IRC logs for #openttd on OFTC at 2026-08-12
⏴ go to previous day
00:47:12 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
00:47:46 *** MinchinWeb[m] has joined #openttd
01:18:26 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
01:20:09 *** WormnestAndroid has joined #openttd
01:32:18 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
01:32:40 *** WormnestAndroid has joined #openttd
01:37:17 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
01:37:33 *** MinchinWeb[m] has joined #openttd
03:12:05 *** Philemon has joined #openttd
03:15:30 *** Phileman has quit IRC (Ping timeout: 480 seconds)
04:08:16 *** izhirahider_ has quit IRC (Read error: Connection reset by peer)
04:08:27 *** izhirahider_ has joined #openttd
04:14:58 *** Flygon has quit IRC (Read error: Connection reset by peer)
05:02:53 *** tokai|noir has joined #openttd
05:02:53 *** ChanServ sets mode: +v tokai|noir
05:10:05 *** tokai has quit IRC (Ping timeout: 480 seconds)
06:01:31 <DorpsGek> - Update: Translations from eints (by translators)
07:44:43 *** Extrems has quit IRC (Ping timeout: 480 seconds)
07:45:23 <_zephyris> peter1138[d]: For treeees, there's no way to force a single tree per tile, right?
07:45:49 *** Extrems has joined #openttd
07:46:38 <_zephyris> You can force a single tree _type_, by having one tree in that class. But you couldn't force a single tree per tile, by having one primary tree and one blank non-primary/non-purchaseable dummy tree in the class, because you'd end up with two of the primary sometimes.
07:47:06 <_zephyris> [Not that you'd really want to do that, but I'm trying to imagine extremes of what grf authors might want to do]
07:48:43 <peter1138> You mean halt the tree growth algorithm so that the tree count stays at 1?
07:57:29 *** wallabra has joined #openttd
08:04:12 <_zephyris> Yeah, that'd be one way to do it.
08:04:55 <_zephyris> I'm imaginint giant redwood trees on a tile, a way to limit that to one or two trees per tile would probably be desirable.
08:07:29 <peter1138> I'm kinda sad that the default attempt is "just make invisible trees"
08:08:12 <peter1138> I mean, it *should* work if self-probability is 0, but it's not really a great idea to go around resolving and drawing invisible sprites. That still takes resources.
08:15:04 <reldred> Stuff like giant redwoods though might be a bit weird with how fast trees currently grow and decay
08:15:33 <reldred> though I'm loathed to say it, almost seems more appropriate for an object 😬
08:15:45 <peter1138> That's why special things should just have extra properties/callbacks instead of hacking with invisible sprites :)
08:37:13 <peter1138> Any need for more than 7 growth stages?
08:37:27 <peter1138> I'm guessing/hoping not :-)
08:38:24 <peter1138> Tree cycle slowdown property?
08:43:46 <_zephyris> I'm thinking how to abuse the system, not sure it's needed
08:44:10 <_zephyris> Better to think about the hacks now, and try to be proactive
08:47:47 <peter1138> Well, if you are ending up with two of the primary sometimes even though self-probability is 0 and the classes contains a usable tree, that seems like a bug.
08:50:05 <_zephyris> I haven't tested it
08:50:57 <_zephyris> I'm in the thought experiment stage, coming up with desirable behaviours which aren't possible or would need a really hacky implementation
08:51:43 <peter1138> In that case I'm not sure why you thought you'd end up with two of the primary sometimes.
08:52:35 <_zephyris> My mistake, point is more that a max # tree/tile might be a nice setting
08:52:56 <_zephyris> To avoid people like me doing silly things like invisible sprites
08:53:23 <peter1138> Yeah, but just ask for that outright.
08:56:32 <peter1138> Growth speed might be interesting but also slightly awkward to get. Not from the generated tree tile spec, but from the tree spec of the last tree on the tile.
09:14:58 <peter1138> Anyway, it's also better not to try to stuff everything into an initial implementation. But yeah that risks people bodging things instead of just asking for stuff.
09:19:51 <peter1138> (But still plan for it :))
09:34:16 <_zephyris> Growth rate I'm not worried, can't see how to bodge that.
09:34:33 <_zephyris> I didn't ask outright because I was thinking about it the wrong way 😉
09:34:52 <_zephyris> But now you've suggested it, #trees/tile would be great!
