IRC logs for #openttd on OFTC at 2021-02-15
⏴ go to previous day
00:05:43 *** andythenorth has quit IRC (Quit: andythenorth)
00:36:08 *** HerzogDeXtEr has joined #openttd
02:10:55 *** Wuzzy has quit IRC (Ping timeout: 480 seconds)
02:19:50 *** Wormnest has quit IRC (Quit: Leaving)
02:36:13 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
03:35:42 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
04:04:46 *** Wuzzy2 has quit IRC (Quit: Wuzzy2)
04:38:47 *** supermop_Home has quit IRC (Ping timeout: 480 seconds)
05:22:08 *** WormnestAndroid has quit IRC (Remote host closed the connection)
05:24:00 *** WormnestAndroid has joined #openttd
07:19:15 *** snail_UES_ has quit IRC (Quit: snail_UES_)
07:30:25 *** andythenorth has joined #openttd
07:55:33 *** sla_ro|master has joined #openttd
08:06:35 *** HerzogDeXtEr has joined #openttd
08:24:08 *** heffer has quit IRC (Quit: heffer)
09:05:56 *** jottyfan has joined #openttd
09:10:02 *** jottyfan has joined #openttd
09:26:00 *** jottyfan has quit IRC (Quit: jottyfan)
09:40:45 *** Eddi|zuHause2 is now known as Eddi|zuHause
10:08:36 *** jottyfan has joined #openttd
11:51:37 <DorpsGek> [OpenTTD/OpenTTD] orudge opened pull request #8675: Fix: [Actions] Use vcpkg to provide libpng on macOS https://git.io/JtXmI
11:52:51 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #8675: Fix: [Actions] Use vcpkg to provide libpng on macOS https://git.io/JtXmY
12:04:20 *** virtualrandomnumber has joined #openttd
12:04:32 *** virtualrandomnumber has quit IRC ()
12:07:21 *** jottyfan has quit IRC (Quit: jottyfan)
12:10:15 <DorpsGek> [OpenTTD/OpenTTD] orudge merged pull request #8675: Fix: [Actions] Use vcpkg to provide libpng on macOS https://git.io/JtXmI
12:27:56 *** Samu has quit IRC (Read error: Connection reset by peer)
12:38:46 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: Change: add support for next/previous railtype global hotkeys https://git.io/JtXOy
12:55:58 <DorpsGek> [OpenTTD/OpenTTD] grossws updated pull request #7851: Change: add support for next/previous railtype global hotkeys https://git.io/JeM3F
12:56:25 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: Change: add support for next/previous railtype global hotkeys https://git.io/JtX3S
12:56:33 *** roadt__ has joined #openttd
13:03:29 *** roadt_ has quit IRC (Ping timeout: 480 seconds)
13:05:27 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
13:05:41 *** WormnestAndroid has joined #openttd
13:07:13 <_dp_> hm, I just found that HQ produces way less cargo than it was supposed to
13:07:57 <_dp_> or who the hell knows actually since it seems hq was supposed to produce as 256 pop house but houses weren't supposed to produce quadratic amounts
13:08:14 <_dp_> anyway, looks like CS bug so call in a feature :p
13:10:26 <_dp_> basically houses produce as population**2 but hq as 4*((population /4)**2) xD
13:11:23 <Eddi|zuHause> and where do you take the 256 figure from?
13:12:40 <_dp_> comment was added later ofc but it seems to match the intention
13:18:21 <_dp_> though I've no idea how comments in r975 come to be
13:24:25 <Eddi|zuHause> so what was meant was that since it's 4 tiles, each tile adds 1/4th of the total, and the "wrong" part is the **2-ness of the formula
13:25:23 <Eddi|zuHause> but how does that compare to regular 2x2 houses?
13:26:58 <_dp_> regular houses only produce from one tile
13:27:35 <Eddi|zuHause> anyway, the whole thing seems fine, except it should apply the same linearization as was done with houses, and "amt = (amt + 1) >> 1" should do "/2" instead of ">>1"
13:28:01 <Eddi|zuHause> _dp_: are you sure? the tileloop will hit each tile of the 2x2 house at different points
13:28:27 <_dp_> pretty sure, one one part of house has population iirc
13:30:03 <_dp_> yeah, you can even inspect that in game
13:31:30 <_dp_> newgrf callback though gets called for each tile so everything is possible
13:32:27 <_dp_> * one one -> only one xD
13:33:51 *** sla_ro|master has quit IRC ()
13:34:36 <Eddi|zuHause> yeah, src/table/town_land.h lists the tiles individually with 3 tiles as 0 population
13:35:19 <Eddi|zuHause> but that's just a specific implementation detail of the original houses
13:35:24 <Eddi|zuHause> not a general spec thing
13:36:37 <Eddi|zuHause> so even without callback, newgrf houses can just set non-zero population for each house part
13:37:21 <_dp_> well, yeah, but it's kind of irrelevant here
13:37:41 <_dp_> I doubt hq was supposed to produce as 128+128 house or smth :p
13:38:30 <Eddi|zuHause> well, if the formula had been linear in the first place, that would be equivalent
13:38:52 <Eddi|zuHause> and possibly easier than trying to figure out which part of the HQ you hit
13:39:50 <Eddi|zuHause> "if (o->location.tile == tile)" <-- well, it might actually be easy enough :p
13:40:09 <_dp_> oh, and btw, recession does nothing to hq production except for the highest level one xD
13:40:55 <Eddi|zuHause> why would lower levels only produce 1?
