IRC logs for #openttd on OFTC at 2025-12-17
⏴ go to previous day
01:01:40 *** tokai|noir has joined #openttd
01:01:40 *** ChanServ sets mode: +v tokai|noir
01:06:12 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
01:06:14 *** WormnestAndroid has joined #openttd
01:08:40 *** tokai has quit IRC (Ping timeout: 480 seconds)
01:30:48 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
01:31:19 *** WormnestAndroid has joined #openttd
02:04:07 *** herms2 has quit IRC (Quit: bye)
02:12:18 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
02:12:21 *** WormnestAndroid has joined #openttd
02:28:44 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
03:02:50 *** Flygon_ has quit IRC (Read error: Connection reset by peer)
03:04:43 *** WormnestAndroid has joined #openttd
03:55:23 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
04:00:46 *** WormnestAndroid has joined #openttd
04:03:31 *** Wormnest has joined #openttd
04:43:01 <DorpsGek> - Update: Translations from eints (by translators)
04:56:04 *** Zathras_7 has joined #openttd
04:59:36 *** Zathras_11 has quit IRC (Ping timeout: 480 seconds)
05:33:51 *** WormnestAndroid has quit IRC (Remote host closed the connection)
05:33:54 *** WormnestAndroid has joined #openttd
07:26:08 <peter1138> maxw defaults to INT32_MAX, maybe just changing that to INT16_MAX will work.
07:29:22 <peter1138> Probably not, the parts that call that won't be expecting line wrapping.
07:51:57 *** Smedles has quit IRC (Read error: Connection reset by peer)
07:52:51 *** Smedles has joined #openttd
08:38:32 <peter1138> Simpler than adding checks.
09:25:49 <_zephyris> Anyone more C++ than me want to check #14868?
09:25:58 <_zephyris> Looks good to me...
10:19:56 *** Zathras_4 has joined #openttd
10:28:14 <xarick> this new version of visual studio is crashrestarting a lot 🙁
11:38:52 <xarick> ``` auto value_iter = std::find_if(this->values.begin(), this->values.end(), [&](const auto &pair) {
11:38:52 <xarick> return pair.second == item;
11:41:13 <xarick> time to give up on the idea of making this->values the primary list
11:47:29 <peter1138> Yeah when you have to search through every time...
11:48:11 <xarick> mr copilot is always suggesting me Boost.MultiIndex
12:02:08 <_glx_> Don't listen to it, we already try to remove Boost from grfcodec
12:18:19 <jfkuayue> Did I become evil for using ChatGPT to plan my trip?
12:58:46 *** Smedles_ has joined #openttd
13:01:42 *** Smedles has quit IRC (Ping timeout: 480 seconds)
13:05:02 *** Smedles has joined #openttd
13:08:01 *** Smedles_ has quit IRC (Ping timeout: 480 seconds)
13:28:16 <jfkuayue> So this is plagiarism and art misconduct
13:30:16 <belajalilija> The way AI speaks gives me the ick
13:32:27 <belajalilija> Idk how to describe it other than being like the AI is trying to fellate you with it’s words
13:33:42 <ahyangyi> belajalilija: "alignment"
13:34:30 <jfkuayue> Although even if I did not generate this myself, I could know it is AI generated from the style
13:34:34 <belajalilija> ahyangyi: Optimised for winter daylight 🤓
13:34:52 <andythenorth> GPT 5.2 is horrific
13:34:56 <andythenorth> it's fully uncanny valley
13:35:27 <jfkuayue> Maybe this plays something
13:36:06 <jfkuayue> But the original text is also, hmm, very deliberately pleasing users
13:36:20 <belajalilija> I use ai sometimes to help me find info on rare or niche vehicles but i feel so for doing it, like when i flirt with someone in English
13:37:06 <andythenorth> and it's currently regressing
13:37:15 <belajalilija> It just leaves one feeling gross
13:37:23 <andythenorth> it has gone from "how surprising, this actually works and is useful"
13:37:49 <belajalilija> Yeah i think we’ve experienced the same thing
13:38:13 <belajalilija> In addition to the fact that the novelty has worn off for me and maybe you too
13:39:00 <belajalilija> Kinda like when you get drunk but then you wanna do something and you’re just like “okay, i wanna be sober now”
13:39:21 <andythenorth> most of what I want it for, would be solved if Apple Dictionary.app also had "robot chat" tab
13:39:31 <andythenorth> I basically want fancier thesaurus
13:39:38 <andythenorth> which is what half the algorithm is
13:39:48 <belajalilija> I mostly just use it as a more advanced Google
13:39:54 <andythenorth> yeah that's unwise
13:39:57 <andythenorth> although google is also bad
13:40:08 <merni> andythenorth: this is... an *interesting* format for a dictionary
13:40:16 <merni> Never seen usage example before the actual definition before
13:40:28 <merni> Oh it's a thesaurus nvm
13:40:38 <ahyangyi> I think it's pretty wise though, not because AI search is getting better, but because Google is getting worse.
