Times are UTC Toggle Colours
00:01:38 <Eddi|zuHause> how exactly do i check that out? 00:01:52 <TrueBrain> git fetch https://github.com/TrueBrain/OpenTTD opengl 00:01:55 <TrueBrain> git checkout FETCH_HEAD 00:02:27 *** DasPoseidon has quit IRC 00:02:31 <TrueBrain> I assumed by your last comment you are using Linux? 00:02:47 <Eddi|zuHause> yes 00:02:53 <TrueBrain> good :) 00:07:18 <TrueBrain> owh, and you might also want to add -ddriver=4 when starting 00:07:22 <milek7_> cursor draws twice 00:07:28 <TrueBrain> it has some juicy OpenGL lines we might need in case it fails to load Eddi|zuHause :) 00:07:49 <milek7_> ie. there are two cursors, one is faster 00:07:50 <Eddi|zuHause> there's a weird mouse trail while reading in newgrfs 00:07:58 <TrueBrain> you will find more bugs, no worries :) 00:08:07 <TrueBrain> but so it works, that is already a nice surprise :) 00:08:16 <TrueBrain> is it also faster than your 10ms? :D 00:08:56 <milek7_> lol, opening performance window kills fps ;D 00:09:42 <Eddi|zuHause> scrolling is abysmally slow 00:09:56 <TrueBrain> I haven't even tried that honestly :) 00:10:03 <TrueBrain> how is the ingame FPS doing? 00:10:27 <TrueBrain> did you mean window scrolling or map scrolling btw? 00:10:33 <Eddi|zuHause> normal fps while screen is stationary is maybe slightly faster 00:10:35 <michi_cc> Scrolling is not really touched by OpenGL though, it's all pure blitter. 00:10:49 <Eddi|zuHause> scrolling the main viewport 00:11:06 <michi_cc> try -b8bpp-optimized, especially if you didn't use 32bpp before. 00:11:13 <TrueBrain> ah, so I am not the only one :) 00:11:22 <TrueBrain> you have to move REAL slow in order for it to do anything 00:11:29 <TrueBrain> I considered it an Windows bug, but if you have it too..... 00:11:52 <TrueBrain> something I will look into tomorrow :) 00:11:56 <milek7_> Eddi|zuHause: do you have mesa driver or nvidia? 00:12:05 <Eddi|zuHause> mesa (amd) 00:12:09 <milek7_> MESA_EXTENSION_OVERRIDE=-GL_ARB_buffer_storage ./openttd 00:12:27 <milek7_> that should fix some problems 00:12:38 <TrueBrain> michi_cc: I am kinda shocked how easy this was and that it runs on other people's machine too :D 00:13:09 <Eddi|zuHause> yeah, with 8bpp blitter, scrolling is faster 00:13:22 <Eddi|zuHause> not exactly "good" :p 00:13:38 <michi_cc> milek7_: One of the problems real game engines have, lots of special casing code to enable/disable stuff based on driver version, GPU vendor etc. 00:15:28 <Eddi|zuHause> milek7_: oh yeah, that's like a million times faster 00:17:38 <TrueBrain> right, enough toying around with this ... tomorrow I will make a .. "more correct" patch :P 00:17:53 <TrueBrain> tnx for testing Eddi|zuHause :) 00:18:47 <Eddi|zuHause> second mouse pointer is irritating at times :p 00:19:00 <TrueBrain> it is a fun gimmick! 00:19:02 <michi_cc> Looks like Mesa is bad with persistent buffer mapping (or maybe only AMD on mesa). I'd appreciate a -d driver=2 output on the PR, just to see vender string and version etc. 00:19:11 <TrueBrain> I think I am going to leave it in Eddi|zuHause , just for shit and giggles :D 00:19:21 <michi_cc> Then it could disable that for mesa by default. 00:19:39 <TrueBrain> michi_cc: and so it begins ... :D :D 00:20:14 <TrueBrain> but seriously, this is awesome work michi_cc :) I have seen more than one OpenGL implementations in OpenTTD over the years .. this one is a lot smoother than any of the others :D 00:20:40 <michi_cc> I don't have any AMD hardware to test around, only Intel and nVidia, which can be a bit of a problem. 00:20:51 <milek7_> dbg: [driver] OpenGL driver: AMD (4.6 (Compatibility Profile) Mesa 21.0.0-devel (git-05ad0f8871)) 00:21:22 <TrueBrain> owh .. seems I removed a few too many lines .. wanted to keep the "3.2" loading in there 00:21:23 <TrueBrain> oops 00:21:25 <Eddi|zuHause> milek7_: https://pastebin.com/Fy6HDJbf 00:21:29 <Eddi|zuHause> michi_cc: 00:21:32 <michi_cc> nVidia is clearly faster with persistent mapping, Intel (at least the one I test on) is potato, potato. 00:22:01 <michi_cc> Eddi|zuHause, milek7_: Could you please make comments on the PR so it doesn't get lost? 00:22:29 <Eddi|zuHause> which pr is that? 00:23:06 <TrueBrain> https://github.com/OpenTTD/OpenTTD/pull/7744 00:27:38 <TrueBrain> right, zzzz time now 00:28:09 <DorpsGek> [OpenTTD/OpenTTD] Eddi-z commented on pull request #7744: Draft Feature: OpenGL video driver https://git.io/JtTs6 00:29:41 <milek7_> hm, I can check on mesa-intel 00:37:46 <michi_cc> Oh, and because this is the first go at Linux, maybe also a past of glxinfo, just to see any other interesting bits. 00:38:17 <michi_cc> E.g. OTTD debug output right now doesn't show any AMD in the output for Eddi. 00:42:05 <Eddi|zuHause> glxinfo output is... very... long 00:42:46 <michi_cc> Can I just have a paste? Doesn't need to be on the PR. 00:43:06 <michi_cc> It's probably showing extensions 00:45:44 <milek7_> that's mine (important part only): https://gist.github.com/Milek7/c288921807e1a84ba4649017763246dd 00:46:32 <Eddi|zuHause> michi_cc: added it to the comment 00:48:17 <michi_cc> Thanks. And good night as well. 00:54:21 <milek7_> michi_cc: https://pastebin.com/raw/6PwVXqqX 01:15:23 <DorpsGek> [OpenTTD/OpenTTD] Milek7 commented on pull request #7744: Draft Feature: OpenGL video driver https://git.io/JtTZ8 01:25:48 *** sla_ro|master has quit IRC 01:34:49 *** Flygon has joined #openttd 01:47:34 *** Progman has quit IRC 02:59:34 *** WormnestAndroid has quit IRC 03:00:10 *** WormnestAndroid has joined #openttd 03:07:25 *** D-HUND has joined #openttd 03:10:48 *** debdog has quit IRC 03:14:21 <reldred> I suddenly find myself posessed with the urge to draw giant megablock style buildings for openttd after playing so much cyberpunk. 03:14:53 <reldred> given most ttrs 'big' buildings are still only 1x1 I think 2x2 with maybe a lil overhang would suffice for a futuristic megablock type building 03:53:13 *** glx has quit IRC 04:12:28 *** supermop_Home has quit IRC 06:33:04 <DorpsGek> [OpenTTD/OpenTTD] orudge opened pull request #8579: Change: Label Mac builds as "macos" instead of "macosx" https://git.io/JtTB6 06:34:30 <DorpsGek> [OpenTTD/website] orudge opened pull request #182: Fix: Add Universal description for modern macOS binaries https://git.io/JtTBi 07:10:40 *** roadt_ has joined #openttd 07:17:01 *** roadt has quit IRC 07:44:54 *** nielsm has joined #openttd 08:49:31 *** DasPoseidon has joined #openttd 08:57:12 <TrueBrain> orudge: so a German person created an info@ account for Steam :P 08:57:30 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8579: Change: Label Mac builds as "macos" instead of "macosx" https://git.io/JtTul 08:59:13 <DorpsGek> [OpenTTD/website] TrueBrain commented on pull request #182: Fix: Add Universal description for modern macOS binaries https://git.io/JtTu0 09:03:55 *** andythenorth has joined #openttd 09:06:34 *** sla_ro|master has joined #openttd 09:16:53 *** Progman has joined #openttd 09:26:27 <andythenorth> yo 09:41:11 *** snail_UES_ has quit IRC 09:50:25 <andythenorth> Timberwolf multiple liveries using company colour combination rules, or not playing :D 09:50:26 <andythenorth> https://gist.githubusercontent.com/andythenorth/94ec0c10570da4835292de62e1b9d5db/raw/582568dff71b42e87d3386ad3e8f1e9576a4555e/gistfile1.txt 09:51:00 <andythenorth> https://grf.farm/iron-horse/2.17.0/html/growler.html 10:01:54 <orudge> TrueBrain: apparently so. Steam's e-mails don't always seem to be getting through though. 10:02:12 <orudge> Can't actually register a new account on another address 10:02:25 <orudge> Will have a try again 10:04:38 <DorpsGek> [OpenTTD/website] orudge commented on pull request #182: Fix: Add Universal description for modern macOS binaries https://git.io/JtTgZ 10:04:51 <TrueBrain> orudge: you can't just recoveri that account? 10:04:55 <DorpsGek> [OpenTTD/OpenTTD] orudge merged pull request #8579: Change: Label Mac builds as "macos" instead of "macosx" https://git.io/JtTB6 10:05:27 <TrueBrain> orudge: what I mean, that we add to the description: "(Universal, including Intel and Silicon support)" or something 10:05:34 <orudge> TrueBrain: that's what I tried to do 10:05:40 <orudge> Yes 10:05:56 <orudge> Don't think that's necessary, honestly 10:05:57 <TrueBrain> when someone arrives at the Get Info dialog, it is a bit late :D 10:06:00 <TrueBrain> k :) 10:06:07 <orudge> Universal is standard terminologh 10:06:10 <DorpsGek> [OpenTTD/website] TrueBrain approved pull request #182: Fix: Add Universal description for modern macOS binaries https://git.io/JtTg8 10:06:20 <DorpsGek> [OpenTTD/website] TrueBrain dismissed a review for pull request #182: Fix: Add Universal description for modern macOS binaries https://git.io/JtTg8 10:06:45 <DorpsGek> [OpenTTD/website] TrueBrain approved pull request #182: Fix: Add Universal description for modern macOS binaries https://git.io/JtTgB 10:06:55 <TrueBrain> orudge: we will direct complaints to you, no worries :D 10:07:06 <TrueBrain> personally I like promoting that we build for Silicon, but, meh :) 10:07:11 <orudge> Happy to receive them ;) 10:07:22 <orudge> We can do that in the release announcement :) 10:07:47 <DorpsGek> [OpenTTD/website] orudge merged pull request #182: Fix: Add Universal description for modern macOS binaries https://git.io/JtTBi 10:36:29 <andythenorth> what shall we do today Pinky? 