IRC logs for #openttd on OFTC at 2023-01-06
⏴ go to previous day
00:03:24 <LordAro> huh, do we do a lot of declaring functions in blocks?
00:07:57 <glx[d]> we use some extern to not include whole headers
00:08:31 <LordAro> yeah, but for the most part not in the function scope, surely?
00:08:36 <LordAro> and those that do can be fixed?
00:11:15 <LordAro> well, just a duplicate block just outside the function ought to do :p
00:11:58 <LordAro> could even avoid the duplicate with something like #define FONT_LOAD_FUNC LoadFreeTypeFont
00:21:27 *** arikover has quit IRC (Quit: ERC 5.4 (IRC client for GNU Emacs 28.2))
00:23:24 <petern> Hmm, got bunged up again...
00:24:24 *** sla_ro|master has quit IRC ()
00:24:46 *** borishiro has quit IRC (Read error: Connection reset by peer)
00:36:56 *** _aD has quit IRC (Quit: leaving)
00:57:42 *** HerzogDeXtEr1 has quit IRC (Read error: Connection reset by peer)
01:06:39 *** Tirili has quit IRC (Quit: Leaving)
01:09:47 <petern> NoDL (pronounced noddle)
02:43:32 *** Wormnest has quit IRC (Quit: Leaving)
03:22:44 <reldred> I pronounce SQL as squeal
03:25:13 <supermop_Home> ugh trying to get midjourney to make some where's waldos
03:30:11 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:39:14 *** D-HUND is now known as debdog
03:40:17 <Pruple> is there a technical reason airports are not the related object of airport tiles?
03:44:40 <Pruple> airports have access to all the base station info about cargos, date built, etc, but you can't actually do anything with it because you can't get to it from the tiles.
04:18:09 *** TROILUS has quit IRC (Quit: Ping timeout (120 seconds))
04:18:21 *** TROILUS has joined #openttd
04:23:03 *** TROILUS has quit IRC (Read error: Connection reset by peer)
04:23:03 *** TROILUS1 has joined #openttd
04:23:08 *** TROILUS1 is now known as TROILUS
06:30:17 <andythenorth> reldred: it's skwul
06:31:11 <andythenorth> emphasis on the non-existent W
06:45:53 <reldred> Skwul does have a nice ring to it though
06:47:47 <dwfreed> I will die on this hill
06:48:18 <andythenorth> and PNG not ping, despite whatever the inventor says
06:48:26 <dwfreed> (S-Q-L is also acceptable)
06:48:42 <dwfreed> yeah, PNG is one I'll always spell out
06:48:54 <andythenorth> nobody says puduf
06:49:24 <andythenorth> I always say D-O-C-X
07:19:07 <Gwyd> I always say jaypeg not J P E G but I'll say J P G not jaypg
07:21:45 <reldred> We have some software called PRTG, I call it purtuhguh
07:22:49 <reldred> Or purrtidge, like porridge
07:22:57 <andythenorth> I always say L O L zee
07:23:15 <reldred> I also hate it. API sucks, uses proprietary database 😒
07:23:34 <reldred> I need something better for the NOC
07:24:44 <reldred> Something better that I can get live data into Grafana with
07:28:36 <dwfreed> reldred: opennms? cacti?
07:35:24 *** sla_ro|master has joined #openttd
07:53:48 <reldred> dwfreed: yeah not fussed what I use as long as it uses a DB I can poll straight from Grafana
07:54:17 <reldred> Played with NetXMS quite a bit, I did enjoy it quite a bit but it's a very very complex tool
07:54:35 <petern> Postgresql is not postgresequel
07:56:38 <reldred> Hmm, cacti uses mysql/mariadb by the looks of it
07:57:42 <petern> Cacti is a pain once you move away from plain network stats
07:57:49 <reldred> we're only really doing ping and snmp currently
07:58:10 <petern> OpenNMS is also a pain. It's all pain.
07:58:19 <reldred> and to be honest I don't want to actually use cacti, I just want something that populates a DB so I can unify stats and alerts etc. in Grafana
07:58:30 <petern> But I have OpenNMS set up.
07:58:44 <reldred> PRTG shits me, it's expensive, it's API sucks, and it uses a proprietary database
07:59:13 <petern> Is it related to MRTG ?
07:59:27 <reldred> I'll do all my graphing, wallboarding, etc. in Grafana once I have Cacti's DB setup as an SQL datasource.
08:02:13 <Rubidium> LordAro: it might also be because those function in block used to be much more comming in header files, if I remember correctly. But then what to filter out and keep in is still up for debate
08:16:41 <Rubidium> though a more interesting question would also be, should we check the Windows and Mac specific code? That would mean two more platforms to run it on, which might not be really wanted in the normal case, so then it would be relagated to someone basically doing it once in a blue moon with a custom branch
08:32:46 *** WormnestAndroid has quit IRC (Remote host closed the connection)
08:35:20 <pickpacket> I play with only the original trains and while I’d love to use some NewGRF with more trains I don’t really see what niche they would fill
08:37:56 <Wolf01> I usually buy the best loco for a generation and stick with that :P
08:45:00 <petern> > This extension is recommended because you have Kubernetes installed.
08:46:54 <petern> > PS C:\Users\peter> kubectl
08:46:54 <petern> > kubectl controls the Kubernetes cluster manager.
08:48:52 <petern> Ah, part of Docker Desktop for Windows. Hmm.
09:52:32 <reldred> is supermop on? Nice work on your sneaky train station addition if you are 🙂
09:53:02 <reldred> this is quite pleasing to look at 🙂
10:17:53 <andythenorth> right now I want to run FFWD a year to see if I have enough trucks / ships on my route
10:18:28 <andythenorth> but that moves the tech tree forward
10:18:54 <petern> That's an interesting idea.