13:40:59 <belajalilija> andythenorth: It gets me what i want, it’s better at finding those niche train facts that don’t appear on Google, i always check everything for sources too, i don’t trust ai to give me correct info
13:41:01 <ahyangyi> Like when reddit adds AI translation and pollutes search results with autotranslated stuff, and Google was... just cool with that
13:41:36 <ahyangyi> So that makes AI search actually useful because you can tell AI to filter out the AI slop in your search result
13:41:51 <ahyangyi> which the non-AI search engines cannot do
13:44:26 <belajalilija> belajalilija: It helps me be a bit more independent in my work given that I’m not as tech savvy as most here or know the best websites to find train info and it means i don’t have to feel like I’m pestering others
13:45:48 <belajalilija> I would stop using it in a heartbeat if Google was less shit though
13:48:05 <belajalilija> belajalilija: There was an option to make the ai mean to you, which i found better as it felt more human, but i can’t find it now
13:48:17 <_glx_> Auto AI translation for learn.microsoft.com is annoying too
13:48:27 <belajalilija> It turned the ai into a depressed gen x goth woman
13:49:20 <_glx_> And I use a browser extension to disable YouTube titles/descriptions auto translate too
13:49:46 <belajalilija> The AIification of YouTube is horrible yeah
13:50:13 <_glx_> It's just annoying when you watch videos in multiple languages
13:50:46 <_glx_> Many humans are not from the USA and know more than one language
13:50:56 <belajalilija> I do question if like
13:51:02 <belajalilija> We’re just getting old
13:52:20 <belajalilija> But at the same time
13:52:35 <belajalilija> I see shit the kids do like the six seven meme
13:52:54 <belajalilija> Because it’s the same nonsense we had just repackaged
14:06:00 <belajalilija> i've decided that when i have kids i'm going to learn their slang so when i pick them up from school i can embarrass them with it and really embrace the dad behaviour
14:12:42 <_zephyris> The enshittificaiton of Google is going fast now. You can search for "github <user> <repo>" and get no hits, despite being a public repo etc.
14:12:44 *** Smedles has quit IRC (Read error: Connection reset by peer)
14:14:36 *** Smedles has joined #openttd
14:25:42 <xarick> this->values.insert({value, item});
14:25:42 <xarick> I think emplace is faster
14:26:57 <xarick> theres also std::make_pair
14:29:03 <_zephyris> Toyland wetlands (rough ground)
14:33:49 <xarick> someone should make toyland a real tileset
14:34:26 <xarick> a tileset with openttd features
14:35:00 <xarick> newgrf as a base tileset, kind of approach
14:36:13 <xarick> oh no, you merged the wrong one
14:36:45 <ahyangyi> Mark wrong ones as drafts :S
14:37:04 <xarick> well still a faster keep list so its fine
14:37:16 <ahyangyi> Then it's not quite wrong 😛
14:37:37 <_glx_> RemoveList is just a little slower for extremelly weird test cases 😉
14:39:05 <LordAro> i'd be interested to see the codegen for that, i'd be surprised if they're any different at all
14:40:42 <xarick> tis merge will heavly conflict with my stuff
14:41:06 <ahyangyi> The updated version looks much cleaner.