10:39:30 <TrueBrain> orudge: I double-checked, the password for info@ for steam was never written down in any of the usual places, it seems :( 10:42:08 <orudge> I managed to reset the password 10:42:14 <orudge> It just wouldn't let me log in 10:42:24 <TrueBrain> what is the error? 10:42:24 <orudge> Haven't translated the German e-mail yet 10:42:35 <orudge> It needed a verification code 10:42:41 <orudge> which never arrived 10:42:56 <TrueBrain> did you put the password in the usual place? 10:43:01 <orudge> Yep 10:43:09 <orudge> The new usual plaxd 10:43:12 <orudge> *place 10:43:23 <TrueBrain> you sure? :D 10:43:35 <orudge> However, that account is registered in DE, not sure if it can be changed to GB 10:43:41 <orudge> since our company is in GB 10:43:49 <orudge> well, unless I forgot to hit Save 10:43:58 <orudge> which is entirely possible :P 10:44:12 <TrueBrain> it is hiding, the Save button 10:44:14 <orudge> Will look into it on Monday 10:44:18 <TrueBrain> I don't like that design 10:44:35 <TrueBrain> well, if I have the password too, I can help you look :) Without it .. not much I can do :P 10:44:38 <TrueBrain> reset it again, at best :D 10:45:43 <orudge> It is in 1password 10:45:59 <orudge> I did save it after all 10:46:04 <TrueBrain> hmm .. in what category? 10:49:19 <orudge> Private, oops 10:49:24 <TrueBrain> :D 10:50:43 <orudge> In fact I put the Apple password in there too it seems 10:50:53 <TrueBrain> the signing one was in the right place 10:51:20 <TrueBrain> but yeah, by defaults it adds it in private, which is a good thing 10:52:13 <orudge> OK, now moved 10:52:28 <orudge> If we can either get the account moved country (Steam suggests that might happen if you make a purchaes from another country) or close it and open a new one, that will work 10:52:39 <orudge> or just get a new one open - I set up a new e-mail address forwarder to use for that 10:52:48 <orudge> If you can't get it to work I can try a new account next week 10:52:54 <orudge> but now I must head off 10:52:58 <TrueBrain> sounds like a plan 10:52:59 <TrueBrain> cheers :) 10:56:20 <andythenorth> hmm this map is too flat 10:56:36 <TrueBrain> that's what she said! 10:59:54 <TrueBrain> "The account name or password that you have entered is incorrect." lol 11:01:17 *** phil[m] has quit IRC 11:01:21 *** phil[m] has joined #openttd 11:03:25 *** patrick[m] has quit IRC 11:03:26 *** patrick[m] has joined #openttd 11:05:24 *** joey[m] has quit IRC 11:05:27 *** joey[m] has joined #openttd 11:05:33 *** linda[m] has quit IRC 11:05:38 *** linda[m] has joined #openttd 11:09:49 *** yoltid[m] has quit IRC 11:09:54 *** yoltid[m] has joined #openttd 11:11:57 *** magdalena[m] has quit IRC 11:11:59 *** magdalena[m] has joined #openttd 11:14:05 *** christoph[m] has quit IRC 11:14:09 *** christoph[m] has joined #openttd 11:14:11 *** Heiki[m] has quit IRC 11:14:15 *** Heiki[m] has joined #openttd 11:16:13 *** johanna[m] has quit IRC 11:16:15 *** johanna[m] has joined #openttd 11:17:38 <andythenorth> why is this game still so addictive after 26 years? 11:17:42 * andythenorth curious 11:18:21 *** karl[m]2 has quit IRC 11:18:25 *** karl[m]2 has joined #openttd 11:20:15 *** glothit7ok[m] has quit IRC 11:20:16 *** glothit7ok[m] has joined #openttd 11:30:51 <TrueBrain> orudge: if we delete the current account, it will go in "hybernation" for 30 days before it is really deleted :) 11:30:57 <TrueBrain> (I am in btw .. took some effort ..) 11:33:50 <TrueBrain> other than emails being weirdly delayed, but sometimes arrive instantly, the account seems in good health. It is indeed linked to Germany, but I do not think that matters for the Partner signup 11:34:48 <TrueBrain> I did have to reset the password again, as for some reason the one in 1password was not working. It is now :P The initial german email is just "your password has changed, if this was you, no action required" email 11:39:03 *** frosch123 has joined #openttd 11:42:26 <Eddi|zuHause> <TrueBrain> orudge: so a German person created an info@ account for Steam :P <-- my guess would be that was planetmaker 11:48:40 <orudge> Thanks TrueBrain 11:49:00 <orudge> Can try partner signup on Monday :) 11:55:42 <TrueBrain> w00p :) 11:56:21 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver https://git.io/JeO8B 12:06:42 *** Samu has joined #openttd 12:15:01 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver https://git.io/JeO8B 12:15:56 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #7744: Draft Feature: OpenGL video driver https://git.io/JtTrC 12:17:06 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8573: Add tile parameter for GSCompany.ChangeBankBalance to show text effect if needed https://git.io/JtTr8 12:20:02 <TrueBrain> michi_cc: linux now fails compiling :( 12:21:50 <Samu> there's something in this PR that just doesn't feel right. https://github.com/OpenTTD/OpenTTD/pull/8513 12:22:47 <Samu> a very flat 256x256 desert map with a max height of 29, it looks worse than mountainous 12:23:21 <Samu> try a max height of 255 12:26:10 <LordAro> Samu: we need more exact than a feeling 12:26:24 <LordAro> TB acknowledges it's not perfect, but it's better than it was previously 12:28:27 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8578: Fix: Stopped ships shouldn't block depots https://git.io/JtTrX 12:31:29 <andythenorth> ok the problem with 8513 in desert and height 255 is that it's not placing rainforest with 12:31:51 <andythenorth> a 256x256 map 12:32:04 <TrueBrain> andythenorth: fully expected and sadly by current design 12:32:19 <TrueBrain> it takes 128 tiles to get to the 255 height ... so yeah, that is a bit of an issue :D 12:32:49 <TrueBrain> we should simplify the normal "start game" dialog, and allow more tuning in an "advanced" mode, honestly .. the way it is now wired in is impossible :) 12:33:18 <TrueBrain> the desert climate is the only one which really has this issue btw, the others work differently 12:33:25 <LordAro> yeah 12:33:29 <TrueBrain> as .. why do the same thing 4 times, if you can at least do 1 time something COMPLETELY different :P 12:33:33 <LordAro> having played around with arctic a bit, i think it's fine 12:34:08 <TrueBrain> desert is fine too if you keep the heightlevel within the dimensions of your map :) 12:34:23 <TrueBrain> (it is the only one that scales something depending on the max heightlevel) 12:34:27 <TrueBrain> not even TGP scaled with that 12:35:14 <TrueBrain> euh, 128 tiles earlier is 255 tiles .. no clue why I divided that by 2 :P 12:35:30 <TrueBrain> I think a 1k map is the first that can reach 255 in height :P 12:35:59 * LordAro tries generating a 4k map with 255 height just for fun 12:36:08 * andythenorth thinks that desert is fundamentally flawed in current design for TGP 12:36:12 <andythenorth> never figured out how 12:36:15 <TrueBrain> I just don't understand max-height .. I understand it was like: well, we might as well do 256 too, but .. it makes no sense 12:36:31 <TrueBrain> TGP, in normal climate, only generates up to 90-ish in height 12:36:37 <TrueBrain> so what are you going to do with the other 160? 12:36:42 <andythenorth> le oof 12:36:57 <TrueBrain> for desert, it adds 1/4th to that value, so it can generate up to .. 12:36:58 <andythenorth> it got merged afaik because forums were saying 'game is dead' a lot 12:37:02 <TrueBrain> @calc 90 + (256 / 4) 12:37:02 <DorpsGek> TrueBrain: 154 12:37:04 <TrueBrain> 150! 