09:45:30 *** andythenorth_ has quit IRC (Quit: Connection closed for inactivity)
10:55:20 <peter1138> > load average: 39.40, 40.77, 40.24
10:59:35 <peter1138> My postgres secondary is out of date, and replaying it is killing it.
10:59:47 <peter1138> I may have to just drop it and start from a fresh backup.
11:00:38 <LordAro> reldred: you never know, it might be a 96 core monstor
11:04:08 <reldred> LordAro: Looks more like a dual ten core twenty thread xeon to me lol
11:04:54 <reldred> I may have had a few r660’s and r650’s with that configuration over the years 🤣
11:05:52 * peter1138 sighs. It's a dual-core i5-4460...
11:07:55 <__abigail> I have 2 12-thread Xeons
11:08:47 <reldred> peter1138: Oh you poor thing 🙁
11:09:14 <peter1138> It's fine as a database secondary, as long as it's not massively out of date.
11:09:38 <reldred> That said my personal cluster is a trio of i5 8500t’s, they’re a little anemic
11:10:00 <peter1138> My main server is a miserly quad-core Xeon, and only 16GB RAM.
11:10:03 <reldred> But, they’re quiet and low TDP
11:10:37 <peter1138> Yeah, there's spare dual-CPU 12-core 256GB RAM behemoth, but I'm not running that.
11:11:36 <peter1138> Maybe I should look at M920qs...
11:11:47 <reldred> Also how are you getting a load of 40 on a dual core cpu
11:12:46 <reldred> I’ve been enjoying my m920q’s. I’m on the look out for some m90q’s though, more modern and more efficient CPU’s and dual nvme as well as pcie slot and a SATA ribbon connector on the main board
11:12:50 <LordAro> i got a load of 1700 the other day
11:12:55 <LordAro> it was quite unhappy with me
11:13:21 <reldred> Already posted it on fedi but the rack mounts arrived from the netherlands
11:13:24 <peter1138> So postgres WAL playback + rsync copying new WAL + dotnet CI job.
11:14:07 <peter1138> The cursed power supply sockets though.
11:15:09 <reldred> I have them velcro’d together in a brick
11:15:22 <peter1138> Those are not poverty spec :o
11:15:58 <reldred> I bought em when they were cheaper
11:16:18 <reldred> shoulda put more ram in them before the apocalypse happened
12:25:25 <andythenorth> `Load Avg: 17.78, 7.04, 4.40`
12:25:31 <andythenorth> during a Horse compile
12:26:02 <andythenorth> but also the AV scan is running, and Apple's full disk indexing at the same time 😐
12:27:04 <andythenorth> horse is quite pythony so can't use many cores
12:29:15 <andythenorth> `Load Avg: 42.20, 15.97, 8.36` OpenTTD compile does a bit better
13:08:40 *** Tirili has quit IRC (Ping timeout: 480 seconds)
13:24:44 <LordAro> eclipse glasses acquired
13:34:25 <peter1138> I will just not look at the sun.
13:35:28 <LordAro> who are you, so wise in the ways of science?
13:47:35 <andythenorth> I seed an eclipse before
13:53:23 <peter1138> There's been a few.
14:02:43 *** Tirili has quit IRC (Remote host closed the connection)
14:38:42 *** xdfsxdfsdf has joined #openttd
14:39:31 <xdfsxdfsdf> Can it be compiled to run on ARM/windows instead of x86-64?
14:44:34 <peter1138> Yeah, we build for 32-bit and 64-bit x86, and also for 64-bit ARM.
15:18:08 *** xdfsxdfsdf has quit IRC (Quit: Page closed)
16:10:25 <peter1138> Another happy customer.
16:39:28 <andythenorth> winner winner chicken dinner
16:39:32 <andythenorth> speaking of lunch also
16:55:55 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
18:03:48 *** gelignite has joined #openttd
18:37:20 <_zephyris> Helps if it's cloudy
18:43:39 <andythenorth> my neighbour had a pinhole box
19:38:42 <peter1138> `Vehicle *v = Vehicle::Get(this->window_number); this->selected_vehicle = v->index;`
19:38:56 <peter1138> Isn't that a long-winded way of doing `this->selected_vehicle = this->window_number:`?
19:41:44 <Rubidium> yeah, but without `static_cast` ;)
20:26:14 <peter1138> I think you made WindowNumber a bit magic, it doesn't need a cast.
21:32:15 <peter1138> The good ol' Ctrl-Z undo function.
continue to next day ⏵
