IRC logs for #openttd on OFTC at 2026-01-23
⏴ go to previous day
02:57:57 *** Wormnest has quit IRC (Quit: Leaving)
04:53:33 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
04:53:35 *** WormnestAndroid has joined #openttd
04:53:56 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
04:53:57 *** WormnestAndroid has joined #openttd
04:54:27 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
04:54:28 *** WormnestAndroid has joined #openttd
05:39:43 *** WormnestAndroid has quit IRC (Remote host closed the connection)
05:39:47 *** WormnestAndroid has joined #openttd
06:00:56 *** ChanServ sets mode: +v DorpsGek
06:00:56 *** ChanServ sets mode: +v blathijs
06:00:56 *** ChanServ sets mode: +v peter1138
06:00:56 *** ChanServ sets mode: +v orudge
07:18:03 *** Smedles has quit IRC (resistance.oftc.net singleton.oftc.net)
07:18:03 *** Flygon has quit IRC (resistance.oftc.net singleton.oftc.net)
07:18:07 *** Smedles has joined #openttd
07:19:09 *** WormnestAndroid has quit IRC (resistance.oftc.net reflection.oftc.net)
07:19:09 *** Extrems has quit IRC (resistance.oftc.net reflection.oftc.net)
07:19:09 *** keisisqrl has quit IRC (resistance.oftc.net reflection.oftc.net)
07:19:10 *** WormnestAndroid has joined #openttd
07:19:10 *** Extrems has joined #openttd
07:19:10 *** keisisqrl has joined #openttd
07:21:10 *** toktik has quit IRC (resistance.oftc.net graviton.oftc.net)
07:25:04 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
07:25:22 *** MinchinWeb[m] has joined #openttd
07:30:47 *** dwfreed has joined #openttd
07:43:12 <andythenorth> I'll test 15145 today
07:53:47 *** Zathras_7 has joined #openttd
08:03:06 *** SmatZ has quit IRC (Server closed connection)
08:03:22 *** avdg has quit IRC (Server closed connection)
08:03:38 *** Guest8291 has quit IRC (Server closed connection)
08:03:53 *** ^Spike^ has joined #openttd
08:03:54 *** Terkhen has quit IRC (Server closed connection)
08:04:27 *** ^Spike^ is now known as Guest15
08:04:32 *** Terkhen has joined #openttd
08:04:32 *** ChanServ sets mode: +o Terkhen
08:09:14 *** Heili has quit IRC (Server closed connection)
08:15:38 *** fonsinchen has quit IRC (Server closed connection)
08:15:54 *** XeryusTC has quit IRC (Server closed connection)
08:16:02 *** fonsinchen has joined #openttd
08:16:32 *** XeryusTC has joined #openttd
08:26:29 *** Smedles has quit IRC (Server closed connection)
08:26:41 *** Smedles has joined #openttd
08:31:10 *** twpol has quit IRC (Server closed connection)
08:43:38 *** welterde has quit IRC (Server closed connection)
08:44:11 *** welterde has joined #openttd
08:47:06 *** KenjiE20 has quit IRC (Server closed connection)
08:47:31 *** KenjiE20 has joined #openttd
08:49:14 *** dih has quit IRC (Server closed connection)
08:50:55 *** dihedral has joined #openttd
08:51:06 *** Yexo- has quit IRC (Server closed connection)
09:00:27 *** kaibaneddy has joined #openttd
09:00:27 <kaibaneddy> who'd have thought
09:02:08 <peter1138> Potential for breaking more stuff than it fixes?
09:05:30 *** Cursarion has quit IRC (Server closed connection)
09:05:35 *** Cursarion has joined #openttd
09:22:06 *** Ttech has quit IRC (Server closed connection)
09:22:17 *** planetmaker has quit IRC (Server closed connection)
09:22:32 *** planetmaker has joined #openttd
09:22:32 *** ChanServ sets mode: +o planetmaker
09:24:48 <pickpacket> I found something interesting. I built a station that didn't accept cargo X but then told trains to unload all X there. Later, as the station was filling up and I realised my mistake, I extended the station coverage to actually cover the industry that accepts X. I expected it to immediately empty the station of all cargo that was piled up, but it
09:25:52 <peter1138> Yes, cargo is actually unloaded from vehicles straight to the industry, not onto the station and then onto the industry.
