IRC logs for #openttd on OFTC at 2026-01-04
⏴ go to previous day
00:00:48 <_zephyris> So the release workflow had been deactivated due to lack of activity...
00:02:00 <_zephyris> And the build and publish steps were skipped on a manual rerun?
00:02:03 <_zephyris> Seems like a tomorrow problem
00:11:18 <_glx_> ah this release workflow runs on schedule, and the schedule has been disabled
00:18:19 <_glx_> and schedule or manual trigger seem to be for nightlies
00:21:23 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
00:21:45 <_glx_> easiest solution is to delete the release and redo it
00:22:21 <_glx_> (with the workflow reenabled)
00:27:00 *** MinchinWeb[m] has joined #openttd
00:32:15 <_zephyris> I guessed that might be the way, but didn't want to risk breaking things!
00:32:33 <_zephyris> Seems to be running OK now
00:56:55 <_glx_> probably easier to put a note somewhere about checking status of the workflow (deactivated or not) before creating a release 🙂
01:02:14 *** tokai|noir has joined #openttd
01:02:14 *** ChanServ sets mode: +v tokai|noir
01:02:38 *** Wormnest has joined #openttd
01:09:26 *** tokai has quit IRC (Ping timeout: 480 seconds)
01:12:48 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
01:13:37 *** MinchinWeb[m] has joined #openttd
01:55:31 <acs121> tabytac: I'll be honest, I believe many are making a useless fuss about this
02:08:36 <belajalilija> There’s things i dislike about the new menu, no small amount is likely only unfamiliarity, but even with that it’s still a big improvement over the old one
02:19:49 <mmtunligit> itd be soooooooo nice if my debugger didnt hang to the point of requiring me to restart my computer when i try to view the state of the local variables for no reason
02:23:50 <mmtunligit> thank fuck that was just a case of me forgetting to add an i++;
02:24:38 <mmtunligit> nvm that wasnt the only thing
03:17:28 *** Wormnest has quit IRC (Quit: Leaving)
07:11:57 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
07:11:58 *** WormnestAndroid has joined #openttd
07:12:02 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
07:12:03 *** WormnestAndroid has joined #openttd
07:12:06 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
07:12:17 *** WormnestAndroid has joined #openttd
07:33:51 <LordAro> mmtunligit: i have been known to `gdb gdb` in the past. that was amusing
07:34:14 <LordAro> having to restart the computer is baffling though
07:34:39 <LordAro> kill -9 is a thing (though even then rarely actually needed)
07:59:28 *** ialokin has joined #openttd
09:41:38 <peter1138> Can we just ban zbase/abase?
09:42:51 <peter1138> Buildings and industries are fine, that's about it.
09:43:06 <andythenorth> yes we could just withdraw distribution of it
09:43:12 <andythenorth> we're not obliged to
09:43:21 <andythenorth> there will be about 12 months of pitchforks
09:49:08 *** definitelynotcheese_ has joined #openttd
09:52:02 <_zephyris> zBase isn't listed for 15+...
09:52:02 <andythenorth> so the pitchfork crowd will assume some sort of conspiracy to hold back zbase because ogfx2 has been criticised
09:52:12 <andythenorth> we might as well just make the conspiracy real, and in the open
09:53:32 <peter1138> NightGFX works quite well too. Better than aBase.
09:53:49 <andythenorth> zbase is not for me
09:53:58 <andythenorth> possible because I already have abase installed?
09:54:17 <peter1138> No, because aBase is a fork not controlled by the same author as zBase.
10:02:33 <ahyangyi> NightGFX is best when not used during gameplay 😛
10:02:46 <ahyangyi> I mean I like it, but mostly for eyecandy purposes
10:05:26 <kuhnovic> NightGFX + completely dark room + screen on max brightness
10:06:33 <michi_cc[d]> People like their (ugly) extra zoom. As soon as OpenGFX2 high resolution is on bananas, I heartily approve killing a/b/c/x/y/zbase with fire 🙂
10:07:26 <peter1138> It's a shame the CZTR and/or "real 32bpp" never did a baseset.
10:07:26 <ahyangyi> Adding an "out of maintenance" notice might help a bit
10:08:35 <peter1138> What's an `18pp` texture... Hmm.
10:09:55 <ahyangyi> 1 something and 8 bits per pixel, I guess?
10:10:48 <ahyangyi> Conveniently it could be 1 cc, though most sprites in the base set do not use company colour at all.
10:12:12 <Rubidium> potentially a wrong transcription of 1Bpp (1 byte per pixel)
10:16:55 <_zephyris> I could put an (incomplete) OGFX2 high def on bananas if it'd reduce pitchforks
10:28:21 <_zephyris> peter1138: I find it amazing how hard it is to motivate people to draw vanilla things
10:29:02 <_zephyris> The motivation for grfs is almost always from new gameplay/objects rather than better vanilla stuff
10:30:39 <peter1138> `const Vehicle *&veh` does look weird :)
10:35:07 <peter1138> Would it make sense to move the `veh` parameter to the `GetDepotVehiclePtData` struct? So there's then only one out parameter that contains three bits of data, instead of 2 out parameters?
10:35:20 <peter1138> And yes, this is probably best as a review comments, oops.
10:41:03 <Rubidium> or maybe just a struct that returns everything, i.e. no out parameters at all
10:41:47 <Borg> pointer to reference? ;)
10:41:57 <Borg> or reference to pointer?
10:42:08 <Borg> I assume the second.. makes more sense
10:42:29 <michi_cc[d]> Stuff binds to the name, not the type, so reference (first level) to pointer (second level).
10:44:11 <michi_cc[d]> Rubidium: The main motivation might have been the write `if (GVDPW() == xx)` for the proper return value, but since there is `if (a = b(); a.x == y)`now, that has become a non-issue.
10:51:33 <peter1138> Rubidium, that too.
11:07:31 <andythenorth> _zephyris: is the dither added automatically from a downscale of the 256 or 64 sprites?
11:24:49 <xarick> openttd doesn't support c++20 entirely?
