IRC logs for #openttd on OFTC at 2021-02-06
⏴ go to previous day
00:01:54 <LordAro> mm, #44 & #45 are missing
00:02:16 <LordAro> that's a really odd approach to take to account flagging, to hide previous issues
00:02:27 <supermop_Home_> 44 i opened in may for my slightly improved maglev station roofs
00:02:30 <LordAro> (unless of course it's been flagged for 9 months)
00:02:39 <supermop_Home_> other people even commented on the issue
00:03:04 <glx> I think they just hide everything by default
00:03:48 <FLHerne> I remember seeing 44 before, but it's 404 now
00:04:28 <glx> ok so I reexported regression pcx, now I think I'll add try catch around pillow calls
00:04:40 <glx> to give nice error message
00:05:27 <glx> wow light mode, I switched to dark mode as soon as it was possible
00:05:42 <supermop_Home_> dark is for autocad
00:05:53 <glx> and I remember this issue
00:05:53 <supermop_Home_> and irc i guess
00:08:23 <supermop_Home_> ugh maybe i should just make a new account?
00:08:46 <glx> I won't do that personally
00:09:41 <supermop_Home_> well this is demoralizing
00:09:54 <glx> I can understand the feeling
00:10:09 <glx> but I think the support will help you
00:10:13 <LordAro> give it at least a couple of days, i'd say
00:11:05 <LordAro> it's not going anywhere ;)
00:11:21 <glx> and trying to get around it may be worse in the end
00:27:35 <supermop_Home_> i guess i’ll go make a cocktail
00:49:52 <Eddi|zuHause> <supermop_Home_> hmm GitHub says my account has been flagged <-- you apparently hit some spambot detection
00:50:32 <Eddi|zuHause> or you used a bad word. like "master and slave"
00:51:52 <supermop_Home_> i never posted any links to outside sites, and in my last post before getting flagged i wrote:
00:52:02 <supermop_Home_> “Presumably if I format them in a new png, I'll need to set out the sprite sizes and locations/offsets somewhere as well right?”
00:52:22 <supermop_Home_> and “Do you think the graphic approach is successful, or should the spaceframe members be thicker and/or denser to better fit the style of the other two stations?”
00:57:34 <Eddi|zuHause> well, nothing anyone of us can do about that
00:58:54 <Eddi|zuHause> anyway, i think making a second account is against their terms of service
01:07:58 <DorpsGek> [OpenTTD/nml] glx22 opened pull request #188: Fix broken regression PCX and enable pillow 8.1.0 with an improved error message https://git.io/Jt2uE
01:14:57 *** jottyfan has joined #openttd
01:27:41 *** Flygon_ has joined #openttd
01:34:30 *** Flygon has quit IRC (Ping timeout: 480 seconds)
01:53:26 *** arikover has quit IRC (Remote host closed the connection)
02:04:54 *** roadt_ has quit IRC (Quit: 暂离)
02:36:10 *** Flygon__ has joined #openttd
02:43:00 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
03:45:42 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:56:26 *** glx has quit IRC (Read error: Connection reset by peer)
03:58:23 *** Wormnest has quit IRC (Quit: Leaving)
04:11:40 *** Flygon_ has joined #openttd
04:18:33 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
04:19:43 *** Flygon__ has joined #openttd
04:26:30 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
04:29:10 *** Flygon_ has joined #openttd
04:36:03 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
04:48:10 *** Flygon__ has joined #openttd
04:55:00 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
05:45:55 *** snail_UES_ has quit IRC (Quit: snail_UES_)
06:06:10 *** Flygon_ has joined #openttd
06:13:00 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
06:22:41 *** Flygon__ has joined #openttd
06:29:30 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
06:34:13 *** Flygon_ has joined #openttd
06:41:03 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
07:16:21 <blelch> Hi! How do I pass a 'no_threads' parameter to the SDL driver?
07:19:52 *** andythenorth has joined #openttd
07:19:52 *** dwfreed has joined #openttd
07:20:53 *** Wuzzy has quit IRC (Quit: Wuzzy)
07:26:05 *** andythenorth has quit IRC (Quit: andythenorth)
07:31:23 *** andythenorth has joined #openttd
07:40:40 *** blelch has quit IRC (Quit: blelch)
07:49:34 <supermop_Home_> i guess i also can’t upload stuff to bananas anymore
08:34:37 <andythenorth> why do we have an explicit refit menu, instead of using vehicle -> infomation -> cargos?
08:34:50 * andythenorth is sure there are actually good reasons for this, like making refit obvious, not hidden
08:44:14 <Wolf01> Add a button in vehicle -> information -> cargos which opens the refit menu?
08:44:45 *** Progman has joined #openttd
08:47:36 *** HerzogDeXtEr has joined #openttd
08:56:09 <andythenorth> I find the refit a bit odd
08:56:28 <andythenorth> if you multiple select vehicles, you get some union of their refits which I don't understand
08:56:39 <andythenorth> not sure if it's OR or AND
08:56:42 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
08:57:11 <andythenorth> this means that it's really poor for the liveries feature
08:57:24 * andythenorth very confused :D
08:57:33 *** HerzogDeXtEr has joined #openttd
08:57:50 <Wolf01> You meant that if I select a flatbed car and a hopper I might get some refits which don't even apply?
