes 09:38 <+bridge> ppl seem to have a similar issue with like 3 different proposed obscene fixes 09:38 <+bridge> i tried any that were applicable and had no luck 09:42 <+bridge> So ist seems to be an issue with git beeing annoying lmao - maybe mv out the .git folder to be sure its not that, i dont remember running into that issue tho. 09:48 <+bridge> nop 09:48 <+bridge> need git metadata for the flake anyway 09:59 <+bridge> chillerdragon: both worked. no, the proxy was compiled into the server, but it roughly only had one intersection point 10:00 <+bridge> chillerdragon: I didn't check the ddnet discord, I did not make an explicit decision not ot 10:00 <+bridge> chillerdragon: I didn't check the ddnet discord, I did not make an explicit decision not to 10:01 <+bridge> ChillerDragon: ssh had a "anyone can log in as root" vulnerability yesterday. patch your server! 10:17 <+bridge> wb heinrich 10:32 <+bridge> thanks 10:33 <+bridge> holy moly. does debian even have patches for that? 10:36 <+bridge> yes 10:36 <+bridge> I already applied them yesterday 10:37 <+bridge> `OpenSSH_9.2p1 Debian-2+deb12u3, OpenSSL 3.0.13 30 Jan 2024` 10:37 <+bridge> LGTM 10:37 <+bridge> is `systemctl restart sshd` enough? 10:37 <+bridge> ssh -v localhost 10:38 <+bridge> smart 10:39 <+bridge> ye lgtm 10:39 <+bridge> the 3 in the end is the fix i assume 10:39 <+bridge> yeah 10:39 <+bridge> well it sounds like they just disabled some code xd 10:39 <+bridge> probably better to update as soon as they apply the real patch 10:40 <+bridge> sounds good to me 10:41 <+bridge> https://cdn.discordapp.com/attachments/293493549758939136/1257617137892069446/image.png?ex=66850ebb&is=6683bd3b&hm=cedb9d0759defd09869c87b56496602266dbc5f91efaea287e99faf612274815& 10:41 <+bridge> my laptop is vulnerable 10:56 <+bridge> My servers ssh version is before the regresshion, so it's safe without upgrading 😄 10:58 <+bridge> "safe" 10:58 <+bridge> xxxxxxxxD 11:05 <+bridge> oh boi 4k players online right now 11:06 <+bridge> i basically didnt open my client this year didnt know the game is still very much alive 11:14 <+bridge> safe from this one specific critical bug, to be more exact 😄 11:14 <+bridge> i just realize 11:15 <+bridge> our company really like pushing releases when i'm not at work 11:15 <+bridge> maybe whenever I'm at work thing are always not ready 11:15 <+bridge> @heinrich5991 when you get around to looking at atleast the teehistorian files local to that server, I can finally consider upgrading and rebooting that server 11:16 <+bridge> local to what server? learath2.info? 11:16 <+bridge> yes 11:16 <+bridge> it has full disk encryption with a handrolled initramfs script, there is a chance that if I upgrade it and something goes wrong I might lose all the data on it 😄 11:21 <+bridge> the script I sent you was buggy, right? 11:29 <+bridge> u didn't quit yet? 11:32 <+bridge> I updated my server, but sshd doesn't restart anymore. I started sshd manually again. but I'm also scared of rebooting ^^ @learath2 11:48 <+bridge> Yes, I think I gave you a full log. It crashes for some reason and dumps just a list of all the files 11:48 <+bridge> It works just fine in a smaller folder though 11:51 <+bridge> ah thanks 11:51 <+bridge> https://cdn.discordapp.com/attachments/293493549758939136/1257634678907605123/image.png?ex=66851f11&is=6683cd91&hm=eb06d163c4e35d3bb414e69d57d1b46368da3c389f27ecb770522e811dd0892c& 11:51 <+bridge> quality software 11:57 <+bridge> name and shame pls 12:00 <+bridge> Grafana Loki 12:41 <+bridge> Why Ukrainian Novice is passworded 🙁 12:41 <+bridge> @heinrich5991 it seems the double message doesn't happen anymore 12:42 <+bridge> Wow, I always wanted