IRC logs for #openttd on OFTC at 2024-06-16
⏴ go to previous day
00:32:39 <Eddi|zuHause> i'd have expected missing return being an undefined behaviour, but how is that resulting in an illegal instruction?
00:59:12 *** squirejames has quit IRC (Quit: User went offline on Discord a while ago)
01:34:01 *** fiddeldibu[m] has quit IRC (Remote host closed the connection)
01:34:12 *** yubvin[m] has quit IRC (Remote host closed the connection)
01:34:12 *** jact[m] has quit IRC (Remote host closed the connection)
01:34:12 *** xyii[m] has quit IRC (Write error: connection closed)
01:34:12 *** magdalena[m] has quit IRC (Write error: connection closed)
01:34:12 *** JamesRoss[m] has quit IRC (Write error: connection closed)
01:34:12 *** thomas[m] has quit IRC (Write error: connection closed)
01:34:12 *** tuxayo has quit IRC (Write error: connection closed)
01:34:12 *** karoline[m] has quit IRC (Write error: connection closed)
01:34:12 *** royills[m] has quit IRC (Write error: connection closed)
01:34:12 *** elliot[m] has quit IRC (Write error: connection closed)
01:34:12 *** Heiki has quit IRC (Write error: connection closed)
01:34:12 *** freu[m] has quit IRC (Remote host closed the connection)
01:34:12 *** menelaos[m] has quit IRC (Write error: connection closed)
01:34:12 *** nolep[m] has quit IRC (Write error: connection closed)
01:34:12 *** philip[m] has quit IRC (Write error: connection closed)
01:34:12 *** hamstonkid[m] has quit IRC (Write error: connection closed)
01:34:12 *** gdown has quit IRC (Write error: connection closed)
01:34:12 *** FelixActually[m] has quit IRC (Write error: connection closed)
01:34:12 *** audunm[m] has quit IRC (Write error: connection closed)
01:34:12 *** igor[m] has quit IRC (Write error: connection closed)
01:34:12 *** paulus[m] has quit IRC (Write error: connection closed)
01:34:12 *** throwawayaccount[m] has quit IRC (Write error: connection closed)
01:34:12 *** shedidthedog[m] has quit IRC (Write error: connection closed)
01:34:12 *** Rdolfs[m] has quit IRC (Write error: connection closed)
01:34:12 *** osvaldo[m] has quit IRC (Write error: connection closed)
01:34:12 *** joey[m] has quit IRC (Write error: connection closed)
01:34:12 *** Guest8757 has quit IRC (Remote host closed the connection)
01:34:12 *** Gadg8eer[m] has quit IRC (Write error: connection closed)
01:34:12 *** tonyfinn has quit IRC (Write error: connection closed)
01:34:12 *** thelonelyellipsis[m] has quit IRC (Write error: connection closed)
01:34:12 *** soylent_cow[m] has quit IRC (Write error: connection closed)
01:34:12 *** Farrokh[m] has quit IRC (Write error: connection closed)
01:34:12 *** phil[m]1 has quit IRC (Write error: connection closed)
01:34:12 *** luk3Z[m] has quit IRC (Write error: connection closed)
01:34:12 *** leward[m] has quit IRC (Write error: connection closed)
01:34:12 *** Heiki[m] has quit IRC (Write error: connection closed)
01:34:12 *** giords[m] has quit IRC (Write error: connection closed)
01:34:12 *** emilyd[m] has quit IRC (Write error: connection closed)
01:34:12 *** CornsMcGowan[m] has quit IRC (Write error: connection closed)
01:34:12 *** blikjeham[m] has quit IRC (Write error: connection closed)
01:34:12 *** Bilb[m] has quit IRC (Write error: connection closed)
01:34:12 *** jeeg[m] has quit IRC (Write error: connection closed)
01:34:12 *** grag[m] has quit IRC (Write error: connection closed)
01:34:12 *** patricia[m]1 has quit IRC (Write error: connection closed)
01:34:12 *** imlostlmao[m] has quit IRC (Write error: connection closed)
01:34:12 *** karl[m]123 has quit IRC (Write error: connection closed)
01:34:12 *** luffy[m] has quit IRC (Write error: connection closed)
01:34:12 *** zzy2357[m] has quit IRC (Write error: connection closed)
01:34:12 *** Elysianthekitsunesheher[m] has quit IRC (Write error: connection closed)
01:34:12 *** VincentKadar[m] has quit IRC (Write error: connection closed)