13:41:19 <_dp_> only highest have production of 2
13:41:31 <Eddi|zuHause> i think your math is off
13:42:08 <_dp_> look carefully, amt is always <= 1 if level != 5
13:42:28 <_dp_> it only enters if on small numbers and then divides it to 0
13:42:58 <Eddi|zuHause> i think your math is off
13:43:10 <Eddi|zuHause> r is between 0 and 255
13:43:20 <Eddi|zuHause> you check r against 256/4/6
13:43:36 <Eddi|zuHause> and then set amt = r/8/4
13:45:18 <Eddi|zuHause> actually, /5 because minimum level appears to be 1
13:45:48 <_dp_> whatever, even 256/4/2 is already 32 :p
13:48:06 <Eddi|zuHause> ok, that is a bit weird
13:53:18 <Eddi|zuHause> anyway, that's a side effect of passengers trickling in by individual tiles in general, instead of industries which deliver in large chunks
13:53:53 <Eddi|zuHause> nothing to really worry about
13:54:16 <Eddi|zuHause> just means in a recession people still need to go places :)
14:24:35 *** andythenorth has quit IRC (Quit: andythenorth)
14:29:41 *** supermop_Home has joined #openttd
14:57:13 *** andythenorth has joined #openttd
15:05:29 <Wuzzy> Spreaking of OpenSFX...
15:05:59 <Wuzzy> when #14 is merged, the official description of OpenSFX will change as well. however, it is translated (lang/ directory)
15:06:18 <Wuzzy> the question is... what to do with all the translations then?
15:11:35 <Eddi|zuHause> i think i said this yesterday: discard the translations, and announce at the forums that new ones are needed
15:16:41 <Eddi|zuHause> i don't think eints integration would be particularly desirable
15:28:13 *** Progman has joined #openttd
15:33:25 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
15:34:07 <Wuzzy> yeah, i had the same idea. sounds the most reasonable
15:51:33 *** gelignite has joined #openttd
16:17:04 <Wolf01> So, yesterday I think I removed too much, but it shouldn't be a problem, the min .net version I still have installed is 4.6.1 and all the stuff seem to work fine, I just needed to update a project which referenced 4.5
16:18:38 <Wolf01> At least now I added ~20GB to the free space :P
16:20:28 <Wolf01> Too bad I can't remove VS2017 too... 2019 doesn't support at all the windows mobile platform
16:23:57 *** Speeder has joined #openttd
16:26:06 <glx> for openttd both should work
16:29:32 <Wolf01> Oh damn... still errors even with VS2019... can't find embed.manifest
16:31:14 <Wolf01> Btw, why cmake compiles on my user folder instead of using the other HDD with a lot of free space?
16:32:27 <glx> I guess you somehow did a clean/rebuild and use source from before 99448eedc (10 days ago)
16:33:00 <glx> cmake defaults to user folder, but it's all configurable
16:33:12 <glx> via GUI or CMakeSettings.json
16:34:13 <glx> until you update source to recent, you can just restore os/windows/openttd.manifest
16:34:32 <glx> cmake deletes it on clean due to a mistake
16:35:24 <Wolf01> Ok, it compiled now and started
16:36:36 <Wolf01> <glx> via GUI or CMakeSettings.json <- I found the folder path (which doesn't even exists) but no other configs related to the compile path
16:38:28 *** gelignite has quit IRC (Quit: Stay safe!)
