IRC logs for #openttd on OFTC at 2022-10-22
⏴ go to previous day
00:19:04 *** Wormnest has joined #openttd
01:41:37 *** Wormnest has quit IRC (Quit: Leaving)
01:46:03 *** milek7_ has joined #openttd
01:47:28 *** milek7 has quit IRC (Ping timeout: 480 seconds)
03:05:50 *** debdog has quit IRC (Ping timeout: 480 seconds)
05:16:33 *** WormnestAndroid has quit IRC (Remote host closed the connection)
05:55:36 *** wallabra_ has joined #openttd
05:59:28 *** wallabra has quit IRC (Ping timeout: 480 seconds)
06:02:34 *** wallabra has joined #openttd
06:05:34 *** wallabra_ has quit IRC (Ping timeout: 480 seconds)
06:07:59 <Bouke> andythenorth: Probably the “max simulation rate” setting defaulting to 2500%.
06:39:49 <Bouke> JGR: Where can I find the Wentbourne savegame? I found a dev post from April 2019 mentioning it, but no download link.
06:51:08 *** sla_ro|master has joined #openttd
06:54:36 <andythenorth> Bouke: yes, I thought I'd changed the setting, but ... mistakes were made
06:58:14 <andythenorth> ok with the correct settings, the Hilbergen save is sustained 85-90x GSF with or without hardware acceleation
06:59:11 <andythenorth> I hardware might be very slightly faster, but I'm just eyeballing it, not proper logging + averages
07:17:52 *** Kitrana has joined #openttd
07:21:52 *** Kitrana1 has quit IRC (Ping timeout: 480 seconds)
07:37:40 <Bouke> And the window is (close to) full screen showing the map as it opens?
07:43:23 <andythenorth> I test by putting the map in one corner
07:43:38 <andythenorth> I can do the numbers for the opening position though
07:46:03 <andythenorth> hardware off, hardware on
07:46:11 <andythenorth> obviously the numbers bounce around
07:47:25 <andythenorth> I am potato / potato on the default mac setting
07:47:51 <andythenorth> it might be significant on the older intel macs
07:48:03 <andythenorth> with optional discrete GPU
07:48:55 *** wallabra_ has joined #openttd
07:54:02 *** wallabra has quit IRC (Ping timeout: 480 seconds)
07:54:02 *** wallabra_ is now known as wallabra
07:54:19 *** Tirili has quit IRC (Quit: Leaving)
07:56:11 <Bouke> This is wentbourne on fullscreen (1680x1050@2 => 3360x2100). ~30% better performance with hardware on.
07:56:33 <Bouke> (This is on the nightly from 2 days ago)
07:59:33 <Bouke> And graphics rendering reduced by 50%: 40,1->19,88.
08:00:09 <andythenorth> hardware on / off
08:00:35 <Bouke> What resolution is this? On a Retina display (@2)?
08:01:30 <andythenorth> yes Retina, it's running windowed at about 1600 x 1050 window size
08:02:09 <Bouke> Ah I think I got misled by the original GFX (non-retina).
08:03:29 <andythenorth> potato / potato on M1 then
08:04:03 <andythenorth> I would like to know the battery / temp effects of each setting, but I can't be arsed to test and log those 😄
08:04:38 <Bouke> Can't you also show the current power draw?
08:05:43 <andythenorth> there's the 'energy impact' tab in Activity Monitor, but it's a crude %
08:05:56 <andythenorth> there might be a power monitor in dev tools
08:06:22 <andythenorth> anyway, it's "interesting" but not interesting enough to try it 🙂
08:09:43 <Bouke> So I assume that unless the game is running at 33 fps/60 fps the power draw would be the same, as the game is trying to run at those speeds. So for Wentbourne I don't think there would be a difference, but for Hillbergen it might.
08:14:30 <Bouke> Eyeballing the "Energy Impact" for Hillbergen, I'm getting higher numbers when HW is on. ~35 HW off, ~55 HW on. Both on integrated gfx.
08:18:06 <andythenorth> if I test with FFWD, impact will be 100% always 😛
08:19:33 <Bouke> Wentbourne is at energy impact ~104. Dunno what the scale is though. Forcing High Performance gfx card doesn't seem to have an impact on FPS interestingly.