09:27:30 <peter1138> Imagine if you had a station servicing a supplying industry with cargo waiting to be loaded onto trains. You wouldn't want that to automatically start moving to a new industry that popped up within the station's catchment.
09:27:34 <pickpacket> hmmm. Is this something worth changing? As in this special case?
09:31:14 <andythenorth> FIRS broke it again
09:31:14 <LordAro> i'd sooner remove the unload all order flag tbh
09:31:45 <LordAro> i can't think of any valid use for it
09:31:55 <peter1138> LordAro, yes, or at least change the wording of it. I think I mentioned that before, but, uh, didn't do anything about it.
09:32:02 <peter1138> Does 15150 break FIRS?
09:32:39 <peter1138> Hmm, doubling checking.
09:33:15 <peter1138> Ah it's fine, an existing translation table is fully cleared (for now.)
09:33:23 <peter1138> Only badges have special treatment.
09:34:32 <andythenorth> all this table stuff arose because I insisted on declaring newgrf cargos in alphabetical order?
09:34:45 <andythenorth> personality-disorder-driven development
09:35:06 <pickpacket> LordAro: unload all is necessary sometimes. I have a station that covers both a tea factory and a tea tree farm, but when trains with tea leaves arrive I don't want them to unload for the factory and then pick up from the farm because there's no factory at the other end of the line
09:35:28 <peter1138> It arose because all sorts of special casing were simplified by making everything use a translation table, whether that was installed by the NewGRF or a default one.
09:35:45 <peter1138> Now I'm de-simplifying it a bit :)
09:35:57 <peter1138> Well, probably it wouldn't've happened without that.
09:36:07 <_zephyris> FWIW I think your PR looks sensible
09:36:47 <peter1138> pickpacket, sounds like you actually want a no-load order, not an unload order.
09:41:08 <pickpacket> peter1138: that's true
09:41:29 *** locosage has joined #openttd
09:41:50 <locosage> so that seems to have been in the spec since forever but the game only stopped copying these flags after 13.4?
09:46:28 <locosage> it broke pcb server grf 😅
09:46:28 <peter1138> Well, broken NewGRFs aside, it's now correct.
09:46:49 <peter1138> Funnily enough, Master Hellish thought there was no NewGRFs in that game.
09:47:17 <peter1138> But that would explain why nobody could recreate it without any.
09:48:03 <locosage> well, none were visible 🤣
09:48:10 *** SpComb has quit IRC (Server closed connection)
09:48:16 <peter1138> It's in the list, no?
09:48:41 <locosage> well, if you don't suspect any newgrfs why open the list
09:48:48 <locosage> even I forgot about pcb server thing
09:49:09 <locosage> but, yeah, I've narrowed it down to a single action0 with just prop 08
09:49:55 <locosage> pcb server doesn't set chuch/stadium flags coz it worked fine before
09:50:05 <peter1138> Should've read the spec :)
09:51:29 <locosage> Well, that goes for openttd devs too 😜
09:51:44 <peter1138> We did, hence it got fixed :)
09:52:55 <locosage> hopefully The Dude can still fix pcb server grf too
09:52:55 <peter1138> As far as breaking things go, it's pretty minor and doesn't prevent games working, unlike say #15116.
09:53:08 <locosage> not that it's particularly necessary grf nowadays
09:54:27 <locosage> all it does for houses is reset authority impact and increase mail acceptance
10:06:00 *** Flygon_ has joined #openttd
10:06:50 *** humanfish451 has quit IRC (Server closed connection)
10:06:50 *** humanfish451 has joined #openttd
10:11:53 *** Flygon has quit IRC (Ping timeout: 480 seconds)
10:33:46 *** Mek has quit IRC (Server closed connection)
10:36:06 *** moll has quit IRC (Ping timeout: 480 seconds)
10:44:42 *** guru3 has quit IRC (Server closed connection)
10:49:30 *** ckb has quit IRC (Server closed connection)
11:28:46 <peter1138> Is it slower because you increase the update rate? Might be best not to do that :)
11:28:56 <peter1138> And... this is IRC not Github, again.
11:29:30 <Borg> uh oh.. so BSPI broken on 14.1 ;> heh ;) thats ok
11:34:04 <peter1138> 14.1 came out a long time ago.
11:37:13 <Borg> im commenting more about issue #15116
11:37:35 <Borg> not sure. I didnt tested.. but comment match :>
11:38:16 <peter1138> BSPI doesn't seem to add any new cargo types?