12:37:07 <andythenorth> not sure there was any other reason 12:37:23 <TrueBrain> but it cannot get any higher really .. so why allow it? Gives a false sense of what is happening 12:37:40 <andythenorth> TrueBrain btw I don't think desert means 8513 is wrong, just trying to turn Samu's 'feeling' into a fact 12:37:56 <TrueBrain> andythenorth: I understand :) 8513 fixes more than it breaks, so I am all good :) 12:38:11 <andythenorth> there is no equivalent problem in arctic 12:38:26 <TrueBrain> so I was thinking: let go of max-height (just make it always 250 or what-ever). Make "terrain types" like flat, hilly, etc with fixes values. Allow anyone to "override" that value in advanced mode, to set it to what-ever it wants 12:38:32 <TrueBrain> that way TGP can generate landscape up to 250 12:38:34 <andythenorth> maybe I should try my 'use temperate for tropic' patch a 3rd time :P 12:38:35 <TrueBrain> if you want to be REALLY silly :P 12:38:40 <LordAro> oof, fully zoomed out 4k map is not fast on this laptop 12:38:46 <LordAro> and that's without anything except terrain 12:38:52 <andythenorth> it really just is better with current tropic TGP rules > dev/null 12:39:06 <TrueBrain> the problem is not TGP perse 12:39:18 <TrueBrain> it is just a combination of weird defaults and non-configurable values 12:39:19 <LordAro> 33ms world tick, with occasional 50ms graphical 12:39:38 <TrueBrain> like the "desert height is 1/4th of MAX height" 12:39:54 <TrueBrain> but okay, we can bitch about this all day .. it has been broken for a few years now .. clearly nobody cares enough :D 12:40:11 <TrueBrain> I might mock up some UIs to reduce the problem soon :) 12:40:39 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8566: Move "town name" selection into map generator GUI https://git.io/JtTon 12:40:42 <andythenorth> original generation has 'interesting' issues with height level 255 in tropic also 12:40:47 <frosch123> TrueBrain: the default value of max-height is 64 or so. 250 is not really wanted 12:41:00 *** sla_ro|master has quit IRC 12:41:08 <frosch123> but yes, if max-height is a lot higher than the landscape really is, stuff breaks 12:41:12 <LordAro> obligatory: what does JGR do? 12:41:21 <andythenorth> with height level 255 on 256x256, original generator makes mangrove swamps of rainforest; rest is desert 12:41:42 <andythenorth> with height level 15 on 25x256, original generator makes 50% rainforest, 50% desert, in 2 halves of map 12:41:51 <TrueBrain> frosch123: default is 30, hence why I do not understand why you can change that to 8 times that value :P 12:42:07 <frosch123> ah, 30, sounds good 12:42:10 <TrueBrain> but okay .. mostly my issue is, that changing that value gives the suggestion it influences terrain generation 12:42:15 <TrueBrain> for 3 climates, it does not 12:42:33 <TrueBrain> (well, it does, but in an inverse way .. it makes hills being chopped off :P) 12:42:37 <andythenorth> make some original tropic maps :P they're quite cool 12:42:38 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver https://git.io/JeO8B 12:42:45 <frosch123> i thought it depends on the total map size? like bigger maps can also get taller? 12:42:54 <TrueBrain> frosch123: indeed 12:42:57 <TrueBrain> but never above the max-height 12:43:09 <TrueBrain> so that setting only makes things worse if it is too low 12:43:13 <TrueBrain> never better by making it higher :D 12:43:19 <TrueBrain> such a mind-fuck :D 12:43:19 <andythenorth> why did TGP need to make it's own re-imagination of tropic landscape? 12:43:48 <TrueBrain> basically, max-height of 250 doesn't give you any hills higher as with max-height of 90 12:43:55 <TrueBrain> landscape will be identical between the two 12:44:02 <andythenorth> original 50:50 rainforest:desert maps look really cool :D 12:44:13 <frosch123> TrueBrain: anyway, the idea of exposing that setting was to give add-ons an idea above what high/low means, so they can detect hills independent of snow 12:44:32 <TrueBrain> frosch123: that would only have made sense if it influences the height TGP worked with :) 12:44:35 <frosch123> stuff like variable-snow-line define snow height as percentage of max-height 12:44:36 <TrueBrain> I would have fully understand that 12:44:42 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8576: Feature: Allow GameScripts to add additional text to Industry view window https://git.io/JtTo2 12:44:59 <TrueBrain> but as said .. no matter what max-height is, everything above 90 generates the same TGP terrain :) 12:45:12 <TrueBrain> so it seems something got lost between idea and implementation :P 12:45:25 <frosch123> TrueBrain: i mean the setting mixes two things: (1) it affecs the map generation somehow. (2) after map generation it affects what is considered high/low 12:45:36 <andythenorth> 8576 might be interesting in future :D 12:45:41 <frosch123> you mostly talk about (1), i talk about (2) :p 12:45:58 <TrueBrain> frosch123: sort of .. my remark is that I would assume that 1 and 2 worked in tandum 12:46:04 <TrueBrain> but 1 works completely independent of 2 12:46:09 <TrueBrain> like .. TOTALLY different :) 12:46:42 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #7744: Draft Feature: OpenGL video driver https://git.io/JtTow 12:47:04 <TrueBrain> with my patch they do effect each other more btw 12:47:21 *** DasPoseidon has quit IRC 12:47:22 *** DasPoseidon has joined #openttd 12:47:37 <michi_cc> TrueBrain: More Height Levels was one of those classical "very well tested" patches. 12:48:09 <TrueBrain> michi_cc: okay, that does help, as that means I don't have to be like: huh? but just assume it was forgotten 12:48:28 <TrueBrain> as I think if we link 1 and 2 closer together, it becomes much more sensible :) 12:49:46 <frosch123> wut, what does "|+" mean in a mediawiki table? 12:50:06 <TrueBrain> I suppressed any knowledge of wikitext 12:50:58 <andythenorth> is 'very well tested' a synonym for 'not tested at all'? 12:51:34 <andythenorth> frosch123 does that span the next td? 12:51:53 * andythenorth pure guessing, wikitext makes me want to delete my internet connection 12:52:16 <TrueBrain> right, lets implement SDL+OpenGL in a proper fashion .. 12:52:38 <frosch123> andythenorth: in this case it was "the user probably meant |-, so lets behave like that" 12:52:49 <andythenorth> "le oof' 12:54:36 <andythenorth> https://user-images.githubusercontent.com/1780327/104812425-f55e4000-57f9-11eb-9afb-33dfe8537297.png 12:54:38 <andythenorth> is nice 12:55:27 <michi_cc> andythenorth: I would translate it more like "it is working very well for exactly the things I do". 12:55:49 <TrueBrain> michi_cc: didn't you mention we shouldn't be using double buffer for OpenTTD? As you enable it for win32? (just asking the question :D) 12:55:55 <andythenorth> we should do a lexicon in contributing.md 12:56:22 <andythenorth> my most recent adventure into 'very well tested' and 'solved' was templates in JGRPP :) 12:56:36 <andythenorth> which it turns out can arbitrarily delete other trains 12:58:15 <andythenorth> maybe tropic is matter of personal taste, but I find this boring to look at and play https://user-images.githubusercontent.com/1780327/104812483-6dc50100-57fa-11eb-9a86-7355fa35fbc1.png 12:58:40 <andythenorth> I supposed the original baseset icon is desert though, on map gen screen 12:58:56 *** nielsm has quit IRC 13:01:09 <andythenorth> on height level 255 with TGP, I think it's actually quite cool, there's proper high desert plateaus etc https://user-images.githubusercontent.com/1780327/104812532-d14f2e80-57fa-11eb-82ee-bc875643794f.png 13:01:29 <andythenorth> not gonna get many farm-type industries though :D 13:03:04 <michi_cc> TrueBrain: You seem to be right. See how good my memory is :) 13:03:11 <andythenorth> original on 255 height level, quite good mangrove-swamps https://user-images.githubusercontent.com/1780327/104812587-1d9a6e80-57fb-11eb-9d5d-2cdcde936370.png 13:03:28 <TrueBrain> michi_cc: well, I mostly wonder if I should enable or disable it, if I can find out how that works :P 13:03:42 <michi_cc> TrueBrain: Funnily, I just tested without double buffering, and is in fact slower even if, thanks to Windows Aero, we don't actually draw to the real screen anywhere. 13:03:57 <TrueBrain> lol 13:04:10 <michi_cc> So I guess enable it, unless Linux does things very differently. 13:04:19 <TrueBrain> michi_cc: with some minor refactoring in Win32OpenGL, I can share that between SDL and Win32 .. is that wanted? 13:04:31 <TrueBrain> would mean it becomes a Mixin, I guess 13:05:05 <TrueBrain> and with minor, I mean, I need a function for SwapBuffer and for UpdateSurface, which Windows does different 13:05:56 <TrueBrain> well, let me start with copy/pasting it, we can refactor after 13:05:57 <TrueBrain> easier :) 13:07:28 <michi_cc> Also, SDL is 99% for Linux and similar, not sure if everything I did in Win32 really holds there. 13:07:56 <TrueBrain> 80% of your Win32OpenGL functions I had to copy/paste :P 13:08:02 <TrueBrain> as they deal with setting up the OpenGL backend properly etc :) 13:09:50 <TrueBrain> haha, you skip palette animation during FF :D Not sure that is a bad thing .. :P 13:10:22 <DorpsGek> [OpenTTD/OpenTTD] ldpl updated pull request #8576: Feature: Allow GameScripts to add additional text to Industry view window https://git.io/JtUMd 13:10:44 <LordAro> TrueBrain: i can see that being useful in an additional "faster forward" mode 13:11:27 <andythenorth> frosch123 is this extra 'scroll to' button intended on vehicle windows? :) https://user-images.githubusercontent.com/1780327/104812751-453e0680-57fc-11eb-9294-2e05197237d3.png 13:11:52 <LordAro> andythenorth: that goes to the station, doesn't it? 13:11:53 <michi_cc> TrueBrain: Huh, were am I skipping this? Must have been always in win32, because I can't remember changing anything there. 13:12:04 <andythenorth> oh it does? 13:12:09 <TrueBrain> let me rephrase: you are doing it on WM_PAINT events, not on every tick 13:12:13 <andythenorth> did there used to be an eye doing that? 13:12:15 *** Progman has quit IRC 13:12:15 <LordAro> andythenorth: i've no idea, what's what i'd expect 13:12:19 <andythenorth> don't think I've ever used that button 13:12:21 <LordAro> that's* 13:12:53 <michi_cc> Well, yes, animating the palette outside painting is a bit pointless. 