08:20:22 <andythenorth> if anything, it's thought that discrete GPU might be slower for the OpenTTD case as more has to be moved over the bus to the GPU
08:20:30 <andythenorth> instead of on the same die as CPU
08:20:39 <andythenorth> but evidence was not found
08:21:01 <andythenorth> ok on M1 the impact is about 25 ish for hardware off and 29 ish hardware on
08:21:18 <andythenorth> the slight increase in performance matches those numbers
08:21:24 <andythenorth> so this is just energy physics
08:21:48 <Bouke> I wonder what Wentbourne's performance would be if it used PBS. (or should I keep that thought to myself? :P)
08:22:23 <andythenorth> on the final generation of intel macbook pros, moving it to the discrete GPU would nerf performance
08:22:47 <andythenorth> as the GPU + CPU is over 200W of TDP, and the chassis only does about 100W
08:22:56 <andythenorth> so it just throttles horrifically
08:22:57 <Bouke> I have one of those as well, could test it.
08:23:37 <Bouke> But I have (and I don't really want to admit this here) modded it to reduce the throttling.
08:24:11 <andythenorth> simply cracking the case open and cleaning the air vents reduces throttling 😛
08:24:28 <andythenorth> they suck in debris and block the fans and exhausts
08:27:06 <andythenorth> ha that seems like "don't use it on your lap" after the mod 🙂
08:27:48 <Bouke> Yep! I mostly use it like a desktop. I would be better off with an actual desktop, but I need to bring it to the office every now and then.
09:00:24 *** HerzogDeXtEr has joined #openttd
09:08:43 <andythenorth> `make` is not `make install`
09:08:54 * andythenorth 10 minutes wondering why changes weren't showing up
09:08:56 <TrueBrain> and now you tell me?!
09:15:45 <andythenorth> what shall we do today Brain?
09:15:56 <TrueBrain> same thing as every night!
09:23:46 <andythenorth> move tiny pixels around
09:35:37 <pickpacket> I'm getting somewhere! Tea tree farm all done. A couple of more tile layouts and the tea factory will be done. Then it's time to start drawing sprites for the teahouse
09:35:59 <pickpacket> That's probably gonna take forever, of course :P
09:38:14 <pickpacket> I have a dream of finishing this weekend, but we'll see
09:53:46 *** michi_cc[d] has joined #openttd
09:53:46 <michi_cc[d]> Quite a lot of people use zBase/aBase, maybe test the OSX graphics speed with a 32bpp base set? I should remember this stuff, but I don't remember what exactly was slower anymore either.
09:54:19 <TrueBrain> I mostly remember a lot of different reports 😄 But it never made it into a ticket .. only on IRC, ofc 😉
10:01:43 *** Limyx826 has joined #openttd
10:01:43 <Limyx826> michiccviaGitHub: About removing Original generation from GUI, can be hide from new game window but accessable from advanced settings.
10:02:41 <Limyx826> Just in case any old player that do play with original generator
10:08:42 *** wallabra has quit IRC (Ping timeout: 480 seconds)
10:10:59 <Bouke> michi_cc[d]: this is with aBase on my Hillbergen game. Quite significant improvement with HW enabled. 102% improved speed factor, and 70% reduction in graphics rendering.
10:27:30 <Bouke> On my old iMac (27” 2012) the difference is also significant: FWD improves from 10x to 25x with HW enabled using OpenGFX at 2560x1440 in my Hillbergen save.
10:31:23 <Bouke> This is on that same iMac with a different savegame. Enabling HW makes the game playable.
10:32:19 <michi_cc[d]> It's entirely possible that some OSX update change this since the original tests, so a PR could be considered.
10:39:17 <nielsm> make a benchmark! generate a game with a huge number of vehicles and run it for 10 seconds, test both options
10:45:46 <Bouke> nielsm: Is this already available somewhere, like hidden command line arguments? If not, we could create such and run it on CI to see perf impact of PRs.
10:49:36 <nielsm> there is the null renderer which can runs the game for a number of ticks, but it doesn't do any graphics at all, not even composition or requesting sprites (as far as I know)
10:50:49 <TrueBrain> michi_cc[d]: do remember that one of the reason to disable it by default, is that MacOS deprecated OpenGL. They froze the implementation, and is stated to not receive any further updates. I remember that being one of the deciders 🙂
10:53:04 <michi_cc[d]> Yes, but even if the OpenGL lib is not changed anymore, the underlying graphics driver will still get updates, so if everything points to it being a good idea now, I'm not against changing it.
10:53:31 <TrueBrain> Yeah, as said, one of the deciders. I also remember different people reporting vastly different results on MacOS 😄
10:53:37 <TrueBrain> but we lack MacOS users to give that a true value, ofc
10:54:27 <TrueBrain> MacOS was one of those platforms that always used 32bpp anyway, not?