10:18:57 <andythenorth> tech tree is supposed to be managed via GS though?
10:20:12 <petern> Freeze calendar time but let economy time continue.
10:20:58 <andythenorth> no progression, or slower progression, but gameplay continues
10:21:32 <petern> I guess that 100% value can be set to 0 🙂
10:23:01 <petern> Seems to be quite slow to change that value though, maybe it does some recalculation
10:23:36 <petern> "Ctrl-click pause to freeze calendar time"
10:24:36 <petern> So yeah, the graphs are in economy time, but show calendar time.
10:25:41 <andythenorth> perfection wasn't achieved yet? 😮
10:26:27 <petern> But I can see that some people will want calendar time graphs
10:26:48 <petern> As long as calendar speed isn't changed mid-game that can work.
10:41:22 <TallTyler> Recalculation should be instant, and check the to-do list in the PR description for unfinished things 😉
10:41:34 <TallTyler> Graphs will be economy time
10:45:52 <Gwyd> andythenorth: This is a use case I hadn't thought of but yeah it would be great for testing and tuning timetables/route capacity
10:46:36 <dP> andythenorth: save, ffwd, load ;)
11:03:28 <petern> TallTyler: I think it's "near instance" but when you spam-click the button to change instead of entering a value, it takes a while.
11:04:08 <petern> 7400% is basically 1 tick per day, nice 🙂
11:05:05 <petern> I assume there's a lot of values that actually end up being the same though.
11:14:30 <TallTyler> Hmm, there’s some onclick code to avoid spamming the button but it shouldn’t be noticeably long I think. How long is “a while?”
11:15:13 <TallTyler> Yes, town growth is economic and already done — I’ll add it to the list. 🙂 Same goes for industry production changes, closure, and spawning
11:15:37 <TallTyler> 7400% is exactly one day per tick 🙂
11:15:55 <TallTyler> If your computer can go that fast, I guess
11:16:06 <TallTyler> Had to choose some maximum value
11:16:39 <petern> Might just be that, but it felt like the palette animation froze for a tiny bit.
11:18:30 <TallTyler> Palette animation definitely isn’t controlled by day-anything
11:18:38 <TallTyler> That sounds like lag
11:18:39 <petern> As 3700% is one day per 2 ticks, there's a lot of values in between that are duplications.
11:19:14 <petern> Of course not, it's more that the game "lags" when changing the setting.
11:19:25 <petern> Hmm, maybe it's just marking the full screen dirty or something 🙂
11:19:26 <LordAro> possible bikeshedding - should display it as "one day per N ticks", rather than XX00%
11:20:02 <LordAro> (with a warning for 200%+ that it might not actually achieve such speeds)
11:20:09 <petern> %age makes sense when you don't know what a tick is.
11:20:29 <petern> With Wentbourne I can't manage 100% 😄
11:20:35 <LordAro> but equally, 7400% is meaningless to everyone
11:20:42 <LordAro> that doesn't know how many ticks in a day
11:20:48 <petern> It means 74 times faster than normal.
11:21:11 <TallTyler> Yeah, 100% is default speed (actually default TTD speed, OpenTTD is slightly slower)
11:21:29 <TallTyler> I don't want to expose ticks to players
11:22:05 <petern> 7400% isn't particularly important, it's just the upper limit, and nobody will be running at the speed.
11:22:27 <LordAro> how about "one day per N real-life seconds" ?
11:22:32 <petern> % makes sense, ticks per day does not really.
11:22:59 <LordAro> "one in-game day per N seconds" (the bikeshed should be red)
11:23:10 <TallTyler> My previous implementation didn’t go any faster than 100% but dP wanted faster speeds for goal games so I figured 🤷
11:23:13 <petern> Is the bikeshed long enough for my XL bikes?
11:23:35 <petern> (I can't use my Thule roof bars because the they're not long enough...)
11:24:36 <LordAro> have you considered being shorter?
11:25:21 <TallTyler> One day per N seconds requires players to know how long default time is, or they’ll be confused when opening the Generate World gui and seeing that setting right there. Way too technical - 100% is less scary 🙂
11:25:50 <TallTyler> (Genworld doesn’t have helptext or tooltips to explain settings, unlike Settings GUI)
11:34:00 <TallTyler> I will await it eagerly 😉
11:37:44 <TallTyler> Even with tooltips, I think a percentage is easier to understand
11:39:01 <andythenorth> petern: challenge accepted? 😛
11:51:11 *** murr4y has quit IRC (Ping timeout: 480 seconds)
12:00:17 <JGR> 7400% is likely to be unplayable, even if the code technically works at that value it may be worth limiting the setting range to less than that
12:01:08 <TallTyler> Any thoughts on a reasonable limit? I can't see anyone needing higher than 100% so my opinion isn't that helpful 🙂
12:01:46 <petern> Faster is useful for testing introduction dates and the like.
12:02:05 <TallTyler> Is it that different from fast-forward though?
12:02:15 <TallTyler> Well, I guess it is
12:02:22 <TallTyler> Because the game itself doesn't run faster
12:04:01 <petern> Oh yes, it is, because it doesn't require so many ticks to get there.
12:05:56 <petern> Hmm, Even limiting to 1000% means lots of values are actually the same.
12:06:20 <TallTyler> I did fractions last time and people didn't like that either 🙂
12:06:47 <TallTyler> I'm not worried about "percentage is too precise to make a real difference"
12:07:19 <TallTyler> Zorg doesn't have merge rights
12:07:25 <dP> TallTyler: It's not that I particularly wanted that but vanilla tech progression is relatively slow for short games so I can see some use speeding it up. To something like one engine every year being the reasonable limit.
