IRC logs for #openttd on OFTC at 2025-09-19
⏴ go to previous day
00:19:41 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
02:27:22 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
03:00:08 *** firepup has quit IRC (Remote host closed the connection)
03:00:21 *** firepup650 has joined #openttd
03:30:12 *** Smedles has quit IRC (Ping timeout: 480 seconds)
03:32:13 *** Smedles has joined #openttd
03:35:42 *** Smedles_ has joined #openttd
03:42:03 *** Smedles_ has quit IRC (Read error: Connection reset by peer)
03:42:33 *** Smedles has quit IRC (Ping timeout: 480 seconds)
03:42:52 *** Smedles has joined #openttd
03:47:45 *** Zathras_11 has joined #openttd
03:51:21 *** Zathras has quit IRC (Ping timeout: 480 seconds)
04:26:04 *** SigHunter_ has quit IRC (Remote host closed the connection)
04:27:04 *** SigHunter has joined #openttd
04:38:33 <DorpsGek> - Update: Translations from eints (by translators)
05:10:13 *** rito12_51026 has joined #openttd
05:10:13 <rito12_51026> I have a question, why custom grf signalls like the JGRPP ones are not implemented in vanilla? Is there any open discussion about that?
05:26:41 <reldred> They’re dependent on a bunch of extra JGRPP features
05:26:53 <reldred> Multi aspect for instance needs realistic braking.
05:27:54 <reldred> And I dare say, most ottd players don’t seem that interested, and us freaks and weirdos who are just play JGRPP.
06:08:16 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
07:33:03 <peter1138> Hmm, should 1.75x be 2x sprites scaled down or 1x sprites scaled up...
07:44:11 *** lobstarooo has quit IRC (Ping timeout: 480 seconds)
07:48:06 <pickpacket> peter1138: try both and see which looks best?
07:48:26 <pickpacket> I guess they're closer in size to 2x?
07:51:58 <peter1138> If the artist has drawn 2x sprites then they'll have more detail than 1x sprites. Therefore, scaling it down will lose some of that detail.
07:53:51 <reldred> yeah I'd probably say don't start using 2x until you hit 2x
07:54:06 <pickpacket> of course. But 1x sprites can't be cleanly resized to 1.75x either.
07:54:55 <reldred> agreed, if I had my way it would be true integer scaling or btfo but I'm not the one writing the patches.
08:01:24 <peter1138> That basically means only adding 3x sprites, and that's only if you consider 1x sprites to be canonical.
08:02:03 <peter1138> Given OpenGFX 2, which has 1x, 2x and 4x sprites in some cases, 3x is only "true integer" if you ignore the fact that 2x and 4x sprites exist.
08:02:28 <peter1138> Scaling 2x to 3x or 4x to 3x is not "true integer"
08:03:04 <peter1138> You'd get 100% - 1x, 200% - 2x, 300% - 1x (*3), 400% - 4x.
08:03:45 <peter1138> 2x * 1.5 is 3x, or 4x * 0.75 is 3x, but neither are integer.
08:08:04 <reldred> Eh, 1x *3 is all I want 😛
08:13:13 <peter1138> lol at OpenGFX 1 using completely different images for some 1x vs 2x sprites.
08:19:13 <locosage> peter1138: scale both and mix them 3:1 🤭
08:36:11 <peter1138> Hmm, 1.5x is just 1x * 3 = 3x then 3x / 2 = 1.5x. All integer... ;p
08:39:50 <locosage> in all integer 3/2 = 1 :p
08:43:51 <peter1138> Ignore padding, or not ignore padding...
08:43:59 <peter1138> (On the cargo icon sprites.)
08:51:02 <peter1138> ((Spot the difference...))
08:57:37 <LordAro> are you sure there's anything different there at all?
09:02:22 *** firepup has joined #openttd
09:04:14 *** firepup650 has quit IRC (Ping timeout: 480 seconds)
09:09:28 <mmtunligit> left side looks better, coal and goods are weirdly spaced on the other
09:09:57 <mmtunligit> also steel and valubles
09:30:19 <_zephyris> peter1138: No scaling sprites please! Neither will look very nice...
09:30:38 <peter1138> Oh jesus fucking christ.
09:30:49 <_zephyris> @me or someone else?
09:31:25 <_zephyris> (no non-integer scaling of sprites)
09:48:34 <peter1138> Might as well remove 2x and 4x icons from OpenGFX 2, otherwise everything gets very inconsistent.