10:58:27 <michi_cc[d]> Yeah, native 8bpp graphic modes aren't supported on the more modern Quartz (they were on the old QuickDraw stuff, but that died with move away from PPC).
10:58:44 <TrueBrain> yeah, okay, so the difference (compared to other platforms) was smaller anyway
10:59:12 <TrueBrain> I am slightly annoyed we didn't write any of this down in the PRs we made 😄 As I know it took a lot of time/effort back then to figure out 🙂
11:07:41 <TrueBrain> just if only everyone was consistent with MacOS / OSX / Mac OS .. 😛
11:08:04 <andythenorth> stuff and thigns
11:08:05 <TrueBrain> `21:49:52 <andythenorth> we might want to advise mac laptop owners not to turn on hardware acceleration maybe `
11:08:14 <TrueBrain> there was something with laptops and the GPU and getting REALLY hot 😛
11:08:31 <andythenorth> yes, then it throttles, so perf. dives
11:08:40 <andythenorth> required running it for more than about 2 mins
11:08:54 <TrueBrain> owh, that was a story too not? That your performance was widely different depending how long you were gaming? 🙂
11:09:32 <andythenorth> on specific intel macbooks yes
11:09:51 <TrueBrain> it is funny how I just remember: it was a shitshow 😛 But yeah .. lack of writing down sucks 😄
11:09:53 <andythenorth> showing my privilege here, but can't mac users just get an M1? 😛
11:10:15 *** WormnestAndroid has joined #openttd
11:10:55 <Bouke> Yeah just share your credit card here and we’ll happily oblige!
11:12:42 <Bouke> Interestingly enabling HW didn’t trigger my MBP to switch to discrete gfx, it continues to use integrated gfx.
11:13:19 <andythenorth> are you on battery or mains power?
11:14:08 <TrueBrain> `12:30:23 <o*** (avoid HL)> Yes, in a more substantial game, I'm getting 60-70fps ish with OpenGL on Mac, but 70-90fps with Cocoa `
11:14:12 <Bouke> Tried both. Also fiddled with Energy settings.
11:14:16 <TrueBrain> (sorry, just digging through the logs to avoid regressions :D)
11:23:38 <TrueBrain> yeah, those are the cases I can find easily in the logs for MacOS:
11:23:38 <TrueBrain> - Runs of more than 2 minutes get a performance hit due to throttling (seems Intel only issue)
11:23:38 <TrueBrain> - More than other OSes, depending on savegame, performance widely differ
11:23:38 <TrueBrain> - Deprecation warning from Apple
11:23:38 <TrueBrain> - Performance gain, compared to other OSes, is lower
11:23:40 <TrueBrain> Hopefully that helps when debating enabling/disabling, to avoid repeating past mistakes 🙂
11:28:06 <TrueBrain> andythenorth: lol; and yeah, now we know (again), we can validate if those points still hold 🙂
11:28:10 <michi_cc[d]> Bouke: We explicitly tell OSX to manage the GPU for us as the default behaviour otherwise would be to just use the discrete GPU for any OpenGL app. For the few OGL stuff OTTD actually does, the integrated GPU is probably a lot more effective regarding power/heating.
11:29:32 <TrueBrain> `21:50:13 <andythenorth> or we could trick the mac into using the discrete GPU `
11:30:20 <michi_cc[d]> There should be an OSX app option to use the discrete GPU (at least that is what google suggests).
11:30:25 <andythenorth> nice trip into last year
11:33:40 <michi_cc[d]> As this discussion only pertains new players anyway, the current state of the world is the most relevant. As soon as you already have a config file, changing our defaults won't matter. And I would definitely not approve of force-changing the setting.
11:34:29 <TrueBrain> hmm, not disagreeing with that stance. I kinda like it even. But how do we figure out what new players use, I wonder?
11:36:57 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
11:37:03 *** WormnestAndroid has joined #openttd
11:37:05 <michi_cc[d]> If it seems that right now that HW acceleration has better results in contrast to earlier in time, this has to be due to some kind of change on Apples sides as we didn't do any. And I'm assuming most users keep somewhat up on OS updates.
11:37:24 <TrueBrain> so N-2 approach, that seems fair 🙂
11:37:33 <TrueBrain> I suggest we check the boxes of the above points I gave, ignoring the 3rd
11:38:01 <TrueBrain> the logs suggest that the integrated GPU didn't actually give any performance boost; so if that has changed, that would be nice ofc
11:38:06 <petern> Afternoon. Should I buy a Mac to test with?
