IRC logs for #openttd on OFTC at 2021-02-20
⏴ go to previous day
00:00:41 <DorpsGek> [OpenTTD/OpenTTD] frosch123 opened pull request #8701: Codechange: rename sound ids to make more sense. https://git.io/JtS3o
00:08:43 <glx> frosch123: maybe also rename CcPlaySound_XXX callbacks
00:17:19 <Timberwolf> 60fps definitely enhances tge "ttain set" feel of sitting and watching stuff potter through a junction.
00:17:40 <DorpsGek> [OpenTTD/OpenTTD] frosch123 updated pull request #8701: Codechange: rename sound ids to make more sense. https://git.io/JtS3o
00:17:44 <Timberwolf> (mobile keyboard fail)
00:19:47 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8701: Codechange: rename sound ids to make more sense. https://git.io/JtSsZ
00:21:45 <DorpsGek> [OpenTTD/OpenTTD] frosch123 updated pull request #8701: Codechange: rename sound ids to make more sense. https://git.io/JtS3o
00:38:11 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
00:55:15 *** Progman has quit IRC (Remote host closed the connection)
01:34:54 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
01:55:18 *** Tirili has quit IRC (Quit: Leaving)
02:18:26 *** Flygon_ has joined #openttd
02:25:15 *** Flygon has quit IRC (Ping timeout: 480 seconds)
03:01:19 <DorpsGek> [OpenTTD/OpenTTD] joestringer updated pull request #8679: Fix: [Cygwin] Fix missing uint definition https://git.io/JtXAu
03:01:48 <DorpsGek> [OpenTTD/OpenTTD] joestringer commented on pull request #8679: Fix: [Cygwin] Fix missing uint definition https://git.io/JtSlt
03:30:03 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
03:31:40 *** Wuzzy has quit IRC (Quit: Wuzzy)
03:40:34 *** Wormnest has quit IRC (Quit: Leaving)
03:41:08 *** jjavah0lic has quit IRC (Remote host closed the connection)
03:42:00 *** jjavah0lic has joined #openttd
04:29:04 *** jjavah0lic has quit IRC (Remote host closed the connection)
04:29:26 *** jjavah0lic has joined #openttd
05:05:27 *** jjavaholic_ has joined #openttd
05:05:28 *** jjavah0lic has quit IRC (Read error: Connection reset by peer)
05:13:43 *** snail_UES_ has quit IRC (Quit: snail_UES_)
07:59:24 *** andythenorth has joined #openttd
08:36:08 <andythenorth> what shall we today pinky?
08:37:36 <TrueBrain> you are missing a word there buddy
08:39:23 <andythenorth> my brain and tpying
08:39:57 * andythenorth is wondering why vehicle info window is so...odd
08:40:49 <andythenorth> - 'cargo' shows current amount
08:41:14 <andythenorth> - 'information' shows build date (for engines only) and current value, but not running cost
08:41:41 <andythenorth> - 'capacities' shows capacities
08:42:02 <andythenorth> - 'total cargo' shows total amount (total capacity), and also for some reason...transfer credits
08:44:09 <andythenorth> so finding out about the load state of a train requires clicking and mentally comparing three different sub-windows
08:44:32 <andythenorth> transfer credits is kind of clearly just a lol 'we don't know where to put this'
08:44:46 <andythenorth> and I can't find out which vehicles in the train are expensive to run :)
08:45:25 <TrueBrain> andythenorth: that you replied to someone who clearly did not read any of the replies ...
08:45:30 <andythenorth> the capacities sub-window also prefixes every value with 'Capacity:'
08:45:34 <TrueBrain> I just completely ignored what-ever he wrote, as clearly he didn't care for what we wrote :)
08:45:50 <andythenorth> I was going to do that, but then I remembered I'm me
08:46:02 <andythenorth> even though I try not to be
08:46:03 *** sla_ro|master has joined #openttd
08:46:37 <andythenorth> so we need to show 'Capacity:' every time even though the other window for 'cargo' doesn't prefix? :D
08:47:08 <andythenorth> in discord I could just post screnshots instead of wall-text
08:52:07 *** WormnestAndroid has quit IRC (Remote host closed the connection)
08:52:20 *** WormnestAndroid has joined #openttd
09:06:15 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
09:06:29 *** WormnestAndroid has joined #openttd
09:10:56 *** sla_ro|master has quit IRC ()
09:18:11 <TrueBrain> owh, lol, the win32 driver is running on a weird fps now :D
09:20:10 <TrueBrain> we cannot use WM_PAINT and dictate fps ourselves :D
09:20:19 *** jjavaholic_ has quit IRC (Read error: No route to host)
09:20:45 <nielsm> hmm... stupid idea for being win32-like:
09:21:09 <nielsm> put the game screen in WM_ERASEBACKGROUND (or whatever it's called) and the mouse cursor on in WM_PAINT
09:21:31 <nielsm> but yeah WM_PAINT should really just do nothing if we're repainting the window on a timer anyway
09:21:35 <TrueBrain> lucky for us, it is perfectly fine to draw outside WM_PAINT :)
09:21:58 <TrueBrain> it just means I have to change how the dirty-rect works
09:22:02 <TrueBrain> but now I have that working
09:22:05 <TrueBrain> win32 is also smooth as silk
09:22:09 *** Eddi|zuHause2 is now known as Eddi|zuHause
09:22:38 <TrueBrain> just need to check I don't make the OpenGL branch have more merge conflicts :
09:26:53 <TrueBrain> seems I cannot avoid this
09:27:00 <TrueBrain> so I will just write a I AM SO SORRY michi_cc LETTER :)
09:27:11 <TrueBrain> there will be some merge conflicts in the win32 video driver :)
09:27:45 <andythenorth> YouTube seems to think...I need easy income fast
09:28:09 <andythenorth> from working online, or as a deliveroo rider, or from secret investment strategies
09:38:42 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #8702: Fix: [Win32] now we are drawing on a tick, no longer use WM_PAINT https://git.io/JtSrT
09:39:02 <TrueBrain> michi_cc: I think it is best you review ^^, both to check if what I did is correct, but also to get the function you have to change in the OpenGL branch :)
09:43:36 *** Progman has joined #openttd
10:14:38 * andythenorth refactoring horses
10:14:45 <andythenorth> so bugs introduce
10:17:13 *** Progman has quit IRC (Remote host closed the connection)
10:27:31 *** jjavaholic has joined #openttd
10:29:38 <TrueBrain> how cute, win32 driver had an option to set display-hz
10:30:11 <Eddi|zuHause> 120Hz ought to be enough for anyone :p
10:30:46 <Eddi|zuHause> <insert "the human eye can't see more than 18Hz anyway" comment> :p
10:31:23 <Eddi|zuHause> (where the number before Hz is effectively a random number)
10:32:53 <andythenorth> same game for audio speakers also?
10:40:28 * andythenorth is going to add some very stupid things to Horse
10:51:01 <TrueBrain> ../src/debug.cpp:48:uint32 _realtime_tick = 0;
10:51:05 <TrueBrain> lol @ location of that variable ..
11:07:36 <andythenorth> hmm shall I do a nice fun detailed mechanic that is very hard to understand?
11:07:53 <andythenorth> OpenTTD tends towards 'crude' :D
11:12:04 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8701: Codechange: rename sound ids to make more sense. https://git.io/JtS6H
11:12:49 <TrueBrain> why doesn't LordAro like confusing numbers?! :o
11:18:45 <andythenorth> simple mechanic: restaurant car in train reduces run costs of all coaches in train by 50%
11:18:55 <andythenorth> due to sale of buns and sherry
11:19:20 <andythenorth> problem: run costs are only about £10k on a train that already earns £100k / year
11:19:41 <andythenorth> if I increase pax coach run costs, players reliably report they can't make any more with pax trains
11:19:58 <andythenorth> due to piss poor networks, and/or cdist transfer diluting income
11:21:59 * andythenorth might do it anyway
11:22:22 <nielsm> need a mechanic where passengers that travel a long distance pay a bonus fare
11:22:43 <nielsm> straight up bonus fare
11:25:22 <andythenorth> can't be done I think
11:25:47 * andythenorth doesn't want to be the 'nothing is possible' guy though :D
11:26:33 <andythenorth> but transfer payments...
11:26:54 <andythenorth> and when things are actually calculated?
11:27:33 * andythenorth hopes to be wrong :)
11:28:14 <andythenorth> negative income would be the most lolz opportunity though
11:28:31 <andythenorth> also income from pax waiting on stations, who are forced to go to the pub
11:28:46 <andythenorth> we kinda lack some of the small fun things that other transport games have as standard
11:28:54 <andythenorth> like special bonus vehicles, and station 'improvements'
11:29:13 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSi2
11:29:40 <TrueBrain> okay, that was a lot harder than I expected, but the result is pretty clean I think. michi_cc I might have overstated that this might help rebased OpenGL ... :(
11:29:49 <TrueBrain> I did cherry-pick some things from the OpenGL branch
11:30:00 <TrueBrain> and did some renaming you were going to do anyway
11:30:04 <TrueBrain> so hopefully that helps a lot :)
11:30:12 <TrueBrain> but it won't be fully trivial .. sorryyyyyyy
11:30:28 <TrueBrain> (I am even willing to rebase btw, if that helps)
11:31:32 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSi2
11:31:35 <TrueBrain> stupid commit-checker :P
11:31:48 <michi_cc> Haven't look at it yet, but it's okay, I'll deal with it somehow :D
11:32:59 <TrueBrain> I am totally unsure what to do with GdiFlush .. it is a real pain in my ass
11:33:09 <TrueBrain> just so you already know :P
11:34:38 <TrueBrain> possibly a really stupid question .. I made "std::chrono::steady_clock::time_point" a variable of a class .. I do not explicitly initialize the value .. is it safe to assume it will be 0?