09:49:13 <peter1138> (Not quite as inconsistent as with OpenGFX 1)
09:57:46 *** ahyangyi has joined #openttd
09:57:46 <ahyangyi> Where are the UI scale applied to?
09:58:02 <ahyangyi> I am under the vague impression that even the station purchase window uses the UI scale in some way.
10:00:01 <ahyangyi> Would it be a solution to allow NewGRFs (basesets included) to provide sprites for all UI scales, wherever it may be used?
10:00:50 <ahyangyi> .. my search-and-replace-fu needs improvement
10:04:15 <peter1138> That's notsdfsdf
10:06:55 *** peter1138 has quit IRC (Read error: No route to host)
10:08:30 *** peter1138 has joined #openttd
10:08:30 *** ChanServ sets mode: +o peter1138
10:09:03 <peter1138> Well. Something make something unhappy.
10:10:44 <peter1138> "all UI scales". Interface scaling is adjustable in 1% increments from 100% to 500%. So no, not really.
10:12:21 <ahyangyi> I was under the impression it was multiples of 25%, sorry.
10:12:27 <peter1138> You can argue that only 25% is relavant because that's what automatic scaling, and using the slider without holding ctrl, limit it to.
10:13:08 <peter1138> Some of you may have a TIL experience here.
10:13:14 <ahyangyi> Anyways, 401 valid values is somewhat within what a switch can do (since nobody would really make 401 different entries)
10:14:04 <peter1138> What does a switch have to do with anything?
10:14:40 <ahyangyi> A switch to choose a version, which OpenTTD then scales to the desired size
10:16:35 <ahyangyi> Since so much of the previous discission is "Should we use the 1x or 2x sprite for the 1.69x UI?"
10:17:05 <ahyangyi> I think we can pick a somewhat sensible default, and if there is desire from NewGRF authors to finetune the behavior, provide a switch interface for that.
10:17:29 <ahyangyi> This way there is no pressure in getting a "perfect" default behavior right now.
10:17:36 <peter1138> I absolutely think leaving this to NewGRF authors is entirely the wrong thing.
10:17:51 <peter1138> Different sets will do different things (most likely nothing.)
10:18:09 <andythenorth> and we haven't discussed railtypes today either 😛
10:18:32 <ahyangyi> Most fonts come with fine-tuned hinting the authors create to ensure they look right at specific px sizes
10:19:12 <peter1138> And if they don't the font is scaled anyway.
10:19:12 <ahyangyi> So I think we can draw some parallels here and admit that graphical creators do care about how their creations get scaled
10:19:26 <ahyangyi> I am not against automatically scaling.
10:19:43 <ahyangyi> And I think scaling is mandatory.
10:20:06 <peter1138> If graphics artists want to provide custom intermediate sizes, then that's no problem, they can.
10:20:16 <ahyangyi> The only thing I am proposing is to give NewGRFs the ability to choose the sprite from which to scale, not whether to scale.
10:20:27 <peter1138> But if we start restricting it to "only integer scaling" then that means that even that is inconsistent.
10:21:54 <ahyangyi> I think that kind of inconsistency is tolerable, since we are talking about scaling 16x16 and 32x32 sprites
10:22:24 <ahyangyi> There aren't infinitely many ways between scaling a 16x16 sprite to 24x24 and 25x25.
10:26:07 <peter1138> Oh damn, the sprite type bits have different meanings between GRF container v1 and GRF container v2.
10:29:44 <peter1138> Confusing there's a bit that says if the sprite is compressed, but that bit isn't tested when decoding a sprite.
10:52:20 <ahyangyi> ahyangyi: Oh, I meant "limiting to integer percent scaling (multiples of 1%) is tolerable", not "limiting to multiples of 1x"
11:14:30 <peter1138> "integer percent" makes no sence.
11:15:29 <ahyangyi> It describes the status quo.
11:15:56 <ahyangyi> "Interface scaling is adjustable in 1% increments from 100% to 500%."
11:16:25 <ahyangyi> And English is not my first language so please do not argue semantics.
11:25:20 *** Zathras_4 has joined #openttd
11:45:18 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
11:45:20 *** WormnestAndroid has joined #openttd
12:02:37 <peter1138> Bilinear filtering. Shame it doesn't work with palette remaps.
12:49:59 <peter1138> Which one looks... less bad?
12:50:10 <peter1138> (I daren't ask "better" :p)
13:01:24 <peter1138> Problem with all the pixel-art specific image scalers is they... only work on integer scaling :)
13:01:40 <peter1138> (AND they do undesirable things.)