11:39:52 <peter1138> And the issue affects 15.0, not 14.1.
11:44:23 <Borg> old one doesnt.. it just touches cargos for custom rating calc..
11:44:30 <Borg> but uses default cargos.
11:44:42 <peter1138> Not a problem then.
11:49:25 <kuhnovic> peter1138: Generating trees during map gen is slower compared to the old ways, but it's not too bad tbh. It's probably a bit slower during runtime as well due to more frequent updates, but tree growth is not really a heavy process to begin with.
11:55:10 *** greeter has quit IRC (Server closed connection)
11:55:42 *** greeter has joined #openttd
11:58:19 <peter1138> Arguably tree growth should be slower, those thigs last years :)
11:58:53 <kuhnovic> It's a game not a simulator 😉
11:59:42 *** rightnut has quit IRC (Server closed connection)
11:59:43 *** rightnut has joined #openttd
12:02:22 <LordAro> kuhnovic: since when?!
12:02:48 <kuhnovic> Could have seen that one coming haha
12:05:08 <kuhnovic> As requested by LordAro
12:06:59 <kuhnovic> Looking at it a year later, I do wonder if i've overcomplicated things a bit
12:54:31 <peter1138> Ah yes, because of course microsoft have the TLD .microsoft...
12:56:59 <kaibaneddy> what went wrong with #15089 , btw?
12:57:46 <kaibaneddy> 15088 feels like a pretty significant UX bug that ought to be fixed asap (in 15.1?)
13:16:39 *** mmtunligit has joined #openttd
13:16:39 <mmtunligit> LordAro: Unload all is very sueful if youre doing 24 hr scheduling and want to make sure all passengers are off the train when you send it to a yard for overnight storage
13:47:38 *** Binette[m] has quit IRC (Server closed connection)
13:47:40 *** Binette[m] has joined #openttd
13:49:52 <_glx_> looks like expected based on the nml code
13:56:28 <peter1138> You mean it's intended to use the railtype of the tile below the bridge?
13:58:04 <_glx_> that's the variable meaning yes
13:59:06 <_glx_> bridge railtype is only known on bridge head
13:59:27 <peter1138> In tunnelbridge_cmd.cpp, line 1684 uses rampsouth and TCX_ON_BRIDGE, but line 1692 uses ti->tile instead.
13:59:36 <peter1138> Probably 1692 should also use rampsouth and TCX_ON_BRIDGE
14:00:09 <_glx_> but how do you determine if you want bridge or under bridge ?
14:02:31 <_glx_> hmm but maybe you're right
14:05:32 <peter1138> Authors making OpenTTD do more work
14:10:11 <peter1138> It's confusing because it's talking about bridges, but of course there's no varactions for bridges yet :)
14:17:00 <peter1138> Urgh at VS trying hard to discourage rebases.
14:17:10 <peter1138> I do not like its git interface :/
14:17:39 <_glx_> anyway you are right, it's actually a mistake in reservation drawing, I'll PR
14:18:24 <peter1138> If the NewGRF wasn't doing silly things it would work of course :)
14:34:02 *** aperezdc has quit IRC (Server closed connection)
14:34:13 *** aperezdc has joined #openttd
14:38:17 <_glx_> funny no newgrf triggered it for more than 10 years 🙂
14:50:44 <peter1138> NML does make it very easy to do slight more inefficient things. Hmm.
14:55:44 *** f_ is now known as funderscore
15:03:49 <_glx_> yup nml is very good for that 🙂
16:02:04 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
16:02:36 *** MinchinWeb[m] has joined #openttd
16:04:43 *** Wormnest has joined #openttd
16:42:51 *** nebulabc has quit IRC (Server closed connection)
16:42:52 *** nebulabc has joined #openttd
16:54:42 *** gelignite has joined #openttd
17:15:39 *** funderscore has quit IRC (Server closed connection)
17:15:40 *** funderscore has joined #openttd
17:20:59 *** ufo-piloot has quit IRC (Server closed connection)
17:21:12 *** ufo-piloot has joined #openttd
17:27:55 *** peter1138 has quit IRC (Server closed connection)
17:28:09 *** peter1138 has joined #openttd
17:28:09 *** ChanServ sets mode: +o peter1138
17:28:27 *** FLHerne has quit IRC (Server closed connection)
17:28:27 *** FLHerne has joined #openttd
17:44:37 <Zathras_7> Climate sub-tropical: Cities that require water to grow but do not have a water tower. is that intentional or could this be seen as a bug?