11:29:22 <ahyangyi> "it" refers to MSVC, not OpenTTD
11:29:38 <LordAro> i seem to recall a discussion about confusing settings
11:30:40 <Rubidium> xarick: your statement about OpenTTD not supporting C++20 entirely is definitely slop, but most likely *your* slop. So again... you don't seem to understand what someone/something writes, or you're ignoring big parts of it
11:31:01 <_zephyris> andythenorth: Dithered from 32bpp 1x zoom, using a custom Floyd-Steinberg based ditherer (which would behave pretty much like any other one in this specific case)
11:32:30 <_zephyris> And the 1x zoom sprite is a simple noise based texture, overlain with a downscaled copy of hand drawn 4x zoom details.
11:33:21 <andythenorth> in temperate and arctic it's prone to creating vertical visual runs
11:33:33 <andythenorth> and temperate is also *very* green 🙂
11:33:50 <andythenorth> tropic is notably different, both the dither, and the variation of base colour
11:33:57 <ahyangyi> I think the vertical visual runs are intentional?
11:34:14 <ahyangyi> Not an artifact introduced by dithering
11:34:39 <andythenorth> it creates an uncanny valley effect which I think some people are noticing but can't describe
11:35:01 <andythenorth> it's quite unsettling, because it looks like the screen is melting 🙂
11:35:07 <_zephyris> Hmm. It is (deliberately) a noise plus vertical motion blur texture
11:35:31 <ahyangyi> I do feel the run in the center column is a bit too strong, I'm almost seeing triangular grids at 1x
11:35:52 <ahyangyi> not sure if it's the uncanny valley andy was talking about
11:36:01 <Rubidium> peter1138: better this way? (still need to test/review it myself though)
11:36:04 <ahyangyi> or we are in two different uncanny valleys
11:36:43 <andythenorth> I'll do a little drawing 🙂
11:38:19 <andythenorth> original aligns this way
11:38:29 <andythenorth> ogfx 1 aligns this way
11:38:47 <andythenorth> ogfx aligns this way
11:39:35 <andythenorth> OpenGFX1 aligns this way
11:39:45 <andythenorth> this will be subjective obvs
11:40:10 <andythenorth> original also has a run of verticals, in a light colour
11:40:12 <ahyangyi> But does OpenGFX2 Classic align in the same way as OpenGFX2 Extra Zoom?
11:40:29 <andythenorth> so it really depends on what the eye processes
11:41:32 <andythenorth> the verticals in original cut through tile edges
11:42:01 <andythenorth> the verticals in OpenGFX2 cut through the corner of the diamond
11:42:47 <ahyangyi> Yeah that's the triangular grid I talked about
11:42:56 <ahyangyi> So we are in the same valley it seems
11:42:58 <andythenorth> OpenGFX2 Tropic is very different perceptually
11:43:25 <mmtunligit> LordAro: It made the debugger hang, then trying to stop the debugger made the ide hang, then that got laggy enough that I just couldn’t use the computer. It’s happened once or twice before and I think it has something to do with uninitialized variables in for loops. I don’t feel like debugging tge debugger so I’ve just chalked it up to “Linux is weird sometimes”
11:43:34 <andythenorth> OpenGFX2 Tropic is phenomenal
11:43:48 <mmtunligit> Oh right that doesn’t ping irc folks
11:44:02 <andythenorth> but also the grass tiles have the verticals cutting the tile edge, not the diamond corner
11:44:54 <ahyangyi> The lesson I learned from those is that the human eye is so good at finding patterns in randomness, so the designer might as well just don't randomize random things, but design the patterns directly.
11:45:02 <ahyangyi> something I knew but am too lazy to care about
11:45:19 <andythenorth> I design FIRS patterns directly
11:45:34 <andythenorth> it's not intentional, it's trial and error
11:46:24 <LordAro> mmtunligit: it did ping irc folks :)
11:46:28 <ahyangyi> Somewhat also similar to stable diffusion does
11:46:49 <ahyangyi> Start from chaos, but when patterns emerge, either suppress them or find a way to use them
11:50:17 <LordAro> mmtunligit: feels suspiciously like a memory issue
11:50:47 <LordAro> debugger probably disconnected but the process continued
11:51:04 <LordAro> htop/system analyser of choice will tell you
11:58:52 <mmtunligit> thanks, next tim it happens ill try to remember
12:02:09 <andythenorth> so the pitchfork crowd think we're just changing things for the sake of changing things
12:02:13 <andythenorth> what if we just made that the goal?
12:06:19 <ahyangyi> Will the pitchfork crowd vote more in Title Screen Competition 2026?
12:13:59 <peter1138> I see the pink arrows on andythenorth's "alignment" pictures but... uh... that's the only thing with a direction?
12:14:53 <andythenorth> dunno it's hard to explain
12:15:07 <andythenorth> it if was water surface, it would be "which way is it flowing?"
12:16:07 <andythenorth> that can be confusing, for example....watching a rising tide against the flow on an tidal river, with a wind moving the surface....
12:16:41 <andythenorth> or when you look at brick wall do you notice the vertical mortar, or the horizontal mortar, or the regular grid of brick shapes?
12:17:54 <ahyangyi> peter1138: principal component, in some sense
12:18:47 <andythenorth> it's just some visual cortex artefact
12:19:09 <andythenorth> my brain spends a lot of time on patterns
12:19:28 <andythenorth> other people just say "I find it confusing" or whatever
12:19:35 <ahyangyi> It's fine. Our ancestors invented all those constellations didn't they.
12:20:59 <ahyangyi> We just do the same stargazing thing, but at OpenTTD ground tiles instead.
12:30:17 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
12:30:18 *** WormnestAndroid has joined #openttd
12:32:36 *** gelignite has joined #openttd
12:39:41 <belajalilija> Would it be in any way possible for vehicle length to appear on the vehicle purchase menu or does the game struggle with articulated vehicles?
12:39:41 <belajalilija> It would be pretty neat for fixed consist vehicles especially
12:40:48 <andythenorth> belajalilija: in the extra info, or in the actual purchase list?
12:41:10 <belajalilija> andythenorth: I mean along with speed, hp, weight
12:41:29 <andythenorth> I added badges to Horse for it, but the weight of opinion was against it
12:41:38 <andythenorth> and sometimes I listen to better counsel
12:41:50 <andythenorth> I thought it would be useful for autoreplace also
12:41:58 <andythenorth> to prevent accidentally changing overall length
12:42:11 <belajalilija> How does it work with badges?