09:00:06 <andythenorth> I'm not sure how it works
09:00:19 <andythenorth> I _could_ read the code :P
09:03:52 <erle-> Do station platforms and depots have path signals integrated or do I need to put some in front?
09:04:14 <nielsm> you need to put signals if you want signals
09:04:46 <nielsm> actually, depots do have signals integrated for exiting trains
09:05:19 <nielsm> I think it functions as a path signal if the route the train wants to take goes to a path signal first, and otherwise it works as a basic block signal
09:07:06 <nielsm> and also, if a train turns around at a station, and _only_ if it turns around, and the route it wants to take goes to a path signal, then the train will try to reserve a path, and will stop until it can reserve the path
09:07:33 <nielsm> so there are some cases where stations behave as if they had a path signal integrated, but it's safer to always build appropriate signals around stations
09:17:07 *** sla_ro|master has joined #openttd
09:17:57 <andythenorth> what about using company colour to set the livery?
09:18:03 <andythenorth> that gives 16 options
09:18:31 <andythenorth> oh wait, I did that already
09:21:04 <andythenorth> what about livery is just random?
09:21:44 <nielsm> DBset did that before, more or less, right?
09:21:54 <nielsm> or that was actually build year of loco determining livery of wagons maybe
09:22:58 <andythenorth> there was a long chat on discord about liveries
09:23:31 <andythenorth> but it goes something like "just use livery callback 19 kthxbai"
09:26:21 *** HerzogDeXtEr1 has joined #openttd
09:28:56 <nielsm> I think secondary color selecting livery would be the most reasonable, given current limitations, unless you specifically want to support trains with multiple mixed liveries
09:29:37 <andythenorth> it kinda works, it conflicts with group liveries a bit
09:29:53 <andythenorth> means you can't choose engine colour separate from wagons
09:30:04 <andythenorth> * group company colours
09:33:20 *** HerzogDeXtEr has quit IRC (Ping timeout: 480 seconds)
09:34:19 <andythenorth> ^ that's about as far as I can push this novelty trick :)
09:34:33 <andythenorth> maximum variety of 'realistic' and 'looks nice' liveries
09:45:44 *** ChanServ sets mode: +v tokai
09:46:40 <nielsm> andythenorth: how about a newgrf system where a vehicle can have a number of purchase options, one purchase option is a selection between two or more alternatives (possibly allow "between one alternative" for "checkbox" options), so in the build vehicle UI you get a new button "Customise" that gives you the purchase options
09:47:39 <nielsm> a purchase option has a title and possibly a flag for "can be retrofitted after purchase too", and then a list of alternatives. an alternative has a graphic and a text
09:48:05 <andythenorth> so more like a full featured customisation?
09:48:06 <nielsm> in the customise window showing the purchase options, there is also a preview at the bottom showing the vehicle graphic and stats that will result from the chosen alternatives for all options
09:48:29 <nielsm> and the purchase options for a vehicle would then be stored as variables available from callbacks
09:48:36 <andythenorth> I get stuck on autoreplace support
09:48:43 <nielsm> also the chosen options for a vehicle model could be remembered
09:49:33 <andythenorth> alternatives included just using a vehicle ID for each variant, and somehow hiding that in the purchase menu
09:49:52 <nielsm> hmm, possibly an alternative for options could have a "not available" flag too, which hides it from the list but reserves the id
09:50:49 <nielsm> maybe identify the options with 4CCs identifiers to allow for some standardisation
09:50:59 <nielsm> both inside and outside vehicle sets
09:51:33 <andythenorth> so far viable routes for the variations seem to be: (1) use IDs (2) just do liveries, similar to my spec draft (3) some more complete customisation layer
09:51:43 <nielsm> so e.g. an option for gearing of a loco (speed vs power) could be named GEAR internally and have a list of standard known gearings
09:51:50 <andythenorth> every time I get stuck on UI though, I have not been able to sketch a single UI design yet
09:52:14 <andythenorth> players (or discord at least) *really* like the refit subtypes UI, and want customisation to work like that
09:52:28 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
09:52:34 <andythenorth> but that relies multiple selection of vehicles, and a union of available subtypes I don't understand
09:52:52 <andythenorth> if I select vehicles A B C, how does a single menu show their customisation choices?
09:52:56 <andythenorth> it's brain breaking
09:53:46 <nielsm> different models? or same model but a number of differently configured ones?
09:53:51 <andythenorth> different models
09:54:06 <andythenorth> it's like trying to unify different schemas
09:54:07 <nielsm> just don't allow customising them then
09:54:28 <andythenorth> but allow multi-select for identical vehicles?
09:54:51 *** HerzogDeXtEr1 has quit IRC (Read error: Connection reset by peer)
09:54:53 <nielsm> possibly even say customisation is a purchase-time-only thing
09:55:12 <andythenorth> I think it's nice to be able to repaint trains without repurchasing them
09:55:22 <andythenorth> although if we had template consists, that wouldn't matter
09:55:41 *** HerzogDeXtEr has joined #openttd
09:55:53 <andythenorth> I tried the templated consists in JGRPP, it's nearly great
09:57:43 <nielsm> but okay, if options have a 4CC code and they are standardised, then if multiple vehicles of different models are selected, you could let the player reconfigure the options the vehicles have in common that are also post-purchase changeable
09:58:35 <nielsm> also only show the choices that are available for all the vehicles selected, matching on "identical string for describing the option"
09:58:43 <nielsm> (may or may not be a good idea)
09:59:18 <LordAro> blelch: -v sdl2:no_threads, i believe
09:59:30 <andythenorth> so the recurring seems to be moving the options into a new newgrf feature
09:59:38 <andythenorth> so they can be shared across vehicles
10:00:24 <blelch> LordAro: oh! so I can pass it as a run-time option to the application? Also, what will that mean? That all the rendering is done on one thread? Will that affect stability & performance?