11:34:50 <LordAro> TrueBrain: you can safely remove all the #if SDL_VERSION_ATLEAST(1, 3, 0) bits from the sdl1 driver, btw
11:34:50 <TrueBrain> or will it be some random value?
11:35:12 <TrueBrain> LordAro: honestly, I think I can safely remove SDL1 :P
11:35:25 <TrueBrain> but yeah, 1.3 never happened, so I might as well
11:35:26 <LordAro> but if you want to make the line removal count look better :p
11:36:26 <LordAro> TrueBrain: time_point is a class, so it will be default constructed, i think
11:36:35 <LordAro> (regardless of what constructors do)
11:38:00 *** HerzogDeXtEr has joined #openttd
11:38:28 <TrueBrain> given it works as it is, I assumed as much, but I was wondering if that was by spec, or just random compilers doing stuff :P
11:39:54 <LordAro> simulate it with a class that prints when it's being constructed and making that a member of another class
11:40:23 <TrueBrain> that doesn't tell me if it is by spec, or random compiler behaviour :)
11:40:32 <TrueBrain> I know it initializes as 0 on GCC / MSVC
11:40:50 <TrueBrain> I was just wondering if that was something that is fine to do, or that it needs to be more explicit :D
11:41:11 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSi2
11:42:29 <TrueBrain> michi_cc: I have not; I haven't even tested screenshots recently :)
11:43:29 <TrueBrain> works "fine" without it, at least :P
11:44:04 <TrueBrain> screenshots are not threaded, are they?
11:44:07 <DorpsGek> [OpenTTD/OpenTTD] nielsmh commented on pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSiF
11:46:02 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSib
11:46:25 <LordAro> TrueBrain: there's an extra warning too :)
11:46:50 <TrueBrain> michi_cc: okay, and this is a class, so we should be fine
11:46:58 <LordAro> michi_cc: you should know better than to use 'random' in such technical contexts :p
11:47:10 <TrueBrain> LordAro: DOH! Especially as I know about it :P
11:47:15 <DorpsGek> [OpenTTD/OpenTTD] nielsmh commented on pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSih
11:47:32 <michi_cc> Sorry, "are initialized to indeterminate values". Very different from random :)
11:48:56 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSPe
11:49:22 <TrueBrain> michi_cc: it seems screenshots are created from the game-tick
11:49:25 <DorpsGek> [OpenTTD/OpenTTD] nielsmh commented on pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSPv
11:49:34 <andythenorth> playing a test game for cargo_age_period that includes transfers :(
11:49:36 <TrueBrain> so I cannot find a way to break it :P
11:50:15 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSPT
11:50:26 <andythenorth> given A<->B<->C route, what happens to income of B->C leg if A->B leg income is increased? o_O
11:50:33 <andythenorth> B->C leg makes less money
11:50:44 <michi_cc> TrueBrain: Well, yes, anything that is called from GameLoop() needs to re-lock the video buffer. You'll only notice it with OpenGL as it zeroes out _screen.dst_ptr.
11:50:58 <TrueBrain> ah, of course, OpenGL does different things there
11:51:02 <andythenorth> but backload C->B paying more will cause B->A to make less money
11:51:11 <andythenorth> so net, all my trains will make same
11:51:12 <TrueBrain> so it is a fix for OpenGL :D
11:51:26 <andythenorth> which is why changing cargo_age_period is having no visible effect
11:52:07 <andythenorth> transfer payments are brain breaking :D
11:52:26 <michi_cc> I'd argue it is a fix in general, as the screen buffer "belongs" to the video driver between UnlockVideoBuffer and LockVideoBuffer. (Yes, I know, none of the existing drivers *do* aynthing with it, but still :D)
11:53:42 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSPq
11:54:05 <TrueBrain> I have to say, I am pretty happy with the video driver code like this
11:54:19 <TrueBrain> and I really love playing OpenTTD at 60fps .. SO SILKY SMOOTH, whhhiiiiieeeeeeeeee
11:55:29 <TrueBrain> and of course we did all that because we are releasing to Steam. That was the motivation right? Scared of bad reviews? Because a game from 1995 runs at 33fps :P
11:56:10 <LordAro> TrueBrain: just need smooth vehicle movement now :p
11:56:22 <TrueBrain> and smooth news window popout
11:56:26 <TrueBrain> but that is really hard to make smooth
11:56:35 <TrueBrain> as you kinda need AA for that
11:56:55 <TrueBrain> which means we have to ditch our 8bpp blitters
11:57:24 <TrueBrain> (anything that you move by pixel, with a set palette, will never feel smooth)
11:58:07 <TrueBrain> well, or just don't support it on 8bpp blitters, I guess
11:58:12 <TrueBrain> no need to be all drastic about it, I guess
11:58:37 <TrueBrain> the news window is funny .. it moves every 15ms (which is interesting as the game was running every 30ms, but okay)
11:58:44 <TrueBrain> so sometimes it moves 2 pixels per game-tick
11:58:49 <TrueBrain> just .. because .. timers
11:58:54 <TrueBrain> and that feels jerky, no matter what you do
11:59:16 <TrueBrain> it now animated on draw-tick
11:59:19 <TrueBrain> which makes it slightly better
11:59:20 <nielsm> I look forward to getting to a place where all the native 8bpp mode and blitters can be excised
11:59:26 <TrueBrain> but there too, you can notice the stuttering of the pixel
11:59:40 <TrueBrain> the only thing I could think off to solve it, was using subpixels
12:00:11 <TrueBrain> will have to look into it more .. as that news popup has to become smooth!
12:00:21 <TrueBrain> (and after that, vehicles, possibly, but first news animation :D)
12:00:36 <TrueBrain> nielsm: well, with OpenGL, that might be worth concidering
12:00:43 <TrueBrain> concidering? Considering ..
12:01:21 <TrueBrain> but that means: reviewing 2 of my PRs, merge, rebase, review of OpenGL PR, merge! :P :D
12:01:26 <TrueBrain> done by end of weekend, yes yes? :D
12:07:43 <andythenorth> I've been testing pax payment rates using 'all trains' window and looking at 'last year'
12:08:00 <andythenorth> but OpenTTD is caching those values and not updating in [cases I can't repro reliably yet]
12:09:24 * andythenorth wonders what the trigger is
12:09:50 <andythenorth> switching group triggers a recalculation / repaint
12:09:54 <michi_cc> TrueBrain: Quick shopping trip first, then I'll look at the video PRs :)
12:15:00 <andythenorth> did anybody try adjusting station rating according to size of connected cdist network?
12:15:24 * andythenorth thinks many things might have been tried in cdist
12:18:04 <andythenorth> it's a good proxy for 'passenger generation'
12:50:07 *** jjavaholic has quit IRC (Quit: Leaving)
12:54:51 <TrueBrain> "It's apparent that you have used AdWords marketing to promote your business in the past". We did?! WHEN?! WHO DID THAT!
12:55:24 <michi_cc> TrueBrain: If I move the mouse fast enough, I get wrong clipping of the cursor. It seems to be a preexisting bug though, as I can see that in versions before the refresh rate as well.
12:55:53 <TrueBrain> well, which video-driver is my real question ofc :P
12:56:38 <TrueBrain> if I move real fast, I just have a hard time finding my mouse
12:57:10 <TrueBrain> what do you mean with wrong clipping exactly?
12:57:50 <michi_cc> If I move the mouse e.g. to the left, every few frames the first x pixel rows on the left are not drawn.
12:58:08 <TrueBrain> how would you spot that :P
12:58:13 <michi_cc> It's really hard to accurately see, it just feels like some cursor flickering.
12:58:45 <TrueBrain> cannot really reproduce, I have to say
12:59:08 <LordAro> the mouse is looking very flickery on your wm-paint branch
12:59:11 <TrueBrain> of course if I move the mouse quick, it feels like there are more than one mouse on the screen :D
12:59:17 <LordAro> just testing to see whether it's the same on master
12:59:46 <michi_cc> LordAro: I checked an earlier commit, and I'd says it is a pre-existing problem.
13:00:07 <michi_cc> Let me see if I can make a video.
13:00:15 <TrueBrain> pre-existing or not, it would be lovely if the mouse is as silky smooth for everyone as it is for me :D
13:00:23 <LordAro> TrueBrain: as in, to my eyes it looks like the mouse is not getting drawn for some frames
13:00:35 <LordAro> (when moving it fast)
13:00:58 <TrueBrain> I see multiple mouses (that is, my eyes makes me believe I do)
13:01:01 <LordAro> that was with your PR, but might be unrelated, might be ancient, might be something else entirely
13:01:11 <TrueBrain> yeah, but my PR fixes many of these issues
13:01:45 <nielsm> actually it may be the right thing to do, to store the last painted game screen as a GDI bitmap, paint it in WM_ERASEBACKGROUND, and paint the mouse cursor in WM_PAINT
13:02:00 <TrueBrain> to be clear, the mouse is always painted
13:02:04 <TrueBrain> every frame has the mouse somewhere
13:02:12 <TrueBrain> just the question is, is it responsive enough :D
13:02:44 <TrueBrain> nielsm: that would be rather laggy
13:03:03 <TrueBrain> you either have to let WM_PAINT drive your responsiveness
13:03:08 *** frosch123 has joined #openttd
13:03:32 <TrueBrain> LordAro: any chance you can record it? And does video capture it? :D
13:03:41 <TrueBrain> mainly to see if you and michi_cc are both seeing the same thing :D
13:03:43 <LordAro> i'm still recompiling master :p
13:04:04 <LordAro> also do some comparisons with the native cursor
13:04:06 <michi_cc> TrueBrain: Okay, it's not a drawing bug. A screen recording is totally fine. I think it is a vsync/tearing problem.