12:42:16 * Borg is testing block signal.. checking it own state.. and sending Wait at PBS signal to the entry signal [; lets see if it will work properly
12:42:59 <andythenorth> belajalilija: it's just a badge?
12:43:00 <Borg> this significantly simplifies RR signals programs
12:44:00 <andythenorth> yeah I deleted Horse length badge
12:45:19 <belajalilija> andythenorth: I’ve not played this game in a year, just making grf xd so I’m unsure how a badge would be useful when there’s so many lengths available
12:45:58 <andythenorth> that's one of the reasons they're not added to Horse
12:46:14 <andythenorth> would it be unit length? Tile length? n/8 length?
12:46:35 <belajalilija> Tile length as in the depot makes the most sense
12:46:49 <belajalilija> The /8 stuff is only useful for devs really
12:48:12 <belajalilija> So in the purchase menu you’d just see something like
12:48:12 <belajalilija> Length: 6,4 tiles
12:48:30 <andythenorth> something like that
12:48:48 <andythenorth> one of the cases made against a badge was that OpenTTD should do it
12:48:56 <andythenorth> because grfs will all do their own weird version
12:50:05 <Rubidium> belajalilija: I think the problem is that NewGRFs provide different stats in 'purchase window' mode versus actually building them, which would result in bug reports about OpenTTD doing it wrong
12:50:28 <andythenorth> that's an argument in favour of badges 🙂
12:50:46 <belajalilija> Rubidium: Could you give an example?
12:50:56 <andythenorth> purchase doesn't know the vehicle length
12:51:04 <andythenorth> so there would have to be a specific callback
12:52:33 <belajalilija> I might just do it myself on any vehicle over 3 tiles long
12:54:41 <_zephyris> andythenorth: I see the temperate pattern more without gridlines, I see the slightly darker vertical corner to corner line
12:55:06 <andythenorth> it's probably tweakable?
12:55:13 <andythenorth> tropic very much doesn't do it
12:57:36 <_zephyris> IIRC arctic and tropic have hand drawn grass tuft and scattered rock details, while (lush) temperate just had the grass tufts.
12:57:57 <_zephyris> But yeah, some pokes to that hand detailing should fix it
12:59:26 <andythenorth> relative contrast within the tile spans a greater range
13:01:32 <andythenorth> I think it might be the reason some of the colour blind users are struggling
13:04:13 <peter1138> belajalilija, number of parts (and their length) is not really possible for the game to know in advance of actually purchasing and attaching them.
13:04:40 <Borg> andythenorth: what are statistics? I mean. what is % of population having any color blinding.. I know that there are a lot of them.. most serve is gray mode.. but its super rare
13:05:17 <andythenorth> I've no idea Borg 🙂
13:05:29 <peter1138> So the game would have to, in the background, build, attach, run calculations that normally happen in a depot, and then destroy it all again. For each engine in the list.
13:06:07 <andythenorth> whereas the compile could just put a badge on it
13:06:50 <andythenorth> ok so the temperate (and arctic) grass does have more sharp luminance boundaries *within* the tile
13:07:04 <peter1138> (And even then it wouldn't necessarily match what you eventually attach it to, because that can affect it.)
13:07:06 <andythenorth> and red-green colour blind users *may* particularly notice that, because they rely more on relative luminance
13:07:28 <_zephyris> About 8% of men, so about 16% of our users...
13:07:30 <andythenorth> but it's not just a colour blind issue
13:07:48 <Borg> _zephyris: thats.. a lot...
13:07:51 <ahyangyi> I'm trying to understand the issue. Is the problem that the texture/contrast can be particularly jarring to colour blind people?
13:08:01 <Borg> do those users have special graphic set ?
13:08:06 <Borg> if not.. maybe they should...
13:08:10 <ahyangyi> 16% is a lot, though the joke might be about the gender ratio
13:08:17 <peter1138> OpenTTD players are not all men o_O
13:08:35 <Borg> peter1138: all not... 99% I guess [;
13:08:58 <andythenorth> this is the green channel for temperate grass
13:09:21 <andythenorth> this is the green channel for tropic
13:09:30 <peter1138> Why is my browser not paying attention to window focus... :o
13:09:32 <andythenorth> this is what's tripping up some users
13:09:44 <ahyangyi> Green contributes the most to luminance even for non-colour-blind people.
13:09:57 <Borg> kuhnovic: the problem is.. there cannot be one
13:10:07 <Borg> quite a lot of variations...
13:10:21 <andythenorth> for temperate blue is adding no spatial information, it's just dither
13:10:22 <ahyangyi> Though it definitely has an ever larger influence when one or two other channels are absent.
13:10:30 <andythenorth> and it's quite contasty
13:10:35 <kuhnovic> Good point. And we need TetrachromacyGFX too.
13:10:44 <rito12_51026> andythenorth: and red?
13:10:48 <andythenorth> whereas for tropic
13:10:55 <andythenorth> blue reinforces the shape
13:10:56 <ahyangyi> We need tetrachromacy screens first
13:11:11 <ahyangyi> Before we worry about the game graphics
13:11:33 <andythenorth> a colour shift of tropic sprite to temperate-ish green
13:12:00 <andythenorth> this is the temperate red channel
13:12:05 <peter1138> What are you trying to prove?
13:12:12 <Borg> a month ago or so. I was doing a little bit of research.. because I was crafting my own colorize() functions..
13:12:13 <andythenorth> this is the tropic red channel
13:12:20 <Borg> came across interesting post:
13:12:27 <peter1138> Is this just analysis of tiles for that sake of it?
13:12:29 <kuhnovic> Tetrachromacy is quite a rare trait that's almost exclusive to women. So that problem is mostly solved by our demographic 😛
13:12:44 <andythenorth> OpenGFX2 Temperate and Tropic make me feel dizzy and sick
13:12:53 <andythenorth> sorry, Arctic, not Tropic
13:12:59 <andythenorth> it's not the sea or the buildings
13:13:01 <andythenorth> so I was curious
13:13:05 <andythenorth> other users are reporting it
13:13:07 <Borg> I play exclusivly using orginal graphics.. its just nice and clean
13:13:23 <andythenorth> yes, it is, but OpenGFX2 Tropic is of equivalent quality to original
13:13:32 <andythenorth> which is a good outcome
13:13:48 <Borg> I never actually explored that part of game.. I mean.. alternative graphics :) just orginal fits the game pefectly
13:13:53 <peter1138> Reporting it? What's the ticket number?