01:34:12 *** temeo[m] has quit IRC (Write error: connection closed)
01:34:12 *** jeremy[m]123 has quit IRC (Write error: connection closed)
01:34:12 *** einar[m] has quit IRC (Write error: connection closed)
01:34:12 *** cjmonagle[m] has quit IRC (Write error: connection closed)
01:34:12 *** citronbleuv[m] has quit IRC (Write error: connection closed)
01:34:12 *** andythenorth[m] has quit IRC (Write error: connection closed)
01:34:12 *** playback2396[m] has quit IRC (Remote host closed the connection)
01:59:19 *** Wormnest has quit IRC (Quit: Leaving)
02:14:02 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:01:24 *** gnu_jj_ has quit IRC (Ping timeout: 480 seconds)
03:17:44 <silent_tempest> Cool I finally got the language filtering working with Fontconfig
04:41:27 <DorpsGek> - Update: Translations from eints (by translators)
05:43:23 *** amal[m] has joined #openttd
05:58:32 *** HerzogDeXtEr has joined #openttd
06:15:35 *** gelignite has joined #openttd
07:54:30 *** SigHunter_ has joined #openttd
07:55:37 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
08:08:45 <silent_tempest> I mean I did see it
08:08:57 <silent_tempest> I don't understand some of the code you have there
09:16:02 *** nielsm has quit IRC (Ping timeout: 480 seconds)
09:47:44 <andriydohniak> same thing with clang btw
09:48:03 <andriydohniak> I think it intentionally generates the illegal instruction
09:48:26 <andriydohniak> if I make this c, instead of c++, it works fine, with no warning
09:49:50 <andriydohniak> It also does print "Res1" 10 times, I just forgot to put \n, so it doesn't flush the cache
10:05:33 <Rubidium> andriydohniak: be happy it's just an illegal instruction. It's defined as undefined behaviour in the C++ specification
10:06:00 <Rubidium> with an exemption for the main function, where omission implies returning 0
10:07:18 <Rubidium> the compiler would be allowed to just return 1 from `function`, making it faster. Practically it could also just add Res1 12 times after eachother and print that in one go.
10:10:52 <Rubidium> and that it works in C is just a happy accident, as missing the return yields undefined behaviour there as well
10:27:48 <andriydohniak> Rubidium: I am happy, but yestarday, I was almost sure that something went catastrophically wrong, and I corrupted the stack pointer. It was late, I was tired, and I thought I would have to debug a nasty memory corruption issue
10:28:18 <andriydohniak> Rubidium: I think C just emplicitly returns 0 in the end of the function
10:29:51 <andriydohniak> Yea, you are right
10:49:29 <andriydohniak> Btw, I used std::optional a bunch in my code, is that allowed?
11:25:53 <andriydohniak> Ok, I fixed all the bugs, and also now when the user clicks on the graph, the closest point gets selected, not just the one first determined to be in bounds
11:26:44 <andriydohniak> So now it is possible to easily select different lines that are close to each other
11:27:28 <andriydohniak> and the same line never gets selected 2 times in a row
11:27:49 <andriydohniak> so if 2 lines are realy close, and you selected the wrong one, click again and the other one will get selected
11:58:28 <_glx_> and if 3 lines are close ?
11:58:55 <andriydohniak> You have to toggle one of them off, or just be really precise
11:59:26 <andriydohniak> there is no clever system, but the closer to the cursor will be selecte
12:21:34 <andriydohniak> I want to attach a demo video to my PR, can github do videos? or do I have to transcode it as a gif
12:23:20 <andriydohniak> Anyway, here is the video in question
12:24:37 <andriydohniak> I forgot to show that filtering also works
12:38:48 *** techmagus has joined #openttd
12:45:22 <peter1139> Oof, that croissant was huge.
12:46:49 <truebrain> and off, I suck at writing Rust. Totally unreadable and horrible code I just produced 😛
12:47:04 <peter1139> C++ is better, see.