13:13:09 <TrueBrain> not disagreeing at all :D 13:13:09 <andythenorth> no button there in 1.10.2 13:13:13 <andythenorth> what does JGRPP do? 13:13:42 <TrueBrain> just not what SDL seems to be doing currently :) 13:13:45 <LordAro> andythenorth: ctrl click or something? 13:14:11 <frosch123> andythenorth: previously "scroll to destination" was "ctrl+click on status bar". i freed that usage of "ctrl" by adding a button that behaves like the other buttons 13:14:43 <michi_cc> TrueBrain: It's not my idea though, it was always this way in GDI PaintWindow 13:14:45 * andythenorth tests 13:14:54 <andythenorth> oh and ctrl-click gets a new viewport 13:14:58 <TrueBrain> michi_cc: well, I take the compliment back in that case! 13:14:59 <TrueBrain> :P 13:15:12 <frosch123> andythenorth: did i add a tooltip? 13:15:24 <andythenorth> yup 13:15:30 <frosch123> no idea how useful the extra viewports are. i never use them. but at least it is consistent with the rest 13:15:56 <andythenorth> I'm potato / potato, but at my current UI zoom / font zoom / baseset combo, the vertical alignment is off 13:16:26 <frosch123> andythenorth: true, but that is the fault of the text 13:16:36 <andythenorth> yeah I'm wondering if I can fix it 13:16:39 <frosch123> the text in the title bar is one pixel lower 13:17:04 <andythenorth> I have long-standing idea to put speed at start of destination string 13:17:09 <andythenorth> as the string is often truncated 13:17:32 <andythenorth> wondering if speed should be split to a separate string? 13:17:43 <andythenorth> next to 'stop/start' 13:17:52 <frosch123> was the setting to show the speed in the status bar deleted? :p 13:18:05 <frosch123> otherwise you can add 4 more options to it :p 13:18:19 <andythenorth> nothing in UI settings for it 13:18:22 * andythenorth looks in JGR 13:18:51 <frosch123> i think a separate widget for the speed would waste space 13:19:00 <frosch123> it's no important when vehicle is stopped 13:19:06 <frosch123> and how often do you look at the speed? 13:19:20 <frosch123> only to check whether it reaches its top speed 13:19:35 <andythenorth> bridges 13:19:47 <andythenorth> not often, but when I want it, I want it 13:20:11 <andythenorth> separate widget I am -1 though 13:20:53 <andythenorth> hmm, does a stopped train in depot need 'ignore signals'? 13:21:03 * andythenorth general vehicle UI questions 13:21:07 <DorpsGek> [OpenTTD/OpenTTD] ldpl commented on pull request #8576: Feature: Allow GameScripts to add additional text to Industry view window https://git.io/JtTK7 13:21:41 <andythenorth> frosch123 generally removal of all those 'location' buttons is great 13:21:54 <andythenorth> my muscle memory for the UI is temporarily broken, but there's just less...noise 13:22:44 <frosch123> andythenorth: https://gist.githubusercontent.com/frosch123/889b6478a6d5d7b10f89d9b014845331/raw/9c965f70b28c599a273ce3809bcdb5b35d50484c/trunc_string_command.png <- probably 10 years old by now :p 13:23:27 <frosch123> andythenorth: two of the buttons change behavior when inside depot 13:23:51 <andythenorth> I have an ulterior motive for that button...for vehicle livery obvs. :P 13:23:58 <frosch123> in theory there is now space to add a button to directly go to the timetable-window, instead of ctrl+orders, but they seem to be unpopular 13:25:11 <andythenorth> frosch123 that truncation seems to valid, is it actually a setting in .cfg? 13:25:16 <andythenorth> to be * 13:26:08 <frosch123> https://github.com/frosch123/OpenTTD/commits/feature_truncate_stringcode <- it's from 2013, good luck syncing that. the string rendering probably changes twice :p 13:26:26 <_dp_> I know 2 usecases for extra viewports - for fast hq placement in citybuilder and for converting stuff on the entire map :p 13:26:29 <andythenorth> I found vehicle_speed cfg item, wonder what that does :P 13:28:27 <frosch123> hmm, no idea what is in that branch i linked. it looks like half is missing :p 13:29:18 <andythenorth> I did do half of the PR for moving speed to start of string 13:29:24 <andythenorth> then I found....extra cases 13:29:33 <LordAro> ooh, "AltSFX" a thing 13:29:36 * LordAro tries it out 13:30:23 * andythenorth playing the game 13:30:32 <andythenorth> forgot all the coverage stuff got added :) 13:30:38 <andythenorth> and the station areas got fixed :) 13:30:46 <andythenorth> so many good changes last 2 years 13:31:40 * LordAro too 13:31:43 <LordAro> giving tropical a try 13:36:05 <andythenorth> so what should 'livery' icon look like on vehicle window? 13:36:17 <DorpsGek> [OpenTTD/OpenTTD] ldpl commented on pull request #8573: Add tile parameter for GSCompany.ChangeBankBalance to show text effect if needed https://git.io/JtT6X 13:36:18 <andythenorth> does it need to look distinct from 'company colour' icon in groups window? 13:41:25 <andythenorth> would we ever do a company colour per vehicle? 13:42:13 <michi_cc> It's not exactly a 'company' colour then anymore, is it? :D 13:43:30 *** iSoSyS has joined #openttd 13:52:54 <andythenorth> it's more like...paint the vehicle 13:54:46 <TrueBrain> michi_cc: you already cherry-picked parts of my work I see? (fixes in opengl.cpp) :D 13:54:48 <TrueBrain> how nice :) 14:03:31 * andythenorth thinks 'set the company colour per vehicle' isn't going to meet the needs of realism newgrfs :P 14:03:32 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8578: Fix: Stopped ships shouldn't block depots https://git.io/JtTiS 14:07:22 <Eddi|zuHause> andythenorth: you can just make a group with only 1 vehicle in it... 14:07:36 <andythenorth> can't make a group per wagon though 14:07:48 <andythenorth> they tend to be attached to trains 14:08:09 <andythenorth> that might be an 'interesting' alternative way to do templates though, vehicle group per vehicle in the consist 14:08:10 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8576: Feature: Allow GameScripts to add additional text to Industry view window https://git.io/JtTiA 14:09:14 <andythenorth> so current icon ideas for livery 14:09:30 <andythenorth> rainbow unicorn; nyan cat; settings cog 14:09:35 <andythenorth> (discord helped) 14:12:21 <DorpsGek> [OpenTTD/OpenTTD] ldpl updated pull request #8576: Feature: Allow GameScripts to add additional text to Industry view window https://git.io/JtUMd 14:13:02 <frosch123> does nml allow variables to start with digits? 14:13:28 <Eddi|zuHause> no 14:13:46 <frosch123> can i delete "256_persitent_registers" then? 14:14:09 <Eddi|zuHause> wtf is that? :p 14:15:03 <LordAro> i feel like i need a gamemode where terraforming is actually prohibitively expensive, or even impossible 14:15:38 <michi_cc> TrueBrain: Yes, I took care of your OCD things :) 14:16:00 <TrueBrain> I have many more, don't worry :) 14:17:32 <andythenorth> frosch123 where even is that? :) I don't find it in nml src or wiki 14:17:59 <andythenorth> oh config flags, missed the missing s 14:18:16 <frosch123> orry :) 14:18:30 <andythenorth> happen to all of u 14:18:50 <_dp_> LordAro, can't you just set terra limit? 14:20:35 <LordAro> _dp_: mm, i suppose i could 14:20:43 <LordAro> basically, i think i want "difficulty presets" :) 14:22:36 <andythenorth> GS :P 14:23:00 <_dp_> I'd vote for general game modes like creative(infinite money and magic bulldozer), realistic(all the crap), etc... 14:23:06 <LordAro> GS can't control game settings :p 14:23:14 <LordAro> oh, definitely want a proper sandbox mode too 14:23:16 <_dp_> can'tit? 14:23:24 <LordAro> i wouldn't have thought so... 14:23:51 <LordAro> besides, a GS for just changing a few settings is massive overkill 14:24:04 <_dp_> prety sure I seen some gs do that 14:25:08 <_dp_> LordAro, not as massive as newgrf for changing settings :p 14:26:13 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver https://git.io/JeO8B 14:26:21 <michi_cc> Pushed a few more potential OCD triggers :) 14:27:10 <LordAro> _dp_: potentially 14:27:35 <LordAro> basically, i want an overhaul of the "new game" environment 14:29:04 <TrueBrain> michi_cc: but but .. THE CONFLICTS! :D (I am kidding, it is fine :D) 14:29:29 <TrueBrain> michi_cc: https://github.com/OpenTTD/OpenTTD/compare/master...TrueBrain:opengl-sdl <- I am taking the slow path, making SDL2 a lot more like Win32, a few lines of code at the time 14:29:38 <TrueBrain> while at it, I find Win32 that I think improves the code 14:29:43 <TrueBrain> this will take a few days to get done :) 14:30:05 <andythenorth> I am considering Free Horse 14:30:07 <andythenorth> for sandbox mode 14:30:17 <andythenorth> bored of players asking for 'cheaper' 14:30:24 <andythenorth> I have made Horse cheaper twice 14:30:54 <andythenorth> with daylength, inflation on, realistic timetables, and trains that only run like once a year 14:30:57 <andythenorth> they need 'free' 14:31:18 <andythenorth> sandbox mode wishlist needed 14:31:44 <_dp_> imo basecosts should just be a setting 14:31:46 <frosch123> i thought the basecost grf was a solved thing 14:31:55 <_dp_> with ability to adjust vehicle prices, cargo incomes, etc 14:32:03 <TrueBrain> michi_cc: in VideoDriver_Win32OpenGL::Paint, you do GetUpdateRect and ValidateRect .. what does this do? Is this Windows way of not redrawing everything? 