13:14:09 <andythenorth> you know they're not in github 🙂
13:14:30 <andythenorth> it's internet anecdote
13:14:33 <Borg> okey.. one thing sucks in orginal graphic.. maglev.. but I never use it.. and I merge maglev -> monorail :D problem solved
13:15:29 <ahyangyi> Anyways, even if people reported it it would have go on OpenGFX2's issue list, not OpenTTD's.
13:15:34 <peter1138> So they're NOT reporting it?
13:16:39 <andythenorth> they aren't reporting it in github
13:16:45 <andythenorth> so it's not a bug report
13:17:39 <ahyangyi> This kind of subjective experience rarely becomes actual bug reports
13:17:49 <andythenorth> I mean...I can file it
13:17:56 <ahyangyi> After all, what can you say? "That graphics look a little off to me"
13:17:57 <andythenorth> I've filed reports on stuff like this before
13:18:03 <andythenorth> no it's measurable usually
13:18:20 <ahyangyi> expecting to be responded that it's PEBKAC?
13:18:43 <andythenorth> we found that the mac palette was distorted by the cocoa driver (?) iirc after weeks of debating colour perception yada yada
13:18:59 <andythenorth> but this is also easy to fix
13:19:13 <ahyangyi> It's measurable only when someone like Andy spends a lot of time to investigate the issue 🙂
13:19:32 <ahyangyi> For most people the bar is too high to report this kind of problem
13:19:51 <peter1138> For most people there is no problem to report?
13:20:14 <peter1138> For some people their brains are wiredly slightly differently and see patterns in things that aren't there?
13:20:31 <andythenorth> steam, reddit both have users reporting that the OpenGFX2 graphics are worse
13:20:40 <andythenorth> and I think there was a discord report
13:20:43 <ahyangyi> If we assume all posts on reddit are jokes...
13:20:47 <peter1138> I mostly attribute that to "I don't like change"
13:20:50 <ahyangyi> well, quite some of them are
13:20:59 <andythenorth> "it's subjective" doesn't cover it, because it's demonstrable in the sprites
13:21:00 <peter1138> Because they're generally coming from OpenGFX1, not original TTD.
13:21:25 <peter1138> So they are used to the dull tones of OpenGFX, not the brightness of TTD.
13:22:18 <peter1138> There are players perfectly happy with zBase, even though it looks terrible, because that's what they're used to.
13:22:42 <Borg> yeah... old habbits die hard :D
13:22:58 <ahyangyi> Anyways, a way to tell between genuine problems and resistance to change might be.. wait for some time?
13:23:25 <ahyangyi> That's not to say we should not discuss the potential issues
13:24:05 <peter1138> But if you have perception issues that only appear with OpenGFX2, then you could at least open an issue ticket or discussion in the relevant place.
13:24:23 <andythenorth> I'm looking for the cause first, kind of curious
13:24:30 <peter1138> Then encourage all the other "reports" that are not actually reporting it, to engage.
13:24:58 <peter1138> Because saying "people are reporting" when they're not... is not useful.
13:25:36 <ahyangyi> "people are saying".
13:26:07 <andythenorth> "people are whining" is what I wanted to say
13:26:12 <andythenorth> but I thought that was less mature
13:28:37 <andythenorth> didn't expect this, this is red channel for original
13:28:55 <andythenorth> that's higher contrast at tile gridline than I would have guessed from looking at the hues in the sprite
13:29:14 <andythenorth> red channel in OGFX1
13:29:40 <andythenorth> red channel in OGFX2, not sure this gives us much insight
13:30:34 <andythenorth> green channel original
13:30:44 <andythenorth> green channel ogfx1
13:30:56 <ahyangyi> What about luminance?
13:31:04 <andythenorth> green channel ogfx2
13:31:29 <andythenorth> again, not sure what to conclude
13:31:46 <ahyangyi> Since red and blue contribute so little to luminance, perhaps large variance in red and blue and small variance in green is the way to achieve colour diversity + a uniform luminance
13:31:51 <andythenorth> conclude that I'm a spanner those are blue
13:34:15 <xarick> ah, screenshot master!
13:34:17 <andythenorth> so green is hottest (highest energy range) in ogfx2
13:34:42 <ahyangyi> Is that any surprise, given the tiles are literraly green grassland? 😛
13:34:52 <ahyangyi> Oh, you are comparing across the base sets
13:34:59 <andythenorth> original composes the grid from the red and blue channels
13:35:07 <xarick> I don't see green, i see gray, maybe I'm colorblind
13:35:16 <andythenorth> grid is present, but highly muted in green channel for original
13:36:46 <michi_cc[d]> Looks to me that original gets most of the contrast from red and blue, while ogfx(2) get it more from green.
13:37:56 <ahyangyi> Hence I'm asking about luminance, does OpenGFX2 feature a larger contrast in luminance?
13:38:42 <ahyangyi> Does turning off the grid lines help?
13:41:10 <andythenorth> ahyangyi: how would you measure luminance? Other than eyeballing levels or curves panels?
13:41:23 <michi_cc[d]> Transform to HSL color space?
13:41:44 <michi_cc[d]> And look at the L channel
13:42:14 <ahyangyi> Or any of the newer color spaces that claim to fit human perception of brightness better
13:42:21 <ahyangyi> But really, any would help
13:44:32 <andythenorth> photoshop doesn't appear to have HSL, but it does have Lab
13:45:17 <andythenorth> lightness channel, original
13:45:36 <andythenorth> lightness channel OGFX1
13:46:10 <andythenorth> lightness channel OGFX2
13:46:35 <andythenorth> so yes OGFX2 temperate is noisier, depending on definition of "noise"
13:46:39 <ahyangyi> Hmm, the contrast in lightness isn't larger, aside from the grid lines
13:46:46 <andythenorth> it has more visual contrast edges to process
13:46:54 <andythenorth> and fewer cues in non-green channels to process them with
13:47:16 <andythenorth> and - this is massive theory - but human brains are tuned to pick apart threat in a world that is predominantly green
13:47:24 <andythenorth> which is a high cognitive focus
13:47:36 <andythenorth> so "the new graphics are more tiring" is plausible
13:47:54 <michi_cc[d]> Hmm, what I take from this is that any patterns on OGFX2 are a lot less visible to me than the Original ones
13:48:57 <michi_cc[d]> The grass part looks quite uniform, which matches andy observation of the brain still trying the hardest do find any pattern.