17:46:32 <Zathras_7> Towns without a water tower somehow seem more reasonable, though.
17:49:08 <talltyler> You have to fund a water tower in those cities.
17:49:34 <Zathras_7> but that is _extremely_ expensive
17:51:22 <LordAro> rito12_51026: intriguing. seems like microbenchmarks though - try larger saves?
17:51:38 <LordAro> also, if you could break up the PR description into paragraphs, that'd make it an awful lot more readable
17:52:53 <andythenorth> should I get 2 more cores to make Horse compile faster?
17:53:07 <andythenorth> the 16 core is on refurb
17:53:17 <LordAro> just get an EPYC server
17:53:24 <LordAro> 128 cores ought to be enough
17:54:00 <andythenorth> yes, especially when 50% of total compile is in single threaded python 😛
17:54:21 <LordAro> fine, Threadripper then
18:02:08 <rito12_51026> LordAro: even better
18:08:58 <peter1138> Splitting tree type across map bytes just to save disk space seems wrong.
18:11:14 <xarick> i have a different idea
18:17:22 <xarick> I don't know where you got this guy from, but I think he's a genius
18:20:55 <ahyangyi> andythenorth: Convert two inches into two CPU cores? MacBook is weird.
18:26:18 <andythenorth> ahyangyi: I did a wrong I guess 🙂
18:26:36 <xarick> the other tree pr is 11263?
18:27:13 <andythenorth> wasn't trees going to just be random per tile?
18:41:24 <_zephyris> Kinda mad, and I kinda love it
18:42:52 <_zephyris> I assume it makes maintenance harder/more confusing
19:19:18 <_zephyris> The reckless disregard for elegance
19:19:53 <_zephyris> .... Yet somehow being elegant through simplicity
19:25:36 <peter1138> I'm not fond of changing the map storage around like this.
19:25:55 <peter1138> There's also no performance metrics, which is more important than size.
19:29:25 <locosage> it sure feels ugly but if that's the only solution we can have I'd still call it an improvement ;p
19:30:38 <locosage> doubt it will have any impact on performance, trees aren't really a performance issue
19:31:27 <locosage> but won't hurt to check ofc
19:41:03 <peter1138> A different (but probably much slower loading/saving) approach would be to store all the tile types, then store all tile data grouped by tile type, so all MP_CLEAR, then all MP_RAILWAY, then all MP_ROAD, etc, etc...
19:42:06 <peter1138> Similar data would then be clustered without relying on hacks like setting things that should be zero to non-zero.
19:43:39 *** Heiki has quit IRC (Server closed connection)
19:44:23 <peter1138> It could even de-pack each tile into a properly save-game described table, though that would likely make things bigger again.
19:47:26 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
19:47:46 *** MinchinWeb[m] has joined #openttd
19:48:39 <_jgr_> Storing each tile together rather than having each sub-field in its own chunk would presumably improve locality issues without needing to change the overall tile ordering
19:49:46 <_jgr_> Or having to change the tile contents themselves
19:55:50 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
19:58:52 *** MinchinWeb[m] has joined #openttd
20:04:39 *** hazardaj_nombroj has joined #openttd
20:05:03 <peter1138> Okay, so I could try that, storing tiles "horizontally" instead of "vertically"
20:05:46 <rito12_51026> Will that improve anything?
20:05:57 <peter1138> I won't know until I try it.
20:07:08 <hazardaj_nombroj> anyone can tell me when STR_INDUSTRY_VIEW_ACCEPT_CARGO strings appear in the game?
20:07:31 <hazardaj_nombroj> is it a newgrf thing, where some industries accept a specific amount of cargo?
20:10:49 <hazardaj_nombroj> there are ones with a suffix and without, and with an amount and without
20:10:58 <hazardaj_nombroj> i've only seen the most basic one in my experiecne
20:12:45 <xarick> find STR_INDUSTRY_VIEW_ACCEPT_CARGO returned 0 results :|
20:13:41 <xarick> it's not in english.txt
20:14:12 <hazardaj_nombroj> it's not an exact name, it's, uh
20:14:32 <hazardaj_nombroj> /STR_INDUSTRY_VIEW_ACCEPT_CARGO(_AMOUNT)?_(NOSUFFIX|SUFFIX)/
20:14:38 <hazardaj_nombroj> there are four strings
20:16:41 <hazardaj_nombroj> probably gonna have to guess how they work exactly
20:16:43 <locosage> _SUFFIX is newgrf-only thing afaict
20:17:11 <hazardaj_nombroj> i've also not seen any industries that only accept a specific amount of cargo
20:22:34 <peter1138> hazardaj_nombroj, yes they require NewGRFs, and then only very few NewGRFs implement this stuff.