11:38:19 <TrueBrain> but it might be good to validate if intel laptops don't do weird shit 😄
11:38:40 <Bouke> Bouke: This iMac is running Catalina (10.15.7), originally released in late 2019. It's the last major version available for this machine.
11:39:15 <michi_cc[d]> Deprecation is irrelevant unless Apple actively sabotages it. Either it keeps working or if Apple decides to remove it sometime in the future and does not sabotage it, the video driver will just not load and fall back to the non-accelerated anyway.
11:39:31 <TrueBrain> michi_cc[d]: that is why I suggest to ignore the 3rd point 🙂
11:39:51 <michi_cc[d]> That's why reading is hard 🙂
11:40:06 <andythenorth> just going to leave here also....the "tests" are barely science 😛
11:40:17 <TrueBrain> so we check on Intel laptops if the problem still exists that andy describes. And we check several savegames to see if they increase performance, over several different Macs
11:40:59 <TrueBrain> make a nice GitHub Discussion out of it
11:41:07 <TrueBrain> and we can refer to it in the future, in case things change (again)
11:41:31 <michi_cc[d]> Some of the data points came from orudge, but I think he switched over to the M1 andy camp, too.
11:41:49 <TrueBrain> yeah, but I am sure our community can gather enough datapoints to get a feel for the landscape
11:42:03 <TrueBrain> it doesn't have to be 100ths of datapoints, ofc 🙂
11:42:16 <TrueBrain> just a bit more than "but it improves stuff on my machine" would be nice 🙂
11:42:20 <andythenorth> I have 3 different intel macs and 3 different m1 macs I could test on right now 😛
11:42:34 <andythenorth> but I am not sure that is what I want to spend my day on 😛
11:42:52 <andythenorth> can't we rent a cloud mac and run some logged tests, and be done with it? 😛
11:42:56 <TrueBrain> so my suggestion: pick a few savegames, an empty one, a busy one, a NewGRF one, I dunno
11:43:04 <TrueBrain> put them in a Discussion, write some lines in how to measure performance
11:43:17 <andythenorth> one of my 'todo' jobs: sell some old intel macs
11:43:38 <TrueBrain> create a nice table of what Mac it is, what OS it is running, and how it did with HW on and off
11:44:21 <TrueBrain> and yes, we should have done that already a few times over, not only for Mac 😛
11:44:37 <andythenorth> we have benchmarked CI tests at work
11:44:45 <andythenorth> they're not perfect, but they are repeatable and baselined
11:44:50 <andythenorth> which wins out over perfection
11:45:01 <TrueBrain> graphical testing over CI too?
11:45:11 <andythenorth> incredibly useful
11:45:15 <TrueBrain> yeah, but not performance?
11:45:27 <TrueBrain> most CIs I know, don't have GPUs available 😦
11:45:34 <andythenorth> yeah this is a web app
11:45:46 <TrueBrain> yeah, so different metrics 😄
11:45:48 <andythenorth> the tests will be varied, but things like jmeter, and selenium run times
11:46:02 <andythenorth> did we ever script our UI? 😛
11:46:28 <TrueBrain> more than once we started an attempt to do performance metrics over a lot of savegames
11:46:32 <andythenorth> we can load and run a savegame over server admin port?
11:46:38 <TrueBrain> I even have a full collection of savegames to do that on
11:46:49 <TrueBrain> but those are all "how fast does the game run" savegames
11:46:52 <TrueBrain> not "how fast does it render" 😄
11:47:31 <TrueBrain> andythenorth: if you want to do performance metrics of: how long does it take to run 10 gamedays with `-vull`, we have complete scripts for that
11:48:09 <TrueBrain> `time ./openttd -snull -mnull -vnull:ticks=1000 -g<savegame> -x -c empty/empty.cfg` would do
11:48:34 <TrueBrain> but yeah, that only works on a the null driver
11:50:15 <TrueBrain> but okay, now the subject changed from GPU benchmarking to game benchmarking 😄
12:09:11 <Bouke> TrueBrain: This looks nice! What’s needed to use that approach for actual GPU testing?
12:11:10 <petern> Will someone go food shopping for me?
12:13:26 <reldred> Who do I have to bribe to get multi tile depots merged? 😎
12:13:41 <petern> You can bribe me with food shopping 😄
12:14:40 <petern> Although I'm sure TB is removing all my rights right now
12:16:01 <reldred> Sounds like I need to bribe TB then
12:16:50 <TrueBrain> petern: no, that deal is fair 🙂
12:17:22 <TrueBrain> Bouke: Access to actual GPUs, for starters
12:17:49 <michi_cc[d]> petern: Isn't that called online delivery?