13:49:03 <ahyangyi> A little not sure about that one, red is much much more alerting than green
13:49:19 <andythenorth> yes, but we have to constantly scan the differences in green
13:49:34 <andythenorth> it's very much theory, I don't know if it's more than a story about living in jungles 😛
13:50:40 <michi_cc[d]> There's a reason the human eye is generally the most sensitive to green, given that slight variations in green used to be the difference between food and poison.
13:50:48 <ahyangyi> I do know that we in general care about differences in green.
13:50:49 <michi_cc[d]> And getting eaten 🙂
13:50:59 <ahyangyi> Hence RGGB bayer filter.
13:52:12 <ahyangyi> But can't figure out much in the screenshot comparison above, besides stylish differences and the obviously larger contrast on grid lines.
13:52:37 <andythenorth> your eye will constantly find the grid corners
13:52:46 <ahyangyi> And that the original graphics having more variance in color (but, as Peter said, people who were complaining about OGFX2 mostly compared it to OGFX1)
13:53:17 <andythenorth> I got into this because I think generally OGFX2 knocks OGFX1 out of the park
13:53:28 <michi_cc[d]> OGFX1 has the gridlines a lot less visible in the green channel, too.
13:53:47 <ahyangyi> Anyways, grid lines are something very easy to fix, since OGFX2 already has the option to disable grid lines
13:54:16 <michi_cc[d]> And you can safely assume that 95% of the players have not found their way to any baseset options (they weren't in the MH video either).
13:54:47 <ahyangyi> That's another issue
13:54:51 <_glx_> michi_cc[d]: at least that means they won't use a/zbase 🙂
13:55:43 <ahyangyi> Just saying it's something that can be tried without modifying OpenGFX2
13:56:01 <ahyangyi> and if it helps, hopefully we can get a new version without much effort either
13:56:04 <michi_cc[d]> I mostly meant the actual parameters for the baseset, Baseset selection itself is probably more well known (and was shown by master hellish).
13:57:16 <_glx_> ah yes the parameters are very new (saw someone trying to use the outdated OGFX2 param newgrf)
13:58:17 <_glx_> oh someone wants a github block
13:58:44 <LordAro> i like how they used 'an' correctly with the pig latin
13:58:47 <andythenorth> _zephyris: sorry about me 😛
14:15:27 <Borg> oh.. in small hoster we could ask for IP of that pig [;
14:18:26 <xarick> is GFX1 still gonna be complete?
14:18:43 <_zephyris> That was pretty special, any gems in the content? Plus, taught me some pig latin.
14:18:57 <_zephyris> xarick: Literally just made 8.0, which is complete...
14:19:49 <_zephyris> andythenorth: No worries, the quality of your sprites proves you're worth listening to!
14:19:57 <LordAro> _zephyris: i'm afraid OGFX2 is all vibe coding
14:19:58 <ahyangyi> Anyways, I managed to submit a Github report against that account for hate speech. Unfortunately their words were too nasty so I did not include a verbatim copy in my report. Hopefully the guys who process Github reports can see their deleted words.
14:20:31 <LordAro> i am amused by the phrase "millstone around the neck based skydiving"
14:20:57 <LordAro> but yes, having looked at the account, dude is not well
14:21:14 <_zephyris> Haha, OGFX2 is worse than vibe coding, it includes chunks of 'I'm teaching myself Python' coding.
14:23:00 <_zephyris> Thanks for handling so swiftly 🙂
14:29:08 <_zephyris> andythenorth: I think the issue is probably lack of intermediate-scale structure in the ground tiles - nothing on the size of ~4 pixels which gives a prominent repeat in addition to the gridlines.
14:33:22 <andythenorth> Tropic grass, green channel only, I find this coherent / easy on the eye
14:33:45 <andythenorth> Temperate grass, green channel only
14:34:03 <andythenorth> reminds me a bit of bump maps from old 3D texturing apps
14:35:11 <andythenorth> lightness mode, temperate
14:35:42 <andythenorth> lightness mode, tropic
14:35:54 <andythenorth> pff zoom level varies sorry 😐
14:37:53 <andythenorth> and a vertical snake
14:48:35 *** MinchinWeb[m] has quit IRC (Remote host closed the connection)
14:49:51 *** MinchinWeb[m] has joined #openttd
14:53:21 *** Tirili has quit IRC (Remote host closed the connection)
14:55:54 <_glx_> unless the picture is not textured at all I think there will always be some kind of pattern (more or less visible of course)
14:58:20 <andythenorth> I think Tropic does a nice job with the pattern
14:58:44 <andythenorth> there are 3 or 4 different patterns, all equally weighted to my eye
14:58:55 <_glx_> yes I can see the pattern, but it doesn't induce ghost lines
15:09:40 <LordAro> hmm, my right headphone has given up
15:11:18 <Rubidium> rito12_51026: 7A seems to be, 64 doesn't seem to be available I'd say
15:12:43 *** peter1138 has quit IRC (Remote host closed the connection)
15:12:58 *** peter1138 has joined #openttd
15:12:58 *** ChanServ sets mode: +o peter1138
15:22:23 <peter1138> Hmm, I guess I missed some fun.
15:26:52 <LordAro> wait, is debug output on stderr?
15:27:11 <LordAro> but "nothing is written out" might suggest that something else is broken
15:28:16 <_zephyris> So much love for sprite 3981
15:30:37 <LordAro> peter1138: you know audio and things, while i'm buying a new pair of headphones, any recommendations for a "cheap and cheerful" DAC?