13:04:24 <TrueBrain> yeah, we are not vsync'd
13:04:28 <TrueBrain> I would say, we should finish OpenGL
13:04:30 <TrueBrain> and test this again
13:05:13 <TrueBrain> it might absolutely be that my screens are 144Hz, that makes me not see what you are seeing :)
13:05:30 * LordAro merely at a plebian 60Hz
13:05:44 <TrueBrain> you can set refresh_rate in [gui] to 30, and check if it is still there
13:05:48 <TrueBrain> the mouse will feel a lot more laggy
13:05:49 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8701: Codechange: rename sound ids to make more sense. https://git.io/JtS1X
13:07:35 <LordAro> frosch123: the comment about SoundFx wasn't visible in the diff :p
13:08:18 <TrueBrain> btw, LordAro , if you check master, check before 60fps to get a fair comparison :)
13:08:24 <TrueBrain> current master on Windows is the worst combination :P
13:08:57 <TrueBrain> (drawing on 60fps, but depending on WM_PAINT an annoying delay between when you see it .. what is fun, that you notice you are constantly overshooting everything when moving with your mouse, as you are over compensating :D)
13:10:01 <LordAro> i think it might be vsync as well
13:10:16 <TrueBrain> in general, OpenTTD has a lot of vsync issues
13:10:19 <TrueBrain> moving the map around quickly too
13:10:26 <LordAro> definitely looks worse than the native cursor
13:10:26 <TrueBrain> you get these annoying horizontal stripes all over the place
13:10:48 <TrueBrain> but as mentioned, with OpenGL that changes anyway
13:10:55 <TrueBrain> so I am curious how that feels :D
13:11:03 <LordAro> it's probably the same cursor issue that was mentioned in the steam discussion
13:11:17 <TrueBrain> yeah ... software cursor has its downsides
13:11:29 <TrueBrain> hardware ones are always better, but ... also more difficult
13:11:41 <TrueBrain> Factorio combines both, for example
13:11:48 <TrueBrain> you have the system cursor + highlights of the game
13:11:58 <TrueBrain> which is a road-not-taken by OpenTTD :D
13:13:36 <TrueBrain> LordAro: but master from a few days ago has the same issue? Just to be sure :)
13:16:30 *** WormnestAndroid has quit IRC (Remote host closed the connection)
13:16:44 *** WormnestAndroid has joined #openttd
13:23:55 <LordAro> TrueBrain: yeah, no real difference
13:24:09 <TrueBrain> no regression, that is good. But still something we should look at
13:24:17 <TrueBrain> after OpenGL, see if it is still there
13:24:21 <TrueBrain> and address it from there :)
13:24:39 <LordAro> i tried recording it with my phone
13:24:47 <LordAro> video was jpeged to hell, couldn't see anything
13:24:49 *** WormnestAndroid has quit IRC (Remote host closed the connection)
13:24:49 <TrueBrain> you cannot really record vsync issues, sadly :)
13:25:01 <TrueBrain> but knowing that is the issue is part of the solution
13:25:35 <DorpsGek> [OpenTTD/OpenTTD] michicc approved pull request #8702: Fix: [Win32] now we are drawing on a tick, no longer use WM_PAINT https://git.io/JtSM1
13:25:49 <michi_cc> As far as I can see, cursor stuff as nothing to do with ^^
13:26:00 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8702: Fix: [Win32] now we are drawing on a tick, no longer use WM_PAINT https://git.io/JtSrT
13:26:05 <TrueBrain> and if it is, we will address it accordingly :)
13:26:29 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSi2
13:27:59 *** WormnestAndroid has joined #openttd
13:28:09 *** WormnestAndroid has quit IRC (Remote host closed the connection)
13:30:14 *** WormnestAndroid has joined #openttd
13:33:02 *** sla_ro|master has joined #openttd
13:33:12 <_dp_> film the screen in slow-mo? :p
13:36:30 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
13:36:44 *** WormnestAndroid has joined #openttd
13:47:07 <andythenorth> so 'vehicles for this station' window is also caching income
13:47:33 <andythenorth> it's not just that the window isn't painting, I can see vehicle sprites change as expected
13:47:52 <andythenorth> but profit is not updated
13:52:11 <andythenorth> RVs also affected, but RVs are updating occasionally (to wrong values)
13:52:16 <andythenorth> trains seem to not update
13:52:55 * andythenorth will make a proper report
14:14:37 *** Tirili has quit IRC (Quit: Leaving)
14:36:27 <DorpsGek> [OpenTTD/nml] FLHerne opened pull request #189: Change: Check that user code doesn't try to use reserved registers https://git.io/JtS92
14:38:15 *** Wuzzy has quit IRC (Ping timeout: 480 seconds)
14:42:37 <andythenorth> oh that's quite cool :)
14:45:39 <TrueBrain> btw, michi_cc , if you want me to cherry-pick some more blobs into my last PR to make OpenGL easier, just let me know. Sometimes I was doubting a bit what to pull in .. like making draw_threaded a member, and stuff like that
14:46:41 <FLHerne> andythenorth: It just switches two arbitrary numbers for different ones that aren't arbitrarily wrong :p
14:46:50 <andythenorth> FLHerne that's why I like it :)
14:46:52 <FLHerne> (and I hope I didn't miss any)
14:47:04 * andythenorth had a huge late breakfast and is sleepy
14:47:45 *** _2TallTyler has joined #openttd
14:48:06 <andythenorth> will I be bisecting to find out why profit display isn't updating?
14:51:36 <FLHerne> I think I broke `black` somehow
14:51:49 <FLHerne> It wants to reformat *everything*, even using the right args
14:53:41 <LordAro> usually means you forgot the line lenght
15:02:56 <DorpsGek> [OpenTTD/nml] FLHerne updated pull request #189: Change: Check that user code doesn't try to use reserved registers https://git.io/JtS92
15:03:07 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler opened pull request #8704: Change: Move toolbar button for cargo payment graph under industries https://git.io/JtSHC
15:05:20 <_2TallTyler> Does anyone know if it's possible add text padding when a window is just as wide as the title text? This is the only thing keeping #8688 in draft form (screenshot in PR).
15:09:35 <FLHerne> andythenorth: fwiw, NML uses its location range from 128 upward, so it almost certainly didn't cause actual problems
15:10:04 <andythenorth> the counting down thing is weird
15:11:50 <FLHerne> "In German, or English, I know how to count down..."
15:11:57 <FLHerne> Why does it count down? :p
15:13:24 <andythenorth> because JFDI coding
15:13:43 <andythenorth> not wanting to work out if registers from 0 up were already used or not
15:13:56 <andythenorth> I housekeep those now, so I would do it differently
15:18:45 <andythenorth> do I need to rerun cmake every build?
15:29:31 <FLHerne> (if you haven't changed things, it'll exit quickly without doing much)
15:34:21 * andythenorth bisecting for when train window stopped showing profit
15:34:43 <FLHerne> Using -GNinja will speed things up quite a bit over the Makefile backend
15:40:35 <andythenorth> I'm in november currently :P
15:40:44 <andythenorth> surprised nobody noticed this before
15:43:43 <andythenorth> lol managed to hit a rev that asserts, super :)
15:48:21 <DorpsGek> [OpenTTD/OpenTTD] michicc approved pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSQN
15:49:07 <andythenorth> train window is updating as expected in dddf885fb400d3ae0a24429d5cdd88f5dd921693 from 7th dec
15:49:15 * andythenorth leaving that here in case I get hit by a bus
15:49:19 <andythenorth> or lose my browser tab
15:57:52 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
15:57:58 *** WormnestAndroid has joined #openttd
16:00:16 <andythenorth> failing by 7b515fa0e287e45aee8281f44ea357498b22020f 25th dec
16:02:51 *** Flygon_ has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:06:02 <andythenorth> lol 21st dec, when mac ffwd got me 8x on an empty map, and the mouse was 30fps
16:06:10 <andythenorth> such changes since then :)
16:08:47 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8703: Deduplicate tick-related code in video drivers https://git.io/JtSi2
16:20:48 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8700: Change: Recolour windows for consistency https://git.io/JtS5V
16:27:39 <andythenorth> lot of marking dirty / force rebuild
16:28:27 <andythenorth> vehicle sprites DO update
16:35:18 *** snail_UES_ has joined #openttd
16:40:18 <DorpsGek> [OpenTTD/OpenTTD] andythenorth opened issue #8705: Vehicle list windows fail to update profit display whilst open, after 584df548f1bdfa27482aa995d151542abe4f0bca https://git.io/JtSd4
16:40:55 <Wolf01> Ahahah hit fast forward with animation off, really nice, most intense 50 years in 30 seconds ever
16:41:27 <andythenorth> ^^ this bug has really fucked with my testing of cargo_age_period :)
16:41:35 <andythenorth> that is a lot of junk testing I did :)
16:41:39 <andythenorth> like 2 hours or so
16:42:13 <Wolf01> You should really train an AI to test that kind of stuff
16:43:07 *** gelignite has joined #openttd
16:44:45 <andythenorth> I have been taking screenshots of profit annually, and then re-running with different cargo_age_periods
16:45:07 <andythenorth> but I had the vehicle window open, left the game on ffwd
16:45:21 <andythenorth> I only noticed when I started seeing identical values year on year
16:47:30 <Wolf01> Oh you meant that bug, sorry, I don't have any attention span :P
16:47:51 <Wolf01> Also, I think I'll take another nap
16:51:35 <andythenorth> TB has tidied up
16:53:49 <DorpsGek> [OpenTTD/nml] frosch123 commented on pull request #189: Change: Check that user code doesn't try to use reserved registers https://git.io/JtSdb
16:55:39 <TrueBrain> andythenorth: you were the one that wanted people to use less words to say the same :P :P :P
16:55:53 <andythenorth> long hash is long
16:58:34 <frosch123> FLHerne: nml has "make black"
17:03:24 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
17:03:56 *** WormnestAndroid has joined #openttd
17:06:57 <DorpsGek> [OpenTTD/OpenTTD] stormcone commented on issue #8705: Vehicle list windows fail to update profit display whilst open, after 584df54 https://git.io/JtSd4
17:09:15 <DorpsGek> [OpenTTD/OpenTTD] andythenorth commented on issue #8705: Vehicle list windows fail to update profit display whilst open, after 584df54 https://git.io/JtSd4
17:09:18 <DorpsGek> [OpenTTD/OpenTTD] andythenorth closed issue #8705: Vehicle list windows fail to update profit display whilst open, after 584df54 https://git.io/JtSd4
17:09:54 <andythenorth> 1.5 hours of bisecting could have been saved
17:10:02 <andythenorth> today is a low productivity openttd day :P
17:10:10 <TrueBrain> but you wouldn't have such fun!