16:38:45 <glx> I set it to ${projectDir}/out/build/${name}
16:40:27 <Wolf01> Oh right, it's everything under "manage configurations" now
16:42:09 <glx> yeah and default configuration is debug
16:42:53 <Wolf01> Yes, I only have that one
16:43:41 <glx> you can clone it and change build type to RelWithDebInfo
16:44:00 <Wolf01> Nah, debug is fine for me :P
16:44:42 <Wolf01> I don't need to benchmark the game, there is Samu for that :P
16:44:50 <glx> I have 17 configurations ;)
16:45:44 <glx> basic x64 and x86 (both debug and release), then another 4 using clang
16:46:28 <glx> mingw64 debug and release, arm64 debug and release
16:46:42 <glx> a debug x64 using vs2017 toolchain
16:47:18 <glx> and x64 non static debug and release
16:47:30 <glx> but I mostly use x64-debug
16:58:44 *** sla_ro|master has joined #openttd
17:02:51 *** smoke_fumus has joined #openttd
17:25:41 *** frosch123 has joined #openttd
17:33:10 *** gelignite has joined #openttd
17:36:50 *** Wormnest has joined #openttd
17:39:08 *** snail_UES_ has joined #openttd
18:01:50 *** Wormnest_ has joined #openttd
18:08:41 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
18:11:31 *** Wormnest has joined #openttd
18:15:33 *** Cursarion has quit IRC (Ping timeout: 480 seconds)
18:17:37 *** Cursarion has joined #openttd
18:18:09 *** Wormnest_ has quit IRC (Ping timeout: 480 seconds)
18:43:00 *** Wuzzy has quit IRC (Read error: Connection reset by peer)
18:47:00 *** Wormnest_ has joined #openttd
18:52:07 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
19:02:03 <DorpsGek> - Update: Translations from eints (by translators)
20:47:53 <LordAro> must be a recent change
20:48:01 <LordAro> possibly the world gen reorganise?
20:48:21 <glx> starting the exe to trigger it
20:54:46 <glx> let's add some breakpoints
21:03:16 <glx> should not be an issue in theory
21:04:10 <Wolf01> Maybe it doesn't create the default?
21:04:25 <LordAro> be nicer if you could get something to warn/error at compile time
21:04:41 <Wolf01> Looking at the comments it doesn't work even with a fresh config
21:04:43 <glx> it doesn't find the setting
21:05:22 <FLHerne> Wolf01: I'm not sure if that comment is "I tried and it didn't work" or "I'm sure it won't work"
21:05:50 <glx> I think I know which commit breaks, checking
21:05:59 <LordAro> it indeed makes no difference though
21:08:33 <Wolf01> Il looks there is something weird on that setting
21:09:16 <Wolf01> Like $var in place of $name and missing $var parameter
21:09:26 <Wolf01> Oh different setting type
21:11:11 <glx> commit in itself looks good, other parts of the code except settings window work
21:11:53 <andythenorth> nice issue title :)
21:12:27 <LordAro> andythenorth: settings causes accident!
21:12:35 <andythenorth> goes it throw out
21:14:55 <LordAro> so what's been forgotten?
21:15:09 <LordAro> what makes a setting "early load"?
21:16:56 <frosch123> "early load" are settings in misc_settings
21:17:03 <frosch123> they are loaded before grfscan
21:17:10 <frosch123> i.e. needed to display the basic intro screen
21:17:23 <frosch123> and probably bootstrap
21:17:46 <glx> confirmed it's a2c3197f424
21:18:49 <frosch123> the zoom settings are the first SDTC_ in misc_Settings
21:19:04 <frosch123> everything else in misc_setting are global variables outside the settings struct
21:19:50 <glx> yup seems they are not in _settings
21:20:03 <frosch123> ah, the crash is because there is no "name="
21:21:15 <glx> well name is replaced by var in the macro
21:25:19 <glx> of course they are in _misc_settings[] now
21:26:13 <frosch123> so none of misc_settings work in console or in the tree?
21:27:24 <glx> hmm something looks wrong in setting table
21:27:43 <glx> let's try a modification in .ini
21:28:07 <frosch123> i see in the debugger, that there is a "gui.zoom_min" in _misc_settings
21:28:32 <frosch123> but GetSettingFromName and friends only care about _settings
21:29:56 <frosch123> so neither console nor tree work for _misc_settings
21:30:14 <frosch123> and i have no idea whether they were intentionally left out :)
21:30:41 <glx> I think it's intentionnal, based on the settings in misc
21:32:53 <frosch123> this looks difficult to me :p
21:33:06 <frosch123> i would now try to merge _misc_settings into _settings
21:33:18 <glx> and it's the first non global put in misc
21:34:44 <glx> I wonder if it even save/load correctly from cfg
21:35:05 <frosch123> the ini load/save is the only place that uses _misc_settings
21:35:41 <frosch123> what would break if we move all _misc_settings into _settings as client settings?
21:35:50 <frosch123> and add some gui_flag or similar for the early-load
21:43:20 <frosch123> i think that's the best solution. other ideas?