15:32:31 <andythenorth> started analysing FIRS sprites now 😐
15:32:50 <andythenorth> the RGB breakdown of this is interesting
15:34:48 <andythenorth> putting it into lab colour, the a, b channels look like wallpaper patterns
15:34:56 <michi_cc[d]> LordAro: The windows EXE is not compiled as a console exe, so it will probably always open its own and not redirect output.
15:36:08 <LordAro> -dmisc=4 maybe? not sure what the category is
15:36:27 *** Smedles has joined #openttd
15:37:56 <michi_cc[d]> LordAro: Okay, it might work with stderr, the release build isn't opening it's own console, that's only the debug version (can you guess how often I launch a release version? 🙂 )
15:38:16 *** Wormnest has joined #openttd
15:45:43 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
15:45:52 *** MinchinWeb[m] has joined #openttd
15:47:05 *** limyx826 has joined #openttd
15:47:05 <limyx826> `openttd.exe -d4 > openttd-log-file.txt 2>&1`
15:47:05 <limyx826> Tried this command, the text file appear but is empty after I exited the game.
15:47:09 <andythenorth> lab colour channels a and b on grass tile is interesting
15:47:17 <andythenorth> but the channel is probably sick of screenshots
15:48:13 <LordAro> limyx826: did the "normal" debug output appear in the console?
15:49:09 <limyx826> `[2026-01-04 23:48:43] dbg: [yapf:3] [YAPFw]- 5 - 36 rounds - 69 open - 35 closed - CHR 0.0% - C 1400 D 0
15:49:09 <limyx826> [2026-01-04 23:48:43] dbg: [yapf:3] [YAPFt]- 28 - 11 rounds - 5 open - 10 closed - CHR 20.0% - C 12766 D 0
15:49:35 *** FLHerne has joined #openttd
15:49:55 <LordAro> limyx826: guess the redirect isn't working for whatever reason then
15:50:20 <LordAro> perhaps -d4,yapf=0 would be enough to hide the yapf output?
15:50:59 <limyx826> The first few seconds I did see the loading of basesets but it quickly get overwritten once the title save is loaded.
15:51:13 <LordAro> you should be able to scroll up...
15:51:34 <LordAro> i'd be surprised if there was enough output that quickly to go beyond the scroll buffer
15:51:44 <limyx826> yapf logs are too many, the earliest logs get overwritten fast
15:52:50 <LordAro> well, `-dmisc=4` then
15:52:59 <LordAro> source suggests it should contain enough output
15:55:04 <peter1138> Hmm, I suppose we could try detecting if vanilla road/rail sprites are overridden but not the bridge overlay sprites, and then ignore as appropriate.
15:57:00 <limyx826> lordaro4353: -d4,yapf=0 works, without yapf logs it is lesser
15:57:43 <limyx826> Though there are quite a number of water logs.
15:59:07 <LordAro> limyx826: think there's still quite a bit missing there
15:59:15 <LordAro> upload it as a file instead?
16:00:16 <limyx826> I checked again the console after launch, turns out I had too many newgrfs that the logs get overwritten fast.
16:00:38 <LordAro> again, cmd has a huge scroll buffer, there should be more text available
16:00:48 <LordAro> thousands of lines, normally
16:06:31 <limyx826> I find that the earliest log that I able to capture via console is sprites replacement
16:06:55 <LordAro> your console is very strange
16:07:13 <limyx826> The initial loading logs is gone once the sprites replacement starts.
16:07:30 <LordAro> are you quite certain there is no scroll bar at all?
16:07:39 <limyx826> ` GraphicsNew: Replacing sprites 0 to 15 of Shore graphics (type 0x0D) at SpriteID 0x1730`
16:07:40 <LordAro> mouse wheel, bar on the right side, etc etc
16:07:46 <limyx826> Top of the scroll bar
16:07:48 *** brickblock19280 has joined #openttd
16:07:48 <brickblock19280> Opengfx2 might cause that?
16:09:27 <LordAro> you could try `-d4 -Q --help` - it still does the baseset lookup, but skips newgrfs and actually running the game
16:14:05 <limyx826> That's odd, if only `-dmisc=4`. The top of my console is cutted off.
16:14:30 <limyx826> `0f9c-Japan-1.0.tar' with 4 new files
16:14:30 <limyx826> [2026-01-05 00:12:00] dbg: [misc:4] Found tar 'C:\Users\Limyx\Documents\OpenTTD\content_download\scenario\heightmap\Singapore_and_Surroundings-0.1.tar' with 4 new files`
16:16:00 <limyx826> That's it. I think I will try to debug it next day. It's past midnight here.
16:16:30 <LordAro> i've never known a console cut off part of a line
16:24:48 <limyx826> @rubidium Got it. The logs is in text file now.
16:25:24 <LordAro> ooh, japanese windows
16:25:39 <LordAro> that can behave very strangely with filehandles etc
16:28:00 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
16:28:17 *** MinchinWeb[m] has joined #openttd
16:28:43 <limyx826> How do you japanes windows? 🙂
16:29:04 <limyx826> How do you know japanese windows? : )
16:29:28 <LordAro> the '¥' being output as a path separator on the console
16:29:42 <LordAro> it's some sort of encoding oddity
16:30:14 <peter1138> Standard for Japanese windows, it's deliberate.
16:31:08 <peter1138> Likely a left over from when DOS codepages were the thing, and of course people don't report bugs for things they're used to :)
16:31:33 <mmtunligit> oh good lord what have i done now
16:31:59 <mmtunligit> waypoint spec seems to change on viewport invalidation? im not debugging thsi shit today
16:32:43 <LordAro> limyx826: we got there eventually, thanks for keeping at it :)
16:32:59 <limyx826> Thanks everyone for the help.
16:33:15 <limyx826> Turns out it is missing a file
16:33:29 <limyx826> Must lost it years ago
16:34:39 <limyx826> Must sleep now, have work to do. :[
16:42:41 <peter1138> mmtunligit, what...?
16:43:33 <mmtunligit> peter1138: im changing some stuff to vectors to enable placing a random item form picker collections and runnning into issues in the process of this
16:44:10 <peter1138> Ah, selected item shenanigans.