14:32:31 <TrueBrain> oeh, and this URL is better to see my work: https://github.com/michicc/OpenTTD/compare/opengl...TrueBrain:opengl-sdl 14:32:32 <TrueBrain> :D 14:32:36 <_dp_> frosch123, pretty sure basecosts can't into vehicle prices 14:33:11 <andythenorth> afaik basecost can't set running cost to 0 14:33:19 <andythenorth> only tiny tiny tiny fractional multiplier 14:33:23 <andythenorth> 0 is demanded imho 14:33:30 <andythenorth> it's a one line change in Horse though 14:33:47 <andythenorth> basically full trainset mode 14:33:56 <andythenorth> place houses, no town limits, vehicles cost 0 14:34:24 <michi_cc> TrueBrain: The GDI comment is better: /* Mark the window as updated, otherwise Windows would send more WM_PAINT messages. */ 14:34:34 <TrueBrain> ah! k :) 14:35:17 <frosch123> _dp_: not individual vehicles, but per vehicle type 14:35:36 *** Tirili has joined #openttd 14:35:44 <DorpsGek> [OpenTTD/OpenTTD] Milek7 commented on pull request #7744: Draft Feature: OpenGL video driver https://git.io/JtTXk 14:36:19 <andythenorth> GS can't credit you in real-time for cost of vehicles you buy, right? GS has no callbacks.... 14:37:08 <_dp_> frosch123, running costs, not purchase costs 14:37:43 <frosch123> _dp_: base costs 15 to 19 14:37:51 <frosch123> (decimal) 14:43:43 *** fnutt has joined #openttd 14:45:04 <_dp_> hm... I do remember I couldn't adjust them for some reason... 14:45:11 <_dp_> mb because they were in callbacks or smth 14:46:24 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8574: Feature: Ships yield when passing through one another https://git.io/JtTXg 14:50:41 <DorpsGek> [OpenTTD/team] fnutt opened issue #125: [nb_NO] Translator access request https://git.io/JtTXr 14:51:16 <andythenorth> people use bouys for ships :O 14:52:23 <DorpsGek> [OpenTTD/team] fnutt commented on issue #125: [nb_NO] Translator access request https://git.io/JtTXr 14:54:02 <LordAro> andythenorth: how very 2013 14:54:17 <LordAro> it always concerns me when people try to self-approve 14:54:40 <LordAro> like... are you actually going to be good at translating if you can't understand the dorpsgek message? 14:56:07 <fnutt> I can see what you are saying yes. Hehe. Sorry. I was looking at past issues with translator requests and saw /approve was written. 14:56:24 <fnutt> I will in the future read it more carefully. 14:57:00 <Eddi|zuHause> <frosch123> i thought the basecost grf was a solved thing <-- no, basecost grf has an incredibly clunky UI with no explanation what anything means 14:57:13 <frosch123> which of the 10? 14:57:18 <Eddi|zuHause> all of them 14:58:04 <LordAro> fnutt: do you feel the message could be improved at all to make it clearer? 14:59:26 <fnutt> To be honest, the message was very clear and easy to understand. No reasons to change it. 15:00:16 <fnutt> I just wanted to be “quick” and skipped reading the message throughout. 15:00:19 <DorpsGek> [OpenTTD/team] LordAro commented on issue #125: [nb_NO] Translator access request https://git.io/JtTXr 15:00:54 <LordAro> fnutt: hehe, fair enough 15:01:50 <LordAro> well the conversation here has satisfied me enough anyway :) 15:02:54 <fnutt> Thank you. 15:04:55 <LordAro> go fix those 61 strings :) 15:05:01 <frosch123> oi, nml has no PR template... now i have to figure out myself what to write 15:05:37 <TrueBrain> words! 15:06:02 <LordAro> typey typey words 15:06:17 <fnutt> It does scares me a tiny bit to see there's just one other no_NB translator for OpenTTD. 15:06:46 <fnutt> I will do my best. Have to look at what has been translated before to be sure to have some sort of consistency. 15:06:53 <DorpsGek> [OpenTTD/nml] frosch123 opened pull request #183: Add support for recent NewGRF additions https://git.io/JtT1J 15:07:56 <DorpsGek> [OpenTTD/nml] frosch123 commented on pull request #183: Add support for recent NewGRF additions https://git.io/JtT1U 15:08:05 *** nielsm has joined #openttd 15:08:28 <frosch123> fnutt: all translators were reset 6 months ago 15:08:44 <frosch123> there are plenty of languages who have no translator currently 15:09:03 <LordAro> fnutt: well, the other translator only joined in November, and they've not done a *huge* amount since 15:09:14 <LordAro> prior to that was June 2019 15:09:40 <DorpsGek> [OpenTTD/OpenTTD] grossws updated pull request #7851: Change: add support for next/previous railtype global hotkeys https://git.io/JeM3F 15:09:46 *** DasPoseidon has quit IRC 15:10:10 <fnutt> Alright :) 15:10:31 *** supermop_Home_ has joined #openttd 15:10:40 <fnutt> Is there some sort of notification system as to when more strings are in need of translation? 15:11:05 <fnutt> (assuming I translate the 59 strings) 15:11:42 <LordAro> 29 languages (of 63) without translators right now 15:11:45 <LordAro> including en_US 15:11:52 <LordAro> supermop_Home_: you wanna get in on that? ;) 15:12:24 <supermop_Home_> ha what does en_us need? 15:12:28 <frosch123> fnutt: no, all notificatons are way more general. better to set a quarterly reminder to check :) 15:13:05 <fnutt> Here's a dumb question - why at all have en_US and en_GB? The difference is so small. 15:13:10 * fnutt finds a corner to hide ;) 15:13:12 <frosch123> supermop_Home_: https://translator.openttd.org/translation/openttd-master/en_US#Missing <- validation, whether to drop some u, or replace some s with z 15:13:15 <supermop_Home_> andy willfully sneaking in weird British vernacular? 15:13:35 <Eddi|zuHause> fnutt: to shut up all the nitpicks about whether to use GB or US spelling 15:13:38 <frosch123> fnutt: is having en_AU reason enough? :p 15:13:51 <LordAro> fnutt: the differences between en_GB & en_AU are even smaller :p 15:13:51 <supermop_Home_> is there en_NZ? 15:13:53 <fnutt> Fair enough 15:14:09 <supermop_Home_> AU has a few weird bits 15:14:11 <Eddi|zuHause> do we have de_AT yet? 15:14:29 <supermop_Home_> but i don't think they'd show up in the game 15:14:29 <frosch123> austrians can't read 15:14:43 <supermop_Home_> en_IN seems like it would be the most needed 15:14:44 <frosch123> but i like how you follow up en_AU with de_AT :) 15:15:09 <supermop_Home_> as some terminology is very different 15:15:57 <fnutt> If I was better in nn_NO, I could have done that one. :p But I think it's better to not start on it. Sort of the same as en_GB and en_US - very similar, but still different. 15:16:06 <frosch123> supermop_Home_: en_AU has "lollies" instead of "sweets" :) 15:16:50 <frosch123> and fairy floss 15:17:07 <supermop_Home_> i wonder what the AU 'tradie' -ie term for chief executive is 15:17:15 <supermop_Home_> Chairie 15:17:23 <supermop_Home_> Executie 15:17:39 <fnutt> "STR_CONFIG_SETTING_LOCALISATION_UNITS_VELOCITY_GAMEUNITS" -> "Game units (tiles/day)" 15:18:03 <fnutt> What is the context of this one? Not entirely sure what units should be translated to. 15:18:26 <fnutt> tiles per day? huh? 15:18:27 <supermop_Home_> frosch123 they seem to use truck more than brits do 15:18:27 <LordAro> https://ss64.com/locale.html quite a lot of possible variants 15:18:40 <frosch123> fnutt: vehicle speed, km/h, mph, tiles/day 15:18:47 <supermop_Home_> its not all lorry down there 15:18:57 <fnutt> Okey, thanks. 15:19:07 *** DasPoseidon has joined #openttd 15:20:23 <supermop_Home_> i am sure these dozens of Caribbean en_ variants are pretty similar 15:21:08 <fnutt> Assuming the other people has translated correctly, I'm basing my translations on https://translator.openttd.org/translation/openttd-master/nb_NO#Correct 15:21:40 <supermop_Home_> do we have en_CA? 15:22:30 <supermop_Home_> also i wonder what the difference between haw and haw_US is 15:23:04 <DorpsGek> [OpenTTD/nml] frosch123 updated pull request #183: Add support for recent NewGRF additions https://git.io/JtT1J 15:23:16 <Samu> I had an idea 15:23:27 <Samu> more height tables! https://github.com/OpenTTD/OpenTTD/commit/fc002100b4d1cf5ae85f6f727e588c564f89c5ff 15:24:05 <supermop_Home_> LordAro if you want me to take a look at US i can 15:24:29 <Samu> then it just needs adjusting the heights for arctic / desert to make tgp create decent maps 15:25:17 <supermop_Home_> frosch123 you can't use "economic inflation"? 15:26:05 <frosch123> https://newgrf-specs.tt-wiki.net/wiki/NML:General#General_variables <- similar variables are named like that 15:27:33 <LordAro> supermop_Home_: would be nice to have *someone* doing it 15:29:55 *** Tirili has quit IRC 15:38:21 <fnutt> Meh... I disagree with some of the earlier translations. :/ "Last year" has been translated to "i fjor", but in reality it should have been "forrige år" if we are to translate it directly (without too much interpretation). 15:38:49 <fnutt> last=forrige, year=år. 15:39:29 <TrueBrain> if you like, and want to pick the brain of the other translator, you can always open a Discussion inside your language team on GitHub 15:40:04 <Samu> nobody likes my idea? 15:42:23 <fnutt> Or I can "just" fix it myself? 