10:00:25 * andythenorth envisages highly detailed UK trains
10:00:38 <andythenorth> which start out 'vacuum brake' then get 'dual brake' then 'air brake only'
10:00:39 <nielsm> hmm... maybe, I didn't think of making it a new Feature
10:01:19 <andythenorth> and then there can be rules "this train can leave depot if it has compatible brakes on enough vehicles"
10:01:34 <andythenorth> I am not even joking, snail has done that with some complex means
10:02:04 <LordAro> blelch: honestly i've no idea what it does
10:02:05 <andythenorth> regearing pax engines to haul freight, for lower running costs
10:02:06 <blelch> LordAro: This is what works: `-v sdl:no_threads`
10:35:33 *** tokai|noir has joined #openttd
10:35:33 *** ChanServ sets mode: +v tokai|noir
10:39:11 <TrueBrain> blelch: normally both the GameLoop and drawing the screen can run at the same time (in 2 different threads). With no_threads option enabled, they are forced to be done one after the other, in the same thread
10:39:32 <TrueBrain> so although the stability is not influence, the performance will for sure degrade
10:40:07 <blelch> TrueBrain: I see... so it will have to be looked into, as to why this setup is causing an issue on Wayland?
10:40:25 <TrueBrain> ideally, nobody has to use no_threads, indeed :)
10:41:18 <TrueBrain> so if you are up for delving into the nuance of a videodriver, please check out what is going on :P
10:41:53 <TrueBrain> or if I can reproduce it, I can help out :)
10:42:07 <TrueBrain> (I have no clue how I would make the SDL_VIDEODRIVER=wayland to work :P)
10:42:38 *** tokai has quit IRC (Ping timeout: 480 seconds)
10:55:19 <TrueBrain> they say Ubuntu 21.04 uses Wayland by default, so lets try the daily snapshot :)
10:58:39 <andythenorth> livestream today?
11:01:56 <TrueBrain> go ahead andythenorth :)
11:02:09 <TrueBrain> funny, people seem to think that if you release a game on Steam, you enable DRM for some reason :p
11:02:29 <TrueBrain> Steam seems to have a rather bad name when it comes to swallowing up games or something :P
11:02:51 <TrueBrain> gives you an idea of what is in players heads
11:06:25 <TrueBrain> I kinda wonder why this is, as many games on Steam do not have DRM and allow cross-play
11:07:14 <TrueBrain> on the other hand, we might have been more clear about this in the news I guess .. shrug open for suggestions :)
11:09:22 <andythenorth> people's gonna people
11:09:34 <andythenorth> and Steam's gonna Steam
11:10:07 <andythenorth> they'll figure it out
11:10:26 <andythenorth> maybe they could explain Steam to me when they do :)
11:10:59 <andythenorth> it's kind of funny, can't open Steam OpenTTD on my mac if any other device on my Steam account has a game running
11:11:16 <andythenorth> no 2 screen life for Steam users
11:11:16 <TrueBrain> you can with other games?
11:11:29 <andythenorth> no idea, I don't understand Steam :)
11:11:41 <andythenorth> I only got it so child #1 could get Stardew Valley
11:11:42 <Wolf01> andythenorth: time to think about a curated list of GRFs guide to put on Steam?
11:12:01 <TrueBrain> nice idea Wolf01 :) As Guide ? :)
11:12:56 <Wolf01> A guide could be enough for now
11:13:19 <TrueBrain> if you want to make one pre-release, let me know, and I can get you a key to unlock the game :) (you cannot write guides it seems without owning the game :P)
11:13:25 <Wolf01> I'm still not able to start a game with the stuff I want because I always forget something
11:14:21 <TrueBrain> awh, seems they lied ... 21.04 daily still uses X11, not Wayland
11:15:37 <milek7> isn't there selector at login screen?
11:15:50 <Wolf01> Only with non proprietary drivers
11:17:03 <Wolf01> I should set up a new ubuntu vm too
11:18:58 <TrueBrain> finally found my way into enabling it .. it is not default yet .. news was misleading :P
11:20:49 <TrueBrain> okay, targeting wayland is indeed pretty broken :)
11:21:07 <TrueBrain> but on wayland using X11 as driver works fine :P
11:22:02 <TrueBrain> the cursor does render .. but it fails in creating a window properly
11:27:26 <Eddi|zuHause> i tried wayland once, something stopped me, don't remember what
11:33:58 <TrueBrain> very odd, with wayland as SDL driver, the window is not movable, resizable, and doesn't even have a title bar
11:34:11 <TrueBrain> like it is a partial-fullscreen or something
11:35:21 <milek7> does SDL support client side decorations?
11:35:39 *** blelch has quit IRC (Quit: blelch)
11:37:02 <milek7> (on gnome, wayland applications are supposed to draw title bar themselves)
11:38:24 <Eddi|zuHause> that sounds like it should be opt-in
11:46:02 <TrueBrain> so long story short, after reading some lovely rants on their bug-tracker, SDL is not ready for Wayland yet
11:46:11 <TrueBrain> in good news, they are actively working on making it so :P
11:47:08 <TrueBrain> at least that explains one part of the problem .. now why doesn't it blitter ..