12:19:49 <TrueBrain> michi_cc[d]: sadly they don't deliver here the same day 😦
12:20:59 <reldred> Same here, Uber eats does limited same day grocery delivery
12:28:13 <andythenorth> deliveroo will do coop, waitrose etc
12:28:17 <andythenorth> petern: no roo? 😛
12:29:11 <andythenorth> ha ha the apostrophe strikes again
12:30:04 <petern> Smart quotes didn't exist in 1995...
12:32:41 <petern> That's the symbol on the BBC Micro...
12:33:39 <petern> (Although it's vertical in MODE 7)
12:39:37 <Bouke> TrueBrain: So each video driver would require such a `:ticks` argument, and we'd `time` the runtime duration? But that only gives the time required to run that amount of ticks (so game speed, right?), how to measure gfx performance?
12:45:17 <Eddi|zuHause> ich finde das ist einer der größten schwachpunkte von ronny, er versteht den sinn von befehlsketten nicht
12:57:08 <TrueBrain> Bouke: as there are many factors involved in what influences drawing speed, I do not think it can be done in a similar way. Where we can make game-performance very measurable, as it is fully deterministic, drawing-performance is not. So it stays very subjective, I think
12:57:34 <TrueBrain> having Discord open or not could already have an impact on drawing-performance, for example
12:59:02 <Bouke> So there's no point in doing this, as the results will not be usable?
12:59:28 <JGR> At some point you will need to make a subjective judgement call
12:59:55 <TrueBrain> exactly; it will not be scientific, but it will be sufficient
13:00:28 <TrueBrain> so what people have been doing, eyeballing the FPS window, is, I think, more than sufficient
13:05:23 <TrueBrain> that is why I suggest to just make written instructions, and a few savegames that widely differ
13:06:26 <TrueBrain> but, that is just me. Others might have other ideas here
13:16:18 <andythenorth> it's trying to find pathological cases no, rather than prove x > y?
13:17:17 <andythenorth> our work performance benchmarking is about verifying that we roughly understand the system, or finding things that are "wtf?"
13:18:05 <andythenorth> the actual IRL workloads are extremely variable, so we don't assume that benchmark results correspond to IRL perf
13:29:20 *** tokai|noir has joined #openttd
13:29:20 *** ChanServ sets mode: +v tokai|noir
13:35:48 <TrueBrain> andythenorth: If only CIs had GPUs 😄 well, there are setups that have that ofc 🙂 but not our CIs 🙂
13:36:19 *** tokai has quit IRC (Ping timeout: 480 seconds)
13:36:24 <TrueBrain> I assume you mostly use it to find regressions?
13:36:50 <andythenorth> we use it to find regressions
13:37:06 <andythenorth> and from time-to-time we compare empirical performance in the live fleet against the benchmarks
13:37:29 <andythenorth> and if we do performance work, it's a verification that making A faster doesn't have side-effect at B
13:37:36 <andythenorth> or at least we know if it does and trade
13:38:20 <andythenorth> these are complex systems, a long way from e.g. me trying to make nml compiles faster 🙂
13:38:40 <andythenorth> so e.g. increasing cache in one layer improves some workloads, but hurts others due to cold-cache
13:38:50 <petern> multithreaded terrain generation?
13:39:15 <andythenorth> petern: use voxels
13:39:56 <JGR> Terrain generation is already very fast
13:40:11 <JGR> It's putting stuff on the terrain afterwards which can take a while
13:40:11 <TrueBrain> Never fast enough 😄
13:40:25 <andythenorth> petern: run 4 different terrains, then have a random battle for each tile to see who wins
13:40:44 <andythenorth> competitive terrain gen
13:41:03 <TrueBrain> Hmm .. Still want to make sharding support in OpenTTD, so you can have truly massive maps 😄
13:41:03 <petern> andythenorth: I know that's a stupid suggestion, but it might be an interesting way to get variety 😉
13:41:27 <andythenorth> regional terrain
13:41:37 <andythenorth> then just substitute segments
13:41:44 <andythenorth> with some kind of edge blending
13:42:33 <andythenorth> ok, lunch: marmite on toast, cheese on toast, or marmite on cheese on toast
13:42:56 <petern> Map sharding suggests you might need town and industry sharding. And maybe vehicle sharding.
13:43:12 <andythenorth> did I ever mention newgrf sharding?
13:43:13 <petern> 4kx4k is not big enough though I guess.