15:42:32 <TrueBrain> sure you can! 15:42:45 <TrueBrain> just be careful it doesn't become a: I want it A, I want it B, between 2 people :D 15:43:40 <fnutt> buzzCraft, are you here with another nick? or someone else speaking norwegian? 15:48:29 <fnutt> http://35.202.151.60:9000/uploads/0fbd5e4e59cf2ac3/Skjermbilde%202021-01-16%20kl.%2016.48.11.png 15:48:54 <fnutt> I clearly see {STRING2}, not {STRING}. :/ 15:49:03 <fnutt> Is this a bug? 15:50:04 <fnutt> Saving it as {STRING}, but the base lang string clearly says {STRING2}. 15:50:43 <Eddi|zuHause> it's more of a misinformation. when the base uses {STRINGn} the translation must use {STRING} instead 15:51:15 <fnutt> Alright 15:53:13 <michi_cc> TrueBrain: I've picked you Win32 cleanups and added another one. I mean, do we really need a Windows 95 specific work-around anymore? 15:53:27 <TrueBrain> it is why I started to remove those comments :) 15:53:28 <Samu> lol, a 64x map using very flat on arctic gets a higher height than the same map on flat 15:53:38 <Samu> you really didn't test this much, did u? 15:55:04 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver https://git.io/JeO8B 15:56:09 *** Flygon has quit IRC 15:56:55 *** Wolf01 has joined #openttd 15:57:12 *** sla_ro|master has joined #openttd 15:57:42 <TrueBrain> slowly pulling this puzzle apart that we call SDL2 :) It is getting close to what Win32 does :D The main difference so far is that on Win32 you can make a DIB before you have a Window .. on SDL this is the other way around :) 16:00:13 *** Wormnest has joined #openttd 16:03:07 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver https://git.io/JeO8B 16:03:12 <michi_cc> Missed a tiny, tiny speck :) 16:07:00 <TrueBrain> I failed to spell usererror correctly 3 times in a row .. that is impressive 16:11:19 <TrueBrain> michi_cc: pushed more Win32 changes you might want to cherry-pick :P It is slowing down ;) 16:11:54 <TrueBrain> just some weird comments, and simplified GetScreenDepth :) 16:13:34 <fnutt> "STR_TREES_MODE_FOREST_SM_BUTTON" -> "{BLACK}Grove" 16:13:34 <michi_cc> TrueBrain: Did you see the comment by Milek7 on the PR regarding SDL_GL_GetProcAddress? Not sure what the proper way there is. 16:13:54 <TrueBrain> haven't looked into that yet, at all. Still working towards re-adding OpenGL to SDL :) 16:13:58 <fnutt> Is this something that groves, or may it grov? 16:14:05 <fnutt> grove* 16:15:43 <fnutt> Never mind it. 16:18:58 <TrueBrain> michi_cc: VideoDriver_Win32GDI::AllocateBackingStore, the "return false" on no change looks a bit odd. Shouldn't that be "return true"? 16:21:04 <TrueBrain> and possibly that usererror should be a return false, to give things a chance to recover? 16:21:08 <michi_cc> True for AllocateBackingStore means "we did something", and it is basically only used in VideoDriver_Win32Base::ClientSizeChanged, because AfterBlitterChange is using the force param. 16:21:55 <TrueBrain> AfterBlitterChange also uses it 16:22:00 <TrueBrain> which, if returned false, reverts 16:22:05 <DorpsGek> [OpenTTD/team] CodCannon opened issue #126: [pt_PT] Translator access request https://git.io/JtTQf 16:22:20 <TrueBrain> so it also seems to indicate "something failed"? 16:22:29 <TrueBrain> owh, force=true 16:22:39 <TrueBrain> so why is it in the "return" value there :D 16:22:39 <michi_cc> There's nothing to recover from, it means the user has somehow managed to start the win32 video driver together with the null blitter. 16:23:28 <TrueBrain> another path is via ChangeResolution, where a "return false" of AllocateBackingStore indicates there was a failure 16:23:43 <michi_cc> No idea, it is all old code. 16:23:44 <TrueBrain> (via MakeWindow) 16:23:48 <TrueBrain> haha :D 16:24:18 <TrueBrain> SDL2 is newer, but has similar weird "double meaning" results :P 16:24:31 <michi_cc> Maybe there used to be a second failure mode in AllocateBackingStore way back. 16:25:24 <TrueBrain> I was mostly wondering if "return false" would be fine for failures, as that function can also do a usererror 16:26:14 <TrueBrain> fuck it, I am going for the usererror 16:26:18 <TrueBrain> if these functions fail, something is really wrong 16:27:32 <TrueBrain> similar with "VideoDriver_Win32OpenGL::AllocateBackingStore" I guess, if gl_rc is not set .. that should never happen, should it? 16:27:55 <TrueBrain> (sorry for all these questions .. it is just easier to talk about them, then to look at it for hours and still failing to understand them :D 16:33:00 <michi_cc> Some of that code was probably really written with old, old systems in mind where e.g. higher resolutions might not be available in 32bpp and so on. Or assuming like 4MB RAM or so. 16:35:07 <michi_cc> TrueBrain: The very first call to VideoDriver_Win32OpenGL::AllocateBackingStore has gl_rc == nullptr :P We can't create the context without a window, and window creation sends a resize message. 16:35:49 <TrueBrain> so you have the same issue as I have with SDL .. I have to change the order of these statements :) 16:35:56 <TrueBrain> Create Window first, AllocateBackingStore after :) 16:37:08 <TrueBrain> but Windows makes it harder on you because of the event 16:37:11 <TrueBrain> yeah, makes sense 16:37:22 <TrueBrain> Video-driver code is most often just one huge mindfuck someone went through once :) 16:37:27 <michi_cc> That's at least how it works on Windows. And as messages flow though Base, adding the check in VideoDriver_Win32OpenGL::AllocateBackingStor was the easiest. 16:40:52 <fnutt> https://translator.openttd.org/translation/openttd-master/nb_NO is now complete. :) 16:41:13 <fnutt> Enjoy! 16:41:25 <TrueBrain> gratz! And tnx :) 16:49:15 <fnutt> 53 PR's on github. uhm.... I see there's some dating back to march 2019. May I ask why they are still hanging around? 16:51:03 <Eddi|zuHause> each one for different reasons 16:51:25 *** Gustavo6046_ has joined #openttd 16:54:56 <andythenorth> because people get sad if I just close them? 16:55:01 *** glx has joined #openttd 16:55:01 *** ChanServ sets mode: +v glx 16:55:27 <andythenorth> Samu what height level for the 64x test? 16:55:44 *** Gustavo6046 has quit IRC 16:55:44 *** Gustavo6046_ is now known as Gustavo6046 16:56:06 <andythenorth> "you really didn't test this much, did u?" no :) 16:56:24 <Samu> 30 16:58:28 <Samu> i'm filling the tables using default values 16:58:39 <andythenorth> how's your combinatorial maths samu? 16:58:43 <Samu> there's some weird results 16:58:52 <andythenorth> mine is not good, I can never remember when to use factorial, npr or ncr 16:59:06 <Samu> { 13, 12, 12, 16, 19, 28, 40 }, ///< Hilly, sub-arctic 16:59:28 <Samu> height for 64 is higher than 128 and 256 16:59:42 <andythenorth> for the same random seed? 17:00:25 <Samu> no, this is the max_height from table, after TrueBrain changes, using default snow height and default max height 17:00:34 <andythenorth> if you want anything other than anecdotal results, you need a script 17:00:37 <andythenorth> do you have a script? 17:00:53 <andythenorth> you could probably patch the game to log, and feed it all the possible combinations of map 17:01:07 <andythenorth> using a stable set of random seeds 17:01:14 <andythenorth> otherwise it's just guessing 17:03:07 <Samu> still need to fill the data for mountainous and alpinist 17:03:22 <Samu> then i'll post the end result 17:04:25 <andythenorth> are you testing with 0% water? 17:06:41 <Samu> im not testing water, i'm currently not looking at the end result 17:07:07 <andythenorth> if you don't have a stable seed, and you don't prevent water 17:07:11 <andythenorth> then the results are junk 17:07:28 <andythenorth> presence of water can limit possible height level 17:07:47 <andythenorth> how many samples are you doing per combination of settings? 17:09:38 <DorpsGek> [OpenTTD/OpenTTD] fnutt opened issue #8580: Norwegian bokmål and nynorsk https://git.io/JtT5Z 17:13:49 <Samu> 1, since I'm only looking at max_height_from_table 17:14:25 <andythenorth> and generating one map only? 17:14:28 <andythenorth> but always from same seed? 17:15:09 <Samu> i think i didn't explain my self too well 17:15:35 <andythenorth> you're logging some output from the TGP function? 17:15:45 * andythenorth makes wild guesses :| 17:17:19 <andythenorth> or we're just reading TGPGetMaxHeight? 17:18:50 <Samu> I am logging the results of max_height_from_table at the last line, https://github.com/OpenTTD/OpenTTD/blob/0a129bc810cddb9f7b684772157ac6e8cbab8297/src/tgp.cpp#L228-L258 17:19:08 <Samu> and i'm testing with default snow line height, 10, and default maxheightlevel 30 17:19:24 *** snail_UES_ has joined #openttd 17:19:36 <Samu> and doing it for all combinations possible, mapsize, terraintype 17:19:55 <Samu> i'm filling a table with the results for arctic and tropical 17:24:23 <FLHerne> fnutt: " if we are to translate it directly (without too much interpretation)" <-- that's not inherently a concern, you should use your language's normal idiom for the in-game meaning 17:26:53 <fnutt> 👍 17:27:17 <FLHerne> Literally translating what the English text says can go wrong 17:27:44 <FLHerne> Particularly because some of the English texts are badly-worded or misleading 17:28:19 <fnutt> Absolutely. Sentence structure in english and norwegian isn't 100 % the same. 17:37:44 <Samu> well, mountainous large maps lost a lot of height 17:42:13 <Samu> well, it would still be capped at 30, but some values i'm getting are in the 20's for alpinist on a 2048x sized map 17:42:58 <Samu> wondering if I should test desert with 255 maxheightlevel 17:43:14 <glx> fnutt: both nynorks and bokmal do exist for openttd 17:44:58 <Samu> only one value remains... alpinist tropical 4096x 17:47:06 <fnutt> Uhm.. you are right.... but... it is not listed in https://translator.openttd.org/translation/openttd-master/nb_NO ? 