20:23:47 <locosage> actually, I completely forgot my own grf does that xD
20:29:24 <locosage> it's for christmas event
20:30:00 <locosage> deliver cargo to unpack gifts and get stuff
20:30:01 <peter1138> Oh. downloading the NewGRF in rito12_51026's scenario crashed OpenTTD :o
20:31:47 <peter1138> -rw-r--r-- 1 petern petern 4354412 Jan 23 18:56 Master-Trees.scn
20:31:52 <peter1138> -rw-rw-r-- 1 petern petern 4016428 Jan 23 20:31 Tile.scn
20:32:01 <rito12_51026> It only uses CZIS from BaNaNaS, forgot to remove it
20:32:02 <peter1138> Smaller improvements.
20:32:11 *** keikoz has quit IRC (Server closed connection)
20:32:32 <xarick> requires toyland tileset it seems
20:32:41 <peter1138> I wonder if it 's something to do with downloading it while in the scenario editor. Probably doesn't get tested much.
20:33:59 <peter1138> I should make sure it works before comparing sizes though :)
20:34:56 <rito12_51026> peter1138: `4016428` That's the size affter saving horizontally?
20:36:05 <peter1138> Got to redo it, silly bug :)
20:38:29 <xarick> why dont you save using all cpu cores
20:39:32 *** belajalilija has joined #openttd
20:39:32 <belajalilija> locosage: what stuff can be given?
20:41:58 <mmtunligit> what is `pip`? I'm trying to install nml but its saying there is no module called pip in my python3
20:42:57 <hazardaj_nombroj> kinda annoyed that the bananas website doesn't offer download links to files
20:44:07 <locosage> xarick: nope, it's for arctic actually :p
20:44:18 <andythenorth> pip is the default python package manager
20:44:48 <locosage> belajalilija: it says on the gift industry, different tiers require different cargoes
20:45:05 <locosage> ah, you mean what you get for unpacking
20:45:17 <locosage> industries, new engines, victory points
20:45:42 <belajalilija> i like this idea
20:46:32 <locosage> It's a cool game mode
20:47:02 <belajalilija> in the game transport giant there was scenarios where you had to deliver things to construction sites to make/open olympic stadiums and whatnot
20:47:29 <belajalilija> i do have a theoretical industry set and that mechanic would work great in it
20:47:45 <belajalilija> so im glad that it's actually possible in openttd
20:49:39 <locosage> it's not exactly possible 😅
20:50:04 <locosage> it requires a patched server for the complete experience
20:50:12 <locosage> newgrf just provides industries that do nothing
20:51:59 <belajalilija> would it still be possible to have industries that require stuff delivered to them before they can function without patches?
20:52:44 <locosage> iirc industry can only pick required cargos on construction so no
20:53:46 <locosage> well, I guess they can still check for certain cargos before producing but ui won't be the best
20:53:51 <hazardaj_nombroj> okay, i did all the missing/outdated esperanto strings, now gotta fix some of the things that annoyed me or are not consistent enough
20:54:07 <locosage> as all accepted cargoes will have to be visible even if you delivered enough
20:55:02 <belajalilija> janky but possible ig
20:55:18 *** Borg has quit IRC (Quit: leaving)
20:55:31 <locosage> maybe there are some ways around that need to check spec
20:55:48 <belajalilija> the idea was for an industry to spawn and then the player needs to deliver say, 1000 steel, 1000 wood and 1000 cement to it (one time) before it can start accepting and producing other things
20:56:15 <belajalilija> i wouldnt even want to actually approach this idea for years, i just wanna know what's possible really
20:57:41 <hazardaj_nombroj> question for y'all: the regular esperanto word for "road", "vojo", applies equally to roads as well as railways and waterways, so i'm instead using "strato" (street) to refer to specifically roads -- would that be okay or should i instead use something like "aŭtovojo" (which would be more like "car road"), in case there _may_ be an actual distinction between roads and streets?