11:47:27 <TrueBrain> I am afraid I already know the answer, but lets see if that is true ..
11:47:43 *** Progman has quit IRC (Remote host closed the connection)
11:48:48 <Eddi|zuHause> hm... i'm missing an option in KMix to show programs that recently had an open sound channel. there's programs that make *palim* at unpredictable moments, and you have 1 second to hit the volume slider, before it disappears again
11:50:15 <Eddi|zuHause> there's no such option in pavucontrol either
11:51:19 <TrueBrain> Unable to make EGL context current (call to eglMakeCurrent failed, reporting an error of EGL_BAD_ACCESS)
11:51:36 <TrueBrain> I am guessing, this is because it draws from a different thread
11:53:12 <milek7> "Can I call SDL video functions from multiple threads? No, most graphics back ends are not thread-safe, so you should only call SDL video functions from the main thread of your application. "
11:55:54 <michi_cc> A GL context can only be current on one thread, so if SDL keeps the context current on the main thread, using it anywhere else won't work.
11:56:30 <TrueBrain> I think it creates the context when it creates the window
11:56:31 <michi_cc> The most simple work-around for OTTD would be to simply disable threaded drawing if SDL using wayland is detected.
11:56:48 <TrueBrain> and hurry up with the OpenGL implementation :P :D
11:57:51 <michi_cc> You can hopefully start by looking at my OSX hack job later today :P
11:59:00 <TrueBrain> I have been wondering if we want this in 1.11 already
12:06:54 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on issue #8029: Black screen on SDL2-enabled OpenTTD, Sway and SDL_VIDEODRIVER=wayland https://git.io/Jv23P
12:08:59 <Eddi|zuHause> "This is most likely also the correct fix for OpenTTD: if wayland backend is detected, disable threading." <-- i'd call this a "workaround" at best
12:11:08 *** virtualrandomnumber has joined #openttd
12:11:33 *** virtualrandomnumber has quit IRC ()
12:13:21 <TrueBrain> no,it is a bit reverse: the way draw thread is implemented in OpenTTD is a workaround ;)
12:13:55 <LordAro> TrueBrain: that redhat link seems wrong? there's no patch there and it's closed as notabug
12:14:29 <TrueBrain> LordAro: oops; fixed
12:18:09 <andythenorth> what if all vehicle properties are just editable in game
12:18:17 <andythenorth> seems to be the most flexible?
12:18:32 <andythenorth> so change speed, weight, TE, capacity, run cost, reliability
12:18:40 <andythenorth> player can just choose the ones they need
12:19:17 <andythenorth> and for the livery, they just paint the pixels they want
12:21:29 <nielsm> do it like A-train, you start with a base model from a manufacturer and can then pay for R&D to make specific improvements, and you'll have to pay prototype costs and get discounts for ordering large production series
12:23:55 <andythenorth> currently I am trying to find all the train sets that use subtype refits the 'livery' callback
12:23:59 <andythenorth> so I can ban them with FIRS 4
12:25:06 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #8648: Fix #8029: [SDL / SDL2] disable draw-thread on wayland SDL video driver https://git.io/Jt275
12:25:19 <TrueBrain> euh .. wrong commit message .. oops
12:25:39 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8648: Fix #8029: [SDL2] disable draw-thread on wayland SDL video driver https://git.io/Jt275
12:27:50 <milek7> >and with EGL you are not allowed to draw in another thread than the context was created
12:27:57 <milek7> that's not strictly true
12:28:20 <milek7> you just need to unbind context from previous thread
12:29:45 <andythenorth> if FIRS probably needs a built in trainset
12:30:11 <andythenorth> then I should add one for v4
12:30:32 <andythenorth> so do I build in OpenGFX+ Trains, or do I just merge Iron Horse into FIRS?
12:30:43 <andythenorth> I am kinda thinking 'just one grf' might be the way forward
12:31:48 <nielsm> default wagons refittable to all the new cargo types? with a setting to control whether they happen, with an "autodetect" that checks for known-compatible sets
12:32:36 <andythenorth> it would eliminate a repetitive source of player confusion
12:32:46 <andythenorth> all they need is class refits on the base wagons
12:34:24 <Eddi|zuHause> is it bad-idea-day?
12:34:38 <nielsm> actually if you do that, also include generic bugfixes like making sure some of every vehicle type remain available also after 3000 years
12:35:09 <Eddi|zuHause> include all vehicle grfs, and a parameter to select one. :p
12:35:12 <Wolf01> Bah... I lost taste again
12:35:26 <andythenorth> Eddi|zuHause what about a UI where you could choose the vehicle grf?
12:35:36 <andythenorth> we could maybe....keep them remotely and fetch them when needed?
12:35:57 <_dp_> yeye, firs mega-grf then a gs to go with it and before you know it you're making a server :p
12:36:33 <andythenorth> I can't merge a GS into FIRS
12:36:46 <andythenorth> so forget GS, just a patched server
12:36:55 <andythenorth> in fact, can I just forget newgrf and patch the base game?