to meet a true Englandian 12:46 <+bridge> just the beta server 12:48 <+bridge> Just unhardcode it, what's the problem 12:52 <+bridge> <0xdeen> I don't see it, is it? 12:54 <+bridge> deen 12:54 <+bridge> I wanna host a trashmap server in usa 12:55 <+bridge> gerdoe u the troll 12:55 <+bridge> :greenthing: 12:56 <+bridge> i thought timakro hosts then 12:56 <+bridge> i thought timakro hosts them 12:56 <+bridge> ya 12:56 <+bridge> ddnet service tho 12:56 <+bridge> ah 12:56 <+bridge> 12:57 <+bridge> i like how youtube recommends me to watch videos i have already watched 12:57 <+bridge> (trance dj mixes) 12:58 <+bridge> 😃 12:59 <+bridge> wanna listen one 12:59 <+bridge> How long is it 12:59 <+bridge> quite a short 12:59 <+bridge> half an hour 13:00 <+bridge> u can send it and I'll listen tomorrow 13:00 <+bridge> it's past my bedtime 13:00 <+bridge> xd 13:00 <+bridge> https://www.youtube.com/watch?v=f6F9Ffc_kx0 13:01 <+bridge> 125p in ddnet SUI 13:01 <+bridge> 128p in ddnet SUI 13:05 <+bridge> I'll fork the DDNet client and remove EVERYTHING /hj 13:05 <+bridge> /hj? 13:11 <+bridge> <0xdeen> @fokkonaut Crashed: https://github.com/ddnet/ddnet/pull/8542#issuecomment-2202783530 I'll disable votes for now 13:12 <+bridge> Handjob obviously 13:15 <+bridge> Good thing I don't use ssh 13:15 <+bridge> half-joking 13:21 <+bridge> its me about to steal components from client side code into server side 13:22 <+bridge> https://cdn.discordapp.com/attachments/293493549758939136/1257657521213145149/image.png?ex=66853457&is=6683e2d7&hm=ed6a33c06518cf0abfa34f0809a006d4212cf343058f2bccbbc0b174899894d7& 13:42 <+bridge> Me Reading your commits :pepeW: :monkaStop: :poggers2: :pepe_holy: :justatest: :MonkaS: :pepe_ew92: :kek: :gigachad: 14:07 <+bridge> Ok, voting can happen yea 14:07 <+bridge> <0xdeen> I think masterserver needs to allow larger json @heinrich5991 ? `2024-07-02 14:00:02 E register/6/ipv4: error response from master: 400: Request body deserialize error: EOF while parsing a value at line 1 column 16383` 14:07 <+bridge> <0xdeen> Happened with ~110 players online 14:07 <+bridge> 110 again right now 14:08 <+bridge> and it works nice 14:09 <+bridge> server disappeared 14:09 <+bridge> <0xdeen> Found it, fixed in the PR 14:09 <+bridge> from list only 14:09 <+bridge> Oh, cool 14:09 <+bridge> ty! 14:10 <+bridge> <0xdeen> Please fix the voting crash though, it seems to happen with a 0.7 player because of the playermapping 14:10 <+bridge> Yes, on it 14:14 <+bridge> https://github.com/ddnet/ddnet/pull/8542/commits/db16907d8cff79958f42e1c0dedfc7361b5e7b0c 14:22 <+bridge> @0xdeen If u want u can set the mapupdaterate to 5, will be quicker to update the map 14:23 <+bridge> but i think for performance it can stay at 15 14:23 <+bridge> dunno 14:23 <+bridge> or 10, whatv 14:32 <+bridge> why does it even have a update rate? 14:33 <+bridge> cuz why would you run it every tick 14:33 <+bridge> useless 14:33 <+bridge> why doesnt it update based on e.g. player movement 14:33 <+bridge> why dont u make ur own xd 14:33 <+bridge> i have my own 14:34 <+bridge> is it based on e.g. player movement 14:34 <+bridge> let's see, havent read that code in \~6 years 14:34 <+bridge> xD 14:34 <+bridge> then probably not 14:35 <+bridge> ChillerDragon: to find out which license to use, where did you copy the code from? probably teeworlds 0.7? 