21:43:48 <frosch123> remove misc_settings, put all of them into _settings. add some flag for early load
21:44:22 <frosch123> that way GetSettingFromName can return a index into _settings for them
21:44:41 <frosch123> and SetSettingValue and IConsoleSetSetting work
21:45:51 <frosch123> there are probably some traps with settings like "graphicsset", which have string data
21:45:58 <frosch123> i believe no other settings have string data
21:46:05 <glx> there's also _gameopt_settings with SC_ category while they are not in the settings window
21:47:12 <frosch123> i think we added SC_ to everything, no matter whether in tree or not
21:47:59 <frosch123> second suggestion: keep _misc_settings, but add an early-load flag to _settings nevertheless, and them move the zoom settings back to _settings
21:49:52 <glx> I'm quite sure the value in [gui] section are not loaded anyway, as misc_settings should read only [misc] section
21:54:36 *** Samu has quit IRC (Quit: Leaving)
21:55:05 <frosch123> no, "misc" is just the default for settings without a "." in their name
21:55:34 <glx> ah yes (looking at the code)
21:56:04 <frosch123> that's where i got the info from :p
21:56:41 <glx> yeah I was in the loading code when you mentioned the "."
22:01:20 <glx> second suggestion looks sane
22:02:32 <glx> and replacing basic_setting and other settings in HandleSettingDescs with a bool parameter in loading proc
22:03:06 <frosch123> do you remember a setting that requires late loading?
22:03:53 <frosch123> except for the obvious GRFLoadConfig, AILoadConfig and GameLoadConfig
22:05:53 <frosch123> so, third solution: load all _settings early?
22:06:05 <glx> looking at all LoadFromConfig() calls
22:06:34 <frosch123> those in 1755-1771 i understand
22:06:55 <frosch123> but no idea why HandleSettingDescs makes such a deal of loading some settings late
22:07:33 <glx> yeah I think it could always load all, except misc, unless at startup
22:08:22 <glx> as anyway they will be reloaded after grf scan
22:08:23 <frosch123> you probably are only allowed to load them once. but i think you can load all of them early
22:08:39 <frosch123> there are some comments about custom currency and other weird things
22:09:37 <glx> I see 3 calls, one in main (the early load), one after grf scan, and a last one for servers
22:10:06 <glx> so basically even the early load could load everything
22:12:25 <glx> it was probably some kind of optimisation to not reload multiple times the same thing
22:13:18 <frosch123> but does it need reloading? can HandleSettingDescs just only run for "minimal"?
22:19:22 <frosch123> night, and good luck :)
22:19:26 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
22:21:43 <glx> oh I think we can always load them,
22:24:18 <glx> and probably all other !minimal settings could be always loaded too
22:24:18 <michi_cc> Probably. In the end, what bad is supposed to happen when loading an integer value twice?
22:25:19 <glx> in workflow I can see load minimal, scan newgrf, load other
22:25:44 <glx> and load other happen after each newgrf scan
22:26:40 <glx> so maybe we just need to take care to not overwrite these 2 gui. value, like we do for newgrf count
22:31:28 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:32:00 <michi_cc> Do you expect the cfg value to change somehow? Otherwise overwriting an integer with itself is totally fine.
22:32:34 <glx> not the cfg, but if user changed the setting before downloading a newgrf
22:33:09 <glx> well it's like other settings in this case I guess
22:42:36 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:45:35 <glx> oh config is always saved on setting change once at least a newgrf scan is done
23:02:44 <glx> hmm it seems moving UpdateGUI() call in ottd_main after grf scan could be enough to not require early loading
23:03:34 <glx> but I guess it's before so the progress bar is properly scaled
23:05:20 <michi_cc> Yeah, if you do it then the grf scan bar will not be scaled and the title game zoom might jump after the scan.
23:09:15 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
23:11:44 <DorpsGek> [OpenTTD/OpenTTD] glx22 opened pull request #8677: Fix #8676: Revert a2c3197f4 and unconditionally load settings "early" https://git.io/JtXSK
23:12:54 *** sla_ro|master has quit IRC ()
23:18:41 *** Progman has quit IRC (Remote host closed the connection)
23:20:03 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8677: Fix #8676: Revert a2c3197f4 and unconditionally load settings "early" https://git.io/JtXSF
23:22:34 *** gelignite has quit IRC (Quit: Stay safe!)
23:24:27 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #8678: Fix #8676: GUI-visible settings may not be part of misc settings. https://git.io/JtX9t
23:24:35 <michi_cc> Well, I let somebody else battle this out :p
23:31:34 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8678: Fix #8676: GUI-visible settings may not be part of misc settings. https://git.io/JtX96
23:50:58 *** azubieta has joined #openttd
continue to next day ⏵