17:11:22 <andythenorth> 2 hours of making newgrf changes using invalid data
17:11:29 <andythenorth> 1.5 hours of bisect due to not checking open bugs :D
17:13:04 <LordAro> and i'd even previously debugged it for you
17:15:30 <andythenorth> these are the things
17:16:01 <andythenorth> at least it wasn't on a livestream
17:16:07 <andythenorth> worst livestream content ever
17:18:47 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8700: Change: Recolour windows for consistency https://git.io/JtSbO
17:21:13 <DorpsGek> [OpenTTD/OpenTTD] andythenorth commented on issue #8575: Profit for each vehicle in vehicle lists aren't being updated regularly https://git.io/JtUMl
17:22:22 * andythenorth has forgotten what the original goal was :D
17:36:34 *** WormnestAndroid has quit IRC (Read error: No route to host)
17:37:09 <_dp_> is a good idea to remove global goals from company goals?
17:37:48 <_dp_> I haven't seen a signle gs that would use both for their indended purpose
17:38:21 <_dp_> or, actually, I could only find one that uses global goals for global goals
17:38:31 <_dp_> all others use them for company league table :p
17:39:54 <_dp_> Only downside I see is it gets a bit more confusing for newbies that don't know there are both global and company goals
17:40:28 *** WormnestAndroid has joined #openttd
17:41:52 <DorpsGek> [OpenTTD/nml] FLHerne updated pull request #189: Change: Check that user code doesn't try to use reserved registers https://git.io/JtS92
17:42:33 <FLHerne> frosch123: ^ included 0xFF
17:43:05 <michi_cc> I might have been bad and also posted in that thread.
17:45:50 <DorpsGek> [OpenTTD/nml] frosch123 commented on pull request #189: Change: Check that user code doesn't try to use reserved registers https://git.io/JtSN4
17:45:54 <frosch123> FLHerne: sorry, did not see this the first time
17:46:49 <andythenorth> so...YACD-Dist revival when?
17:46:58 <andythenorth> now the game is fast...we can make it...better?
17:47:50 <FLHerne> frosch123: Huh, I realised it would apply to perm storage, and then apparently misread the docs
17:54:24 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8700: Change: Recolour windows for consistency https://git.io/JtSAk
17:56:42 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
17:57:10 *** WormnestAndroid has joined #openttd
17:57:45 *** _2TallTyler has quit IRC (Quit: Page closed)
17:57:55 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #8701: Codechange: rename sound ids to make more sense. https://git.io/JtSAO
18:01:07 <DorpsGek> [OpenTTD/OpenTTD] frosch123 merged pull request #8701: Codechange: rename sound ids to make more sense. https://git.io/JtS3o
18:02:33 *** arikover has joined #openttd
18:03:12 <TrueBrain> I see what you did there frosch123 :P
18:04:35 <frosch123> now everyone thinks i did something bad :(
18:04:56 <DorpsGek> [OpenTTD/nml] FLHerne updated pull request #189: Change: Check that user code doesn't try to use reserved registers https://git.io/JtS92
18:06:36 <DorpsGek> [OpenTTD/nml] frosch123 approved pull request #189: Change: Check that user code doesn't try to use reserved registers https://git.io/JtSAb
18:13:59 <DorpsGek> [OpenTTD/OpenTTD] stormcone commented on pull request #8700: Change: Recolour windows for consistency https://git.io/JtSxO
18:15:24 <DorpsGek> [OpenTTD/nml] FLHerne merged pull request #189: Change: Check that user code doesn't try to use reserved registers https://git.io/JtS92
18:18:41 <andythenorth> is there a vehicle var 'time since last station'?
18:18:49 * andythenorth still trying to make restaurant cars a thing
18:20:48 <andythenorth> 80+ seems to have Cargo transit time, in +185 ticks (~2.5 days) units
18:21:10 <frosch123> that's since loading
18:21:15 <frosch123> there is also time since service :p
18:21:38 <frosch123> if you define "servicing" as "restocking food"...
18:22:46 * andythenorth considering restaurant car that reduces running costs, but only if route is long enough
18:22:59 <andythenorth> nobody buys a sandwich on the train for 10 tile journey
18:23:35 <andythenorth> vehicles-affect-rating would still be the most useful / logical thing to do
18:23:38 <andythenorth> but is that just boring?
18:27:00 *** otetede has joined #openttd
18:32:13 *** Progman has joined #openttd
18:32:44 <frosch123> you could reduce running cost when servicing often
18:32:56 <frosch123> it's cheaper to run the refridgerator
18:33:17 <frosch123> and if the impact is significant enough, it may incentivate to build servicing track layouts
18:33:41 <frosch123> maybe if the food is stored too long, passenger are poisoned, and you need to pay for them?
18:34:08 <andythenorth> we need event-driven arbitrary income cb
18:34:30 <andythenorth> events: servicing, unload, load, crash
18:34:34 <andythenorth> what did I miss? :P
18:34:44 <TrueBrain> to suggest this 15 years ago :P
18:35:00 <andythenorth> ideally, everything about payment and rating would be different :P
18:35:34 <andythenorth> as it is not ideally, what lolz can we do?
18:35:44 <andythenorth> nothing is 'ideally'
18:36:11 <andythenorth> frosch123 I like the servicing idea, but then I'd have to actually service trains :D
18:42:54 *** otetede has quit IRC (Quit: Nettalk6 - www.ntalk.de)
19:00:53 <DorpsGek> - Update: Translations from eints (by translators)
19:10:35 *** Tirili has quit IRC (Ping timeout: 480 seconds)
19:12:08 <frosch123> i pressed a wrong button :p
19:12:57 <frosch123> TrueBrain: i missed the deploy button, and pressed "ready for review", not sure i can change it back :p
19:13:06 <frosch123> no idea why i was allowed to press it in the first plae
19:13:19 <TrueBrain> you pressed a button where?
19:13:30 <TrueBrain> on a PR, from draft?
19:13:33 <frosch123> ah, i found the revert button
19:13:36 <TrueBrain> context is missing, so I am guessing :)
19:13:54 <frosch123> i changed 8693 from draft->ready->draft
19:13:55 *** jottyfan has joined #openttd
19:14:00 <TrueBrain> yeah, to draft is hdiden
19:14:04 <frosch123> but the "ready->draft" button was in a completely different place
19:17:13 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8693: Feature: allow changing the game-speed other than 1x and warp-speed https://git.io/JtSjO
19:18:19 <TrueBrain> frosch123: currently I did the reverse: CTRL+CLICK enables/disables FF
19:18:24 <TrueBrain> otherwise people will never find the window honestly
19:18:46 <frosch123> added a second suggestion
19:19:04 <TrueBrain> you are not the first to suggest that :)
19:19:12 <TrueBrain> the short summary :)
19:19:16 <TrueBrain> it was easier to implement a dropdown
19:19:22 <TrueBrain> but it is really crappy without new icons
19:19:35 <TrueBrain> you have no real visual feedback of the speed the game is trying to run on, was my main issue
19:19:55 <TrueBrain> but I have some other ideas for the GUI I am going to try .. see if they feel a bit better :)
19:20:25 <TrueBrain> will be a lot of trial and error :D
19:21:33 <frosch123> i don't understand your "multiplayer" TODO item. do you want to allow changing speed in multiplayer?
19:22:14 <frosch123> non-dedicated server only?
19:22:25 <frosch123> i mean: who is allowed to change speed?
19:22:25 <TrueBrain> no, I had a dedicated server in mind for that
19:22:40 <TrueBrain> could be like pause
19:22:44 <TrueBrain> but I guess server-owner for now
19:23:00 <andythenorth> how would I normalise a cargo age malus to vehicle speed?
19:23:08 <andythenorth> malus unit is days
19:23:20 <andythenorth> vehicles have different speeds :P
19:23:37 <andythenorth> so vehicle stats progression tends to eliminate the malus :P
19:24:02 <TrueBrain> frosch123: I am really just experimenting with this, see if anything sticks, and is useful
19:24:25 <TrueBrain> so I was thinking of making "set game_speed 25" possible, so the game would run 1/4th of the speed
19:24:39 <TrueBrain> no clue if it works or if it is even enjoyable
19:25:16 <frosch123> it's as useful as youtube slow playing
19:25:30 <frosch123> you use it to look at details, to stop trains in specific positions, debugging...
