IRC logs for #openttd on OFTC at 2025-09-01
โด go to previous day
01:47:17 *** Wormnest has quit IRC (Quit: Leaving)
03:10:20 *** Zathras has joined #openttd
03:13:44 *** Zathras_11 has quit IRC (Ping timeout: 480 seconds)
03:36:50 <bigyihsuan> I wonder what would it take to add 4x4 grid for towns
03:37:05 <bigyihsuan> Would houses even be placed in the middle 4 squares?
03:37:20 <bigyihsuan> Sometimes I see empty holes in the middle square in 3x3
04:42:14 <DorpsGek> - Update: Translations from eints (by translators)
06:56:24 <peter1138> _glx_, so we don't actually need to store size at all because new/delete has it hidden away already?
07:06:14 <DorpsGek> - Add: summary for week 35 of 2025 (by OpenTTD Survey)
07:09:04 <andythenorth> what's after 15? ๐
07:09:11 <andythenorth> the internet has many hopes and dreams about it
07:32:39 <peter1138> Hmm, someone's been making branches on upstream again ;-)
07:42:30 <peter1138> Monday morning, 18 minutes to go. Do I have enough time to update to Debian 13? ;p
07:46:54 <peter1138> I wonder if I should wipe Windows from my work laptop.
07:48:06 <peter1138> I'm not sure it is even my work laptop any more, as I don't tend to use it. 13" display and laptop keyboards are not great for actually doing work.
07:56:49 <reldred> Lol, my new work laptop is locked down like fort bloody knox
07:56:59 <reldred> it's also not my problem if things don't work on it
07:57:29 <reldred> god I love working for big hole
07:57:53 <reldred> some of the policies and procedures are a bit onerous but eh
08:00:44 *** mindlesstux has quit IRC (Ping timeout: 480 seconds)
08:42:16 <peter1138> CodeQL is slow these days ;(
08:47:16 <_zephyris> talltyler: I started writing notes on this... Then stopped and forgot. IMO sync _only_ saves, plus maybe config.
08:47:21 <peter1138> `const_cast<const void **>(reinterpret_cast<const void *const *>())` That's somewhat terrifying.
08:49:12 <peter1138> Fallback fonts: It looks like on MacOS, bold fonts are excluded "Skip bold fonts (especially Arial Bold, which looks worse than regular Arial)"
08:49:58 <_zephyris> Sharing (ie copy/paste) save games works well. Can download missing online content, warns about missing local content. So, good enough for the vast majority of users.
08:50:27 <peter1138> On Windows, it looks like the fallback font finder doesn't consider font weight at all.
08:50:49 <_zephyris> Config takes a little more thought. A few settings are much more computer-specific, eg gui scale, track pad scroll. Naive config sync is OK, but ideally split out very computer/hardware specific settings to local only.
08:50:53 <peter1138> And with font-config, it explicitly looks for the heaviest font it can find.
08:51:32 <peter1138> Can anyone confirm this behaviour on Windows and/or Mac OS? If it picks a fallback font, is normal weight, or bold?
08:53:22 <peter1138> (Because _zephyris did a good job with the built-in scalable font, you'll need to pick a non-latin language to test it these days.)
08:55:05 <_zephyris> Vietnamese is always a good test case... Gives a regular weight font for me (looks like arial?)
08:55:50 <_zephyris> If you set the font in cfg to an unknown font (eg. Foo), does that trigger the same fallback? That gives regular Arial ish for me
08:57:10 <peter1138> Hmm, on Linux if you do that it just falls back to the sprite font.
08:58:13 <peter1138> CJK fonts also have various weights.
09:03:19 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
09:14:28 <LordAro> reldred: 'big hole' :D
09:19:48 <reldred> LordAro: we dig holes and sell the things we find
09:22:30 <andythenorth> he works for Minecraft
09:22:56 <reldred> the children, they yearn for the mines.
09:23:43 <LordAro> 'big hole employee confirms child labour'
09:29:50 <peter1138> Hmm, dynaplug tyre plugs are quite expensive :o
09:35:00 <peter1138> I left OpenTTD running with the original music going. There's a bit where it pitch bends and it goes a bit whacky, because pitch bend can be variable. (It's supposed to be +- 2 semitones.)
09:37:22 <peter1138> Heh, the tracks that are broken.