12:47:04 <truebrain> When in doubt, just add `Arc<Mutex<>>` around things 😄
12:56:34 <andriydohniak> truebrain: What are you doing in rust?
12:56:48 <andriydohniak> Because in a lot of cases you can just use channels
13:03:14 <andythenorth> truebrain: Did you try GPT? 😛
13:09:11 <andriydohniak> andythenorth: GPT is HORRIBLE with rust
13:09:17 <andriydohniak> Absolutelly HORRIBLE
13:09:26 <andriydohniak> beyond a hello world, it sucks!
13:10:39 <andythenorth> Do we know what it’s good with? Python is good. AWS config seems good. I assume it’s good at js / node. HTML could be better imho
13:11:21 <andythenorth> It’s quite neat watching it write python then execute it
13:11:43 <andriydohniak> andythenorth: It is pretty good at doing small scope + popular things
13:12:05 <andriydohniak> if the python codebase has a bunch of big classes, it will suck
13:12:30 <andriydohniak> but building a no custom code < 100 loc python code is not that hard for it
13:12:46 <andythenorth> HTML, it’s much worse at traversing the DOM than it should be, given what it can do now with xml parser
13:12:54 <andriydohniak> also it sucks with smaller rust crates / rust macros, because the syntax might change
13:13:24 <andriydohniak> andythenorth: Well HTML DOMs are kinda big, so we run into the context issue
13:13:37 <andythenorth> How long has it been able to write code then run it? I only saw that this week
13:13:57 <andriydohniak> andythenorth: For a long time, ~3-6 month now
13:14:09 <andriydohniak> Basically when the GPT applets were introduced
13:14:45 <andythenorth> I didn’t see it until I bought API credit, although that might be coincidence
13:15:12 <andriydohniak> andythenorth: I think you had to enable it
13:15:17 <andriydohniak> but now it's on by default
13:16:23 <andriydohniak> cmon, approve me, quickly so you don't see the quality of my code! 🤣
13:26:29 <LordAro> andriydohniak: i almost want to ask you to do a self-review
13:26:45 <LordAro> look at the diff, see what you can see
13:26:52 <LordAro> because i can see a lot :p
13:27:10 <andriydohniak> that is not good...
13:28:43 <andriydohniak> Ok, I will collect a list of stuff I think is wrong, and you will add what I miss, ok?
13:28:57 <talltyler> Commit checker is failing for both commit messages and for invalid tab usage
13:29:17 <talltyler> You can click on it in the actions report to see what it doesn’t like, same with unused strings
13:29:39 <andriydohniak> I was told to let the unused strings fail, and correct it later
13:30:28 <talltyler> Yes, I think we have reached “later” 😄
13:32:42 <andriydohniak> is auto allowed in for loops?
13:33:46 <andriydohniak> ok, also forgot to put const where possible :)
13:37:30 <andythenorth> Can we teach GPT our code style?
13:37:30 *** gelignite has quit IRC (Read error: Connection reset by peer)
13:37:45 <andythenorth> Then let it review 😛
13:38:21 <andriydohniak> andythenorth: No need, clang-format exists
13:38:40 <andriydohniak> and I think you can add custom lints to it or cppcheck, but I don't remember
13:39:18 <LordAro> clang-format has no idea what to do with our gui definitions
13:40:14 <andythenorth> It’s not that I think GPT *can* do all these things. It’s just curiosity.
13:41:03 <andythenorth> And I seem to have a thing about pushing some technologies to find what breaks 😛
13:42:11 *** gelignite has joined #openttd
13:45:10 <andriydohniak> LordAro: I think you can put a special comment that disables it for a section
13:46:05 <andriydohniak> //clang-format off
13:46:05 <andriydohniak> <indent how you want here>
13:46:05 <andriydohniak> //clang-format on
13:46:16 <LordAro> yeah, but it's a pain to have to do so for every one of them
13:46:31 <andriydohniak> that is true, but it is still nice once you have it
13:46:32 <LordAro> there's a load of other things like that too
13:46:40 <LordAro> last time anyone checked, anyway
13:46:57 <andriydohniak> It is pretty configurable
13:47:22 <andriydohniak> but yea, for example the saveload tables will probably also be wrong
13:47:36 <andriydohniak> but I think it can allign them automatically
13:48:21 <LordAro> iirc there were also issues with OTTD not really having a line length limit
13:48:47 <LordAro> you set no limit, and it makes all lines as long as possible, you set any other limit, and it rewrites the other half of the code
13:49:09 <LordAro> it's been a while since i tried though, they seem to add new options every release
13:56:07 <andriydohniak> I don't have much experience with it either
13:56:13 <andriydohniak> What is going on here?