19:25:45 <frosch123> for regular playing the vehicle movement will look ugly :)
19:26:01 <frosch123> 7.5 fps vehicle movement is ugly :)
19:26:16 <TrueBrain> something I also would like to fix :)
19:26:20 <TrueBrain> but .. something something .. 8bpp :P
19:26:39 <TrueBrain> but mainly, I want to find a nice looking GUI for game-speed
19:26:46 <TrueBrain> and I have something in mind, but I have to custom draw it
19:26:55 <TrueBrain> so I keep postponing starting with it :P
19:26:57 <TrueBrain> which is really funny :D
19:27:02 <frosch123> many games just have two hotkeys: slower, faster
19:27:29 <TrueBrain> I know various of ways games do it .. from <tab toggles speed> betwen 1x, 2x, 4x and 8x
19:27:42 <TrueBrain> to a debug-setting only
19:29:23 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8700: Change: Recolour windows for consistency https://git.io/JtSjo
19:29:51 <andythenorth> Warcraft 1 just has increments with arrows, very 1995 GUI
19:30:20 <TrueBrain> (with a vertical line on that horizontal)
19:30:50 <frosch123> make it flashing red when it cannot keep up
19:30:56 <frosch123> show cpu temperature
19:31:10 <TrueBrain> the first suggestion, some visual feedback, is good .. but after that you are pushing it :P
19:32:56 <TrueBrain> I will also push the dropdown solution btw, so people can look at that too (well, not via a PR, but I will figure out something)
19:33:04 <TrueBrain> and if anyone has any other ideas, I am all ears :)
19:34:57 <frosch123> acoustic feedback! play a sample of some ship diesel under different loads
19:35:25 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #8707: Codechange: [OSX] Only keep a total dirty rect for drawing. https://git.io/Jt9eI
19:35:30 <TrueBrain> sorry, I should have said: "useful ideas" :)
19:35:39 <TrueBrain> michi_cc: I was wondering about that ... :D
19:35:40 <michi_cc> andythenorth: ^^^^^^ This PR might or might not do anything for you.
19:35:54 <michi_cc> Damn, wrong PR title :)
19:36:08 <andythenorth> I will try shortly, in the middle of cargo aging tests :)
19:36:45 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9eO
19:37:21 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9ec
19:38:43 <TrueBrain> frosch123: would be fun if the simulation cannot keep up, you here explosion sound
19:39:35 <frosch123> there are wind and heavy-wind samples
19:39:56 <TrueBrain> anyway, just to check, do you like this idea at all? Or you think it is stupid? :)
19:40:22 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9ea
19:40:25 <frosch123> it really depend on the final gui
19:40:50 <frosch123> it's like transparency options: most of the time you only press "x" to toggle your common presets
19:41:05 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9ei
19:41:08 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9eP
19:41:08 <frosch123> i don't expect people to change FF speed all the time. rather toggle their usual preset
19:41:24 <TrueBrain> good point, something to keep in mind
19:42:10 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9e1
19:43:03 <frosch123> you could call it "toolbar", like WC_TRANSPARENCY_TOOLBAR
19:43:57 *** _2TallTyler has joined #openttd
19:43:58 <frosch123> then you can also make it DARK_GREEN (i already forgot which colour you used)
19:44:40 <TrueBrain> same as all the other toolbars
19:44:46 <TrueBrain> as I like things being the same :)
19:44:56 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9eN
19:45:32 <TrueBrain> michi_cc: I don't understand the framerate measure thing, so I am just a bit scared :) I was fiddling a bit with it with SDL, as it also counts the time when there is nothing to do
19:45:37 <TrueBrain> giving a much lower value than it really should
19:45:44 <TrueBrain> but I found out I don't really grasp what it is doing exactly :P
19:45:49 <TrueBrain> hence the annoying questions :)
19:45:58 <TrueBrain> I also do not know what we want the value to show
19:46:07 <TrueBrain> like .. "worst" time in a second?
19:46:42 <andythenorth> michi_cc any particular way to test this PR?
19:46:53 <DorpsGek> [OpenTTD/OpenTTD] nielsmh commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9ve
19:47:12 <andythenorth> the game is already fast enough again on mac, I don't trust subjective 'feels faster' any more
19:47:24 <michi_cc> andythenorth: Run it. And see if cursor movements feels more immediate.
19:47:46 <andythenorth> cursor is already instant everywhere I try
19:47:56 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9vT
19:49:06 <michi_cc> andythenorth: It's entirely possible that OSX controls physical screen updates anyway. Then you'd not see a difference.
19:49:42 <TrueBrain> so with your PR all video drivers are now driven by us, the video output .. that is pretty sweet :)
19:50:28 <andythenorth> is there any possibility that this makes FFWD 2x as fast as a few days ago michi_cc ?
19:50:32 <andythenorth> otherwise something else has
19:50:43 <andythenorth> I benchmarked a specific save in multiple areas of the map
19:51:07 <michi_cc> I don't think so. It might have been TB's earlier driver refactor.
19:51:32 <TrueBrain> less dirty-draws might be faster?
19:51:37 <andythenorth> yeah I loaded the wrong save
19:51:43 <andythenorth> was the wrong game date
19:52:08 <andythenorth> numbers are back where I expected
19:54:21 <DorpsGek> [OpenTTD/OpenTTD] andythenorth commented on pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9vz
19:56:19 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9eI
19:57:09 <TrueBrain> I am so happy you don't mind fixing my OCD :D
19:57:50 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9vd
20:03:27 <andythenorth> now that we have lots of performance, we can have YACD back, right? :P
20:05:54 <TrueBrain> yet-another-collosal-deathwish?
20:06:40 *** _2TallTyler has quit IRC (Quit: Page closed)
20:14:48 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #8707: [OSX] Draw immediately instead of waiting for the event loop. https://git.io/Jt9eI
20:14:56 <TrueBrain> really cool michi_cc :D
20:25:53 *** jottyfan has quit IRC (Quit: jottyfan)
20:41:54 <supermop_Home> alright, work more on log flume?
20:43:07 <andythenorth> now that you fixed ogfx
20:44:26 <supermop_Home> the side boards of the flume could use a bit more love
20:44:57 <supermop_Home> mostly need to get the log offsets dialed in, but hard to do when one cant see the log
20:45:19 <supermop_Home> then add fun splashes
20:46:21 <supermop_Home> then figure out how to apply same abuses to make a nice Alweg
20:48:59 <supermop_Home> killing time while work computer runs batches of renders
20:54:18 <supermop_Home> temperate ogfx hotel also needs help
20:58:55 <glx> TrueBrain: seems there's a missing "override" in sdl2_v.h
20:59:12 <TrueBrain> didn't I fix that already?
20:59:19 <TrueBrain> I remember I was like, yes, I knew that, let me fix that
20:59:24 <TrueBrain> boy, I am getting old :(
20:59:56 <glx> well latest CI run reports it
21:00:12 <andythenorth> hmm special pax cargos
21:01:11 <TrueBrain> I am not doubting the CI nor you glx :)
21:01:22 <supermop_Home> andythenorth foamers?
21:01:33 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #8708: Fix ec1dd0bf: missing override causing compiler warnings https://git.io/Jt9Tf
21:03:30 <DorpsGek> [OpenTTD/OpenTTD] glx22 approved pull request #8708: Fix ec1dd0bf: missing override causing compiler warnings https://git.io/Jt9TL
21:04:24 <TrueBrain> I really cannot believe I didn't fix that .. owh well
21:05:10 <LordAro> i even told you about it and everything
21:05:53 <TrueBrain> and I even acknowledged it!
21:05:55 <TrueBrain> that is the worst part
21:06:01 <TrueBrain> no clue what happened between that and ... well .. merge
21:06:09 <TrueBrain> I really am getting old :'(
21:06:23 <andythenorth> same reason I bisected for a known bug
21:06:37 * LordAro ponders fixing that bug
21:06:52 <TrueBrain> don't do anythingfdrastic please
21:11:17 <andythenorth> restaurant car randomly reduces run cost of other vehicles in train
21:11:30 <andythenorth> to reflect that people may or may not eat a lot of pies
21:11:40 <andythenorth> the main thing is this absolves me from detailed testing :)
21:13:00 <TrueBrain> glx: did you see that windows-2016, vcpkg cache step fails?
21:13:34 <glx> hmm wrong tar used again ?
21:13:47 <TrueBrain> another one for sure
21:13:52 <TrueBrain> but the error is odd as fuck :P
21:14:04 <TrueBrain> it uses the GNU tar, which windows-latest doesn't
21:14:30 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8708: Fix ec1dd0bf: missing override causing compiler warnings https://git.io/Jt9Tf
21:14:46 <TrueBrain> I do kinda agree with Lukka btw that we should stop using these custom build-steps, and just use cmake --build
21:14:52 <TrueBrain> which does all the MSVC stuff for us
21:14:55 <TrueBrain> doesn't solve vcpkg btw
21:15:08 <TrueBrain> but okay .. bit done with GHA for a while .. nothing is broken, so who cares really :)
21:17:58 <glx> indeed very weird failure: /usr/bin/tar: c\:\vcpkg: Cannot stat: No such file or directory
21:18:20 <TrueBrain> I think the GHA expects Windows tar
21:18:31 <glx> oh it's because windows-2019 has tar
21:18:45 <glx> but not previous versions
21:19:11 * andythenorth quits playing tanks
21:19:28 <andythenorth> when the Definitely Balanced Matchmaker gives you teams of 40% players
21:19:46 <LordAro> andythenorth: didn't have anyone else to carry you?
21:20:06 <LordAro> (that's me playing tanks)
21:20:19 <andythenorth> every loss is everyone else's fault
21:20:35 <LordAro> havne't played WoT in quite some months, maybe i should give it a go
21:20:40 <andythenorth> I only play Blitz
21:20:56 <andythenorth> hmm, I'll need some random bits on reload
21:21:02 <andythenorth> frosch123 random_switch is 'no'?
21:23:00 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
21:23:38 *** WormnestAndroid has joined #openttd
21:32:10 <andythenorth> are RV speed units half?