20:59:15 <belajalilija> my theoretical industry set already has things like schools, hospitals, shopping mall and nuclear power plants as sort of end game industries, it would be nice to create more involvement in their construction
21:00:43 *** urdh_ has quit IRC (Server closed connection)
21:00:46 <locosage> I don't recall there being any difference between "car" roads in openttd. they may look different in towns but they're still called roads
21:01:29 <locosage> but you can check some road type newgrfs that add highway road types to see if that still looks adequate
21:01:47 *** SigHunter has quit IRC (Server closed connection)
21:01:47 *** SigHunter has joined #openttd
21:01:52 <locosage> "highway" streets may not be the best pairing
21:07:03 <hazardaj_nombroj> huh, "aŭtovojo" seems to instead be reserved for highways specifically
21:07:54 <locosage> though if road type situation is anything like rail type one any hopes of sanity is already lost there 😅
21:08:11 *** skee has quit IRC (Server closed connection)
21:12:14 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
21:12:44 *** MinchinWeb[m] has joined #openttd
21:14:54 <hazardaj_nombroj> gonna go back to fixing the annoyances
21:15:09 <Rubidium_> andythenorth: what does openttd -ddriver=9 show with #15145?
21:15:30 <ahyangyi> One way to say a road in Chinese is a "horse road".
21:17:29 <andythenorth> Rubidium_: ./openttd -ddriver=9
21:17:29 <andythenorth> [2026-01-23 21:17:06] dbg: [driver:1] Successfully loaded blitter '32bpp-anim'
21:17:59 <andythenorth> then the macOS 'no suitable video driver found' dialogue
21:18:43 <hazardaj_nombroj> another question: would "entrepreno" ("enterprise") be a good substitute for "industry" when it is used to describe a single place/building producing resources? because "industrio" specifically means "industry" as in "a branch of the economy"
21:18:59 <hazardaj_nombroj> the russian translation seems to use "предприятие", which is also more like "enterprise"
21:19:34 <locosage> hm, looks promising with no changes to map array
21:20:50 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
21:21:22 <locosage> it's ok to call a single factory "предприятие"
21:23:11 <hazardaj_nombroj> so far, i'm using "entrepreno" (enterprise) for single factories, mines, oil rigs, etc., and "industrio" for things like "industry chains" that describe how things work in the entire game
21:24:06 *** MinchinWeb[m] has joined #openttd
21:29:04 <andythenorth> hmm CMakeCache.txt though has `COCOA_LIBRARY:FILEPATH=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/System/Library/Frameworks/Cocoa.framework`
21:29:48 <housemd123123> alright. so ive tried everything that i could find online for help with the NewStations thing. again, Newstations wont show the textures on all of the stations.
21:33:50 <Rubidium_> andythenorth: that's weird. What difference in video drivers are there when you do ./openttd -h with the patched version and an (older) working version?
21:34:40 <Rubidium_> it almost looks like the driver isn't even being registered for some reason
21:35:38 <peter1138> When you say "NewStations" do you mean any custom NewGRF stations, or specifically MB's NewGRF called "NewStations"?
21:36:17 *** hazardaj_nombroj has left #openttd
21:36:54 *** DorpsGek has joined #openttd
21:36:54 *** ChanServ sets mode: +o DorpsGek
21:37:22 <peter1138> And also, you might fancy posting a screenshot of what you mean.
21:38:00 <_jgr_> housemd123123: If you're getting purple artefacts you'll likely need to use the toggle palette button in the NewGRF window
21:40:30 <housemd123123> peter1138: yes. i will.
21:41:03 <ahyangyi> Were you replying to JGR?
21:41:31 *** herms2 has quit IRC (Server closed connection)
21:47:55 *** DorpsGek_vi has quit IRC (Server closed connection)
21:47:57 *** DorpsGek_vi has joined #openttd
21:48:11 *** andythenorth has quit IRC (Server closed connection)
21:48:11 *** andythenorth[d] has joined #openttd
21:50:56 <andythenorth[d]> this build of 15145 has
21:50:56 <andythenorth[d]> List of video drivers:
21:50:56 <andythenorth[d]> cocoa-opengl: Cocoa OpenGL Video Driver
21:50:56 <andythenorth[d]> dedicated: Dedicated Video Driver
21:50:56 <andythenorth[d]> null: Null Video Driver```
21:51:06 <andythenorth[d]> I'm curious why that's not starting, it looks correct?