13:56:25 <peter1138> You're loading a float with a double.
13:56:36 <andriydohniak> so 0.85 is a double?
13:56:59 <andriydohniak> So there is no way to define a constant float?
13:57:04 <andriydohniak> or do you have to cast it
13:57:30 <andriydohniak> `const float cargo_rect_scale_factor = (float)0.85`
13:57:41 <andriydohniak> but I can just make it a double
13:57:52 <peter1138> There is, but you've used a constant double not a constant float.
13:58:33 <andriydohniak> Ahh, you can put a `f` at the end
14:05:54 <andriydohniak> and that does not include the CI warnings, add that to the list
14:06:00 <andriydohniak> I think I finished reviewing my own code
14:07:13 <LordAro> the other main codestyle one is that opening function { should be on its own line
14:07:30 <andriydohniak> Ahh, I didn't see that
14:08:00 <andriydohniak> I am gona go touch some grass, go for a walk, come back, and fix all the stuff I found
14:08:49 <andriydohniak> LordAro: is that a code suggestion? public? or pub, like go for a drink?
14:09:12 <peter1139> 300 mile bike ride later...
14:09:15 <andriydohniak> but I will make myself a cup of tea
14:09:51 <andriydohniak> peter1139: I love biking, but here in Madrid the infrastructure for it is worce then my code
14:10:11 <peter1139> As a roadie, roads are infrastructure :)
14:10:35 <andriydohniak> I am not brave enough to ride on public roads
14:11:25 <_glx_> it's sometime safer to be on plain road than on badly designed dedicated lanes
14:11:45 <andriydohniak> yea, if it's not well separated I am out
14:18:10 <LordAro> i find that's a somewhat self-fulfilling attitude
14:21:13 <andriydohniak> not sure what you are talking about, but if you are talking about biking, I really can't drive close to the cars, because I don't know the driving rules, and if it's not well separated, I need to have a drivers license to drive near cars, just to know what they are about to do
14:22:09 <LordAro> learning the rules of the road is not the same as learning to drive
14:22:30 <andriydohniak> I am only 19, and I hope that If I get a job, I will be able to buy myself a good bicycle, and just use it on park roads to relax
14:22:33 <peter1139> Plenty of drivers don't know the rules of the road.
14:23:00 <LordAro> assume they're going to do something stupid until proven otherwise
14:23:07 <peter1139> Like the one this morning, overtaking me on solid double white lines around a blind bend.
14:23:18 <LordAro> peter1139: ah, they must have the see-round-corners package
14:23:26 <LordAro> lots of people have that
14:24:21 <LordAro> regardless, more people on bikes is a good thing, because it means fewer people in cars
14:24:27 <LordAro> which also makes things better for those in cars
14:24:59 <andriydohniak> Also the roads around the area where I live are pretty big, so drivers don't like to slow down much
14:25:19 <andriydohniak> and there is a culture of night time drunk driving
14:25:32 <andriydohniak> but some people like to start early so that is also a big danger
14:31:31 *** toktik is now known as Guest9793
14:31:55 *** Guest9793 has quit IRC (Remote host closed the connection)
14:53:35 <merni> peter1139: heh, thinking lines mean anything
14:54:02 <merni> I'd never survive a week as a pedestrian here if I expected drivers to follow road markings :p
15:17:40 <talltyler> Not many corners in the grid-pattern city where I live, but every car and delivery truck comes equipped with “park in the bike lane” flashing lights
15:56:23 <johnfranklin> Driving on highways, not flashing when changing lanes is a norm.
15:57:35 <peter1138> Flashing sounds dangerous. I just indicate.