14:41:10 <_glx_> RemoveList is different, old way was iterating the to "source", new way iterates the "target"
14:41:39 <ahyangyi> Algorithm changes typically cannot be "optimised away" 😛
14:42:24 <_glx_> KeepList is also different, used to be copy the list, remove the things to keep from the copy, remove the things in copy from target
14:52:19 <_zephyris> xarick: What does this mean?
14:53:09 <_glx_> newgrf can already redefine a baseset
14:54:12 <ahyangyi> So, did you mean a redesigned toyland, or to make the other three climates also toyland? 😛
14:55:49 <xarick> a tileset that could expose openttd features (usually newgrf features= that go beyond the regular tilesets.
14:56:48 <xarick> like... roads with speed limit, trams
14:57:04 <_glx_> there's newgrf for that
14:57:34 <_glx_> baseset can only replace TTD original sprites
14:58:01 <_zephyris> Well, there's very few protections, but it shouldn't do silly things
14:58:33 <_glx_> baseset is similar to static newgrf
14:58:35 <xarick> but something more official
14:59:03 <_zephyris> Do you mean like preset newgrf collections?
14:59:09 <_zephyris> It has been suggested
14:59:37 <_zephyris> Scenarios minus heightmap
14:59:42 <_glx_> since every user can use a different baseset it should not have any effects on the game logic
15:01:11 <xarick> oh RemoveItems is inside the hpp file?
15:01:50 <jfkuayue> Was oliver twist born in midland
15:02:38 <jfkuayue> Seventy miles from London, and he passed barnet
15:07:34 <xarick> i was actually considering
15:07:50 <xarick> for these keep/remove value
15:08:00 <xarick> iterate from the values list
15:08:29 <xarick> use lower_bound/upper_bound
15:08:53 <xarick> with the std::numeric_limits::max/min for the .second in the pair
15:10:31 <xarick> stuff is already sorted by value
15:11:03 <xarick> so it would save a few iteration cycles and comparisons
15:11:33 <xarick> it would have a range where to start and finish
15:25:53 <_zephyris> Wrote and trained a custom ML pixel art scaler to help with OpenGFX2... Good? Kinda... Faster than just pushing the pixels? Well...
15:48:39 <xarick> my regression test got stuck in an infinite loop 🙂
15:48:49 <xarick> how long does it take before it times out?
16:02:36 <xarick> oh, so i need to restart windows
16:03:07 <_jgr_> Why not just kill the process in the usual way?
16:03:19 <xarick> that's cheating but ok
16:03:40 <_jgr_> And restarting Windows isn't ? 😛
16:04:19 <LordAro> sounds like my coworkers - "unexpected build failure? let's just git clean -fxd the whole thing and start from scratch!"
16:05:52 <xarick> restarting visual studio didn't close it
16:06:18 <xarick> really have to end task 🙁
16:07:00 <_glx_> kill regression_regression.exe, not VS
16:07:25 <xarick> i forgot to advance iterator
16:07:35 <xarick> so it was stuck in an infinite
16:08:53 <LordAro> ...have you always restarted your computer after writing an infinite loop?
16:10:15 <rito12_51026> Does "git clean -fxd" mean git, clean with forced smile, please?
16:13:22 *** Wormnest has joined #openttd
16:13:47 <Rubidium> LordAro: don't you just buy a new computer?
16:19:06 <ahyangyi> rito12_51026: I force-push with `git push -fuq`.
16:21:36 <_jgr_> Using that habitually sounds like living dangerously to me
16:22:55 <ahyangyi> It's admitting f**k-ups, not a willingness to create more
16:27:41 <_glx_> but usually if regression is stuck in a loop it's easy to kill the process
16:28:35 <_glx_> for CI the timeout is important because we can't kill processes
17:09:49 <peter1138> "My computer was stuck in a loop, so I bought a new one"?