12:07:43 <TallTyler> His suggestions are non-binding
12:08:00 <dP> But my main request with all these time changes is to keep the ability to run game at exact same speeds as it is now
12:08:08 <TallTyler> dP: Suggestions on a max speed that would work for your needs?
12:08:39 <TallTyler> Would it be 100% or 90%, since moving tick speed from 30ms to 27ms is 10% faster?
12:09:09 <dP> well, if you change tick speed it would be neither :p
12:09:54 <petern> TallTyler: Also here 😉
12:10:02 <dP> I guess it would work if I change tick speed on server but it would be nice to have that option offitially
12:11:43 <TallTyler> I really don't want to add a "legacy tick speed" setting, because that would mean all the real-time strings are wrong
12:12:14 <TallTyler> Having duplicate strings, graph code, etc., is not really a desirable option either
12:12:16 <dP> they'll be wrong either way :P
12:12:23 <dP> imaging someone pausing the game :p
12:12:37 <TallTyler> Well, real-time if you don't pause or fast-forward 😛
12:12:47 <TallTyler> But I think that's understandable to most people
12:12:53 <JGR> Duplicate strings is not that big a problem, just make the unit name a parameter
12:13:34 <TallTyler> True, but the units are only identical for months - days are two seconds and years are twelve minutes, so the numbers need converting as well
12:13:43 <TallTyler> Guess I could do a function for that which checks the setting
12:14:36 <dP> JGR: just call it "period", would also look way better than real-time values
12:14:45 <dP> who the heck measures industry production in minutes :p
12:15:31 <TallTyler> You can answer the endless "how long is a period" comments in Discord channel #openttd-help, Reddit, and everywhere else 😛
12:15:46 <dP> better than answering how long is minute :p
12:15:56 <TallTyler> Satisfactory measures production in real-time
12:16:06 <TallTyler> Not sure about Factorio but it wouldn't surprise me
12:16:52 <petern> How long is a piece of string?
12:17:07 <petern> Can I have a long weight please?
12:17:18 <TallTyler> Everyone knows how long a minute is, it's the only way to avoid having to deal with two player-facing calendar systems
12:17:54 <dP> why not just count production in months?
12:18:53 <TallTyler> Because that makes no sense if calendar progression is frozen entirely
12:21:17 <dP> I suspect many "realism" players would be unhappy to find their production in minutes just because someone wanted to freeze time :P
12:21:43 <JGR> It's more that a lot of players can't cope if things don't match what they see in the tutorial videos
12:22:02 <TallTyler> `spacebar_heating.png` ?
12:22:39 <JGR> I'd be inclined to have a time setting with something like "traditional" and "modern" options
12:23:09 <JGR> In traditional mode, time scaling isn't supported at all, to avoid all the problems with UI strings and such
12:23:38 <TallTyler> I'll see if I can work out a clean, maintainable way to support both
12:23:50 <JGR> When users complain, you can just say "use traditional mode" which resolves the problem quickly
12:25:04 <dP> TallTyler: ```The main goals of the official branch are:
12:25:04 <dP> Stay faithful to the original gameplay from Transport Tycoon Deluxe``` 😜
12:25:22 <TallTyler> 27ms per tick is truly faithful to TTD 😛
12:25:45 <dP> yeah, in that part I disagree with official goals xD
12:26:10 <dP> mainly because I have like a 10 years of scores to maintain
12:26:34 <TallTyler> I don't know if you can pick and choose when to hold others to the official goals
12:27:08 <TallTyler> Traditional mode might have to be 27ms, because I'm not sure if it's smart to have something so low-level be a setting
12:28:08 <JGR> I think the UI changes are likely to be much more important than 27ms vs 30ms, which users likely won't notice
12:32:11 *** WormnestAndroid has joined #openttd
12:33:58 <andythenorth> Pruple: “please cut off a nanosecond and send it over to me”
12:34:21 <andythenorth> the goals are adjustable 😛
12:34:25 <andythenorth> but not trivially
12:36:26 <TallTyler> If I can support both systems, I'll probably have it automatic where 100% uses months and anything else uses minutes
12:36:38 <TallTyler> Saves adding another setting for people to find
12:39:36 *** gnu_jj has quit IRC (Remote host closed the connection)
12:40:12 <dP> JGR: yeah, after all I can change tick speed on server and I guess game will have to support that either way. but at least would be nice to have sane ui if I do that
12:41:29 <TallTyler> The only problem with compatibility mode is that months don't line up - economy months are always 30 days, so there are only 360 days in a year
12:42:27 <TallTyler> So "production last month" updates at a different time than when the actual calendar rolls over to a new month
12:42:33 <TallTyler> I don't see an easy way to fix this
12:43:20 <TallTyler> One could argue that UI is not gameplay, so this still meets the goal 😛
12:43:32 <dP> TallTyler: Imo it's more reasonable to keep months for everything that is not that close to frozen
12:44:06 <dP> but as long as game keeps enough values it can chose either way for visual representation
12:44:36 <dP> well, for frozen "enough" is infinity, but that can be ignored)
12:46:00 <Eddi|zuHause> TallTyler: might be confusing that months go in and out of phase
12:46:43 <dP> why does economy time even have months?xD
12:46:51 <dP> if it's in ticks and minutes
12:48:00 <TallTyler> Internally, you mean?