12:37:35 <Eddi|zuHause> anyway, the least-worst thing you could do would be to add iron horse as dependency on bananas, so people will at least have it downloaded (they could still just not activate it)
12:37:44 <andythenorth> I think OpenTTD compiles faster than Iron Horse
13:09:33 <DorpsGek> [OpenTTD/OpenTTD] michicc approved pull request #8648: Fix #8029: [SDL2] disable draw-thread on wayland SDL video driver https://git.io/Jt2Fz
13:09:49 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8648: Fix #8029: [SDL2] disable draw-thread on wayland SDL video driver https://git.io/Jt275
13:09:52 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain closed issue #8029: Black screen on SDL2-enabled OpenTTD, Sway and SDL_VIDEODRIVER=wayland https://git.io/Jv23P
13:09:54 <michi_cc> Okay, shopping, then back to OSX mouse and key events :)
13:11:20 <michi_cc> I'm hoping to accidentally fix andy's stuck mouse, but as I can't reproduce it for me, I have no idea if my changes will do anything.
13:11:39 <TrueBrain> think of it this way: it wouldn't get worse, so ... :P
13:15:58 <michi_cc> I tried out some free trial for some cloud mac service last year. Frankly, the remote desktop experience left a LOT to desire. I don't think this is better than VMware locally.
13:16:27 <TrueBrain> I tried to install macOS on VirtualBox the other day .. that was not a pleasant experience either :P
13:16:31 <TrueBrain> have yet to try VMWare :)
13:17:18 <michi_cc> You need to obtain an "unlocker" (and maybe a prepared disk image), but then VMware just runs (on Intel, for AMD you need some config hackery).
13:17:33 <TrueBrain> yeah, for VMWare I noticed you need to binary change VMWare executables
13:17:35 <TrueBrain> that was a bit scary :P
13:17:53 <michi_cc> Normal VMware pro contains all code from Mac VMware Fusion, they just can't legally enable it.
13:19:13 <michi_cc> Use appropriate unlocker, and VMware on Intel CPU will just run. Graphics are a bit laggy as there is no accelerated display driver, but otherwise it runs quite okay.
13:19:39 <andythenorth> 'graphics are a bit laggy' sounds like my Mac
13:19:51 <andythenorth> everything runs about 0.1s behind where it should
13:23:31 *** jottyfan has joined #openttd
13:36:03 <erle-> which AI is recommended for road vehicles?
14:00:48 *** WormnestAndroid has quit IRC (Remote host closed the connection)
14:01:01 *** WormnestAndroid has joined #openttd
14:05:54 *** frosch123 has joined #openttd
14:14:35 *** ChanServ sets mode: +v tokai
14:15:34 <andythenorth> so...revert NRT, do BGT?
14:16:26 <supermop_Home_> i mean nrt does just about everything needed,
14:16:40 <supermop_Home_> other than drawing catenary sprites in the correct order
14:16:49 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
14:17:01 <glx> but catenary was broken before it seems
14:17:02 *** WormnestAndroid has joined #openttd
14:17:12 <andythenorth> nrt was a mistake, which is my fault :P
14:17:43 <glx> main issue is spec came from implementation I think
14:18:18 *** iSoSyS has quit IRC (Remote host closed the connection)
14:18:33 <andythenorth> oh that's a neat summary
14:22:43 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
14:22:43 <glx> I can see why it was done that way for stations (because ttdpatch had some limitations)
14:25:09 <supermop_Home_> glx i meant that one of the ways NRT falls short is that it doesn’t fix the catenary problem
14:25:41 <glx> but it's not the fault of NRT as I understand it
14:26:41 <supermop_Home_> i am extra bummed about GitHub thinking i am spam now that i realize it means i can’t upload updates to my grfs to bananas
14:29:20 <glx> oh of course, hidden account means you can't login to bananas
14:29:49 <glx> let's hope support will react fast
14:39:47 <DorpsGek> [OpenTTD/nml] frosch123 commented on pull request #188: Fix broken regression PCX and enable pillow 8.1.0 with an improved error message https://git.io/Jt2N5
14:46:35 <DorpsGek> [OpenTTD/nml] glx22 updated pull request #188: Fix broken regression PCX and enable pillow 8.1.0 with an improved error message https://git.io/Jt2uE
14:49:00 <DorpsGek> [OpenTTD/nml] frosch123 approved pull request #188: Fix broken regression PCX and enable pillow 8.1.0 with an improved error message https://git.io/Jt2NN
14:57:46 <DorpsGek> [OpenTTD/nml] glx22 merged pull request #188: Fix broken regression PCX and enable pillow 8.1.0 with an improved error message https://git.io/Jt2uE
15:11:11 <DorpsGek> [OpenTTD/nml] frosch123 approved pull request #179: Fix: optimising switches could change variable scope https://git.io/Jt2As
15:11:49 <frosch123> optimisation is quite complex :)
15:12:12 <glx> yeah, optimisation is reduced now
15:12:36 <glx> but at least it's less broken
15:28:03 <frosch123> yay, climbing up the rabbit hole. let's see whether i can fix access to town persistent storage :p
15:33:58 <glx> yeah #173 is quite complex to read
15:34:25 <frosch123> oh, don't read it. it's completely out of sync
15:34:43 <frosch123> there came like 5 other PR out of it
15:39:07 <frosch123> in my mind there was more :p
15:39:20 *** jottyfan has joined #openttd
15:40:19 <TrueBrain> the snow ... is coming .... the snnoowww ... is cooommminnngggg :D
15:40:43 <supermop_Home_> more snow here tomorrow
15:40:48 <glx> and sand in some parts of the country
15:40:53 <supermop_Home_> today is nice and sunny though
15:41:57 <glx> yeah high wind in the sahara means sand in southern europe :)
15:42:06 <supermop_Home_> a sand … storm?