17:11:56 <xarick> from now on, i'm testing with garbage collection
17:14:24 <ahyangyi> peter1138: "keep buying computers until I break, that's how I learn about loops"?
17:41:26 <_zephyris> My old boss used to buy a new laptop when the desktop got full, and filed them in a bookcase
17:43:01 <LordAro> _zephyris: when the disk got full, rather than number of icons on the desktop
17:44:50 <_zephyris> He did understand that it wasn't literally full, but it was genuinely a deciding factor...
17:44:59 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
17:45:16 *** MinchinWeb[m] has joined #openttd
18:00:37 <xarick> i can finally get rid of my slow valuate branch!
18:01:47 <talltyler> I knew an amateur photographer who'd buy a new SD card every time she filled one up
18:04:14 <LordAro> if even you no longer see a use for it, i think we're probably fine without it
18:04:38 <xarick> yeah, now that valuate is suspendable 😐
18:04:54 <xarick> i don't need to hack into Valuate myself
18:10:08 <_zephyris> I'm sure young people judge me for buying spinning rust instead of using iCloud or something
18:14:27 <peter1138> Wasn't LordAro buying up stocks of 16TB drives the other day?
18:15:16 <peter1138> clang++: error: unable to execute command: Abort trap: 6
18:17:59 <Rubidium> it wouldn't amaze me when a retry of that just builds
18:22:42 <_glx_> oh even if clang++ didn't abort there's a unused variable 🙂
19:20:54 <ahyangyi> That post feels a bit confused
19:21:34 <ahyangyi> Uh, I meant the tt-forum one, though the other is probably more confused.
19:35:20 <andythenorth> did we delete gradual loading?
19:35:26 <andythenorth> or did I imagine it existed?
19:35:41 <andythenorth> my trains load one vehicle at a time, but OP in another channel they load all at once
19:35:52 <andythenorth> there's no setting for it
19:39:07 <peter1138> There is a setting for gradual loading.
19:39:24 <peter1138> There is also improved_load.
19:39:59 <peter1138> I don't know why we removed settings from appearing in Expert of advanced settings.
19:45:32 <andythenorth> players change them then report bugs
20:13:43 <xarick> some times don't make sense though
20:15:08 <xarick> i need to see how am i adding items
20:15:32 <xarick> aux_backup[GSBase.RandRange(range)] = 0;
20:15:39 <xarick> main_backup[GSBase.RandRange(range)] = 0;
20:16:32 <xarick> this is the thing i don't get about std::map/set
20:16:50 <xarick> clearing 18 million items took 1 second, but clearing 11 million takes 12
20:17:20 <xarick> ah, i remember I enabled garbage collection
20:22:42 <xarick> i forgot, but what is garbage collection for
20:34:03 <peter1138> Collecting garbage.
20:45:20 <rito12_51026> I think it is used for the same purposes as in other interpreted languages
20:48:01 <xarick> if objects are no longer needed, why collect them to delete them later, just delete them the moment they're not needed
20:52:34 <rito12_51026> When an object becomes unreachable the script might be doing an important task, to not slow it down the object is deleted in the "meantime"?
20:57:20 <_jgr_> If the performance test is just doing scriptlist stuff it should not be generating that many squirrel objects items that wound need to be traced by the GC
21:33:50 *** toktik has quit IRC (Remote host closed the connection)
21:53:43 <xarick> trying to rewrite how I execute the tests
21:54:56 <xarick> wanted to make them all sub 10 minutes max
21:55:30 <xarick> otherwise I waste all day waiting for results
22:05:23 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
22:08:58 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
22:09:33 *** WormnestAndroid has joined #openttd
22:47:23 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:58:28 *** toktik has quit IRC (Remote host closed the connection)
23:23:09 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
23:23:22 *** WormnestAndroid has joined #openttd
23:35:38 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
23:42:28 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
23:43:25 *** MinchinWeb[m] has joined #openttd
continue to next day ⏵