12:48:37 <Eddi|zuHause> the current "production last month" has the weird quirk that it doesn't line up with the 256-tick production periods, so you randomly have 8 or 9 production cycles in a month, showing fluctuating production that doesn't actually fluctuate
12:49:44 <JGR> Months are not all the same length in real life either, that is something that most players can be expected to unconsciously understand
12:50:50 <TallTyler> Yeah I've never been bothered by monthly production fluctuating somewhat. But if the "last month" UI doesn't update when the calendar does, I expect all sorts of bug reports
12:52:30 <dP> I already talked about it in #7575, game can keep production history in some tick-besed period and recalculate to months whenever it needs that for gui
12:53:18 <petern> "economy month" is measured in ticks
12:53:18 <Eddi|zuHause> there isn't really a reason that it has to be "month" in the gui
12:53:18 <dP> it won't be 100% accurate but it's better than production per minute imo ;p
12:53:52 <petern> (or probably days per tick, which is still ticks)
12:55:01 <petern> If you made 1 economy month be ~2252 ticks it would roughly line up with calendar months, when on 100%.
12:55:54 <TallTyler> Or we could change industry production speed to match months, to really spice things up 😛
12:57:05 <TallTyler> In any case, I think changing the accuracy of how production is tracked is a separate PR
13:00:58 <dP> it related to how you show it in the ui though
13:01:24 <dP> merging #10322 with minutes and switching back to months in separate pr is doable but awkward
13:02:18 <TallTyler> It won't be ready to merge for quite some time
13:03:48 <Eddi|zuHause> TallTyler: if you change industry production cycles you'll get a grf author revolt
13:04:06 <Eddi|zuHause> (meaning around 2 people)
13:04:19 <TallTyler> It wasn't a serious suggestion 🙂
13:08:37 <TallTyler> Okay, can someone remind me how to rebuild regression results? I'm running the `regression (utility target)` and getting
13:08:37 <TallTyler> CMake Error at C:/Users/tyler/Documents/GitHub/OpenTTD/cmake/scripts/Regression.cmake:49 (message):
13:08:37 <TallTyler> Regression did not output anything; did the compilation fail?
13:40:29 <Eddi|zuHause> interesting solution to the ufo problem
13:54:09 <glx[d]> TallTyler: I just run ctest from visual studio
13:54:55 <glx[d]> If regression fails, the output is stored in build dir
13:57:12 <TallTyler> `Errors while running CTest` is different than regression failure, right?
13:57:26 <TallTyler> Was I supposed to delete regression test results before running the test?
13:57:52 <TallTyler> The output is only stored as .log files, not as the usual regression output
14:04:49 <glx[d]> Regression result is written to regression_xxx_output.txt, but in your case it errors before this step because output is empty
14:05:12 <glx[d]> Possible cause is edit in step
14:06:07 <glx[d]> As openttd.exe if not modified by editbin won't output anything
14:06:57 <TallTyler> How would I fix it?
14:07:44 <TallTyler> I haven't run regression since I upgraded from VS2019 to VS2022
14:08:00 <TallTyler> And a few other things broke when I did that
14:08:13 <glx[d]> Cmake should tell if it found editbin
14:09:12 <glx[d]> Could also be a crash during regression before the AI start
14:09:33 <glx[d]> Though assert should output
14:10:24 <TallTyler> CMake Settings shows a path for `EDITBIN_EXECUTABLE`, if that's what you mean
14:11:32 <glx[d]> If there's a path it should be fine then
14:14:47 <glx[d]> Both test failed the same way, maybe openttd crashed
14:16:27 <glx[d]> But as they run in CI it's improbable
14:18:05 <glx[d]> First step would be to try with master
14:20:54 <andythenorth> FML left my game running while I did work 😦
14:21:02 <andythenorth> lost 10 years of roleplay
14:21:21 <TallTyler> Hmm, I could fix that 🤔
14:21:22 <andythenorth> I guess I reload last save
14:21:28 <andythenorth> you could fix it? 😮
14:21:38 <TallTyler> Not retroactively 😛
14:21:56 <andythenorth> I can retroactively fix it by reloading most recent save, which I fortunately have 🙂
14:31:29 <petern> `Test not available without configuration. (Missing "-C <config>"?)`
14:33:49 <petern> ```CMake Error at I:/src/OpenTTD/cmake/scripts/Regression.cmake:15 (message):
14:33:49 <petern> Regression test stationlist does not exist (tip: check regression folder
14:33:49 <petern> for the correct spelling)```
14:34:13 *** nielsm has quit IRC (Ping timeout: 480 seconds)
14:34:49 <TallTyler> Are you having the same problem on your machine?
14:35:05 <petern> I don't know if it's the same problem. I never bother with tests :p
14:35:23 <petern> Ok, Regression.cmake is looking for ai/{test}/test.sav
14:35:26 <TallTyler> I only found it because my PR is failing CI for regression
14:36:21 <glx[d]> Cmake copies regression files into build
14:36:33 <petern> My cmake begs to differ.
14:39:29 <petern> Okay, in my normal cmake building it just builds the openttd part.
14:40:48 <glx[d]> It's done at the end of root CMakeLists.txt
14:41:34 <petern> No idea, I just have a button in VS Code that starts debugging OpenTTD, and it doesn't do anything with regression.
14:42:19 <petern> Found this, it works from here.
14:42:26 <petern> Although regression fails obviously.
14:43:11 <glx[d]> Looks like .sln version
14:43:13 <petern> Those regression tests are not very useful tbh.
14:43:21 <petern> VS Code doesn't do .sln
14:44:18 <petern> Built in unit-test stuff works too, lol
14:44:40 <petern> 3: 1: + ERROR: Registering two scripts with the same name and version'
14:46:17 <glx[d]> Yeah local files may be an issue
14:46:49 <petern> Apparently I had unpacked renewed village growth at some point.