21:32:22 <TrueBrain> there was something odd with RVs
21:32:38 <andythenorth> RVs are speed in mph * 3.2
21:32:51 <TrueBrain> there it is done too :)
21:33:15 <glx> I can try to move away from run-cmake
21:33:30 <TrueBrain> I found cmake --build . to work excellent
21:33:35 <TrueBrain> guess we should do that on all platforms tbh
21:33:41 <TrueBrain> instead of "make" and all the other types
21:33:44 <andythenorth> hmm can I multiply a running cost by random 1-10% in nml?
21:33:46 <TrueBrain> put CMake in control :)
21:33:47 <michi_cc> Indeed, RV speed is mph*3.2 while the bridge newgrf prop seems to fixed to mph*1.6
21:33:50 <andythenorth> or do I have to do it all on returning results?
21:34:10 <glx> but run-vcpkg simplifies caching
21:34:12 <andythenorth> and how DO random triggers work? :P
21:36:42 <LordAro> hmm, there are a lot of /2 surrounding roadtypeinfo max_speed
21:38:31 <frosch123> rv and ships have a factor /2
21:38:47 <FLHerne> Just how many recolour_sprites can one thing need?
21:38:55 <frosch123> aircraft used to have a different factor, but now use the train unit
21:39:31 <andythenorth> FLHerne funny you say that
21:39:40 <andythenorth> Horse is relatively conservative
21:39:43 <frosch123> andythenorth: nml does not help much with random_actions :)
21:39:55 <andythenorth> FLHerne I don't do cargo colouring with recolour_sprites
21:40:10 <andythenorth> FLHerne I'm always scared of them :P
21:40:22 <frosch123> we need to change nml to provide a better random_switch
21:40:31 <andythenorth> I want to randomly reduce run costs for the whole consist if it has a restaurant car
21:40:46 <andythenorth> I'll probably need to pick the random bits out of the restaurant car vehicle with var 0x61
21:40:49 <frosch123> what part of that is "random"?
21:40:53 <supermop_Home> I've put a tiny little pink spotlight over each cocktail in this rendering to make them look more appetizing
21:41:01 <andythenorth> it should vary on load, or reload or something
21:41:05 <andythenorth> or 32 day callback
21:41:16 <supermop_Home> digital cocktails better than meatspace cocktails?
21:41:24 <andythenorth> supermop_Home unlikely
21:41:48 <andythenorth> frosch reduction would be 1-25% or something
21:41:51 <frosch123> andythenorth: the front engine can check whether a restaurant car is present (if you have a unique id for it)
21:41:56 <andythenorth> yup, got that part
21:41:57 <FLHerne> But most players will have multiple trains with restaurant cars, so your random factor will average out and be meaningless
21:42:01 <frosch123> it does not matter which random bits you use
21:42:18 <FLHerne> (of course, most players will never notice even if they have one)
21:42:20 <andythenorth> there are other things that might be using triggers already
21:42:22 <frosch123> why do you want to use those of the restauarant car? are they more random than other bits?
21:42:57 <andythenorth> I've had unwanted behaviour from triggers interacting 'unexpectedly'
21:43:01 <andythenorth> unexpected to me :P
21:43:19 <frosch123> triggers do not trigger when the random_switch is a callback
21:43:21 <andythenorth> I wanted to use some that are in safe sandbox
21:43:23 <frosch123> that's their main problem
21:44:02 <frosch123> if you want to use triggers, yuo have to insert a dummy random_switch at the top of the graphics chain
21:44:12 <frosch123> give it the bits it shall rerandomise
21:44:33 <frosch123> and then only define one case to continue unconditionally
21:44:59 <frosch123> then add a random_switch without triggers into the callback, and somehow get nml to make them use the same random bits
21:45:11 <andythenorth> sounds simple :D
21:45:13 <frosch123> probably by adding a "dependant" thing
21:45:31 <frosch123> it would be nicer if nml would do that magic :)
21:45:49 <andythenorth> how do I specify which bits to rerandomise? I am overlooking that somewhere in the wiki
21:46:10 <frosch123> it rerandomised those bits, which the switch uses :p
21:46:20 <frosch123> so it depends on which bits the switch uses
21:46:53 <andythenorth> I have a feeling that is magically abstracted for me :P
21:47:11 <andythenorth> I don't really understand random switch, I've been just picking bits out manually when I need them
21:47:58 <andythenorth> hmm, the actual feature could be random per vehicle, doesn't really matter
21:48:03 <andythenorth> player can't see run cost anyway
21:48:14 <TrueBrain> owh, this reply on "the forum thread" is hilarious :)
21:48:15 <andythenorth> could just use the vehicle's bits
21:48:27 <andythenorth> I don't really care what they are
21:48:33 <andythenorth> or if they get retriggered
21:48:54 <andythenorth> this whole restaurant car feature...it's not trolling exactly...but nobody should be looking too closely at the gameplay value P
21:50:02 *** jottyfan has joined #openttd
21:50:15 <andythenorth> TrueBrain so are there some giant pictures there proving that max map height didn't work as expected?
21:50:57 <andythenorth> seems to be proving that map gen was broken in 1.10.3
21:51:09 <TrueBrain> it proofs people have no clue what max-height does, yes :) And they are not wrong to not understand :P
21:51:26 <FLHerne> andythenorth: I still have no idea why you think it should be random
21:51:39 <andythenorth> FLHerne which 'it'? :)
21:51:54 <FLHerne> It takes a feature that no-one will notice, and turns it into a feature that no-one will notice and also makes everything even more complicated
21:52:09 <FLHerne> Your restaurant bonus
21:52:12 <LordAro> road vehicles appear to always take the shorter route, regardless of bridge speed
21:52:21 <andythenorth> FLHerne making it random means I don't have to set up a dedicated test game, and test 1900, 1930, 1960, 1990, 2020
21:52:22 <FLHerne> LordAro: There's a bug for that
21:52:28 <andythenorth> over different lengths of route
21:52:33 <LordAro> FLHerne: no, that bug is for roadtypes
21:52:38 <andythenorth> random is random
21:52:39 <LordAro> this is standard roads
21:53:04 <LordAro> and the code (linked above) suggests that it should be accounted for
21:53:22 <FLHerne> LordAro: Eh, I guessed if the pathfinder did speed limits before someone would have put the roadtypes into it
21:53:22 <supermop_Home> LordAro maybe the pathfinder knows the slow bridge is the more scenic route
21:53:29 <FLHerne> Maybe that isn't how it doesn't work :p
21:53:40 <LordAro> FLHerne: that's the roadtypes bit, yes
21:53:50 <LordAro> but this is (likely) pre-roadtypes
21:53:54 *** Samu has quit IRC (Quit: Leaving)
21:53:58 <andythenorth> TrueBrain that Alpinist picture though - that's not a defensible outcome IMHO :)
21:54:21 <andythenorth> could just be very selective screenshot, I'll test
21:54:29 <TrueBrain> I am happy to see the terrain generator now does what I expect it to do :)
21:54:46 <FLHerne> andythenorth: I *really* don't understand how randomizing the effect means you have to test *less* :-/
21:55:00 <TrueBrain> but I love the map with 70 max height and 80 snow
21:55:07 <TrueBrain> it shows once again how silly the max-height setting is :)
21:55:13 <andythenorth> FLHerne don't have to 'balance'
21:55:20 *** m811 has quit IRC (Read error: Connection reset by peer)
21:55:27 <TrueBrain> I still kinda want to make it do what people expect it to do, but it will be a rabbit hole weeks deep :)
21:56:14 <FLHerne> andythenorth: But with any appreciable number of vehicles it'll average out to some constant value anyway, and that doesn't need balancing less than an actual constant?
21:56:20 <supermop_Home> FLHerne I think his thinking is 'if it is always wrong, I never have to check that it is right'
21:56:31 <FLHerne> While being harder to test because statistics
21:56:37 <andythenorth> you are both correct
21:57:08 <andythenorth> it will average to 10! / 10 or something if it's 1-10%?
21:57:11 * FLHerne flails hands helplessly
21:57:13 * andythenorth not very statistics
21:57:30 <supermop_Home> LordAro that temperate hotel... :'(
21:58:08 <andythenorth> Alpinist with low snowline is definitely sub-optimum
21:58:16 <TrueBrain> what makes me giggle most about the forum thread, honestly, of all the changes we accepted in the last 3 months ... THIS is the one that causes a scene :)
21:58:19 <andythenorth> it's not unplayable
21:58:19 <FLHerne> andythenorth: It'll average to 5.5%
21:58:26 <TrueBrain> I would have given 20 other PRs I merged I expected this to happen with
21:58:30 <TrueBrain> but not this one :D
21:58:34 <andythenorth> TrueBrain this is why this kind of stuff doesn't get touched
21:58:36 <TrueBrain> "This is the sword I die on" :D
21:58:40 *** arikover has quit IRC (Ping timeout: 480 seconds)
21:58:41 <supermop_Home> TrueBrain quick change lots of other stuff while people are distracted!
21:58:50 <TrueBrain> supermop_Home: sssttt, I already did :P
21:58:51 <andythenorth> most devs here are still recovering from 'you can no longer flip vehicles in depot by default'
21:58:57 <supermop_Home> time to remove ships
21:58:57 <andythenorth> which was like 2008 or something
21:58:57 <FLHerne> andythenorth: Assuming even distribution
21:59:10 <supermop_Home> no one will notice til 1.14 or something
21:59:15 <FLHerne> Which, given how grfs do randomness, is a stupid and wrong assumption
21:59:23 <TrueBrain> well, though love for LC, but I stand by the choice we made, and I will die on that sword too :P
21:59:56 <FLHerne> supermop_Home: All my networks will break :-(
22:00:38 <supermop_Home> "we changed the subarctic generator back to 1.10.3, but now all subarctic grass is pink"
22:00:50 <andythenorth> FLHerne make a counter proposal?