17:47:37 <glx> https://translator.openttd.org/translation/openttd-master/nn_NO 17:48:17 <fnutt> Sure sure, but... something ain't right here since there are no language string for it? 17:49:40 <glx> STR_NETWORK_LANG_NORWEGIAN is just for network games IIRC, it's to indicate the spoken language on the server 17:49:42 <fnutt> glx: Is your username on github glx22? 17:49:46 <glx> it is 17:50:10 <fnutt> Alright. That makes sense. 17:50:48 <glx> and yes translating without the context is not easy :) 17:50:52 <fnutt> Based on this I guess I can close the issue. Thanks for the help. 17:52:08 <glx> at least we try to use descriptive string IDs, but it's not always enough 17:53:36 <DorpsGek> [OpenTTD/OpenTTD] fnutt closed issue #8580: Norwegian bokmål and nynorsk https://git.io/JtT5Z 17:54:27 <glx> oh you could have left the original message I think 17:54:47 <fnutt> I could, but it was just stupidity by me. Wanted to remove it. 17:55:30 <glx> it was a good question, not knowing where the string was used 17:58:52 <Samu> TADA!! https://github.com/OpenTTD/OpenTTD/commit/61c745223757a9c9f5565e1d0b63ebb3fac2605a 17:59:20 <fnutt> Zero translators for nn_NO. *sight*. I'm not qualified to do the translations correctly. 355 strings in need of translation. :/ 17:59:45 <fnutt> *sigh* 18:00:15 <glx> yeah, some languages don't have translators 18:02:47 <Samu> considering max_heightlevel, any value above 30 is capped to 30 18:06:00 <LordAro> fnutt: please don't blank issue text, the history is useful 18:06:01 *** Gustavo6046_ has joined #openttd 18:06:24 <LordAro> and it's in the history anyway 18:06:27 *** Gustavo6046 has quit IRC 18:06:27 *** Gustavo6046_ is now known as Gustavo6046 18:06:39 <LordAro> just a comment explaining why it's invalid would do fine 18:07:51 <Samu> i think i need to test extreme values, like a retarded snow line height of 253, max height level of 255 18:08:05 <Samu> need to get a bigger picture of the issue 18:15:53 <Samu> wow, this looks broken 18:16:32 <Samu> set a snow height of 253 and a max height of 255 18:16:43 <Samu> now generate a 4096x4096 sub-arctic very flat 18:17:24 <Samu> it looks.... worse than alpinist 18:18:09 <andythenorth> lol tree generation time on 4096x4096 18:18:18 <andythenorth> trees are so stupid, they should just be procedural :) 18:18:34 * andythenorth 1 min later, map gen finishes 18:18:40 <andythenorth> nope, still going 18:18:51 <andythenorth> 128000000 individual trees to place 18:19:03 <andythenorth> instead of 256 pre-computed variants 18:19:27 <Samu> there were no suitable bla bla for forests :p 18:20:07 <andythenorth> same you make a very good case here for the removal of More Height Levels 18:20:15 <Samu> hide the trees 18:20:46 <Samu> it's not "very flat" at all 18:20:57 <Samu> it's "to the moon" height 18:21:00 <andythenorth> ok seems to me 253 snowline and 255 max height is working correctly 18:21:12 <andythenorth> the map is ugly AF and I wouldn't want to play it 18:21:15 <andythenorth> but it's correct 18:21:24 *** Gustavo6046 has quit IRC 18:21:40 <TrueBrain> did you just really say: I entered some insane values and am disappointed in how it looks? :D 18:21:53 <andythenorth> it's working correctly 18:22:00 <TrueBrain> guess a nice example of why adding settings for everything is silly :P 18:22:22 <andythenorth> this is a matter of how people's brains are ordered 18:22:34 <andythenorth> if player sets max height 255, surely they want 255 no? 18:22:52 <Samu> map generated no snow 18:22:53 <andythenorth> but now these reports are that map should be both 255 high, AND very flat 18:23:33 <andythenorth> wait you might be right 18:23:45 <andythenorth> 4096x4096 is really so slow 18:23:53 <andythenorth> but demand is for 16kx16k 18:24:37 <Samu> honestly, it should try to at least make it look "very flat" 18:25:09 <andythenorth> what, a 255 high plateau? 18:25:24 <TrueBrain> andythenorth: but their machine can handle it !!!!!11111oneone 18:25:33 <andythenorth> TrueBrain it's very well tested 18:25:36 <Samu> it makes no difference which terraintype is set 18:25:45 <andythenorth> Samu confirmed no snow 18:25:48 <andythenorth> ok revert PR pls 18:26:25 <andythenorth> this really is cool AF, TGP has made incredibly ugly very high arctic maps 18:26:33 <andythenorth> with ugly ugly slopes 18:27:07 <Samu> it only made it that high because of the snow line height being set that high, it praticall ignored the "very flat" preset 18:27:17 <Samu> feels wrong 18:27:27 <andythenorth> shrug 18:27:38 <andythenorth> you can argue order of operations forever 18:27:47 <andythenorth> 255 height > snowline > map flatness 18:27:50 <andythenorth> is what's happening 18:27:59 <andythenorth> lol 100 tile long slopes 18:28:47 <TrueBrain> its amazing, you can step out of this channel for hours, and come back to the exact same discussion, not an inch further :) 18:29:01 <andythenorth> mac FFWD on 4096x4096 map, 1x :P 18:29:02 <andythenorth> lol 18:29:03 <glx> anyway using "incompatible" settings you can't expect "working" map ;) 18:29:21 <TrueBrain> if you shoot yourself in the foot .... boy it hurts :) 18:29:40 <andythenorth> TrueBrain can I combine 'rm -r' and '*' ? 18:30:12 <andythenorth> shall I try? 18:30:23 <Samu> but, other than this specific issue, look at the results using default values 18:30:36 <glx> you can andythenorth, but be carefull where :) 18:30:42 <Samu> https://github.com/OpenTTD/OpenTTD/commit/61c745223757a9c9f5565e1d0b63ebb3fac2605a 18:30:48 <TrueBrain> don't forget the f andythenorth ! :D 18:31:38 <Samu> you can see for instance, that 512x size hilly generates higher levels than 512x size mountainous 18:31:58 <andythenorth> glx I was thinking of "cd /" 18:32:02 <andythenorth> does that work? 18:32:26 <glx> hmm probably will require "sudo" too ;) 18:32:36 <andythenorth> I run as root all the time 18:32:39 <andythenorth> be fine right? 18:32:47 <andythenorth> for the record, I don't :P 18:32:54 <Samu> or that a 2048x sized mountainous has higher height than a 4096x sized mountainous 18:33:19 <andythenorth> glx shall we talk about livery spec? :P 18:33:32 *** Gustavo6046 has joined #openttd 18:33:53 <andythenorth> https://gist.github.com/andythenorth/636908e3498197124782c6daf40de6ae 18:36:58 <Samu> I was going to propose a revert on that PR, and use 3 more tables for heights, manually inserting the heights on the table ourselves as we see fit 18:37:11 <Samu> 1 height table for each landscape 18:38:33 <Samu> i also believe very flat shouldn't be too flat for arctic/desert, but when working with extreme values, it outputs very bad maps 18:40:50 <DorpsGek> [OpenTTD/OpenTTD] DorpsGek pushed 1 commits to master https://git.io/JtTx3 18:40:51 <DorpsGek> - Update: Translations from eints (by translators) 18:41:06 <Samu> as for desert vs rainforest in sub-tropical, I have another proposal 18:43:46 <Samu> btw, lowering snow_line_height from 15 to 10 is the best thing out of that PR 18:51:09 <Samu> Must think 18:51:56 <Samu> i think I'll go with the desert/rainforest proposal first 18:52:04 <Samu> will make a PR soon 18:52:41 * andythenorth considers if industries can use liveries 18:52:42 <andythenorth> :P 18:52:45 <TrueBrain> wait, does that mean you closed other PRs?! :o 19:04:46 <andythenorth> I could do it? 19:04:49 <andythenorth> if that helps? 19:13:59 <andythenorth> silent agreement? o_O 19:25:07 <TrueBrain> pretty sure he is silent for other reasons ;) 19:51:14 *** Gustavo6046 has quit IRC 19:58:52 <Eddi|zuHause> wait... so if "livery" means "some user-defined visual choice"... why would industries have those, when there is not even a dialog window to choose an industry layout? 20:04:56 <andythenorth> because we haven't made it yet! 20:14:26 <andythenorth> can you tell these 2 ships apart visually? https://user-images.githubusercontent.com/1780327/104821950-62dc9180-5837-11eb-8d3f-6df43f91aa0c.png 20:14:33 <andythenorth> or do they look the same? 20:14:42 <TrueBrain> one is clearly bigger than the other 20:15:02 <TrueBrain> one has yellow things in the front, white thing around the middle thing, stairs on the right .. 