15:59:07 <_glx_> drivers not using indicators seems to be correlated with the cost of the car
15:59:27 <johnfranklin> I mean the left/right turn flashlight, which is hardly to be seen to be used here when changing lanes
15:59:57 <_glx_> or when they use it they already started the move
16:00:10 <andythenorth> classic, in order to avoid writing Squirrel, I've now refactored core FIRS python
16:00:20 <andythenorth> definitely not a rabbit hole
16:05:23 <_glx_> since 2021 vehicles >3.5T have stickers related to blind spots, but cars and bikes still don't take care of it
16:17:08 *** YourOnlyOne has joined #openttd
16:17:28 *** YourOnlyOne has quit IRC ()
16:22:54 *** YourOnlyOne has joined #openttd
16:23:01 *** techmagus has quit IRC (Quit: Leaving)
16:23:23 *** YourOnlyOne is now known as Guest9804
16:25:19 *** techmagus has joined #openttd
16:25:37 <johnfranklin> “Bicycles are often stolen in United Kingdom”, is it true?
16:25:53 *** techmagus is now known as Guest9805
16:26:47 *** YourOnlyOne[m] has joined #openttd
16:29:58 <pickpacket> Aren't bicycles often stolen everywhere?
16:40:43 <peter1139> 826/1240 monsters. Getting there.
16:40:57 <peter1139> At least with this map I can get them in-fighting a bit.
16:53:41 <michalc5706> Right now it can fetch the latest version of a piece of content (and its dependencies). But I would like to make it expose the full MD5 of content, and offer functionality that allows someone to use a previously extracted md5 to fetch the exact same version, even if newer versions of that pieces of content have been released.
16:53:41 <michalc5706> I guess I'm asking - would this be ok? Could it even depend on the license of the piece of content? e.g. I would have thought that if it's GPL v2-ed for example then it's all good, but it'll be murkier if it's "Custom" say. It looks like the license is available via the API, so it could certainly behave differently depending on the license.
16:57:57 <peter1139> _glx_, problem is that should be a warning for the lorry driver. They are the ones in control of the badly designed vehicle, after all.
16:59:18 <_glx_> I think they know their blind spots, and it's impossible for them to notice when someone is there
17:00:36 <_glx_> educating others about this fact is important too
17:17:44 <peter1139> 1240/1240 dead, nice.
17:18:26 <andriydohniak> Is it a hard rule to spell color as colour in the codebase 🤣
17:18:46 <peter1139> No, it's very easy, you just spell colour correctly.
17:29:15 <peter1139> British game, British spelling.
17:30:12 <andriydohniak> Yea, this is the my first time seeing a British codebase, and I can say I am happy to code in here
17:31:58 <peter1139> Why did you tag your PR with [translation]?
17:33:44 <andriydohniak> JustLinuxUserviaGitH: I still need to rename the commits
17:33:48 <andriydohniak> don't run CI yet
17:34:00 <andriydohniak> peter1139: because I touched english.txt file
17:34:06 <andriydohniak> I added more strings there
17:34:10 <andriydohniak> and removed some
17:34:26 <peter1139> That doesn't require a tag.
17:34:39 <andriydohniak> Ok, gona remove it then
17:39:09 <andriydohniak> Ok, now *I think* the commit names are ok
17:39:14 <andriydohniak> and everything else too
17:43:53 <andriydohniak> Wow! your CI is amazing! I didn't even notice that you can just play the PRed game in the browser with wasm!
17:51:51 <truebrain> _glx_: Impossible how? By fitting sensors? By having "auto-stop" functionality in their vehicles?
17:51:51 <truebrain> Always felt to me that they were just too cheap to invest in the problem, so instead they "educated" the rest of the world. Such an inside-out solution 😛
17:53:41 *** ufo-piloot_ has quit IRC ()
17:53:45 <andriydohniak> The GCC CI failed, but it's not my fault
17:54:22 <andriydohniak> What are Annotations?
17:54:56 *** ufo-piloot has joined #openttd
17:55:08 <peter1138> Warnings and errors.
17:55:52 <peter1138> In graph_gui.cpp you've put includes in the wrong order.
17:55:59 <peter1138> sfdafx.h must come first.