09:40:22 <reldred> I shouuld play some more D2, I'm surprisingly close to finishing act 5 in hell difficulty with my melee (technically crossbow) sorceress
09:45:07 <peter1138> Gosh, why would you play a 25 year old game............
09:45:33 <reldred> technically I'm playing the remaster that came out a few years ago
09:45:42 <reldred> tis a really good remaster :>
09:47:44 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzzโฆ)
10:10:15 <peter1138> Hmm, am I missing the bit in the Windows fallback font stuff that allows/rejects very light or very heavy fonts?
10:43:35 *** toktik has quit IRC (Remote host closed the connection)
10:49:53 *** dh1 has quit IRC (Ping timeout: 480 seconds)
11:03:56 *** dh1 has quit IRC (Ping timeout: 480 seconds)
11:04:00 <peter1138> If there are no mac os users to test things, does that mean we don't need to support mac os?
11:11:32 <andythenorth> buy one of my used macs ๐
11:16:08 <cu-kai> i played openttd on macOS for the longest time without realising hardware acceleration was disabled
11:16:16 <cu-kai> (going back a few years now, before i realised it was a thing)
11:16:29 <andythenorth> last time I tested, it's faster without hardware acceleration
11:16:44 <cu-kai> definitely the other way around in my experience
11:16:54 <andythenorth> results are quite variable
11:17:00 <cu-kai> it was only recently that i wondered why it made the machine heat up & fans spin
11:17:07 <cu-kai> mind you, that isn't difficult on an i9 2019 model
11:17:14 <andythenorth> I had that model
11:17:23 <andythenorth> the fans clogged with fluff
11:17:33 <andythenorth> they suck in everything through the bottom vents
11:17:58 <peter1138> HW acceleration is very unlikely to have enough of an impact to affect that.
11:18:23 <cu-kai> i was recently tempted by an M2 max model on ebay, so i could continue to tinker with FOSS on a modern mac platform... but it's ยฃ1.7k or something
11:18:39 <cu-kai> it definitely felt more sluggish with HW accel disabled
11:18:40 <peter1138> All it does is offload palette animation, at the expensive of more data being moved around.
11:18:57 <cu-kai> could be placebo i guess, i'll test it again some time
11:19:52 <andythenorth> I have an M1 16" MBP to sell
11:19:53 <peter1138> I guess the other thing is HW acceleration allows vsync to be used.
11:20:07 <andythenorth> and an M1 13" MBP with a cracked bezel (but the screen is fine)
11:20:21 <cu-kai> the only reason i was looking for a max model was, compile times lol
11:21:19 <andythenorth> I bought an M4 Max, the OpenTTD compile time is wild
11:22:20 <cu-kai> how would you say it is vs the older machines?
11:24:32 <andythenorth> both the M1 are incredibly good value
11:24:42 <andythenorth> the M4 was twice as fast for the things I do
11:25:07 <andythenorth> but previously I went from the high end i9 to M1, and that was transformational
11:28:38 <LordAro> the container that contains our AI experimentation has had a service try to restart 14000 times in the last 4 days. For Reasons, this includes redownloading an entire pytorch dependency tree every time
11:28:53 <LordAro> the container is now taking up 800G
11:45:04 <_zephyris> Feels like an nml release would be useful, especially once station bridge shenanigans are sorted.
11:45:28 <talltyler> Agreed, it would help GRF authors test beta3 features.
11:45:39 <talltyler> (Station bridges and blocked pillars)
12:06:54 *** WormnestAndroid has joined #openttd
12:14:56 *** toktik has quit IRC (Remote host closed the connection)
12:17:10 <andythenorth> Was nml badges PR merged?
12:17:26 <andythenorth> (On my phone, hard to check)
12:19:45 <peter1138> No, it needs rebasing to smush it all together.
12:28:16 <rito12_51026> How this bitmask for pillars flags should work?
12:35:16 <peter1138> Look at how station tile flags work? You already copied (possibly erroneously) part of that.
12:45:36 *** orudge` has joined #openttd
12:45:37 *** ChanServ sets mode: +o orudge`
13:09:56 <peter1138> will there ever be a hex tile mode?
13:10:30 <peter1138> I can confidently say... potentially probably maybe unlikekily not.
13:10:47 <LordAro> why do i seem to recall some sort of hacky mockup?