14:38 <+bridge> seems it simply checked the player distance on the fly 14:38 <+bridge> was enough for 256 players on a cheap vps 14:39 <+bridge> so u have ddnet code 14:39 <+bridge> xd 14:39 <+bridge> it was for fng vanilla based 14:42 <+bridge> https://github.com/Jupeyy/dd-pg/blob/af04025138549ef64d70a9b0b748b18654571fec/game/shared-game/src/entities/character/pos.rs 14:42 <+bridge> this would be even better 14:42 <+bridge> could directly also be used to solve the O(n²) problem 14:42 <+bridge> without breaking physics 14:43 <+bridge> rust is so ugly 14:43 <+bridge> :kek: 14:43 <+bridge> c++ ftw 14:43 <+bridge> i guess porting 165 lines of rust code to cpp isnt the world 14:44 <+bridge> > `HashMap`, `LinkedHashMap`, `Pool` 14:44 <+bridge> not trivial though, I'd say 14:45 <+bridge> sure, pool is just to save heap allocs, would also work fine with heap allocs or stack allocations (esp. if max. number of clients are known at compile time) 14:45 <+bridge> linkedhashmap, i dunno why i used that xd 14:46 <+bridge> maybe at some points i was iterating over the entries 14:46 <+bridge> but doesnt look like i still do that 14:48 <+bridge> and for ddnet physics it would make sense to use a btreeset instead of hashset to keep the ID order correct 14:48 <+bridge> (which is only interesting when to emulate the for i \< MAX\_CLIENTS loops in the gamecore) 15:02 <+bridge> How does a hashset solve the O(n^2) issue? Do you use a position hash that hashes nearby people in the same bucket? 15:03 <+bridge> yea 0.7 I think and a bit of huffman-py and TeeAI xd 15:03 <+bridge> And then what? I just copy the license? 15:04 <+bridge> Huffman-py is also bsd and TeeAi is gpl 15:05 <+bridge> basically: 15:05 <+bridge> i calculate the tile of the tee (a tee is *almost* 1 tile, so that fits quite nice) and add its id to a hashset 15:05 <+bridge> 15:05 <+bridge> later in character core: 15:05 <+bridge> i get all ids in a certain radius (so basically all tiles in that radius) 15:05 <+bridge> the "real" radius calculation is still done 15:05 <+bridge> just for fewer entries 15:05 <+bridge> since the tile based radius check ofc isnt perfect 15:06 <+bridge> @heinrich5991 I also took the packer code 1:1 15:06 <+bridge> whats the problem here? 15:06 <+bridge> im ootl 15:06 <+bridge> when quadtrees? 15:06 <+bridge> i think that approach is better 15:06 <+bridge> And swarfy also looked at the TeeAI code for his type script Huffman 15:06 <+bridge> Ah the grid approach, where does a hashmap come in? 15:06 <+bridge> bcs updating the quadtree isnt really cheap 15:06 <+bridge> simply the tile index 15:07 <+bridge> @heinrich5991 you just wanna merge licenses together or what? At which point is a rewrite enough to decouple the license? What license is teeworlds even? Isn’t that bsd like too? How would I give proper credit? 15:08 <+bridge> Ah to get what character is on the tile 15:08 <+bridge> yeah 15:08 <+bridge> in "normal" case this is very fast, since tees usually dont stack 15:09 <+bridge> in worst case it's as bad as the current O(n²) plus the overhead of this idea 15:09 <+bridge> so in practice even a bit worse 15:09 <+bridge> ChillerDragon: teeworlds has zlib license. I just used that instead of your bsd license, because you wanted a simple solution. you're allowed to look at other code, preferably only to figure out how it works. copying the code requires you to stay with the original license (or a compatible one) though. you could e.g. create a new implementation by just looking at docs, e.g. libtw2 docs (this is one reason I created them) 15:11 <+bridge> is TeeAi mine? 15:11 <+bridge> gpl pog 15:11 <+bridge> (btw you are right, a grid works much better than a quadtree with the amount of moving things we have) 15:11 <+bridge> Yes you said assume it’s gpl 15:11 <+bridge> true 15:11 <+bridge> It has no license file 15:11 <+bridge> ill add the license 15:12 <+bridge> done 15:12 <+bridge> I guess I need this license in a bunch repos too now 15:12 <+bridge> Took me two days to rewrite Huffman-py to get rid of gpl in twnet\_parser 15:12 <+bridge> xd 15:13 <+bridge> And now heino says I’m violating licenses again -.