14:46:55 <TallTyler> So if I have two gamescripts with the same name and version, it'll break it?
14:47:05 <petern> 1/1 Test #3: regression_stationlist ........... Passed 2.09 sec
14:47:20 <petern> The other one fails though.
14:47:36 <glx[d]> But it will output in this cas TallTyler
14:47:48 <petern> Yes, it will output if that's the cause.
14:47:55 <glx[d]> Your fail was no output at all
14:53:21 <glx[d]> Maybe we need an option to only scan scripts next to cfg
14:54:31 <glx[d]> Doesn't matter for CI, but for local testing it's different, like the newgrf scan skip
14:55:29 <petern> Anyway, in reality the regression failed only because the game is now different.
14:55:42 <TallTyler> Right, but I do want to fix it 😉
16:02:03 <Samu> version 13 is a disgrace :(
16:03:05 <Samu> AI settings don't get fixed and they're yet pushed as release candidate
16:06:33 <LordAro> sorry Samu, everyone except you doesn't view it as a priority
16:06:48 <LordAro> also, "release candidate" just means "feature freeze"
16:06:55 <LordAro> it doesn't mean that there will be no further changes
16:08:29 <michi_cc[d]> TallTyler: If you are still stuck on the regression, an VS update sometimes has editbin in a different path and then at least I always have to delete the cached path in cmake and re-run cmake config.
16:14:13 *** TROILUS3 has joined #openttd
16:14:14 *** TROILUS has quit IRC (Remote host closed the connection)
16:14:20 *** TROILUS3 is now known as TROILUS
16:18:22 <supermop_toil> andythenorth: what do you think the elevation angle of the sun is?
16:26:20 <Samu> wow someone posted a crash report for my AI
16:26:52 <Samu> I can replicate the crash, but I wonder why it happens
16:27:51 *** TROILUS has quit IRC (Read error: Connection reset by peer)
16:28:03 *** TROILUS has joined #openttd
16:28:12 <Samu> some vehicle capacity from eGRVTS v2.0 NRT - r207 returns 0 when it wasn't expected to be 0
16:29:58 <andythenorth> supermop, varies according to the mood of SF
16:31:24 <supermop_toil> SF is all over the place
16:31:34 <supermop_toil> whats the mood of andy?
16:53:26 *** Wormnest has joined #openttd
17:18:43 <TallTyler> michi_cc[d]: That fixed it, thanks
17:32:59 *** HerzogDeXtEr has joined #openttd
17:58:39 <Samu> holy crap, the issue is bigger than i expected
17:59:04 <Samu> GetBuildWithRefitCapacity again giving wrong capacity values
18:07:32 <petern> Even the normal player build window does that. I blame callbacks.
18:10:34 <andythenorth> supermop angle is quite high
18:11:09 <andythenorth> there are no shadows cast on the roof by the low front wall on that building
18:11:10 <JGR> I could upstream my fix for that issue when I get some time
18:11:20 <Samu> well, seems to be a bug in openttd :(
18:11:44 <andythenorth> I don't know about sun angle measurements, but I'd guess about 70 degrees or something
18:12:00 <andythenorth> I'd say it's between 11am and 2pm in June in the UK
18:12:36 <Samu> the command detects that no refitting is needed, so refitting_capacity is filed with the wrong cargo type
18:13:09 <Samu> it's a road vehicle that I am trying to get the capacity, refitted to passengers
18:13:45 <Samu> no refit is required as it's the default cargo, but refitting_capacity gets the capacity of mail which is 0 :(
18:16:02 <Samu> this is so confusing to follow
18:17:35 <Samu> equals 0, meaning it doesn't have capacity of anything?
18:20:27 <petern> Removing that function makes sense, because the function itself doesn't....
18:20:52 <TallTyler> An upstream would be nice 🙂
18:21:13 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
18:21:33 <Samu> it's a horse and a carriage
18:21:47 <Samu> it's articulated vehicle, the horse carries 0 passengers
18:23:05 <petern> The write-only medium of IRC.
18:23:16 <TallTyler> Instead of getting all worked up, you could just cherry-pick those commits to fix the problem 😛
18:24:06 <JGR> They will conflict, some more work is needed for that
18:24:19 <TallTyler> Is it time to start a NoDL game to find all the things that need to be fixed? I think maybe yes.
18:30:16 <TrueBrain> Today in "hunt all the bugs", Tyler is looking for problems with code he wrote .. excited .. what will he notice today!
18:30:24 <TrueBrain> (read this in a creepy voice please)
18:31:50 <andythenorth> can you do one for all of us?
18:32:47 <TrueBrain> ran out of funds for that
18:34:30 <Eddi|zuHause> TallTyler: you think that is ever possible to "fix all things" with daylength?
18:35:17 <JGR> Eddi|zuHause: You can asymptotically approach the absence of bugs with time
18:35:50 <Eddi|zuHause> JGR: yes, but the track record with that concerning daylength is not very good
18:36:50 <Rubidium> Eddi|zuHause: sure, document it in the NewGRF specs and it's essentially fixed :D
18:36:53 <JGR> Policy disagreements are but the sane as bugs
18:46:26 <DorpsGek> - Update: Translations from eints (by translators)
18:47:49 <TallTyler> andythenorth: petern and anyone else who wants to join, there's a NoDL server online if you want to join in
18:47:59 <andythenorth> do I need the client?
18:48:05 <petern> What branch name do I need? 😄
18:48:38 <TallTyler> Time is frozen to see what breaks
18:50:19 <petern> Also it's not the same revision as the PR.
