IRC logs for #openttd on OFTC at 2026-10-17
โด go to previous day
00:11:57 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:12:16 *** WormnestAndroid has joined #openttd
00:25:52 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:26:05 *** WormnestAndroid has joined #openttd
00:27:45 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:28:07 *** WormnestAndroid has joined #openttd
00:28:57 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:29:08 *** WormnestAndroid has joined #openttd
00:29:33 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:30:00 *** WormnestAndroid has joined #openttd
00:30:38 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:31:04 *** WormnestAndroid has joined #openttd
00:39:18 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:39:45 *** WormnestAndroid has joined #openttd
03:59:38 *** Philemon has joined #openttd
04:03:11 *** Phileman has quit IRC (Ping timeout: 480 seconds)
04:53:57 *** dwfreed has joined #openttd
05:00:45 *** tokai|noir has joined #openttd
05:00:45 *** ChanServ sets mode: +v tokai|noir
05:07:46 *** tokai has quit IRC (Ping timeout: 480 seconds)
05:11:42 *** larndey has joined #openttd
05:17:41 <larndey> Alright my friend owes me five bucks B)\
05:31:39 <larndey> Might listen to that later?
05:35:06 <larndey> Well anyways.. anybody got any tips of searching for a specific mod?
05:43:31 <belajalilija> larndey: I think so
07:41:57 *** skee has quit IRC (Quit: WeeChat 3.5)
08:07:04 *** larndey has quit IRC (Remote host closed the connection)
08:12:07 *** skee has quit IRC (Quit: WeeChat 3.5)
08:19:07 <andythenorth> doesn't work currently
08:22:01 <reldred> 4CC, what a world we live in ๐
08:26:27 *** guru3 has quit IRC (Remote host closed the connection)
08:28:11 <andythenorth> if anybody can spot my mistake please say ๐
08:28:34 <reldred> im just happy its a thing
09:00:46 <reldred> wait did they roll their own android port/fork, get a crash, then have the nerve to raise an issue on the parent project
09:02:10 <reldred> hmm its an empty repo they've just used to attach the apk to
09:07:01 *** Flygon has quit IRC (Read error: Connection reset by peer)
09:27:21 <andythenorth> is 16075 actually returning the new palette index?
10:17:27 <andythenorth> ternaries eh ๐
10:45:21 <_zephyris> Double ternary hurts my head
10:48:28 <_zephyris> I think the issue is that the bool condition for the outer ternary should be `(force_pal || (this->seq[i].pal & PALETTE_MASK) == 0)`, ie. if the palette is forced or no palette is set.
11:24:18 <andythenorth> test grf if helpful
11:24:22 <andythenorth> hmm will attach to PR
11:37:18 *** jfs_ has quit IRC (Ping timeout: 480 seconds)
11:57:43 *** zyliety has joined #openttd
11:57:43 <zyliety> Iโm trying to update water rocks in my landscape grf, but it seems I canโt get it working even with the latest nml.
11:58:08 *** jfs_ has quit IRC (Ping timeout: 480 seconds)
11:59:48 <reldred> also peter you should talk to zyliety about trees, zyliety has drawn lots and lots of trees
12:10:25 <_zephyris> I don't think sea rocks have nml support yet, could be wrong though...
12:12:42 <_zephyris> Fallback for the sea sprites is the snowy rock overlays, so if you have snow in those sprites you'll see it on the water.
12:12:43 <reldred> how'd openttd.grf get updated then? I caught distinct sprite ID's just using the sprite aligner tool, or is openttd.grf all nfo in the backend?
12:12:58 <reldred> ah I didn't check agianst snowy rocks
12:13:20 <_zephyris> openttd.grf uses nfo, not nml
12:13:31 <reldred> well that explains that then
12:16:09 <_zephyris> zyliety: Short answer is to remove the snow from the rock overlays, should work fairly well then.