- 15:13 <+bridge> is twnet_parser closed source 15:13 <+bridge> no 15:13 <+bridge> No but I want it to be bsd 15:13 <+bridge> https://en.wikipedia.org/wiki/Clean_room_design 15:13 <+bridge> It’s on Gitlab 15:13 <+bridge> woops, wrong link, but also good 15:13 <+bridge> https://gitlab.com/teeworlds-network/twnet_parser/ 15:14 <+bridge> why not use gpl 15:14 <+bridge> I want freedom for my users 15:14 <+bridge> GPL is good for users, bad for devs 15:14 <+bridge> Freedom includes closed source and mixing with other licensed code 15:15 <+bridge> thats what gpl does best for users 15:15 <+bridge> closed source removes freedom for users tho 15:15 <+bridge> > MIT/X license, BSD licenses (in the three-clause and two-clause forms, though not the original four-clause form), MPL 2.0, and LGPL, are GPL-compatible. 15:16 <+bridge> 0BSD, best for devs, good for users, but can be bad if used in closed source 15:16 <+bridge> The users of a network protocol library might want to build a closed source project. If they can not do it because it’s gpl they are less free. 15:16 <+bridge> He wants freedom for HIS users, not his users users 😄 15:16 <+bridge> only best for devs wanting to create closed-source projects 15:16 <+bridge> for devs wanting to create open-source software, gpl can be better 15:17 <+bridge> they have to know about limitations of GPL 15:17 <+bridge> but if you want to support devs wanting to create closed-source projects, then gpl is indeed incompatible with it 15:17 <+bridge> that alone makes it worse 15:17 <+bridge> I like code with no strings attached. Free to copy paste, edit. No obligations what do ever. 15:17 <+bridge> then why do you use the bsd license? 15:17 <+bridge> i guess we value different things 15:17 <+bridge> I like code with no strings attached. Free to copy paste, edit. No obligations what so ever. 15:17 <+bridge> that requires some things 15:17 <+bridge> I thought that’s the one xd 15:17 <+bridge> chiller use 0BSD 15:18 <+bridge> Finally something I kinda agree with chillerdragon on 15:18 <+bridge> use C0 15:18 <+bridge> Lmao 15:18 <+bridge> lmao 15:18 <+bridge> CC0 sucks bcs europe rights xd 15:18 <+bridge> 0BSD is my new love 15:18 <+bridge> xd 15:18 <+bridge> agpl is my love 15:18 <+bridge> https://opensource.org/license/0bsd 15:18 <+bridge> also coolest logo 15:18 <+bridge> also an accepted open source license 15:18 <+bridge> reference? I don't think so 15:18 <+bridge> https://cdn.discordapp.com/attachments/293493549758939136/1257686941105651763/AGPL-1.png?ex=66854fbd&is=6683fe3d&hm=b85ab9619f6d953afd90fefd52b5f34b8f620372a326b65da0861b04294e244f& 15:18 <+bridge> CC0 is compatible with europe's right. are you referring to "public domain" maybe? 15:19 <+bridge> uhrheberrecht 15:19 <+bridge> you cannot really do public domain 15:19 <+bridge> urheberrecht 15:19 <+bridge> 0BSD lgtm might use it in the future 15:19 <+bridge> bsd 3 clause is do whatever you want but keep this notice, and don't use my name as marketing ALSO YOU CANT SUE ME IF THIS DESTROYS THE WORLD 15:19 <+bridge> CC0 is not public domain fully but its like public domain 15:19 <+bridge> use llvm license 15:19 <+bridge> ew 15:19 <+bridge> the llvm dev live reaction: 15:19 <+bridge> the only thing i hate about llvm is the license and that they provide a C api on usage effort only 15:19 <+bridge> xd 15:20 <+bridge> > Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 15:20 <+bridge> > Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. 