22:01:01 <andythenorth> because digging through all the random triggers looks less and less appetitising
22:01:22 <andythenorth> I was going to do % cost reduction on % loaded
22:01:31 <supermop_Home> andythenorth some station rating hack probably the best?
22:01:33 <andythenorth> because more passengers = more eccles cakes sold
22:01:37 <andythenorth> supermop_Home not possible
22:01:40 <supermop_Home> andythenorth that will desync
22:01:48 <supermop_Home> based on my experience with trams
22:02:07 <supermop_Home> changed cargo decay based on % loaded
22:02:21 <andythenorth> I do loading sprites on % loaded often
22:02:41 <supermop_Home> idk I never play MP but changing the cargo decay stat while out of depot caused it apparently
22:03:14 <supermop_Home> it seems like it shouldn't but I got a handful of reports of it
22:03:36 <andythenorth> worth knowing, because I was about to do just that
22:03:50 <supermop_Home> maybe its a bug?
22:03:55 <Timberwolf> Interesting. I do just that, maybe I should test an MP game.
22:04:00 <andythenorth> TrueBrain it's bad form to quote forum posts behind people's back, but this is the funniest "without respecting other styles of play than one's own"
22:04:09 <andythenorth> given how little OpenTTD devs mostly play
22:04:20 <Timberwolf> (Only for the preservation railways)
22:04:55 <andythenorth> FLHerne so I can just take the % reduction as the static value that the random would have averaged to? :D
22:05:05 <andythenorth> saving a lot of reading about random :)
22:06:56 <supermop_Home> the switch was inspired by Melbourne where E class trams (and B class) can be a comfy light rail train to ride way out into the suburban hinterlands off peak, but are miserable at crush load
22:07:11 <andythenorth> it's a good idea :D
22:07:23 <andythenorth> if only cargo_age_period wasn't so crappy to test and balance
22:07:42 <FLHerne> andythenorth: Yes, that was what I meant
22:07:45 <supermop_Home> and its realistic, because I observed a lot more people "forgetting" to tap on with their farecards which it was full and standing
22:08:04 <FLHerne> If it's randomised per-train, anyone with more than a handful will get the same result either way
22:08:19 <supermop_Home> randomized per month
22:08:27 <andythenorth> hmm is there a floor under cargo aging? I think there is
22:08:33 <andythenorth> yes the chart shows that
22:08:40 <supermop_Home> the menu in some months being more appetizing than others
22:09:40 <andythenorth> "the time component will never be lower than 31"
22:09:57 * andythenorth wonders if the aging gets reset when transferring?
22:10:20 <andythenorth> and if cargo is aged on stations
22:11:13 <Timberwolf> If you're one of the first people on the train at rush hour, it's excellent.
22:11:16 <Timberwolf> Otherwise... it's not.
22:11:54 <andythenorth> wow so not commuter
22:12:30 <Timberwolf> They were very end-of-life by that point.
22:12:36 <supermop_Home> I recall even in 2012 some subsurface stock on circle or something that had essentially what looked like compartments to me
22:12:43 <Timberwolf> Every so often you get a seat that's completely collapsed.
22:13:05 <Timberwolf> supermop_Home: The really old stuff had some lateral seating.
22:13:30 <Timberwolf> I think a few units with wooden floors survived as late as 2010-ish.
22:13:33 <supermop_Home> this had little partitions between the 'booths'
22:13:57 <supermop_Home> yeah I seen those either in 2011 or 2012
22:15:24 <supermop_Home> contrast with hikari rail star in japan - when a 700 series shinkansen get cascaded to a more local / commuter level of service, they rip out the interior and replace with a more spartan one
22:16:02 <Timberwolf> That's more normal here. South West Trains completely refurbished all the 455s/456s that got cascaded to them.
22:16:30 <DorpsGek> [OpenTTD/OpenTTD] ldpl opened pull request #8709: Change: Don't show global goals in company goal windows https://git.io/Jt9tF
22:17:25 <supermop_Home> the rail star is still a shinkansen tho, so here I am with my Green Car rail pass stuck in essentially cattle class from Kobe to Fukuoka for hours as the train has no green car...
22:17:26 <TrueBrain> right, lets see if this reply helps calming LC down a bit :)
22:18:37 <DorpsGek> [OpenTTD/OpenTTD] nielsmh commented on pull request #8709: Change: Don't show global goals in company goal windows https://git.io/Jt9qJ
22:18:43 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8709: Change: Don't show global goals in company goal windows https://git.io/Jt9qT
22:18:57 <TrueBrain> same thought, different solution :)
22:20:39 <_dp_> I was thinking to just add that line in the gamescript
22:21:14 <supermop_Home> TrueBrain just rename max height to 'clamp' or 'cutoff' or something?
22:21:23 <_dp_> Button would be nice I guess but where to put it? in the title bar?
22:21:26 <TrueBrain> or ... remove it, I know, shocker :D
22:21:31 <TrueBrain> _dp_: other windows do this
22:21:41 <TrueBrain> that is why I went there
22:21:47 <TrueBrain> I would never have done it myself :D
22:21:53 <supermop_Home> TrueBrain I am sure some maniac wants those ugly chopped mountains
22:22:12 <supermop_Home> but it definitely should not be there in the new map window
22:22:15 <TrueBrain> supermop_Home: yeah ... would that give another nice thread? :D
22:22:25 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8709: Change: Don't show global goals in company goal windows https://git.io/Jt9qs
22:23:01 <supermop_Home> "1.11 beta forces players to play with pointy mountains!!!!"
22:24:19 <_dp_> max height is kinda useful for generating super flat maps
22:24:29 <frosch123> wow, i didn't even know there is a global goals window
22:24:33 <TrueBrain> for doing the exact opposite the name suggests, yes :)
22:24:33 <_dp_> well, if min is lowered because default 15 isn't xD
22:24:41 <frosch123> i always just pressed the button, and did not check the dropdown
22:24:49 <TrueBrain> frosch123: :D I had the same :P
22:26:06 <supermop_Home> rename it "Mountain chop factor"
22:26:29 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8709: Change: Don't show global goals in company goal windows https://git.io/Jt9qR
22:27:30 <TrueBrain> supermop_Home: "hop chop this value makes mountains go flllaaaatttttt
22:27:57 * andythenorth makes TrueBrain flat
22:27:57 <TrueBrain> no, honestly, I want to do either of these two: either make max-height auto-detected from the mapgen, and/or make max-height influence the mapgen
22:28:00 <andythenorth> and then supermop_Home flat
22:28:10 <TrueBrain> but I have to look into the code more to make a more educated opinion
22:28:14 <_dp_> frosch123, I couldn't find a single gs that would use both global and company goals for their intended purpose
22:28:26 <TrueBrain> _dp_: define "intended purpse"
22:28:26 <_dp_> one uses global goals but not company
22:28:53 <_dp_> TrueBrain, that uses global goals to show global goals :p
22:29:02 <_dp_> all but one show company rankings there :p
22:29:06 <TrueBrain> yeah, just trying to avoid confusion here :)
22:29:50 <supermop_Home> Need a GS with the global goal to flatten all mountains
22:29:57 *** Tirili has quit IRC (Quit: Leaving)
22:31:41 <supermop_Home> my biggest gripe about terrain generation is the weird pyramid holes up in mountains and valleys that do not have an outlet
22:32:06 <TrueBrain> ideally we get a student that rewrite it for college credits :)
22:32:11 <supermop_Home> or when a river stops in a 6 tile lake like 4 tiles away from the ocean
22:32:28 <supermop_Home> OpenTTD LLC hires an intern?
22:35:15 <frosch123> _dp_: sure, but in that case i still only want to press the goals toolbar button to see the goals
22:35:38 <frosch123> the dropdown is useful to see goals of other companies, just like you can see vehicles of other companies
22:36:05 <frosch123> but the global goals in inconvenient to reach, for how often i would want to look at them, if a gs has global goals
22:37:10 <_dp_> frosch123, so you think it's convenient to scroll down each time to get company goals? :p
22:37:33 <_dp_> and there are way more gamescripts that use global goals for rankings rathere than actual goals
22:37:41 <frosch123> you said, no gs has both, so you never scroll
22:38:06 <_dp_> both in a sense of actual goals
22:38:14 <_dp_> plenty use global goals for other stuff
22:39:21 <frosch123> ah, now i understand those screenshots
22:39:41 <frosch123> i hate citybuilders
22:40:43 <TrueBrain> as long as you understand those numbers :P
22:40:45 <LordAro> i'm not entirely certain, but it feels like the cost for a road vehicle might be too high
22:41:01 <LordAro> pathfinder cost for a road vehicle turning a corner*
22:41:35 <LordAro> road has no speed limit, tube bridge has high speed limit, wood bridge has low speed limit
22:41:54 <LordAro> i would expect the road vehicle to take the road
22:42:01 <LordAro> but it carries on over the wooden bridge
22:42:31 <TrueBrain> it enjoys the view?
22:42:35 <LordAro> (train works as expected, takes the faster bridge)
22:42:49 <supermop_Home> TrueBrain that's what I said!
22:43:22 <TrueBrain> LordAro: turning a corner is expensive after all :)
22:43:31 <TrueBrain> from a PF point of view, the other tunnel takes 2 more tiles
22:43:37 <TrueBrain> doesn't that balance out the speed limit?
22:43:51 <TrueBrain> what if you put the wooden bridge on equal road-length?