15:42:24 <supermop_Home_> do do do do doot do do do do doot do do do
15:42:32 <TrueBrain> yeah, we get those sand storms in the summer sometimes .. really annoying ....
15:42:41 <TrueBrain> but tonight and tomorrow .. they expect 20cm of snow
15:42:44 <TrueBrain> so that will be fun :)
15:42:50 <supermop_Home_> doot do do do do do doot do do do do d
15:42:52 <glx> not a storm but weird orange/yellow sky
15:43:01 <TrueBrain> we get the sand on the cars and shit :P
15:43:05 <supermop_Home_> well you can still listen to sandstorm
15:57:22 *** jottyfan has joined #openttd
16:12:21 *** Flygon_ has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:18:51 *** tokai|noir has joined #openttd
16:18:51 *** ChanServ sets mode: +v tokai|noir
16:25:45 *** snail_UES_ has joined #openttd
16:25:48 *** tokai has quit IRC (Ping timeout: 480 seconds)
17:02:14 *** sla_ro|master has quit IRC ()
17:58:58 <supermop_Home_> andythenorth i get 96kmh tank cars around 1930, but the 120kmh lemon also comes out then?
17:59:13 <andythenorth> sounds interesting
17:59:35 <andythenorth> oh some of the 1930s freight engines are 75mph
17:59:50 <andythenorth> they're intended to last up until 1980 if you want them
17:59:57 <andythenorth> sort of BR Standard Classes
18:00:14 <andythenorth> stuff and things
18:02:30 <supermop_Home_> now time to re-timetable this mainline for 144kmh expresses
18:03:09 <supermop_Home_> and maybe cascade the 120kmh thunderers as locals
18:07:42 <andythenorth> Horse maximum micro-management
18:10:56 <supermop_Home_> that’s the whole point no?
18:11:24 <supermop_Home_> already have several million quid in 1931,
18:11:57 <supermop_Home_> but its more fun to try to reuse trains wherever possible that wholesale replace everything
18:12:31 <supermop_Home_> build little connecting chords on the outskirts of town to transfer stock from one line to another
18:14:37 <supermop_Home_> question is whether this big pacific is overkill,
18:17:42 <andythenorth> 'winning' is finding a use for the streamliner
18:18:31 <supermop_Home_> testing timings by assigning the new train to a ‘Works’ group with bright yellow livery
18:19:08 <andythenorth> RTC red and blue
18:22:36 <Timberwolf> Anyone had an `Assertion failedat line 111 of ... src\core/pool_type.hpp: index < this->first_unused` in 1.11 beta?
18:23:33 <glx> hmm without call stack it's hard to tell :)
18:24:15 <Timberwolf> Yeah, it's an annoying place for the assert to happen as it's a genuine "could be anything" place.
18:24:58 <Timberwolf> Anything useful I can do with a call stack that only has memory addresses?
18:25:38 <glx> you should have a .dmp then
18:25:44 <Timberwolf> (64-bit 10, to be specific)
18:26:29 <Timberwolf> Yep, I do indeed. Anywhere useful to load that to figure out what happened?
18:26:37 <glx> not self built version I guess, so open an issue, then we can look at it
18:26:38 <supermop_Home_> acceleration a little slow - time to test the pegasus
18:28:31 <Timberwolf> I only have one from a self-built version, a friend has it from the normal beta so let me see if I can get his crash.dmp
18:29:34 <glx> ah yes if you built your self only you can get useful info from your dmp
18:34:12 <glx> but I can explain you how to do it if you want
18:36:29 <Timberwolf> I think I've figured most of it out, but must have rebuilt things since the last crash as VS won't load the executable or symbols file.
18:37:14 *** Wormnest has joined #openttd
18:37:39 <glx> basically you need openttd.exe that generated dmp, and corresponding pdb (created when compiling openttd)
18:38:32 <andythenorth> oof I wanted to add a tile to CHIPS
18:38:40 <glx> you used release or relwithdebinfo ?
18:38:51 <andythenorth> it's all GRM and C pre-processor macros :)
18:39:31 <glx> ok so it won't be possible
18:40:37 <glx> probably explains why you only have addresses in crash.log
18:41:39 <Timberwolf> Oh well. Build with more sensible settings, hope it can be triggered again.
18:42:37 <supermop_Home_> hmm Pegasus only save 3 days over the whole trip vs strongbow
18:44:45 <glx> usually to "read" a dmp for official build, I get the zip from website, then uncompress it, then I get the matching pdb and put it next to openttd.exe (pdb must be renamed to openttd.pdb)
18:45:15 <glx> then I put the dmp from issue next to openttd.exe, then I open dmp in VS
18:45:50 <glx> of course I need to checkout the correct source files
18:46:45 <glx> but with matching exe, pdb and source, you can open the dmp and get the call stack
18:47:50 <glx> and sometimes values of variables (but that's a non reliable bonus)
19:03:18 <andythenorth> oof is today the day to learn about GRM?