13:43:13 <TrueBrain> petern: If you truly want to scale, yes
13:43:30 <TrueBrain> I wss thinking of an MMO server
13:44:05 <TrueBrain> But you need to hand over vehicles and viewports somehow
13:44:29 <andythenorth> but all at 256x256 😛
13:44:32 <andythenorth> the correct size
13:44:48 <andythenorth> single player bigger maps are sharded 😛
13:44:50 <petern> Hmm, I wonder, does this -1 mean to 'fix' an off-by-one, or does it mean to move up by 1 pixel...
13:44:57 <andythenorth> can we go shard-per-core?
13:45:50 <petern> How long does it take for one griefer to destroy a 100kx100k map?
13:47:02 <petern> Basically, basing your offsets off step_size, but your starting y position is already +1 scaled pixel, so step_size should be at least -1 scaled pixel. Hmm.
13:48:41 <petern> I'm being picky & perfectionist 😉
13:49:04 <petern> The -1 comment and the step_size thing are related
13:54:26 <Bouke> Can I pass in-game console commands from say a bash script? Use-case: extract “fps” output using a script.
13:55:22 <TrueBrain> There is also console scripting, but it is .. Well, it is. And that is it 😄
13:56:45 <Bouke> Thanks, will look into that!
13:59:47 <TrueBrain> You can also execute scripts from console
14:00:31 <TrueBrain> What might work is that you preprogramm some scrollto and run fps between them to get a runtime idea of what the fps is of a savegame
14:03:03 <Bouke> So my overall idea here is that I have a bash script that’ll launch ottd with some savegames and predefined config and run the fps command at an interval and aggregate/avg those values.
14:04:12 <nielsm> <TrueBrain> Admin port? <-- admin port only works against a multiplayer server
14:04:47 <nielsm> which is why there are many things it's specifically not a solution to
14:04:48 <TrueBrain> Still want to remove the difference between SP and MP .. wish I had more time as resource 😄
14:06:26 <TrueBrain> Bouke: Keep in mind by default the refresh rate is capped, and you cannot enable FWD by scripts
14:06:52 <TrueBrain> The cap can however be disable (or made very large) by config
14:10:59 <andythenorth> full-time OpenTTD when? 😛
14:11:13 <andythenorth> we should have included that coin-miner, then we could rent player CPU time to third parties
14:11:23 *** m3henry has joined #openttd
14:11:44 <andythenorth> how about we do a startup so we can sell it and retire to do OpenTTD?
14:12:07 <petern> OpenTTD-Discord bridge when?
14:13:13 <petern> Do I need to scale the "checker" pixels? Hmm
14:17:06 <TrueBrain> andythenorth: Suggestions?
14:17:23 <TrueBrain> m3henry: Don't fix what isn't broken!
14:17:48 <andythenorth> TrueBrain: how about a game about trains?
14:18:28 <andythenorth> but on the blockchain
14:18:35 <TrueBrain> Now we are talking!
14:19:56 *** m3henry_ has joined #openttd
14:20:26 <m3henry_> But it always shows empty
14:20:31 <m3henry_> Seems pretty broken to me
14:21:54 <TrueBrain> m3henry_: Well, we cannot fix things we don't own / host; but what we do works fine, so *shrug* 🙂
14:22:06 <Bouke> TrueBrain: So can I trigger console scripts externally? Or do I have to put a script named “autoexec.scr” in the user’s OTTD folder?
14:22:27 <TrueBrain> (In other words, use weblogs.openttd.org/openttd/ and stop using 3rd party logs :p)
14:23:12 <TrueBrain> Bouke: There are no external triggers as far as I know in SP
14:25:13 *** m3henry has quit IRC (Ping timeout: 480 seconds)
14:28:41 *** m3henry has joined #openttd
14:36:54 <petern> Is Allegro used by anyoine?
14:37:47 <m3henry> TrueBrain: Looks like the link on the wiki needs updating to point at that then
14:38:36 <TrueBrain> m3henry: Go for it!
14:42:24 <m3henry> Done, didn't relaise the wiki was open for all to edit
14:44:40 <TrueBrain> What else are wikis good for 😄 cheers!
15:14:22 *** virtualrandomnumber has joined #openttd
15:14:37 *** virtualrandomnumber has quit IRC ()
16:04:06 <petern> Heh, nobody spoke here while I went shopping
16:06:24 <andythenorth> that is the best
16:06:36 <andythenorth> container shipping game?
16:07:04 <JustANortherner> Where is the bridge tho?