22:44:06 <TrueBrain> (by moving the depot 1 tile to NE)
22:44:25 <TrueBrain> tunnel? bridge, lol
22:45:49 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:46:39 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:47:26 <TrueBrain> I don't know how much the wooden bridge limits the speed
22:47:30 <TrueBrain> but 2 more tiles is hefty
22:47:35 <TrueBrain> on such a small stretch
22:47:59 <LordAro> it's quite significant
22:48:47 <TrueBrain> @calc (127 - 32) * 8
22:49:15 <TrueBrain> hmm .. that seems a bit overpriced, yes
22:49:28 <TrueBrain> depending on how the penalty of speed-difference is calculated, ofc
22:49:50 <LordAro> by my extremely rough counting, it takes an extra 9s for it to cross the wooden bridge vs tube bridge (on same distance route)
22:50:07 <DorpsGek> TrueBrain: 5.984251968503937
22:50:16 <TrueBrain> 2 tiles extra count for 6
22:50:21 <TrueBrain> corners or not, that is just a bit too agressive
22:50:59 <andythenorth> hmm if restaurant cars work what else can I do?
22:51:01 <andythenorth> observation car?
22:51:31 <TrueBrain> LordAro: guess the question is, is it corners that have a too high penalty, or is it speed limit that is calculated weirdly :)
22:51:51 <LordAro> i'm pretty certain the speed limit is fine
22:51:56 <LordAro> (it's the same one as for trains)
22:52:19 <TrueBrain> I dunno how our PF works, but can you dump H-values?
22:52:31 <TrueBrain> does it have a nice place where it shows its heuristic value?
22:53:13 <TrueBrain> as that should be easy to see ofc .. 2 busses on 2 roads with different bridges
22:53:18 <TrueBrain> and check the cost of the path for both
22:54:11 <supermop_Home> andythenorth certainly bar car
22:54:27 <supermop_Home> trolley service?
22:54:30 <LordAro> still picks wooden bridge
22:54:42 <LordAro> maybe road tiles as a whole are too expensive?
22:54:59 <supermop_Home> build sandwich shop at station for passengers to take to-go
22:55:05 <andythenorth> supermop_Home station tiles :P
22:55:11 <TrueBrain> or speed limit is not addressed sufficiently :)
22:55:13 <andythenorth> a whole other transport game mechanic we lack :D
22:55:26 <TrueBrain> dumping the cost of both routes will give you an indication LordAro :)
22:56:04 <LordAro> TrueBrain: i think that's what i've got with my "segment x -> y" statements
22:56:11 <supermop_Home> slider for cleaning funding to reduce the amount stations smell like urine
22:57:19 <supermop_Home> "the ammonia cargo rating for this station is horrible because the station smells"
22:57:30 <TrueBrain> LordAro: your screenshot only shows me the cost of 1 of the paths
22:57:41 <TrueBrain> 83a5 to 8ca5 is not shown :)
22:59:21 <TrueBrain> LordAro: does that other bridge have a minimum speed?
22:59:47 <FLHerne> frosch123/LordAro: I see that POP_WORD is 'deprecated' in favour of SKIP in grfstrings.py, but in the wiki there's only POP_WORD and SKIP isn't listed
22:59:53 <FLHerne> Should I change the wiki?
22:59:55 <TrueBrain> the penalty for speed is the difference between max vehicle speed and max bridge speed
22:59:59 <LordAro> nothing has any minimum speed, afaict
23:00:21 <FLHerne> (tbh, I find POP_WORD clearer about what it does...)
23:00:29 <TrueBrain> default tile is 100 penalty
23:01:15 <TrueBrain> so every tile on the wooden bridge is as 2 normal tiles, I guess
23:01:32 <FLHerne> ^ actually I should ask Yexo, it was his commit 10 years ago :p
23:01:40 <frosch123> FLHerne: i don't know the status of nml, but in ottd all those positional things are deprecated
23:01:54 <frosch123> they make no sense in translations, when translations can reorder parameters anyway
23:02:23 <frosch123> POP_WORD is probably deprecated, because NML uses 32bit ints for everything, so words don't fit
23:02:52 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
23:03:17 <TrueBrain> int max_veh_speed = v->GetDisplayMaxSpeed(); <- "Display" always scares me a bit
23:03:18 *** WormnestAndroid has joined #openttd
23:03:22 <andythenorth> restaurant car becomes observation car if at back of train? o_O
23:03:26 * andythenorth could, code supports it
23:04:08 <frosch123> FLHerne: so my theory: nml had POP_WORD first. then copied SKIP from ottd, and then ottd dropped SKIP :p
23:04:45 <TrueBrain> LordAro: penalty on rails is a lot higher for speed difference btw
23:05:24 <FLHerne> frosch123: andy is using SKIP in at least one thing, which is why I tried to look it up in the first place
23:05:33 <frosch123> FLHerne: nml also has "str_pos". so it's okay to deprecate both SKIP and POP_WORD
23:05:44 <frosch123> assuming it works :p
23:06:56 <FLHerne> andythenorth: all the hidden features
23:07:32 <TrueBrain> LordAro: I would make 2 unconnected bridge
23:07:36 <TrueBrain> so you can really see the difference in the score
23:07:41 <TrueBrain> now it is hard to deduce what the PF did
23:08:00 <LordAro> TrueBrain: unconnected bridge? i don't follow
23:08:12 <TrueBrain> just a simple road, depot, bridge, depot
23:08:15 <LordAro> you mean without any junctions?
23:08:24 <TrueBrain> I cannot see now if it picked one bridge over the other
23:08:30 <TrueBrain> so the scores don't really indicate what it thought each segment was
23:09:38 <LordAro> trouble is, it needs a junction to do any pathfinding at all :p
23:09:48 <TrueBrain> just not with bridges
23:09:56 <TrueBrain> long long route around it
23:10:02 <TrueBrain> like via another town :P
23:10:11 <TrueBrain> road: segment_cost += 1 * (max_veh_speed - max_speed);
23:10:17 <TrueBrain> rail: extra_cost += YAPF_TILE_LENGTH * (max_veh_speed - max_speed) * (4 + tf->m_tiles_skipped) / max_veh_speed;
23:10:30 <TrueBrain> rail has a huge penalty on speedloss
23:11:15 <TrueBrain> glx: that already looks a lot less magic :D
23:11:30 <TrueBrain> some newlines and EOF trick for "cmake .." line will also help
23:11:47 <glx> not powershell friendly :)
23:11:57 <TrueBrain> switch the shell to bash :P :P
23:13:03 <TrueBrain> 100 difference in penalty, lol
23:13:10 <TrueBrain> the length of a single tile
23:13:15 <TrueBrain> no wonder you get shitty results
23:13:47 <TrueBrain> as I expected a higher difference
23:14:14 <TrueBrain> but I think if you change that penalty to "YAPF_TILE_LENGTH * (max_veh_speed - max_speed) * / max_veh_speed;", the result might already be better
23:16:46 <TrueBrain> what is "wrong" about the current code, it fully depends on in what unit "max-speed" is
23:16:52 <TrueBrain> it is not normalized towards any penalty
23:16:57 <TrueBrain> which is a bit odd, to say the least
23:17:14 <TrueBrain> means that by luck (or in this case, no luck) it competes with penalties
23:17:49 <TrueBrain> that said, the speeds you give, and given max_speed should be in ~km/h, it should have been okay-ish
23:18:32 <LordAro> the ones that come out of GetSpeedLimit are "rv units"
23:18:42 <TrueBrain> what is the value of that?
23:18:47 <LordAro> 3.2 * kmh, as per andy
23:18:56 <LordAro> (i.e. double trains, for some reason)
23:19:16 <TrueBrain> int32 max_speed = e->GetDisplayMaxSpeed(); // km-ish/h
23:19:19 <TrueBrain> guess comment is a lie
23:19:24 <TrueBrain> wouldn't be the first time :P
23:19:33 <TrueBrain> anyway, if you can dump those values around that place
23:19:38 <TrueBrain> might give us a better insight
23:20:47 <TrueBrain> (I assumed you were debugging YAPF btw :P)
23:22:11 <TrueBrain> so the bridge is 64, and the vehicle is 1216?
23:22:31 <LordAro> tubular is 1216, i think
23:22:38 <TrueBrain> ah, that makes more sense
23:23:13 <TrueBrain> debugging what "F.m_tiles_skipped" is when max_speed < max_veh_speed might also give an interesting result
23:23:30 <TrueBrain> as, and this is just a hunch, it might be that it only calculates this penalty for 1 tile, instead of all tiles
23:23:32 <LordAro> it's not touched by GetSpeedLimit, anyway
23:24:47 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
23:24:55 <LordAro> which is... something to do with the length of the tunnelbridge
23:25:15 <TrueBrain> that would fit better by the problem .. you can validate it by changing the bridge length, and trying both roads again
23:25:24 <TrueBrain> or debug the code :D
23:25:39 *** WormnestAndroid has joined #openttd
23:26:15 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
23:26:33 <TrueBrain> yeah, PF seems to skip bridges too
23:26:46 <TrueBrain> (I am doing this all via github.com btw ... playing Factorio :P)
23:27:01 <TrueBrain> so that speed penalty for road is very odd
23:27:08 <TrueBrain> but be careful, if you fix this, you limit the way people can play the game
23:27:17 <TrueBrain> just so you are aware :D :P :P :P
23:32:18 <glx> hmm I may retry to do vcpkg stuff without external action
23:32:40 <andythenorth> genuine UK Gov department logo
23:33:01 <DorpsGek> [OpenTTD/OpenTTD] ldpl updated pull request #8709: Change: Don't show global goals in company goal windows https://git.io/Jt9tF
23:35:29 <TrueBrain> LordAro: yeah, your images clearly show the speed penalty was only taken over a single tile
23:35:43 <TrueBrain> so that explains that moving your faster bridge 1 further away makes it skip it
23:49:56 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
23:50:19 *** WormnestAndroid has joined #openttd
23:58:49 *** andythenorth has quit IRC (Quit: andythenorth)
continue to next day ⏵