13:12:25 <_glx_> might need a mix of `ctt_list`, `station_tile_flags` and `byte_sequence_list` (for a function accepting prop_num, an array of odd length and customisable error message)
13:13:33 <_glx_> *even lenght (for people on irc)
13:15:56 <rito12_51026> peter1138: Ye๏ปฟ but what flags should it contain?
13:17:27 <peter1138> If you don't know perhaps it's best to let someone else handle this.
13:35:20 <peter1138> (It is documented in almost the same place as prop 20 though)
13:38:44 <peter1138> ((Although there's no suggested naming for each flag. The OpenTTD source has some names.))
13:40:03 <rito12_51026> `enum class BridgePillarFlag : uint8_t {
13:40:03 <rito12_51026> /* Corners are in the same order as Corner enum. */
13:40:03 <rito12_51026> CornerW = 0, ///< West corner is obstructed.
13:40:03 <rito12_51026> CornerS = 1, ///< South corner is obstructed.
13:40:04 <rito12_51026> CornerE = 2, ///< East corner is obstructed.
13:40:04 <rito12_51026> CornerN = 3, ///< North corner is obstructed.
13:40:06 <rito12_51026> /* Edges are in the same order as DiagDirection enum. */
13:40:06 <rito12_51026> EdgeNE = 4, //< Northeast edge is obstructed.
13:40:08 <rito12_51026> EdgeSE = 5, //< Southeast edge is obstructed.
13:40:08 <rito12_51026> EdgeSW = 6, //< Southwest edge is obstructed.
13:40:10 <rito12_51026> EdgeNW = 7, //< Northwest edge is obstructed.
13:40:39 <_glx_> flags will go in global_constants.constant_numbers
13:43:45 <_glx_> oh another repo requiring some muting of the bot
13:46:16 *** dh1 has quit IRC (Ping timeout: 480 seconds)
15:15:50 <LordAro> the linters failing is a sign you should set up your editor properly, not just spam commits until it works ;)
15:16:40 <rito12_51026> LordAro: I should fix proper editor and do not use gedit
15:18:44 <peter1138> No offence but this is quite a farce.
15:19:25 <LordAro> peter1138: we were all new once
15:22:04 <_glx_> or use the makefile (`make black` and `make flake` are present for a reason)
15:35:20 <peter1138> Yes, it's not been updated from the original fixed-length fields.
15:44:16 <rito12_51026> LordAro: What should I put in black settings?
15:44:56 <rito12_51026> I have tried `-l 120 --exclude "(ply|nml/actions/action2var_variables.py|nml/actions/action3_callbacks.py)" --check --diff nml` and it did not run
15:46:44 <LordAro> define "it did not run"
15:47:10 <_glx_> I just use make before commit, simpler
15:48:21 <rito12_51026> PyCharm showed `Black: Failed to format file action0properties.py`
15:48:50 <LordAro> i suspect what it actually means is 'black returned an error'
15:48:55 <LordAro> which it will have done with the --check flag
15:49:03 <rito12_51026> and `Usage: python -m black [OPTIONS] SRC ...`
15:49:21 *** orudge` has quit IRC (Ping timeout: 480 seconds)
15:49:52 <LordAro> follow the same process as reproducing any other bug - simplify and add options until you work out what's broken
15:50:15 <_glx_> makefile uses `BLACK_OPTIONS=-l 120 --exclude 'action2var_variables.py|action3_callbacks.py|ply'`
15:50:35 <LordAro> having that 'nml' in the pycharm options is probably doing something undesirable as well
16:21:04 <ahyangyi> Anyways, would it be easier to move the many `black` options into `pyproject.toml`?
16:21:29 <LordAro> (something something uv/ruff everything)
16:24:40 <ahyangyi> I do use that combo for my new projects, but I'm fine with my older projects to stick to whatever installer/formatter/linter they happen to be using.
16:27:30 <ahyangyi> Only exception being grf-py which currently manually pins all dependencies, which I want to split into broad ranged `pyproject.toml` and a precise lockfile.
16:27:41 <ahyangyi> ( but too lazy to actually do that and send in a PR )
16:33:26 <rito12_51026> As you use uint16_t for size of array I thought that it can store 65535 elements
16:34:30 *** Wormnest has joined #openttd
17:10:13 <michi_cc> peter1138: Lot's of things can have changed on macOS since I wrote the font stuff, but at least back then for many of our languages Arial Bold was always the first match, and at least on macOS 10.whatever, it IMNSHO rendered a lot worse than normal Arial.