18:50:40 <TallTyler> Hm, not sure how to sync those
18:51:07 <andythenorth> thanks Apple, you updated Xcode
18:51:18 <petern> If you've made changes, push them, if not then, er, it should be same revision.
18:53:46 <andythenorth> hmm I'm still dragging around a local patch to format.h
18:53:56 <andythenorth> should I be filing a bug or something? 😛
18:54:47 <Samu> i don't think i need to fix anything in the AI
18:55:03 <Samu> once the bug is fixed in openttd, it should go away
18:57:32 <andythenorth> what am I fetching? 10322?
18:57:41 <TallTyler> Not yet, let me fix it
18:57:48 <TallTyler> Rebase it, specifically
19:02:17 <TallTyler> detached my head, stand by
19:03:01 <glx[d]> detached usually during rebase
19:03:52 <TallTyler> Yes, I made the fatal mistake of accepting GitHub suggestions in the browser and then trying to pull them to my local copy
19:04:29 <TallTyler> Trying to rebase them into the right commits triggered merge conflicts on everything so I tried to pull a previous commit to abort, and now I don't know where I am 🙃
19:04:50 <glx[d]> pulling should be fine, unless local and pushed were different
19:05:44 <glx[d]> first thing to do is `git rebase --abort`
19:06:06 <TallTyler> What I already accepted merges?
19:06:34 <TallTyler> There was a whole string of merge conflicts and I did `git rebase --continue` on several of them
19:07:03 <TallTyler> I barely understand gitk but it looks wrong
19:16:00 <TallTyler> Okay, I did a hard reset to the PR, now it just has to build all over again
19:20:14 <TallTyler> Okay, server round 2 is up, andythenorth petern
19:28:46 <petern> ITL Houses 2.2 is not on bananas
19:29:03 <TallTyler> Typical lazy NewGRF devs
19:31:44 <petern> And finally... server password
19:32:20 <TallTyler> Although I thought I didn't set one...user errorr
19:32:35 <andythenorth> hmm version mismatch
19:32:53 <andythenorth> I build a branch instead?
19:32:56 <petern> Needs to be clean build
19:33:33 <andythenorth> ok let's try again 😛
19:34:13 <andythenorth> do you have any brie?
19:34:27 <TallTyler> If I leave the passworld field blank maybe there's no password?
19:34:29 <andythenorth> what about stilton?
19:34:45 <petern> There is no password now 🙂
19:35:43 <andythenorth> 1e44c71241b91152c6c6061ffafb2def2e8d2975 ?
19:35:48 <andythenorth> version mismatch
19:36:02 <TallTyler> Do you have your own fixes you mentioned above?
19:36:02 <petern> If it's m1e44 then it's wrong
19:36:25 <andythenorth> is this because I have local patches?
19:37:36 <petern> Yes, if you have changes it's not the same version.
19:38:32 <andythenorth> ok I have some admin to do
19:41:32 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
19:43:38 <andythenorth> meh github diff formatting is crap, I'll attach it as a file
19:44:30 <andythenorth> no diffs allowed
19:44:51 <andythenorth> github is amazing, until it's lol
19:45:52 <andythenorth> ok attached a diff
19:46:45 <glx[d]> It's something on your system, you're the only one affected it seems
19:50:01 <andythenorth> hard to know where to start
19:50:08 <andythenorth> the only SDK is the Apple provided one afaict
19:50:19 <andythenorth> there's no weird path stuff to libs or headers or anything
19:51:35 <andythenorth> reldred: or EmperorJake or orudge can any of you build on macOS 12.6.2 with XCode 14.2?
19:56:57 <TrueBrain> there is one constant in OpenTTD, that andythenorth fails to build it on his machine 😄 (sorry :P)
19:57:23 <andythenorth> this is a constant in all software I build
19:57:36 <TrueBrain> at what point do we conclude it is you, not me? 😄
19:57:43 <andythenorth> _sometimes_ it's because somebody put weird fixes in for one project
19:57:55 <andythenorth> but I no longer allow people to type random shit in my conf
19:59:42 <andythenorth> dunno I could delete my drive and reinstall windows?
19:59:50 <andythenorth> isn't that how most problems get fixed?
20:03:44 <TrueBrain> it is the reinstalling part that worries me 😛
20:04:36 <andythenorth> ok what is format.h?
20:05:02 <andythenorth> and why does it appear to be validating Apple's SDK?
20:05:38 <andythenorth> maybe that's just coincidental
20:06:26 <andythenorth> this is irrelevant to the error, yes?
20:06:26 <andythenorth> `/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/math.h:155:5: note: expanded from macro 'isfinite'
20:06:26 <andythenorth> ( sizeof(x) == sizeof(float) ? __inline_isfinitef((float)(x)) \
20:06:40 <andythenorth> it's just a coincidence that similar lines follow all the errors?
20:13:56 <andythenorth> maybe I fixed it? 🧐
20:46:50 <petern> Does anyone use the NewGRF rail station filter?
20:47:21 <petern> I think it's useless as it only searches the station classes, not the stations.
20:48:06 <TallTyler> It would be nice if it searched the station names though
20:56:47 <TallTyler> Interesting the graphs work without me fixing them
20:57:16 <TallTyler> Must be based on ticks or they already got moved to `OnNewEconomyMonth()` or the like
20:57:24 <TallTyler> Obviously labels are totally wrong
20:58:09 <TallTyler> Dates especially since economy years have 5 fewer days...that adds up pretty quick from year 0
20:58:20 <TallTyler> That's why the graph is several years in the past
20:58:27 <petern> If I was to change the station UI, would it fall under the "you mustn't ever change the ui" rule?