16:31:44 <petern> Hmm, what happened to the NewGRF y-position thing...
16:42:19 <Bouke> TrueBrain: Alright, I've created draft instructions to kick this off. I invite you all to provide some feedback on this draft before I post the discussion on GitHub. The goal of this discussion would be to make an informed decision on whether to enable hardware acceleration by default. https://gist.github.com/Bouke/482c4f33898cc1dbc198bdbdb5eb6bce
16:58:06 <Bouke> Bouke: I’ll add running latest nightly to the instructions as there’s a crash with 32bpp in 12.2. Also: people should report their gfx base set.
17:07:03 *** Wormnest has joined #openttd
17:09:30 *** esselfe has quit IRC (Remote host closed the connection)
17:21:49 *** esselfe has joined #openttd
17:34:05 *** gelignite has joined #openttd
18:20:08 *** esselfe has quit IRC (Quit: Lost terminal)
18:30:22 <TallTyler> What are everybody’s thoughts on 10058? That’s been talked about for a while but somebody finally did it.
18:33:53 *** esselfe has joined #openttd
18:58:18 <DorpsGek> - Update: Translations from eints (by translators)
19:26:11 <petern> Fix my code, or watch Lego YouTube?
19:41:24 *** wallabra has joined #openttd
19:42:11 <TrueBrain> TallTyler: My only comments is that I don't like the custom height of the buttons 😄 I don't see why they can't just be normal height .. they look weird now :p
19:45:15 <TrueBrain> Bouke: Nice! Much more scientific then I would have done it, so I can only applaud that 😄 tip: make the title read MacOS first, as people will otherwise not read it and you get all kind of OS reports. Not a bad thing .. but yeah 😄
19:53:06 <pickpacket> I’ve finished the tea farm and the tea factory 😊 Will work on teahouses tomorrow.
19:53:32 <pickpacket> andythenorth: is it possible to make my newgrf compatible with firs?
19:54:49 <andythenorth> pickpacket: 'kinda'
19:56:45 <petern> Oh, that beer ran out 😦
19:57:21 <andythenorth> goes it release Horse?
20:00:56 <petern> goes it release chunky bevels?
20:12:27 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
20:58:44 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:03:34 * petern cracks open the gouda instead
21:22:40 <TrueBrain> We need a design guide 😄
21:37:35 <petern> newgrf window has a lot of shittyness
21:38:15 <petern> I think I have a patch to fix that though 😉
21:42:31 <andythenorth> I have written 2 UI design guides
21:42:36 <andythenorth> but not for OpenTTD 😛
21:56:41 <petern> Just add two buttons to the bottom of the window and leave the rest alone, bosh, job done.
21:58:37 <petern> From a usability perspective moving buttons around that lots of people are used to isn't particuarly helpful
22:04:13 <TrueBrain> Only spelled wrong once? Shocker
22:04:25 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:04:48 *** wallabra_ has joined #openttd
22:05:02 <TallTyler> I had to be really careful not to misspell it in the commit or PR title 😛
22:07:10 <TrueBrain> Guess 10058 would be better if AI and GS was split first honestly ..
22:07:32 <TrueBrain> GS in the AI window was a lazy move, and doesn't actually make sense :p
22:08:00 *** Arastais has joined #openttd
22:08:00 <Arastais> TrueBrain: I can do that tbh, if that's agreed upon
22:08:26 <Arastais> but I don't want ot start a more drastic change without a more unanimous opinion first
22:08:41 <petern> Smaller issue, what about the Play Heightmap window, should that not get the same treatment?
22:09:06 <petern> What about the Scenario Editor, where NewGRFs are pretty important to get right/.
22:09:20 <TrueBrain> Arastais: If there is one thing we all agree on, is that we always disagree about GUIs 😄
22:09:40 <Arastais> petern: it does, i made the same changes to the play heightmap window. There's two pairs of images for both the original design and newer one.
22:10:21 <petern> Cool, I'm skipping back and forth and just being blind.
22:10:31 <petern> I really hate that massive Generate button though
22:10:38 *** wallabra has quit IRC (Ping timeout: 480 seconds)
22:10:38 *** wallabra_ is now known as wallabra
22:11:00 <petern> (Not that it matters what much what I think :P)
22:11:01 <TrueBrain> I once designed a new window in total, as I hate that window even how it is now :p
22:11:29 <Arastais> petern: so you prefer this but with a single height row of buttons?
22:13:38 <Arastais> TrueBrain: I honestly think this is a good idea/compromise if the dedicated Configure Game Script button is widely disliked
22:14:17 <petern> Single height, possibly with gaps instead of being joined, and yes, we should split the AI/GS settings windows.