17:46:24 <Rubidium> LordAro: would be nice if we could do uv/ruff style changes on the main C++ code as well :D
17:47:27 <LordAro> (see the many previous attempts with clang-format)
17:49:12 <peter1138> I'm using frosch's .clang-format still, for the most part it's pretty usable.
17:54:42 <andythenorth> I think I know the answer....but
17:54:55 <andythenorth> there's no way to hint to the pathfinder about a preferred railtype for a train?
17:55:22 <andythenorth> consists of mixed RAIL / ELRL engines will happily use RAIL when an ELRL alternative exists, if the lead engine is RAIL
17:55:27 <andythenorth> then lower power
17:55:30 <LordAro> only speed based & compatibility afaik
17:58:46 *** dh1 has quit IRC (Ping timeout: 480 seconds)
18:07:59 <andythenorth> railtype subtypes?
18:08:04 <andythenorth> railtype variants? ๐
18:08:31 * andythenorth exploring railtype label semantics
18:11:31 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
18:13:05 <talltyler> Mixed RAIL / ELRL engines sounds like something has already gone wrong, or you're just roleplaying something I haven't heard of
18:13:22 <fairyflossy> talltyler: DVT up front, ELRL engine in the back
18:14:19 <talltyler> The solution is for us to add support for unpowered DVTs/cab cars/etc leading trains, not faff around with railtypes ๐
18:14:33 <andythenorth> but the lead vehicle would still be RAIL
18:14:48 <andythenorth> so the consist would then route to an unpowered section
18:15:15 <talltyler> Presumably the pathfinder would look at powered vehicles, not just whatever is first
18:16:31 <talltyler> I realize none of this helps you right now ๐
18:18:01 <talltyler> Does it? That looks like something else. Your DVT is actually powered on RAIL, but with very little power.
18:19:40 <peter1138> Someone else deal with it.
18:22:44 <andythenorth> talltyler: hmm yes
18:23:25 <peter1138> michi_cc, I was mainly concerned about the inconsistency of it all... some prefer regular, some prefer super heavy.
18:23:27 <talltyler> I will look at unpowered driving vehicles, but need help with the tech debt of consists first ๐
18:23:48 <talltyler> (complexity of work, not amount of it)
18:23:54 <andythenorth> FWIW I haven't encountered the DVT issue in game, and I don't think it's necessarily wrong
18:24:14 <andythenorth> at some point 'play stupid games, win stupid prizes' has to be applied
19:30:01 <michi_cc> peter1138: Well, regular weight is consistent with Win32 at least, as far as I know `EnumFontFamiliesExA` will default to regular weight if it exists for the font family.
19:30:37 <peter1138> FontConfig is "lol, you didn't specify so you get anything"
19:42:46 *** tokai|noir has joined #openttd
19:42:46 *** ChanServ sets mode: +v tokai|noir
19:43:14 *** pickpacket0 has joined #openttd
19:45:20 *** wallabra2 has joined #openttd
19:46:18 *** wallabra has quit IRC (Remote host closed the connection)
19:46:19 *** guru3 has quit IRC (Read error: Connection reset by peer)
19:46:19 *** herms has quit IRC (Remote host closed the connection)
19:46:19 *** Yexo has quit IRC (Ping timeout: 480 seconds)
19:46:19 *** herms0 is now known as herms
19:47:09 *** dh1 has quit IRC (Remote host closed the connection)
19:48:06 *** dihedral has quit IRC (Ping timeout: 480 seconds)
19:49:01 *** pickpacket has quit IRC (Ping timeout: 480 seconds)
19:49:01 *** pickpacket0 is now known as pickpacket
19:49:21 *** tokai has quit IRC (Ping timeout: 480 seconds)
21:05:08 *** ChanServ sets mode: +v tokai
21:08:50 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:12:04 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
21:21:19 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
21:21:21 *** WormnestAndroid has joined #openttd
21:21:33 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
21:21:35 *** WormnestAndroid has joined #openttd
21:21:38 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
21:21:38 *** WormnestAndroid has joined #openttd
21:28:50 *** keikoz has quit IRC (Ping timeout: 480 seconds)
continue to next day โต