21:51:53 <andythenorth[d]> a working version has
21:51:53 <andythenorth[d]> List of video drivers:
21:51:53 <andythenorth[d]> cocoa-opengl: Cocoa OpenGL Video Driver
21:51:53 <andythenorth[d]> cocoa: Cocoa Video Driver
21:51:53 <andythenorth[d]> dedicated: Dedicated Video Driver
21:51:55 <andythenorth[d]> null: Null Video Driver```
21:53:19 * andythenorth[d] researching how deprecated Cocoa OpenGL is
21:55:39 *** Rubidium_ has quit IRC (Server closed connection)
21:55:44 *** Rubidium has joined #openttd
22:01:55 <_glx_> andythenorth[d]: according to a recent comment to a very old issue, it should run fine with `-v cocoa-opengl`
22:02:45 <_glx_> (both with master and the PR)
22:03:46 <_glx_> but that doesn't help to solve the `cocoa` issue
22:05:41 <andythenorth[d]> _glx_: confirmed
22:05:43 <Rubidium> it won't autoprobe the cocoa-opengl if you disabled hw accel, but it's weird that cocoa doesn't register at all
22:06:44 <_glx_> might require use of debugger and breakpoints
22:07:07 *** housemd123123[d] has joined #openttd
22:07:07 *** housemd123123 has quit IRC (Server closed connection)
22:07:47 <Rubidium> andythenorth[d]: you're sure 317dc71 is the commit that broke it?
22:08:01 <andythenorth[d]> I haven't properly bisected
22:08:48 <andythenorth[d]> I'll build 317dc71
22:12:11 <Rubidium> I think it's 666073fc7b that broke it, and I hope that update to #15145 fixes it
22:14:50 <_glx_> oh good candidate indeed
22:15:11 <michi_cc[d]> Yeah `- static FVideoDriver_CocoaQuartz iFVideoDriver_CocoaQuartz;` should do it. Oops 🙂
22:17:13 <Rubidium> 317dc71 will probably still show cocoa in openttd -h, but the debug log will probably say probing it failed
22:18:06 <andythenorth[d]> 317dc71 started for me, which was puzzling
22:18:10 <andythenorth[d]> 666073fc7b did not
22:18:22 <andythenorth[d]> I am testing 15145 again
22:18:58 <michi_cc[d]> The version check is probably totally fine, it's just that all these elaborates circles aren't needed if you are strictly 10.10+
22:19:33 <andythenorth[d]> ok for 15145 ```List of video drivers:
22:19:33 <andythenorth[d]> cocoa-opengl: Cocoa OpenGL Video Driver
22:19:33 <andythenorth[d]> cocoa: Cocoa Video Driver
22:19:33 <andythenorth[d]> dedicated: Dedicated Video Driver
22:19:33 <andythenorth[d]> null: Null Video Driver```
22:19:54 <_glx_> CI targets 10.13 anyway `MACOSX_DEPLOYMENT_TARGET: 10.13`
22:21:03 <_jgr_> There seems to be a lot of stuff changed in 666073fc7b for a supposed documentation improvement?
22:21:16 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:27:16 <_glx_> hmm I don't why it added a new define specific to doxygen when it could have added WITH_COCOA to the doxygen config
22:33:39 *** Tirili has quit IRC (Quit: Leaving)
22:33:55 <peter1138> -rw-rw-r-- 1 petern petern 5475336 Jan 23 22:27 'Wentbourne Transport, 3739-09-06c.sav'
22:33:58 <peter1138> -rw-rw-r-- 1 petern petern 5952748 Jan 23 22:33 'Wentbourne Transport, 3739-09-06d.sav'
22:34:04 <peter1138> Hmm, grouping by tile type not so good.
22:34:30 <peter1138> c is just saving all tiles in order, d is grouped by tile type.
22:34:44 <peter1138> Not using anyone elses code, I was already fiddling with it.
22:36:07 <locosage> hm, wonder what's the easiest way to confirm that with my code as it doesn't load vanilla saves 😅
22:38:58 <peter1138> locosage, set the chunk type for the unneeded chunks to CH_READONLY, then leave them in the list of handlers.
22:39:16 <peter1138> (And remove the Save() function)
22:40:00 <peter1138> Any views on #15150
22:48:17 <locosage> hm, I get smaller wentbourne
22:49:05 <peter1138> Well your approach is different...