13:29:34 <ufiby> Good afternoon! A strange waypoint behavior has been detected. Master 20250919
13:33:49 <LordAro> this is not the correct place to report bugs
13:37:28 *** audigex has joined #openttd
13:37:28 <audigex> LordAro: Maybe tell them where the correct place is?
13:38:29 <LordAro> audigex: i could, but they're not new
14:26:35 *** Flygon has quit IRC (Remote host closed the connection)
14:38:42 *** Smedles has quit IRC (Read error: Connection reset by peer)
14:39:32 *** Smedles has joined #openttd
15:00:29 <peter1138> Yeah, technically the flow for stations could break if you don't run a Test before Exec, but that shouldn't... happen.
17:20:38 *** Wormnest has joined #openttd
18:05:43 *** lobstaroooo has joined #openttd
18:12:45 <peter1138> Hmm, yeah, I suppose assert there is a bad idea, even if it should always be fine.
18:43:41 <peter1138> Different approach that doesn't "randomly" call Allocate twice.
19:10:11 *** f_ has quit IRC (Ping timeout: 480 seconds)
19:51:07 *** lobstaroooo has quit IRC (Ping timeout: 480 seconds)
19:53:11 <peter1138> I seem Ruby's going well.
20:19:45 *** SigHunter has joined #openttd
20:32:28 *** yiffgirl has joined #openttd
20:32:28 <yiffgirl> i've been sitting on that one-line change to add the city icons to the town directory for the past month or so. i should probably stop sitting on it
20:36:16 <peter1138> Bring it up in 5 years time. "I've got a patch for that."
20:36:26 <peter1138> Rust in git. Why. :o
20:38:36 <yiffgirl> peter1138: wait, seriously ?
20:39:55 <peter1138> They appear to want it.
20:46:39 <peter1138> I don't get it personally, but.
20:48:59 <yiffgirl> i'm a massive fan of rust AND a massive fan of git so for me this news is the best i've had all week.
20:50:17 <peter1138> Not so great for the distribution maintainers that support systems that rust isn't available for.
20:50:21 <yiffgirl> evidently not a big enough fan that i know what they're up to, but still
20:51:32 <jfkuayue> So, how to use one sentence to describe rust for a noob who only knows NML and very basic python…
20:52:15 <peter1138> "Completely unrelated"
20:53:00 <peter1138> It's a programming language with some truely bizarre syntax decisions, along with stange namings.
20:53:52 <yiffgirl> you come around to the rust way of thinking eventually
20:54:05 <peter1138> e.g. Rust's "equivalent" of a std::shared_ptr is "std::rc::Rc" or "std::sync::Arc", depending on threadsafetyneededness.
20:55:33 <peter1138> But hey, it's totally safe.
20:57:13 <_jgr_> Large businesses, governments, etc have started to cotton on that memory safety vulnerabilities everywhere is a bad thing
20:58:16 <_jgr_> Using Rust is a fairly concrete way to make that go away
21:00:48 <_jgr_> That, or things like Go, which is what my team at work is into
21:01:50 <yiffgirl> the one real beef i have with rust right now is the laissez-faire attitude to package management. npm "six hundred vulnerabilities a day" was not an instruction manual
21:05:50 *** ChanServ sets mode: +v tokai
21:12:45 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
21:25:21 <yiffgirl> right, updated to master. do i change the (City) in the town window too, or leave that for when the explanation of what a city is gets added?
21:28:48 <yiffgirl> the ttf font feels so blurry at low scales...
21:29:12 <peter1138> It is, it doesn't get scaled well.
21:31:17 <peter1138> Hmm, let's try without hinting.
21:33:07 <peter1138> Oh, actually, it seems better for me these days, and disabling hinting did nothing.
21:53:15 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:09:51 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:24:49 *** klibber has joined #openttd
22:25:06 *** Wormnest has quit IRC (Quit: Leaving)
22:26:42 <klibber> Hi guys, tried to build with gcc/g++ 11 but failed and upgraded to 13 and it worked. I thought maybe there should be a check in the cmake or something for compiler version or what do you think ? :)
22:33:52 *** Wormnest has joined #openttd
22:45:09 <peter1138> > Every compiler that is supported by CMake and supports C++20, should be able to compile OpenTTD. As the exact list of compilers changes constantly, we refer to the compiler manual to see if it supports C++20, and to CMake to see if it supports your compiler.
22:52:42 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
22:58:12 <_glx_> and compilers may lie about their support of C++20
23:02:21 *** dh1 has quit IRC (Ping timeout: 480 seconds)
23:05:31 *** klibber has quit IRC (Remote host closed the connection)
continue to next day ⏵