17:08:09 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
17:08:49 *** MinchinWeb[m] has joined #openttd
17:10:29 <_glx_> had to update requirements 😉
17:11:10 <_glx_> for now I just removed 3.9 and didn't had others, else that would block all existing PRs
17:53:44 <peter1138> No nightly builds since last year.
17:56:00 <peter1138> Either since the year rolled over, or since the 15.0 release :)
17:56:09 <_glx_> indeed, both eints and nightly were disabled
17:57:26 <_glx_> that's because automated schedule
17:57:54 <_glx_> and no changes in the repo itself
18:06:02 *** Tirili has quit IRC (Quit: Leaving)
18:17:25 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
18:18:27 *** MinchinWeb[m] has joined #openttd
18:29:35 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
18:35:56 *** MinchinWeb[m] has joined #openttd
18:36:47 <xarick> according to mr copilot perception of openttd CODINGSTYLE.MD, he started inserting a space between template and the open < bracket.
18:36:56 <xarick> `template <typename T, typename Tsomething, int N, uint8_t Tnumber_of_something>
18:37:14 <xarick> was I wrong all along?
18:39:56 <xarick> dang, I was wrong all this time
18:40:07 <andythenorth> which do you believe, you, or the dice-rolling algorithm?
18:41:07 <xarick> the document doesn't say it explicitly that a space is required, just an example
18:44:21 <xarick> well... openttd is consistenly adding a space
18:45:13 <Rubidium> andythenorth: I believe neither, but the latter is probably more often right than the former :(
18:45:27 <andythenorth> it has more training data
18:47:55 <xarick> i asked mr copilot to document the entire BPlusTree after I threw the entire CODINGSTYLE.MD his way
18:48:13 <xarick> he did it! inflated the file from 2200 lines to 2600
18:49:24 <xarick> it was the longest response I had from it. Had to wait 3 minutes to have it all written
19:15:59 <andythenorth> most of the LLM hype is that MS, Google etc can employee engineers for cheaper, preferably in low-cost jurisdictions, and the public markets pay for that cost saving
19:16:24 <andythenorth> because an LLM has more training data than a roll-the-dice "let's just guess" develope
19:16:45 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
19:18:58 *** MinchinWeb[m] has joined #openttd
19:25:54 <andythenorth> what's the badge limit?
19:26:01 <andythenorth> Iron Horse uses 5328 currently
19:26:18 <peter1138> Iron Horse is doing something wrong, then.
19:27:40 <andythenorth> it's more than I expected TBH, I'd have guessed at 2000 or so
19:28:13 <peter1138> The design assumption was something like < 100...
19:29:05 <andythenorth> 3974 of them are variant group IDs, which is probably a bug
19:30:42 <mmtunligit> that was much easier to fix than i expected thank god, just had some if conditions done incorrectly
19:34:49 <andythenorth> yeah 3974 badges only used for debugging
19:40:54 <andythenorth> ok 1337 badges in Horse
19:40:57 <andythenorth> is that leet somehow?
19:41:45 <ahyangyi> 5328-3974=1354 though
19:42:02 <ahyangyi> Where are the remaining 17 badgers
19:42:05 <andythenorth> well maybe I miscounted somewhere 😛
19:45:09 <xarick> what is this stupid error about
19:45:24 <xarick> linux and mac complain, windows is fine
19:46:19 <andythenorth> those are unusual separators to my eye
19:46:45 <andythenorth> `#include "core/container_func.hpp"`
19:47:40 <ahyangyi> We have quite a few unusual separators today 😛
19:47:47 <ahyangyi> (there was also `¥`)
19:50:19 <xarick> I'm interested in the doxygen document, wondering what I need to do to have it dtected
20:18:13 <_zephyris> Where did we get to with vanilla badges?
20:23:45 <xarick> CI doesn't build hpp files?
20:26:55 <rito12_51026> Is there act 0xff ?
20:29:24 <peter1138> 0xFF is a data block.
20:30:29 <xarick> i can't find my file being built
20:36:29 <xarick> core/bplustree_type.hpp not found... how does this work?
20:36:58 <xarick> how is it compiled then?
20:37:30 <_glx_> it's a header, it's never directly compiled
20:38:28 <_glx_> it's something known since C and any language using includes
20:40:19 <_glx_> only cpp files are compiled
20:40:39 <xarick> compiler is quick to point the errors on them if they exist..
20:43:42 <michi_cc[d]> _zephyris: In geopolitical discussions about flags.
20:44:40 <mmtunligit> ohhh i sure do love compiler errors
20:44:52 <mmtunligit> i think i will untangle that tomorrow
20:45:08 <_glx_> best errors involve templates
20:47:10 <mmtunligit> ive got the same two pairs of custom endianbufferreader/writers in waypoint_cmd.h and station_cmd.h and it yells at me when theyre in both, and when theyre only in one, but i get different errors depending on which files have them
20:48:12 <mmtunligit> probably just need to more explicitly borrow them from one in the other but i have no idea how id do that or which one to borrow from and my mental capacity for such things is waning
20:48:31 <andythenorth> someone ship a vanilla badge grf
20:52:07 <xarick> ScriptList does not require it
20:59:24 <outofcreativity> _zephyris: I'm responsible for the OpenTTD port on OpenBSD and wanted to include the new opengfx2 sets with 15.0. However the tar file inside the zip file is named with upper and lowercase letters. This was different in the past and brakes the extract script. Could you please revert to all lowercase lettering or is there a reason for the name change? :-)
21:03:17 <xarick> isn't CodeQL "AI slop"?
21:04:08 <peter1138> No, it's static analysis, though it may have been contaminated since the initial configuration.
21:04:43 <andythenorth> I mean LLMs have to be jammed into everything currently
21:04:51 <andythenorth> or no funding can be made available
21:07:21 <_zephyris> outofcreativity: Pop an issue on the github repo and I'll take a look ASAP, most likely a change from using the new release workflow for the first time
21:08:18 <andythenorth> well most of the online 15.0 hate seems to have stopped
21:08:37 <andythenorth> I think I've seen double digits of grumbling, and about 3 crash outs
21:09:02 <andythenorth> but the Steam 15.0 post has 2074 upvotes
21:09:36 <andythenorth> some of the known and predictable squeaky wheels have squeaked
21:09:44 <andythenorth> bingo card partially filled
21:10:43 <peter1138> What kind of script is needed anyway? The .tar file doesn't need to be extracted.