22:14:47 <TrueBrain> I don't dislike the button, but I do think it is too specific.. First press one button to set the GS, then another button to configure it .. that doesn't actually help in amount of clicks 😄
22:15:14 <Arastais> TrueBrain: good point, i guess I can get to work splitting the windows
22:15:15 <petern> (Also, NewGRF on the right, so it matches the order on the main menu)
22:15:58 <petern> Also I should fix my own code first
22:17:00 <Arastais> but I still dont think theres an agreement on height/layouts of the buttons, so in the meantime i think it would be good to get opinions on. Specifically:
22:17:00 <Arastais> 1) single or double height?
22:17:00 <Arastais> 2) generate button on bottom or integrated like the original design?
22:17:00 <Arastais> 3) gap between the 3 buttons?
22:18:28 <petern> We should have a vote, it'll be 52/48
22:18:43 <petern> And then we can through it out and just put Boris Johnson back.
22:19:08 <petern> Content Window: all separate buttons
22:19:14 <petern> Multiplayer Window: all separate buttons
22:19:16 <TallTyler> No opinion on 1 or 3, but I'd like to see at least a mockup of the Generate button at the bottom. Having it in the middle is fine now because the map edge selector is weird, but with more buttons the Generate button placement looks pretty arbitrary
22:19:22 <petern> AI window, buttons are all joined
22:19:31 <petern> NewGRF window: wtf was I doing?
22:22:19 <andythenorth> I did one years ago, it was horrific
22:22:31 <andythenorth> has everyone tried fixing newgame?
22:23:09 <TallTyler> I actually quite like the general idea
22:23:44 <TrueBrain> andythenorth: Ignoring the colours (yet another colour schema), not bad 🙂
22:24:07 <TallTyler> The colors are a bit ugly and the tabs might look better taking up the full width of the window, but the interaction flow looks to be an improvement
22:24:57 <andythenorth> I did lots of versions
22:25:01 <andythenorth> none of them stuck
22:26:08 <petern> NWidget(WWT_TABS) then...
22:26:35 <TrueBrain> Tabs are the future! :p
22:33:09 <TrueBrain> Not even a picture?! Tssk :p
22:34:41 *** sla_ro|master has quit IRC ()
22:36:20 <petern> (But it's not commit-before-compile)
22:36:34 *** gelignite has quit IRC (Quit: Stay safe!)
22:36:41 <petern> Apparently when I change header files I have to wait ages for compile 😦
22:36:58 <petern> Therefore I think we should remove header files. Probably put all source into one file.
22:37:56 <Arastais> petern: god forbid u change english.txt 😓
22:38:23 <petern> The branch I had just switched from does, yes.
22:41:00 <TallTyler> I'm working on adding a scrollbar to the Company Infrastructure window (#9056)...where does it go? Does it need converting to a matrix first?
22:42:04 <petern> You could turn it into a WWT_MATRIX (not a NWID_MATRIX) and change how it looks.
22:42:44 <petern> Or you can replace all the separate widgets with one giant panel and draw it all in that, taking account of scrollbar (like the goal / story book windows, and probably others)
22:43:12 <petern> Or I have a patch somewhere that lets us simply add a scrollbar and keep the rest as it is, and it just works.
22:44:02 <petern> TrueBrain: added pics 😉
22:45:05 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
22:50:05 <Arastais> TrueBrain: for this, inthe code, should I reuse the AI parameter configuration window or create a new similar window
22:50:23 <TallTyler> I initially just wanted to standardize text colours between these two windows but realized while I'm at it, I should fix the lack-of-scrollbar issue
22:50:48 <Arastais> Arastais: specifiaclly im referring to `AISettingsWindow`
22:51:57 <Arastais> also, should I do all this in a new file or keep it in `ai_gui`
22:52:36 <TallTyler> That said, the company window uses gold headlines and white text, so maybe I need to change the finance window colours to match the company and infrastructure windows...
23:05:19 *** _aD has quit IRC (Remote host closed the connection)
23:36:02 *** wallabra_ has joined #openttd
23:40:08 *** wallabra has quit IRC (Ping timeout: 480 seconds)
23:40:08 *** wallabra_ is now known as wallabra
23:55:13 *** wallabra_ has joined #openttd
23:57:03 *** wallabra has quit IRC (Ping timeout: 480 seconds)
23:57:03 *** wallabra_ is now known as wallabra
continue to next day ⏵