12:18:05 <reldred> just about bed time here so probably
12:19:58 <_zephyris> Cumberland sausage sandwich
12:27:43 *** rito12_13 has joined #openttd
12:27:43 <rito12_13> nikogra1viaGitHub: They failed the test :(
12:28:32 <rito12_13> they also code in polish
12:32:52 <zyliety> _zephyris: If that's the case, would it affect the existing snow rocks in arctic?
12:37:47 <peter1138[d]> _zephyris: Fallback is the non-snowy overlays. Maybe someone has put snow in them.
12:38:27 <peter1138[d]> The first set of overlays is non-snowy.
12:39:18 <peter1138[d]> There is a comment in our NFO that says they are unused, however they are used in deserts.
13:48:17 <tony_pixel> peter1138[d]: I rebased, reapprove pls
16:11:14 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
16:12:40 *** MinchinWeb[m] has joined #openttd
16:50:02 *** MinchinWeb[m] has quit IRC (Remote host closed the connection)
16:50:50 *** MinchinWeb[m] has joined #openttd
17:07:09 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
17:08:22 *** MinchinWeb[m] has joined #openttd
17:46:49 <andythenorth> does the outer if need `HasBit(this->seq[i].pal, PALETTE_MODIFIER_ALT_COLOUR)`
17:47:33 <andythenorth> "we have PRs at home"
18:05:43 *** aperezdc has quit IRC (Remote host closed the connection)
18:27:25 *** aperezdc has joined #openttd
18:34:04 <andythenorth> is `force_pal` the crash and news palette?
19:10:29 <_zephyris> Not sure about news, but pretty sure it is for crashed vehicles.
19:49:06 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
19:49:24 *** MinchinWeb[m] has joined #openttd
20:08:02 <andythenorth> I've run out of action 2 IDs ๐
20:10:16 <andythenorth> ah it will be random trigger switches ๐
20:10:26 <andythenorth> they span the whole grf and eat IDs
20:20:10 <_glx_> reduce scope is possible
20:22:55 <andythenorth> wonder if I can re-randomise bits without this strange chaining
20:27:05 <andythenorth> does nmlc automatically resolve which bits to use in `random_switch`
20:27:38 <andythenorth> nfo spec suggests the bits for RandomAction2 can be explicitly defined
20:31:57 <andythenorth> oof 2 of these wagons consumes 18 concurrent spritegroups ๐ฎ
20:32:20 <audigex> Yeah I've found I had to re-emit some global switches to avoid running out of concurrent IDs
20:33:05 <andythenorth> this is very expensive just to randomise containers on load ๐
20:33:12 <audigex> One thing that's caught us out with random_switch lately is that it seems to use the same bits by default unless you specifically set an offset with getbits(), so if you chain two random switches then the second can't actually select from all possible outputs
20:33:45 <audigex> It makes sense when I think about it, but isn't very intuitive initially
20:34:00 <andythenorth> hmm maybe this isn't random triggers, I just nerfed that off in the whole grf, and saved 1 ID
20:46:19 <_glx_> procedures also eat IDs
20:53:33 <audigex> The worst offender for me used to be old janky push-pull, ditching that when the new reversing functionality was added saved about 60 IDs
20:56:38 <andythenorth> I think in this case it's because I let the robot write a template
20:56:50 <andythenorth> robots are now good at statistically emitting syntax
20:57:09 <andythenorth> but the statistics don't understand context
20:58:07 <audigex> Depends what you tell/ask them, tbf - if you specifically explained to it about action2 IDs and asked it to prioritise the limits, it'd probably do a better job
20:58:46 <_glx_> if you have a lot of procedures used by a lot of switches it could be a good idea to group them and use an intermediate switch with indirect calls (but it's a pain to do in nml)
20:59:00 <audigex> They tend to ignore things they don't know about... but so do humans, my Action2 optimisation was terrible too, because I didn't know to optimise for it until I hit the limit
20:59:20 <audigex> _glx_: Indirect calls?