15:20 <+bridge> it must also go into the binary and you cannot remove attribution 15:20 <+bridge> llvm's license actually the only thing why it got so popular 15:20 <+bridge> "keep this notice" 15:20 <+bridge> already way worse than teeworlds' zlib license 15:20 <+bridge> we could have many gcc based compilers if not gpl 15:20 <+bridge> 0BSD or Apache license for me 15:20 <+bridge> well it got popular because apple uses it 15:20 <+bridge> its funded by apple too 15:20 <+bridge> and because gpl was anti-extensions 15:20 <+bridge> also because the code is more readable 15:20 <+bridge> No way you tought of licenses while creating docs xd 15:20 <+bridge> and modular 15:20 <+bridge> than gcc 15:20 <+bridge> i mean its more popular than you think 15:21 <+bridge> gcc u mean 15:21 <+bridge> yes, gcc 15:21 <+bridge> I did. I wanted libtw2 to be a clean-room implementation 15:21 <+bridge> llvm is rly popular yes 15:21 <+bridge> its a good piece of software 15:21 <+bridge> compiler wise its the epitome of research imho 15:21 <+bridge> like you can't make proprietary software if u set gpl as license 15:21 <+bridge> specially MLIR 15:21 <+bridge> llvm would die at its start 15:21 <+bridge> or someone would remake it 15:21 <+bridge> i dont think its true 15:21 <+bridge> gcc is gpl and its widely used 15:22 <+bridge> "way worse" is very subjective, the zlib license only allows binary distributions to not include the notice, that's all that's extra 15:22 <+bridge> gcc is gpl but u can compile propietary software btw 15:22 <+bridge> i don't really see many gcc extensions made by companies 15:22 <+bridge> the compiled stuff has nothing to do with gcc 15:22 <+bridge> im talking about tooling, not about compiling foreign programs 15:22 <+bridge> nvcc uses gcc 15:22 <+bridge> nothing subjective here. chillerdragon said he wanted the least restrictions for users. I'm not sure what you're talking about 15:22 <+bridge> intel's compiler is moving to llvm 15:23 <+bridge> cuz intel suxs 15:23 <+bridge> ofc companies will prefer llvm now 15:23 <+bridge> but u can get success with gpl 15:23 <+bridge> blender is GPL 15:23 <+bridge> blender is not llvm xd 15:23 <+bridge> The "way" is what makes it subjective. It is worse, exactly 1 clause worse 15:23 <+bridge> im talking about llvm and only llvm compilant tools 15:23 <+bridge> that wasnt the point 15:24 <+bridge> then let's say "it's obviously worse" 15:24 <+bridge> fair. then let's say "it's obviously worse" 15:24 <+bridge> did i say gpl kills every software it sticks to 15:24 <+bridge> i said llvm would die if used gpl 15:24 <+bridge> llvm license allowed stuff like nvvm to be born 15:24 <+bridge> nvvm is propietary blegh 15:24 <+bridge> https://docs.nvidia.com/cuda/nvvm-ir-spec/ 15:24 <+bridge> i like amd tends to be more open 15:25 <+bridge> I kinda doubt this. Have you ever touched the gcc codebase? 15:25 <+bridge> Imagine there was some bullshiting going on about antibot being closed source because teeworlds was super infectious lgpl++ which if linked against a struct of same shape statically or dynamically or even just through a reverse proxy forces you to publish source code as rom disk on request within 24 working days 15:26 <+bridge> ye and it stinks i know 15:26 <+bridge> llvm is easier to contribute and use thats the simple reason 15:26 <+bridge> I think only 10 people on earth would be able to do anything with that codebase, in comparison llvm is very pretty 😄 15:26 <+bridge> and its made to be used for other stuff than C based langs 15:26 <+bridge> could be prettier though 15:26 <+bridge> i think the effort can be commended 15:27 <+bridge> do u know how big llvm code is 15:27 <+bridge> its insane 15:27 <+bridge> ye i know 15:27 <+bridge>