17:56:18 <Rubidium> andriydohniak: two things... stdafx.h must be the first include and safeguards.h must be the last include
17:56:54 <peter1138> Ah yes, that one was removed, probably because it's "not used" which is the point of the safeguards 🙂
17:56:55 <Rubidium> the fact that the others work is because of precompiled headers messing with things, which is why we disabled precompiled headers on one of the faster platforms
17:58:20 <andriydohniak> Should I do it in every file?
17:58:42 <peter1138> Your own files should include stdafx.h and safegaurds.h, yes.
17:59:02 <peter1138> We also don't usually include system headers in each file, because stdafx handles most of them.
17:59:33 <andriydohniak> What do you mean system headers?
17:59:53 <andriydohniak> Ahh, like stlib?
17:59:56 <peter1138> Anything included with <> instead of ""
18:00:32 <peter1138> Otherwise we'd need a huge list of them in every source file...
18:07:31 *** YourOnlyOne[m] is now known as YourOnlyOne
18:15:52 <andriydohniak> I am encountering a bunch of warnings, like this one, but not in the code I changed.
18:26:16 <andriydohniak> and I checked, and the master branch just doesn't have those errors
18:26:21 <andriydohniak> what is happening?
18:27:03 *** Flygon has quit IRC (Read error: Connection reset by peer)
18:33:23 *** virtualrandomnumber has joined #openttd
18:33:37 *** virtualrandomnumber has quit IRC ()
18:44:07 <andriydohniak> It was the extra imports I added ...
18:44:31 <andriydohniak> I didn't know how the imports worked in your project, and that you had a very specific system on how to manage it
18:45:53 <peter1138> Ah, the C++20 feature that we don't use.
18:48:42 <andriydohniak> I don't think I did imports
18:52:27 <andriydohniak> How do I run the Annotations workflow locally?
18:58:53 <Rubidium> the annotations workflow on itself is not that interesting, it's the input of the annotations workflow that's important... and that's the different compiles since the annotation workflow just parses the build logs
18:59:46 <Rubidium> I think the whole build might also be running when you create a PR in your own fork, so you could use that for some experimentation
19:01:42 *** yiffgirl has quit IRC (Ping timeout: 480 seconds)
19:01:42 *** kamnet has quit IRC (Ping timeout: 480 seconds)
19:01:42 *** alfagamma7 has quit IRC (Ping timeout: 480 seconds)
19:01:47 *** godfther has quit IRC (Ping timeout: 480 seconds)
19:01:47 *** talltyler has quit IRC (Ping timeout: 480 seconds)
19:01:47 *** _jgr_ has quit IRC (Ping timeout: 480 seconds)
19:01:52 *** locosage has quit IRC (Ping timeout: 480 seconds)
19:01:52 *** georgevb has quit IRC (Ping timeout: 480 seconds)
19:01:53 *** yeah_its_gloria has quit IRC (Read error: Connection reset by peer)
19:01:53 *** ahyangyi has quit IRC (Read error: Connection reset by peer)
19:01:53 *** sherlockholmes1904 has quit IRC (Read error: Connection reset by peer)
19:01:53 *** andriydohniak has quit IRC (Read error: Connection reset by peer)
19:01:53 *** truebrain has quit IRC (Read error: Connection reset by peer)
19:01:53 *** kuhnovic has quit IRC (Read error: Connection reset by peer)
19:01:53 *** wensimehrp has quit IRC (Read error: Connection reset by peer)
19:01:53 *** DorpsGek_vi has quit IRC (Read error: Connection reset by peer)
19:01:53 *** binglunge has quit IRC (Read error: Connection reset by peer)
19:01:53 *** merni has quit IRC (Read error: Connection reset by peer)
19:01:53 *** michalc5706 has quit IRC (Write error: connection closed)
19:01:53 *** jfs has quit IRC (Write error: connection closed)
19:01:53 *** andythenorth has quit IRC (Write error: connection closed)
19:01:53 *** _glx_ has quit IRC (Write error: connection closed)