20:59:55 <andythenorth> I have machinery to group the procedure calls into a common chain
21:00:07 <andythenorth> that provides a much more sparse graph
21:00:08 <_glx_> you call a procedure with index to another procedure and params
21:00:34 <andythenorth> I don't know the technical name for this type of tree thinning
21:00:37 <andythenorth> but I'm sure there is one
21:01:11 <_glx_> that way you can have hundreds of procedures not eating IDs
21:01:45 <_glx_> but with nml and different number of params it's painful
21:02:06 <andythenorth> eh, sounds like a dispatcher ๐
21:02:10 <andythenorth> I don't do that, but I could
21:02:27 <will_marshall_> Out of curiosity, are we talking about something that's a visible stat in the nmlc summary?
21:02:47 <_glx_> yes it's the concurrent action 2 ID
21:03:06 <will_marshall_> Action2 ID? Not Action2 Registers?
21:03:58 <_glx_> it's "Concurrent spritegroups"
21:04:41 <andythenorth> we could also increase the limit
21:04:44 <will_marshall_> Yeah, I'm familiar.
21:05:33 <_glx_> andythenorth: main issue is 0xFF is not reserved
21:06:17 <_glx_> else ID would have become an extended byte a long time ago
21:06:55 <will_marshall_> So there are some (nfo-only?) projects that might not work?
21:07:07 <andythenorth> grf version bump
21:07:26 <will_marshall_> just scan the corpus of newgrfs to see if anything actually allocates an action2 in slot FF?
21:08:07 <_glx_> nml does, it starts at 0xFF then goes down
21:11:09 <_glx_> funny part is action3 already supports 16bit IDs
21:13:36 <andythenorth> fixed my issue by using the solution I already wrote for this ๐
21:41:37 <andythenorth> will_marshall_: someone had a scanner - rito12_13 maybe?
21:41:55 <audigex> andythenorth: Yes please
21:41:55 <audigex> It definitely feels a bit low for modern newGRF development where people are doing more advanced stuff, more templating etc which allows fancier random switches etc
21:49:40 <_glx_> andythenorth: no need to scan, all newgrf written in nml will use 0xFF ๐
21:51:20 <_jgr_> It would have to be opt-in, in which case you might as well roll it into the next version bump, whenever that will be
21:52:17 <_jgr_> Which then does create awkward issues for backwards compatibility and so on
22:03:54 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
22:04:58 *** MinchinWeb[m] has joined #openttd
22:05:02 *** _meja_ has quit IRC (Quit: User went offline on Discord a while ago)
22:09:35 *** jfs_ has quit IRC (Ping timeout: 480 seconds)
22:26:53 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:54:49 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
22:56:45 *** MinchinWeb[m] has joined #openttd
23:14:30 *** Twofish has joined #openttd
23:26:03 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
23:28:24 *** MinchinWeb[m] has joined #openttd
23:28:44 <belajalilija> I have a feature request but a hopefully very simple one
23:29:38 <belajalilija> I know recently the red/green toggle was made to resemble a physical switch for colour blind accessibility reasons
23:30:19 *** Twofish has quit IRC (Remote host closed the connection)
23:30:28 <belajalilija> When something is on it still has a green bit, but when off it is grey, could it be made red instead of grey?
23:30:52 <belajalilija> So it would give the best of both worlds
23:31:12 <belajalilija> If it would be better for me to say this in a github thingy could you link me to it?
23:33:25 *** Twofish has joined #openttd
23:53:49 *** SigHunter_ has joined #openttd
23:54:44 *** SigHunter has quit IRC (Quit: No Ping reply in 180 seconds.)
23:58:27 <belajalilija> here's a mockup i've made using BI (brain intelligence) with creative assistance from MS Paint
23:59:00 <tony_pixel> as in I (the person) and I - Intelligence
23:59:12 <belajalilija> <a:catyes:1333226855255244861>
continue to next day โต