20:58:42 <TallTyler> I missed that rule somewhere 😛
20:59:00 *** TROILUS has quit IRC (Read error: Connection reset by peer)
20:59:26 <petern> It's several "years" in the past because it's not calendar time 😉
21:00:21 *** TROILUS has joined #openttd
21:01:25 <Pruple> petern: while you're there, want to get rid of the airport categories? 😛
21:02:24 <Pruple> it's so tedious having to select small, large, etc from the dropdown when there's only a couple of airports in each category anyway, why not just list them all to begin with?
21:03:47 <andythenorth> BECAUSE CATEGORIES
21:03:49 <TallTyler> NewGRFs can add them though, as long as they use one of the predefined state machines. OpenGRF+ Airports has five small airports, and five helicopter airports
21:03:53 <andythenorth> WE NEED CATEGORY VARIANTS
21:04:50 <petern> Airports are a bit big to show a small preview of each :/
21:05:09 <petern> I wonder if it's "allowed" to draw a zoomed out version.
21:05:40 <Pruple> petern: or even cropped / representative 😛
21:05:57 <TallTyler> I don't quite understand how that's related
21:06:06 <TallTyler> It just has dirt varients, basically
21:06:25 <TallTyler> And a seaplane port
21:06:48 <Pruple> it has "old" and "modern" variants, wouldn't need to do that if you could change the appearance based on year built 🤷
21:07:03 <Pruple> but even the way it is, I think it's still a small enough number that people could handle them all being in the same list
21:07:20 <petern> I can fiddle with 10324
21:08:20 <JGR> The template error messages that the new command system generates if things aren't just so are almost comically unhelpful
21:08:55 <JGR> It should all be working though
21:09:15 <TallTyler> Does it also fix the script API for #9954
21:09:30 <JGR> I presume so, I'm not going to write an AI to check it
21:09:31 <petern> Okay, so station tiles are directly the station
21:09:42 <petern> But airport tiles are not?
21:09:59 <TallTyler> JGR: I imagine we can ask Samu to check, since they seem particularly concerned about it
21:10:11 <Pruple> because stations are old and weird
21:10:38 <petern> With stations the parent is the nearby town.
21:10:56 <Pruple> and airports have a fixed layout, like industries
21:11:06 <TallTyler> You could do something different with the station based on if the town is a city, I guess
21:11:37 <Pruple> with industries, the town is the parent of the industry, and the industry is the parent of the tiles
21:12:30 <petern> Yes, you can't access the parent of the parent can you?
21:12:51 <Pruple> not directly, although industries have permanent storage that could be used to pass info to the tiles
21:12:53 <TallTyler> I think you can indirectly via procedures
21:13:35 <TallTyler> I haven't done NewGRF stuff in a while but I think may have done that with industries when doing...something
21:14:21 <Pruple> the nml spec suggests airports also have permanent storage, but I don't know if that's true 🙂
21:14:27 <petern> Just wonder if accessing the town of a station of an airport tile is useful
21:14:39 <JGR> I doubt that that would work, the self/parent type is relative to the resolver executing the callback
21:15:06 <JGR> Which is tied to the feature type
21:15:13 <petern> Pretty sure it's possible to added a grandparent type
21:15:32 <petern> But no point if it's not useful
21:15:35 <TallTyler> Oh, I think I had to mix SELF and PARENT and used procedures for that -- not grandparent
21:15:51 <TallTyler> That does work though 🙂
21:16:01 <petern> "possible to add", as in change openttd, not in newgrf
21:17:16 <andythenorth> hmm shall I rebuild my OS then?
21:17:20 <glx[d]> andythenorth: maybe attach the full message in the report
21:17:23 <JGR> Adding grandparents might be a GRFv9 thing as the action2 bit format would need to be changed/extended
21:18:45 <petern> There's no real defined format, it's just if it's those special values then it uses the type instead of the standard "real" group
21:19:59 <petern> I wonder if any newgrf uses anywhere near that many numloaded/numloading slots
21:20:22 <petern> (Don't make one just to make sure there is a conflict :D)
21:21:03 <petern> Hmm, "self scope, parent scope, relative scope"
21:21:15 <petern> Only for randomization
21:26:05 <andythenorth> ok if I used make with -j flag some of the errors are mangled due to processes interleaving their output to stdout
21:26:20 <andythenorth> and if I don't use -j, the build stops after the first set of errors
21:26:59 <andythenorth> I should have googled 🙂
21:27:13 <andythenorth> but doing the right thing has been the meta all week, and I've run out
21:29:59 <petern> andythenorth: why would you want it to continue?
21:30:12 <andythenorth> so I can paste the full output
21:30:43 <petern> It's better to only paste the full output of one error.
21:31:53 <glx[d]> yeah it's in the header, next errors will be the same
21:32:12 <andythenorth> but the same pattern repeats
21:32:32 <glx[d]> could be related to include paths (not the first time with macos
21:33:52 <andythenorth> it's quite plausibly my OS is fucked
21:34:15 <glx[d]> I guess it's not including the right math.h
21:34:45 <petern> `xcrun --show-sdk-path`
21:35:17 <glx[d]> and try verbose build, so the full command is printed
21:38:01 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:38:20 <TallTyler> FYI - NoDL server is off because dinnertime, not because crash 🙂
21:38:24 <andythenorth> `OpenTTD.andythenorth/build (master↑13|…) % xcrun --show-sdk-path
21:38:24 <andythenorth> /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk`
21:38:30 <glx[d]> comment in the PR Samu
21:39:21 <andythenorth> oh is it time to guess CMAKE flags to the SDK again?