19:03:37 <Timberwolf> glx: Aha, I have all the necessary magic on my desktop computer (including the same crash)
19:03:37 <andythenorth> CHIPS is so magical, there are no IDs for the tiles
19:04:01 <Timberwolf> Looks like company pool, this was on line 2433 of road_cmd.cpp (`Company * c = Company::Get(_current_company)`)
19:12:36 <Timberwolf> OK... so I think this happens when a town expands to a half-tile road which needs converting, and therefore there's not a valid company to update the infrastructure counters for (and thus tries to get from outside the pool)
19:21:06 <Timberwolf> Yep, exactly that.
19:34:12 <TrueBrain> ha, Steam auto-updated tonight \o/ :D w00p
19:51:13 <LordAro> Timberwolf: stick it in an issue or it will be forgotten
19:58:24 <DorpsGek> [OpenTTD/OpenTTD] Shemilt opened issue #8650: Assertion failed at line 111 of D:\a\OpenTTD\OpenTTD\src\core/pool_type.hpp: index < this->first_unused https://git.io/Jtakq
19:59:16 <DorpsGek> [OpenTTD/OpenTTD] mattkimber opened pull request #8651: Fix 0125892: Don't crash when towns upgrade road tiles during expansion https://git.io/Jtaks
20:00:56 <Timberwolf> LordAro: I went straight to "fix", worth an issue as well or is the PR enough?
20:02:45 <DorpsGek> [OpenTTD/OpenTTD] mattkimber commented on pull request #8651: Fix 0125892: Don't crash when towns upgrade road tiles during expansion https://git.io/Jtakz
20:04:08 <LordAro> Timberwolf: PR is enough
20:04:09 *** qwebirc51815 has joined #openttd
20:04:25 <LordAro> don't suppose you can check if it's related to #8650 though? :p
20:04:53 <Timberwolf> I'll download and see :)
20:06:59 <Timberwolf> Yep, it calls the affected code.
20:07:39 <Timberwolf> Let me double-check the crash dump to make sure.
20:08:54 <andythenorth> bananas is quite displeased with my readme.txt, changelog.txt
20:09:01 <andythenorth> "Could not recognise this file; possibly the extension is wrong?"
20:10:33 <andythenorth> just line endings or something?
20:10:45 <TrueBrain> are you sure it really is named readme.txt and it is in the root?
20:11:42 <TrueBrain> re.compile(r"(readme|changelog)(_[a-z]{2}(_[A-Z]{2})?)?\.txt$") <- you got to love regex :D
20:11:44 <andythenorth> oh it's the silly nesting
20:11:55 <andythenorth> legacy makefile crap
20:12:18 <andythenorth> tar has a folder in it
20:12:23 <andythenorth> grf is found, other files are not
20:14:50 *** ChanServ sets mode: +v tokai
20:15:47 <glx> Timberwolf: if it's the same issue you can add a reference to #8650 in the description of your PR with any of the magic closing words
20:18:38 *** qwebirc51815 has quit IRC (Remote host closed the connection)
20:19:55 <Timberwolf> Done, it does appear to be the same issue.
20:21:33 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
20:34:46 *** jottyfan has joined #openttd
20:39:22 <andythenorth> seems GRM is just counting?
20:43:08 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8651: Fix 0125892: Don't crash when towns upgrade road tiles during expansion https://git.io/JtaLj
20:43:23 <glx> really not tested, just an idea
20:45:51 <Timberwolf> glx: I'll have a look, see if there's a good reason (or not) why the original avoids that method in that one case.
20:54:34 <Timberwolf> I think it's trying to avoid the update of ownership if you're building into or converting a tile which is not currently owned. I'll see if that is the case in game.
20:56:58 <DorpsGek> [OpenTTD/OpenTTD] nielsmh approved pull request #8608: Codechange: [NewGRF] Make it more explicit which parts of the TTD veh… https://git.io/Jtatb
20:57:12 <glx> we're in case where owner == _current_company so should not change ownership
20:58:30 *** Progman has joined #openttd
21:04:14 *** tokai|noir has joined #openttd
21:04:14 *** ChanServ sets mode: +v tokai|noir
21:07:58 <DorpsGek> [OpenTTD/OpenTTD] nielsmh commented on pull request #8519: [OSX] Render screen in full native resolution on HiDPI displays. https://git.io/JtaqP
21:10:16 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8519: [OSX] Render screen in full native resolution on HiDPI displays. https://git.io/JtaqF
21:11:09 *** tokai has quit IRC (Ping timeout: 480 seconds)
21:13:48 *** erle- has quit IRC (Quit: Leaving)
21:18:31 *** welterde has quit IRC (Quit: WeeChat 2.9)
21:29:23 <andythenorth> oof let's become lawyers
21:30:10 <TrueBrain> its a nice article :)
21:30:26 <frosch123> at least the first comment points it out :p
21:30:49 <TrueBrain> yeah, that made me giggle
21:31:00 <TrueBrain> the second is also funny: but the trademark! :)
21:31:12 <nielsm> I also tweeted at them now
21:31:22 <nielsm> well, responded to their tweet of the article
21:32:03 <frosch123> the screenshot says 1.1.0 :p
21:49:14 *** jottyfan has joined #openttd
21:49:52 *** Samu has quit IRC (Quit: Leaving)
21:54:30 <michi_cc> andythenorth: I might have solved your OSX performance problems.
21:54:45 <andythenorth> I switch to Windows? :o
21:54:45 <michi_cc> Not using OpenGL, BTW.