21:11:36 <outofcreativity> _zephyris: thanks! lets see if my login still works, haven't used github since 10? years. in general i don't like it very much. will try this the coming week.
21:12:50 <outofcreativity> _zephyris: i've tested 15.0-rc3 and it worked nicely
21:16:05 <peter1138> > Update has broken FIRS mod's arctic chemical plant placements and the graphics look like upscaled vomit. I'm reverting.
21:16:09 <peter1138> It's always FIRS' fault :D
21:16:50 <LordAro> peter1138: andy you look different today
21:17:13 <andythenorth> it's not because FIRS 5 has flawed logic for chemical plant placement in arctic
21:18:03 <andythenorth> what about newgrf feature for main menu definition?
21:18:33 <andythenorth> we'd have to load grfs before title screen though
21:20:50 <_zephyris> outofcreativity: Thanks - saves me from forgetting, and making sure it's clear exactly what the issue is.
21:21:37 <_zephyris> peter1138: I enjoyed that one too.
21:22:30 <andythenorth> I suspect some players are confused and think OpenGFX2 *is* Toyland style
21:22:33 <peter1138> People wanting to go back to "TTD" graphics, because OpenGFX2 is "cartoony"... I am guessing they never actually played with the original TTD graphics...
21:22:58 <andythenorth> I think it's probably just shock value of the Toyland title screen
21:23:36 <peter1138> Ah maybe that is the "cartoony" bit.
21:25:29 <xarick> question about #define, is it `#<space>define`, or `#<tab>define` or `#define` or it depends?
21:26:21 <andythenorth> we did get one pitchfork about signals 🙂
21:26:24 <andythenorth> we can blame michi_cc[d] 🙂
21:28:43 <Rubidium> xarick: just don't do it that way
21:29:45 <cu-kai> shit, no wonder german trains are always late
21:30:51 <xarick> how would I do it Rubidium
21:31:22 <andythenorth> I do suspect they're seeing the title screen and yoloing out
21:31:33 <andythenorth> and the climate options are gone from the main menu also
21:31:41 <andythenorth> so it probably looks like we went cartoon
21:31:59 <andythenorth> wonder what Toyland is like for all the aBase users
21:32:04 <andythenorth> it's massively popular
21:32:19 <xarick> don't worry, you can release 16 on aprils fools
21:32:55 <andythenorth> not downloading 270MB of aBase to find out
21:33:03 <andythenorth> I only have a 1TB
21:33:07 <michi_cc[d]> andythenorth: And even that is wrong, when a track has for example Selbstblock (which used to be quite common with mechanical interlockings), signals default to green, even in Germany.
21:33:21 <andythenorth> someone else pointed that out I think later on Steam
21:33:41 <andythenorth> "always" and being confidently wrong
21:36:22 <_zephyris> andythenorth: 15 can be New Coke, 16 can be Coca-Cola Classic.
21:40:30 *** Borg has quit IRC (Quit: leaving)
21:40:47 <andythenorth> dunno, I probably shouldn't be screenshotting single users 😛
21:41:10 <andythenorth> but generally there is a lot of appreciation
21:41:32 <andythenorth> 2000 people have shown appreciation on Steam
21:41:36 <andythenorth> recent reviews are good
21:41:42 <andythenorth> lots of nice comments
21:42:42 <_zephyris> I'm impressed with the number of positives, given normal tendency to complain it looks very positive really
21:44:49 <andythenorth> lots of nice features
21:45:05 <andythenorth> generally a nice playerbase, every community has a few dickheads
21:45:52 <andythenorth> abase Toyland actually works
22:05:08 <reldred> Complainers are usually always the minority
22:05:56 <reldred> its just that they usually feel excessively compelled to voice their complaints whereas a lot of people if they're content don't see a reason to pipe up.
22:08:36 <andythenorth> in some cases there is a disorder I think
22:09:16 <andythenorth> and in others, poor choices 😛
22:09:34 <reldred> Some people are just cunts, disorder or not.
22:10:36 <talltyler> I think the game being free helps a lot with the negative comments. At least two thirds of the comments I see about the game I work on in my day job seem to be upset that it costs money.
22:11:23 <talltyler> Steam comments are generally the most unhinged, worse even than YouTube somehow
22:28:19 <michi_cc[d]> Maybe. Or maybe you are just imagining things. Who knows 🤷♂️
22:46:19 <xarick> just saw a video about licenses and copyrights... it just baffles me
22:46:25 <andythenorth> I do wonder if we should have kept plain text passwords, and just forced them to be 4 generated words or something
22:46:55 <andythenorth> I don't play multiplayer currently though, so I might STFU 🙂
22:48:50 *** patrickvarela has joined #openttd
22:48:50 <patrickvarela> hey, guys, I know we can't change the game speed on multiplayer, but I was wondering if it is possible to create a patch for this to be possible and what do I need to change. Like citymania do. I tried cmgamespeed but it seems it is not working currently with the last openttd version.
22:48:50 <patrickvarela> Not sure if anyone ever tried this, tho
22:53:07 <xarick> am I allowed to keep my copiloted-through bplustree in a github branch? must I delete the branch?
22:54:37 <xarick> in my fork of openttd, that is
22:59:01 <peter1138> You can change the speed of the game via the minutes per year setting of wallclock mode.
23:01:28 *** Flygon has quit IRC (Read error: Connection reset by peer)
23:02:08 <patrickvarela> peter1138: limited to minimum of 12 minutes per year, right? idk if it is possible to go faster 🤔
23:02:21 <patrickvarela> peter1138: thanks for replying, mate
23:06:37 <belajalilija> reldred: Exactly
23:07:10 <belajalilija> I for one welcome our new update overlord
23:10:27 <talltyler> It’s not possible to go faster than 12 minutes per year, no.
23:10:55 <talltyler> But you can go as slow as one real week per year, or freeze the calendar entirely (0 minutes per year).
23:11:35 <xarick> there is a fastforward speed multiplier setting, I think you can change that to below 1x
continue to next day ⏵