22:49:07 <locosage> somehow it's even smaller after I noticed I wasn't saving high byte of m2 xD
22:49:38 <peter1138> I went for using the saveload system as designed, instead of shoving manually encoded byte streams in it :)
22:49:50 <peter1138> (Which is what the existing map chunk handlers do, to be fair)
22:50:44 <locosage> I suppose the main difference is that I save type and heightmap separately
22:51:51 <locosage> I get 5044892 on wentbourne vs 5750748 master
22:52:20 <locosage> did you re-save it in the latest master before comparing btw?
22:52:51 <locosage> it's larger than original wentbourne by quite a bit
22:53:26 <peter1138> master is, with my PR (and your version) it's not.
22:54:24 <locosage> yeah, by type it's smaller
22:57:19 <housemd123123[d]> peter1138: heres those images
22:57:50 <locosage> oh, wait, am I reading it right that you're just saving tile data as a struct, not grouping tiles by type like I do?
23:02:22 <peter1138> I tried grouping, and it's larger.
23:03:10 <locosage> well, not in my version...
23:03:33 <peter1138> 22:49 <@peter1138> I went for using the saveload system as designed, instead of shoving manually encoded byte streams in it :)
23:06:35 <locosage> may be a bit of re-design is needed then ;p
23:06:53 <locosage> what I'm doing is not that different from existing map chunk handlers anyway
23:07:52 <peter1138> Not really, we don't necessarily design the savegame format to be as small as possible.
23:08:49 <peter1138> e.g. out of all of our userbase, there's basically only one person complaining about file size.
23:09:39 <locosage> did you count players complaining they can't join servers? ;)
23:10:44 <locosage> 30% is not insignificant, can be the difference, especially on default settings
23:12:09 <locosage> also I bet everyone can appreciate having their collection of saves 30% smaller ;)
23:12:35 <squirejames> housemd123123[d]: Is it a palette issue? Windows vs DOS?
23:12:51 <housemd123123[d]> squirejames: its windows.
23:13:11 <squirejames> What I mean is, i've had similar errors myself from having the wrong palette selected
23:13:16 <peter1138> In the NewGRF settings window, there's a "palette" option, toggle that.
23:13:30 <peter1138> There are no "missing textures", which is why your searching didn't give you any help.
23:13:39 <housemd123123[d]> there is no "palette" setting
23:14:54 <mmtunligit> my entire saves folder isnt even whole gigabyte and its got like at least 30 4k^2 saves in it i really dont think savegame size is an issue
23:15:01 <squirejames> (i'm honestly slightly chuffed I got the answer right. All those years of messing around with TTDPatch/OpenTTD etc and I finally could go "ooh! I know how to fix it")
23:15:42 <belajalilija> squirejames: id thought it but i assumed they'd already checked
23:16:06 <peter1138> #15159 is smaller savegames and smaller code, so that's a win.
23:16:22 <squirejames> belajalilija: Fair point, taking the victory where I can though 🙂
23:16:45 <peter1138> (And it uses the standard system)
23:17:03 <peter1138> ((Although not going as far as dissecting the bits of each tile :)))
23:19:57 <locosage> it's larger saves than my branch by a lot
23:20:04 <locosage> actually on my tree test it's larger than master
23:20:25 *** Flygon_ has quit IRC (Read error: Connection reset by peer)
23:22:38 <locosage> which is basically just a `newgame 0`
23:36:59 <locosage> hm, keeping MAPT and MAPH chunks in 15159 makes it better on new games but worse on wentbourne
23:39:08 <xarick> oh, I'm already using lzo
23:39:30 <_glx_> don't, lzo is the worst
23:39:53 <_glx_> if you ignore uncompressed
23:40:05 <locosage> It's worst by size but fastest iirc
23:40:24 <_glx_> if you want fast don't compress 🙂
23:44:19 <locosage> lzo actually offers quite a good balance between speed and size
23:45:25 <locosage> though for whatever reason I'm getting 1mb/s download speed on servers lately so lzma is still the best
23:45:34 <locosage> not 10mb/s that I calculated for in that pr
23:47:37 <locosage> btw, I've started recording some connection timings on citymania servers so hopefully will have enough data to compare different compressions soon
23:47:45 <locosage> at least lzma vs zstd
23:48:18 <locosage> though it's already pretty clear that for me it makes almost no difference
23:53:05 <xarick> oh, and i should also test master, lol
23:54:51 <locosage> you can also test my branch it's likely the best compression
continue to next day ⏵