19:01:53 *** emperorjake has quit IRC (Write error: connection closed)
19:01:53 *** audigex has quit IRC (Write error: connection closed)
19:01:53 *** _zephyris has quit IRC (Read error: Connection reset by peer)
19:01:53 *** salut3585 has quit IRC (Write error: connection closed)
19:01:53 *** stewardcollin has quit IRC (Write error: connection closed)
19:01:53 *** notluke2578 has quit IRC (Write error: connection closed)
19:01:53 *** reldred has quit IRC (Write error: connection closed)
19:01:53 *** exceptik has quit IRC (Write error: connection closed)
19:01:53 *** j0anjosep has quit IRC (Write error: connection closed)
19:01:53 *** peter1138 has quit IRC (Read error: Connection reset by peer)
19:01:53 *** silent_tempest has quit IRC (Write error: connection closed)
19:01:53 *** bootmii has quit IRC (Write error: connection closed)
19:01:53 *** muxydugoulp has quit IRC (Write error: connection closed)
19:01:53 *** bigyihsuan has quit IRC (Write error: connection closed)
19:01:53 *** ahym has quit IRC (Write error: connection closed)
19:01:53 *** michi_cc has quit IRC (Write error: connection closed)
19:01:53 *** johnfranklin has quit IRC (Write error: connection closed)
19:02:14 *** DorpsGek_vi has joined #openttd
19:03:21 *** andriydohniak has joined #openttd
19:03:21 <andriydohniak> Ok, I ran the same commands as the CI runs (sorry for wasting your resources) and I think we are good this time
19:13:46 <dwfreed> andriydohniak: github actions is free for open source, so the only resources you're wasting are github's :D
19:18:39 <andriydohniak> Ok, I thought you were paying for it
19:18:45 <andriydohniak> I feel much better now
19:28:18 *** andythenorth has joined #openttd
19:28:18 <andythenorth> FIRS compile (with primed caches): 8 seconds
19:29:48 <andriydohniak> Can somebody with permissions approve the CI
20:16:10 <andriydohniak> It was a type conversion issue that only windows compiler noticed :(
20:27:40 <andriydohniak> Yea, the CI id definitely not free, too many runners and I think some dedicated servers as well
20:29:48 <dwfreed> no, github actions is free
20:33:16 <andythenorth> can I actually GPL v2 something GPT has written?
20:34:31 <dwfreed> you cannot validate the original copyright of the data it pulled from, because you don't even know the data it pulled from
20:34:55 <andythenorth> "does OpenAI grant copyright permissions to use this method under GPL v2?"
20:35:03 <andriydohniak> dwfreed: github actions are free for open source projects, but it is very limited, and I think this CI uses more then github actions, and github actions just trigger something else
20:35:14 <andythenorth> "Yes, OpenAI grants you the necessary permissions to use, modify, and distribute the code provided by me under the terms of the GPL v2 license. This means you can incorporate the method into your GPL v2 project without any legal issues. Here is a statement to clarify the licensing for the code provided:"
20:35:29 <dwfreed> andythenorth: and that answer could be completely pulled out of its ass
20:35:54 <andythenorth> OpenAI lawyers would have to account for that
20:36:12 <andythenorth> they could have guard-railed this question
20:36:20 <dwfreed> but sure, take *legal* advice from something that's not any better than a few million matrix multiplications
20:36:26 <andythenorth> License Grant for Code Provided by OpenAI's ChatGPT
20:36:26 <andythenorth> The code snippets and methods provided by OpenAI's ChatGPT in this interaction are licensed to you under the terms of the GNU General Public License, version 2 (GPL v2). You are free to use, modify, and distribute this code in accordance with the terms of the GPL v2.