21:39:58 <petern> Dunno, I don't have a working mac.
21:40:05 <Rubidium> andythenorth: is there a cmath.h or so in /usr/local/include?
21:42:58 <andythenorth> hmm I was about to start trying random stackoverflow suggestions
21:43:07 <andythenorth> but that's the problem, not the solution
21:43:55 <Samu> didn't test triggering CmdRefitVehicle, let me build some MAIL aircraft
21:44:36 <petern> Heh, that links to the one I linked to 🙂
21:52:46 <andythenorth> hmm seems I need to delete and reinstall XCode
21:53:00 <andythenorth> google search for how to do that suggest I need to buy MacCleaner or something
21:53:18 <andythenorth> apparently it's created junk files that slow down my Mac and needs an urgent fix
21:57:56 <andythenorth> ok so now I have nothign
21:58:06 <andythenorth> no clang, no git, no toolchains 😛
21:59:23 <andythenorth> 80 hours to install
21:59:29 <andythenorth> writing progress bars is lol
22:06:34 <petern> michi_cc[d]: As airport tiles can't even access airport information... Hmm.
22:07:07 <petern> But I don't know if that should be included with airport tiles (like for stations) or done as airport tile parent.
22:09:29 <andythenorth> fucking SO overflow random advice 😛
22:09:35 <andythenorth> I now have no dev environment at all 🙂
22:12:14 <Pruple> petern: I think ultimately the answer to "is town information useful to station tiles" is "not really". Can we have airport tiles parented to airports without getting bogged down in a grand scheme for reworking all of newgrf? 🙂
22:12:41 <petern> There's no spec for what an "airport" is.
22:13:13 <petern> There is a spec for stations, but that's the railstation spec.
22:13:49 <glx[d]> and airport and station share some variables
22:14:26 <petern> But the code the rail part and the base part is all mingled together.
22:15:14 <glx[d]> funny clicking on airport tile doc points to industry tile
22:17:28 <petern> No, it's the same as airport tiles
22:18:43 <petern> Not sure where that 0x40 is handled though 😮
22:19:11 <andythenorth> ok after rebuilding xcode same error
22:19:22 <andythenorth> so either I have broken headers / paths
22:19:23 <petern> No, there is an airportresolver already
22:19:26 <andythenorth> or Apple broke the build
22:19:53 <petern> I missed it becuase it's not in a header.
22:20:35 <glx[d]> andythenorth: try verbose build
22:21:07 <petern> The airport resolver has hardly anything in it though, nothing like the (rail)station resolver.
22:21:13 <petern> But it is there I guess.
22:26:28 <Samu> is it just me or are weights being miscalculated also?
22:28:41 <Samu> 1 tonne = 256 tonnes in the "fixed" gui
22:32:22 *** esselfe has quit IRC (Remote host closed the connection)
22:32:25 <andythenorth> is verbose build a cmake flag now?
22:32:46 <glx[d]> just set VERBOSE env var
22:33:36 <glx[d]> or `cmake --build . -j $(sysctl -n hw.logicalcpu) --verbose` like in the CI run I did 🙂
23:00:42 <glx[d]> `clang: warning: argument unused during compilation: '-mno-sse4' [-Wunused-command-line-argument]` funny, it should only be there for non arm64
23:02:21 <andythenorth> oh a reproducible crash 🙂
23:05:55 <andythenorth> ok I will file that tomorrow 😛
23:06:06 <andythenorth> brain checked out for the day
23:06:13 <andythenorth> clean repro 3 times though
23:07:06 *** Samu has quit IRC (Quit: Leaving)
23:08:02 <andythenorth> replace vehicles in nested groups
23:08:13 <andythenorth> multiple ways to trigger it
23:08:18 <glx[d]> the only obvious diff I can spot is ```
23:08:18 <glx[d]> -isystem /opt/homebrew/include
23:08:18 <glx[d]> -isystem /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX13.1.sdk/usr/include
23:08:18 <glx[d]> ``` while CI has `-isystem /usr/local/share/vcpkg/installed/x64-osx/include`
23:08:38 <andythenorth> ^ ok that's quite a plausible cause though?
23:08:49 <andythenorth> I can patch cmake to a different local path?
23:09:03 <glx[d]> I guess one of the homebrew lib does the blah/usr/include thing
23:09:47 <glx[d]> CI doesn't use homebrew for libs
23:11:00 <andythenorth> decimal.hh looks suspicious
23:11:05 <andythenorth> I have no idea what any of this does but
23:11:14 <andythenorth> `#undef iscanonical /* math.h */
23:11:14 <andythenorth> #undef isfinite /* math.h */
23:11:14 <andythenorth> #undef isinf /* math.h */
23:11:14 <andythenorth> #undef isnan /* math.h */
23:11:14 <andythenorth> #undef isnormal /* math.h */
23:11:16 <andythenorth> #undef issubnormal /* math.h */
23:11:16 <andythenorth> #undef iszero /* math.h */
23:11:18 <andythenorth> #undef isspecial /* ctype.h */`
23:11:52 <andythenorth> maybe unrelated, I'm just searching for things at this point
23:13:51 <glx[d]> one or more lib probably add /opt/homebrew/include, and the sdk/usr/include
23:15:39 <andythenorth> ok I need to sleep 🙂
23:15:47 <andythenorth> thanks for the help so far
23:16:10 *** esselfe has joined #openttd
23:17:16 <glx[d]> anyway I'd blame a lib for doing what it's not supposed to do
23:18:03 <glx[d]> or homebrew, as it is fine with vcpkg
23:26:10 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:36:32 *** sla_ro|master has quit IRC ()
continue to next day ⏵