21:54:45 <TrueBrain> don't say that unless you mean it!
21:55:02 <michi_cc> Watch this space in a few minutes.
21:55:22 <TrueBrain> and now: okay, I am off to bed, bbbyyyeeeee
21:55:51 <michi_cc> Got some cleanup first.
21:56:11 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #8518: [OSX] Replace FreeType by native glyph rasterization https://git.io/JLxK8
21:56:33 <michi_cc> Somebody please either look at that or close it ^^.
21:56:46 <LordAro> needed someone to actually test it, last i looked
21:56:53 <LordAro> if andy says yes, i'll merge it :p
21:59:57 <michi_cc> orudge wanted to test it as well.
22:05:04 <michi_cc> Let's see if the CI will yell at me :D
22:06:14 <glx> commit checker is not happy
22:06:22 <TrueBrain> that is a lot of code removal :D
22:06:26 <michi_cc> Eh, formatting problem in a code part that is completely deleted later on :p
22:07:06 <TrueBrain> and that fixes performance, you say?
22:07:26 <michi_cc> The last commit fixes performance, everything else is just for show :)
22:07:48 <michi_cc> It is entirely possible to cherry-pick that one.
22:07:53 <glx> change of API helps with performance I guess
22:08:35 <TrueBrain> lol .. that is crazy michi_cc :D
22:09:00 <michi_cc> Quartz (like most compositing window managers) renders views into a bitmap. We also render into a bitmap. Why not just give Quartz that bitmap?
22:10:04 <michi_cc> Layer-backed views where not a thing when the video driver was written.
22:10:24 <TrueBrain> and especially macOS, had a lot of driver changes over the years
22:10:34 <TrueBrain> TRY IT OUT andythenorth ! We need to know!
22:10:35 * andythenorth should learn more about macOS drawing :P
22:11:40 <andythenorth> I worked out that even running OpenTTD paused causes WindowServer to sit at 30% on this mac, usually it's about 8-10%
22:11:51 <andythenorth> ok let's try this PR
22:14:40 <andythenorth> slow linking is slow
22:15:53 <michi_cc> If it doesn't do anything for you, your Mac really is broken :)
22:16:10 <michi_cc> In VMware, fast forward goes from like 40 fps to 200 fps.
22:16:15 <andythenorth> Full Animation works in FFWD
22:16:51 <glx> I think we have validation from andythenorth :)
22:17:00 <andythenorth> it's got some issues
22:17:10 <andythenorth> but the performance is substantially increased
22:17:19 <michi_cc> I might have broken lots of things on the way :)
22:17:32 <andythenorth> moving the map with the mouse is...unreliable
22:17:34 <andythenorth> it skitters / jumps
22:18:06 <andythenorth> maybe it's just much faster than previously :P
22:18:29 <michi_cc> Can you try a diffent scrolling setting, just to check?
22:18:43 <nielsm> what if you use the keyboard to scroll?
22:18:51 <andythenorth> keyboard seems fine
22:19:02 <andythenorth> I'm just trying to find a binary of master to compare FPS with
22:19:11 <michi_cc> RMB scrolling was never very reliable/fluid for me on VMware, which is probably why I didn't see any difference.
22:20:35 <andythenorth> I'll post some numbers
22:21:30 <andythenorth> slow linker is slow :P
22:23:19 <michi_cc> Now with less lopp :)
22:25:10 <andythenorth> michi_cc same savegame, same map location, same zoom. Recent master runs about 2.8x on ffwd, full animation. 8652 is doing about 6.5x
22:26:53 <andythenorth> with full animation off, recent master is pushing around 8.5x, 8652 is doing about 12.5x
22:30:47 <andythenorth> not sure it's made much difference to WindowServer CPU use
22:31:06 <TrueBrain> still a good improvement for a few lines of code :P
22:31:52 <LordAro> other than the mouse thing
22:32:07 <andythenorth> that just looks like some scroll rate thing
22:32:14 <michi_cc> andythenorth: Can you try a scroll setting that doesn't lock the mouse?
22:32:17 <andythenorth> I assume there's some kind of 'move n'
22:33:11 <andythenorth> panning at window edges seems to work ok
22:33:42 <andythenorth> panning without locking mouse position works great
22:34:20 <andythenorth> yeah it's just the two settings with mouse lock that are skitterish
22:34:57 <andythenorth> if I move the mouse tiny amounts, it works
22:35:12 <andythenorth> is there a scroll rate var?
22:44:41 <andythenorth> can we declare winning?
22:49:23 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
22:56:29 <nielsm> I think it is yes, can't remember the reason though
23:03:08 <andythenorth> michi_cc so this PR \o/
23:03:18 <andythenorth> mouse thing will be resolvable I bet :)
23:03:36 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:09:52 *** andythenorth has quit IRC (Quit: andythenorth)
23:15:41 <DorpsGek> [OpenTTD/OpenTTD] mattkimber updated pull request #8651: Fix 0125892: Don't crash when towns upgrade road tiles during expansion https://git.io/Jtaks
23:16:33 <DorpsGek> [OpenTTD/OpenTTD] mattkimber commented on pull request #8651: Fix 0125892: Don't crash when towns upgrade road tiles during expansion https://git.io/JtaZo
23:16:45 <Timberwolf> glx: Yes, that works and is cleaner.
23:50:37 *** nielsm has quit IRC (Ping timeout: 480 seconds)
continue to next day ⏵