20:36:33 <andythenorth> product warranty
20:36:40 <andythenorth> I wonder what the ToS say
20:37:07 <andriydohniak> the debate is not weather AI code is GPLv2, it is weather AI code can be proprietary
20:37:36 <_jgr_> I expect that OpenAI would "grant" you anything you wanted if you posed the question right
20:38:12 <andriydohniak> Yea, you can't rely on changGPTs answer about anything
20:38:30 <andriydohniak> but in this case it is true
20:41:37 <andythenorth> comes down to extent of applicable law
20:42:01 <andythenorth> and whether there is a copyright here that I can rely on for a copyleft license
20:44:27 <dwfreed> either ask a real lawyer, or just don't
20:44:34 <dwfreed> like, this is the problem with AI
20:45:28 <_glx_> andriydohniak: no, we just use actions for the CI, nothing external
20:45:41 <dwfreed> technically the preview deployment is cloudflare pages
20:45:55 <_glx_> but preview build is hosted on a paid service yes
20:45:56 <dwfreed> but that's a fixed cost, not variable based on how much it's used
20:46:51 <_glx_> anyway preview is only enabled for PR when appropriate
20:47:20 <andythenorth> dwfreed: no case law, no precedent, no UK lawyer is going to want to advise on this for a hobby project, unless they're particularly fascinated by GPL
20:47:47 <dwfreed> gee, I wonder why that is
20:47:53 <dwfreed> because it's a legal minefield
20:48:12 <andythenorth> not really, it's just awaiting case law
20:48:25 <_glx_> all AI stuff is legal minefield
20:49:20 <dwfreed> AI is ultimately creating derivative works of items it does not have license to create derivatives of, and the parody fair use exemption does not apply
20:50:24 <andythenorth> the OpenAI terms put the responsibility on the user
20:50:47 <dwfreed> exactly, and it's legally impossible to actually resolve that
20:50:57 <_glx_> the only safe way is to train your AI only with your own work
20:51:14 <dwfreed> because you can't even know all of the data sources it derived its answer from
20:51:42 <andythenorth> can the plaintiff know for certain that their work was the source of a derivative work?
20:51:57 <andythenorth> (yes, in some cases)
20:52:01 <andriydohniak> Can _glx_ run the free CI? :)
20:52:02 *** nielsm has quit IRC (Ping timeout: 480 seconds)
20:52:14 <andriydohniak> the payed one is already finished
20:52:16 <andythenorth> GPT will reproduce, e.g. scripts from films and novels quite accurately right now
20:52:48 <andythenorth> anyway, I'm GPL-ing python and squirrel from GPT
20:53:04 <andythenorth> it will be an interesting legal case, in an alternative universe
20:53:20 <andriydohniak> Yea, but It is only trained on open source material, so if it reproduces a chnk of that in another open source GPL project it's fine, the problem is if it reproducec gpl code in a non gpl project
20:53:40 <andythenorth> it's trained on more than open source
20:53:54 <andythenorth> that's why it can accurately tell jokes about Hitchhikers Guide to The Galaxy
20:54:05 <_glx_> andriydohniak: when it's the first PR github is "annoying", requiring manual validation before run, but it's to prevent abuse from yet unknown contributors
20:54:28 <_glx_> you can also open a PR on your own fork to run the CI
20:55:12 <andriydohniak> _glx_: annoying! That is me! :)
20:55:42 <andriydohniak> This will teach me to carefully think about int data types. I was spoiled by rust
20:56:18 <andriydohniak> and includes too
20:56:23 <andriydohniak> I am soo spoiled
20:56:38 <_glx_> anyway this github safety things mostly applies to organisations using paid action plans
20:56:59 <_glx_> it's just a side effect for free plan users
20:57:47 <andythenorth> I wouldn't declare methods-in-methods, some python people get fussy about it, and I find it harder to reason about
20:59:11 <_glx_> oh nml declares full class in methods
21:06:48 <andriydohniak> Why is CodeQL's build so much slower then all the other builds?
21:17:26 *** SigHunter has joined #openttd
21:17:43 <andriydohniak> What is github codeql? And is it gona roast me?
21:18:02 *** gelignite has quit IRC (Quit: Stay safe!)
21:18:40 <andriydohniak> Can I run the openttd's codeql hook locally?
21:19:43 <andriydohniak> ok, I can run it locally
21:21:12 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:29:59 <andriydohniak> YESSS!!! Green CI
21:58:21 *** YourOnlyOne has quit IRC (Quit: WeeChat 4.4.0-dev)
22:12:05 *** wensimehrp has joined #openttd
22:12:05 <wensimehrp> tt-wiki down? loading is slower than usual
22:13:59 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:22:31 *** truebrain has joined #openttd
22:26:50 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
22:33:36 *** talltyler has joined #openttd
22:33:36 <talltyler> Probably a feature, yes 🙂
22:51:48 *** toktik has quit IRC (Remote host closed the connection)
continue to next day ⏵