20:15:09 <andythenorth> not just further away? 20:15:12 <TrueBrain> this is a silly game "find the differences", it is pretty obvious 20:15:20 <andythenorth> I have learnt to check 20:15:29 <andythenorth> some people would see those as identical 20:15:45 <TrueBrain> I know nothing about gfx, don't put any value to what I say about it :P 20:15:52 <TrueBrain> it has red, yellow and white 20:15:53 <TrueBrain> yes, the same 20:15:58 <TrueBrain> confirmed :) 20:16:19 <andythenorth> thought so 20:16:25 <andythenorth> 2 holds, one chimney 20:16:30 <andythenorth> they both look like boats! 20:19:43 *** Progman has joined #openttd 20:36:09 *** snail_UES_ has quit IRC 20:36:54 *** gnu_jj_ has joined #openttd 20:41:19 *** gnu_jj has quit IRC 20:42:58 *** Wormnest has quit IRC 21:01:07 *** gelignite has joined #openttd 21:04:07 <andythenorth> lol mac performance 21:04:33 <andythenorth> CPU 1 of 8 cores is busy 21:04:42 <andythenorth> discrete GPU is 50% workload 21:05:04 <andythenorth> World of Tanks Blitz (3D game using Metal API) is supposed to run at 60fps 21:05:14 <andythenorth> it runs at 30fps with all settings on 'low' or 'ultra low' 21:06:10 <andythenorth> GPU is 4GB Radeon pro 5500M 21:06:44 <glx> so mobile version 21:06:49 <andythenorth> yair 21:07:01 <andythenorth> iPad runs the same game at 60fps with most effects on 'high' 21:07:10 <TrueBrain> michi_cc: it is funny I run in the exact same issues as Win32 .. in this case: AllocateBackingStore is called before the context exists :D It aligns pretty nicely :) 21:07:47 <andythenorth> ipad screen is half the pixels at best 21:07:48 <michi_cc> Well, SDL on Windows at least has to do exactly the same stuff :) 21:09:15 <TrueBrain> hihi, my screen flickers black sometimes .. pretty sure there is an issue with the palette :D 21:10:59 * andythenorth considers writing Apple a stern note 21:11:51 <TrueBrain> owh, not even the palette ... just not drawing to the screen, lol 21:12:33 <TrueBrain> ah, calling SDL_UpdateWindowSurface is a bad idea :D 21:13:58 *** WormnestAndroid has quit IRC 21:13:59 <TrueBrain> 8bpp-optimized is a lot quicker than 40bpp-anim .. was that expected, I forgot .. me checks PR 21:14:04 *** WormnestAndroid has joined #openttd 21:14:47 <TrueBrain> ah, no, nevermind 21:14:49 <TrueBrain> depends on what is on screen 21:15:04 <glx> 8bpp being faster seems correct to me 21:15:49 <TrueBrain> performance of both is within the noise :P 21:16:30 <glx> and I guess palette animation with opengl can use GPU, while other drivers are all cpu 21:20:51 <TrueBrain> michi_cc: found a bug .. if you change resolution with win32-opengl, it crashes :D (not dragging your Window, selecting another resolution!) 21:21:07 <TrueBrain> I noticed, as it crashes SDL too :P 21:21:36 <michi_cc> Hmm, will test :) 21:21:46 <glx> ah not manual resize, but changing in option window ? 21:21:49 <TrueBrain> yeah 21:21:57 <TrueBrain> seems to be called in a weird moment :) 21:23:03 <michi_cc> Not crashing for me :( 21:23:15 <TrueBrain> ah, AllocateBackingStore does a ReleaseVideoPointer but not a GetVideoPointer if it was released 21:23:38 <TrueBrain> michi_cc: with 40bpp-anim btw 21:23:44 <TrueBrain> it crashes because anim_buffer is NULL, sort-of 21:25:46 <michi_cc> Ahh, I see the problem. OpenGLBackend::Resize is not mapping the anim buffer if persistent mapping is not used. 21:25:50 <michi_cc> Fix incoming shortly. 21:26:21 <TrueBrain> sure! 21:27:25 <TrueBrain> right, figured out what functions should be class-based .. now I only have to clean up the palette functions for the non-opengl part .. 21:27:31 <TrueBrain> palette is a bit of a mess in SDL :) 21:29:37 <TrueBrain> on 2k resolution, video output goes from ~1.4ms to 0.14ms (SDL vs SDL-OpenGL) 21:29:37 <TrueBrain> lol 21:29:53 <TrueBrain> (in 32bpp mode) 21:29:59 <glx> nice improvment 21:30:35 <TrueBrain> in 8bpp it is even worse: from 3ms (SDL) to 0.03ms (SDL-OpenGL) 21:30:36 <TrueBrain> haha 21:31:18 <milek7_> that's not counting gpu time, though 21:31:49 <glx> but that means cpu has more free time to do game state stuff 21:33:39 <TrueBrain> let me put it in stats that give less weird responses: my ff FPS goes from ~150 fps to ~1500 fps with 8bpp; same map, same everything 21:34:07 <TrueBrain> and full animation on ofc, as fuck yeah 21:35:16 <glx> seems cpu uses the free time :) 21:35:18 <TrueBrain> this makes fullscreen a lot more playable for me :D 21:35:52 <TrueBrain> but okay, my machine is not really low-end or anything (whistles innocent) .. so we will have to see how it hold up on other hardware :D 21:36:01 <TrueBrain> first, we need to finish it :) 21:37:57 *** DasPoseidon has quit IRC 21:38:11 <TrueBrain> there is a weird blob of code in the SDL driver that Win32 doesn't have .. hmmm 21:40:00 <TrueBrain> ah, via Windows WM_PAINT trigger a paint 21:40:03 <TrueBrain> SDL just does it .. often 21:42:58 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver https://git.io/JeO8B 21:43:12 <michi_cc> Look at VideoDriver_Win32OpenGL::AllocateBackingStore again now. 21:44:05 <TrueBrain> ha, I wrote a near identical fix there for SDL to try it out :D 21:44:42 <TrueBrain> means I am starting to understand this stuff a tiny bit :P 21:45:09 <michi_cc> TrueBrain: I don't have the exact number offhand, but even with the Intel GPU on my laptop, I still have a very noticeable speedup (x3 or something like that I think). 21:45:19 <TrueBrain> what amuses me, is that std::max(a, 1) throws an error with a being an uint .. like ... can't you figure this out compiler? Really? 21:46:03 <TrueBrain> hmm, random crash 21:46:49 <TrueBrain> nope, still crashing on me michi_cc :( 21:47:06 <TrueBrain> in very weird places :P 21:47:31 <michi_cc> Then SDL seems to do something differently. 21:47:38 <TrueBrain> no, with win32-opengl 21:48:09 <TrueBrain> "openttd.exe -snull -mnull -ddriver=4 -r 1024x768 -vwin32-opengl -x -c test.cfg -b 40bpp-ani" <- where "test.cfg" does not exist 21:48:30 <TrueBrain> then Game Options, screen resolution, 2560x1440 21:48:40 <TrueBrain> lower resolutions do work! 21:48:52 <michi_cc> Persistent buffer mapping or not? 21:48:56 <TrueBrain> yup 21:50:07 <TrueBrain> does work on SDL :o 21:50:18 <TrueBrain> (with your new AllocateBackingStore ofc) 21:50:30 <michi_cc> Works for me, with the exact args as you. 21:51:16 <TrueBrain> btw, something completely different: is there a difference between "this->VideoDriver_Win32Base::Stop" and "VideoDriver_Win32Base::Stop" ? 21:51:53 <glx> none IIRC 21:52:04 <TrueBrain> similar to how you can leave out "this->" in many places, I assumed? 21:54:03 <michi_cc> Yeah, our usage of this-> is completely superfluous, unless you really have a local variable shadowing a member variable. 21:54:19 *** Samu has quit IRC 21:54:49 <TrueBrain> persistent mapping or not, crashes on highest resolution 21:55:41 <glx> using this-> in most case is to know with call a member function and not a global scope one 21:55:47 <TrueBrain> okay, going to full screen was a huge mistake 21:55:54 <TrueBrain> crashes hard :P 21:57:04 <TrueBrain> seems random memory corruption 21:57:48 * glx is waiting for symbols, iptv tends to steal bandwidth 21:57:53 <TrueBrain> starting with "-r 3000x3000" produces the same result 21:59:17 <TrueBrain> SDL places the window in a funny place on that resolution :D 21:59:27 <TrueBrain> a problem for another time .. pretty sure it is frosch123 's code doing that :D 22:00:09 <TrueBrain> michi_cc: I have a hard time tracing this issue .. on slightly-lower-than 2560x1440 resolutions, it runs for a bit, but crashes shortly after 22:00:45 <michi_cc> Something is corrupting the heap at very high resolutions. 22:01:04 <TrueBrain> seems so, yes 22:01:57 <TrueBrain> I think it is not inside your opengl.cpp :) 22:05:31 <TrueBrain> I am rebuilding everything just to be sure it is not some weird glitch in the way I do incremental builds :) 22:05:47 <TrueBrain> (I am controlling MSVC via Git Bash ... it sometimes does weird shit) 22:06:12 <milek7_> compile with addresssanitizer 22:06:15 <TrueBrain> too bad, didn't fix anything :P 22:06:20 *** Gustavo6046 has joined #openttd 22:07:21 <michi_cc> No, I can reproduce it, I just don't see any obivous cause. 22:09:14 <TrueBrain> owh, pfew, that makes me feel less alone :D 22:09:29 <TrueBrain> I couldn't spot one either tbh .. not many things are influences by resolution 22:09:43 <frosch123> TrueBrain: i cannot start ottd with a bigger window than the screen 22:09:53 <frosch123> sdl makes it smaller 22:10:07 <frosch123> well, no idea whether sdl, or the window manager 22:10:14 <TrueBrain> frosch123: lol .. it indeed finds the nearest "valid" resolution, which is over my max size :P 22:10:34 <frosch123> oh, you mean fullscreen? 22:10:37 <TrueBrain> no no 22:10:39 <TrueBrain> windowed 22:11:12 <frosch123> sounds like i have to read 5 hours of backlog to understand the issue :p 22:11:17 <TrueBrain> not really 22:11:23 <TrueBrain> I just star tthe game with "-r 3000x3000" 22:11:30 <TrueBrain> SDL makes that a bit smaller 22:11:34 <TrueBrain> and puts the window offscreen :D 22:12:26 <TrueBrain> dbg: [driver] SDL2: using mode 3000x1421 22:12:45 <TrueBrain> but okay, I know where in the code it does this, so I can take a look and see if it is worth fixing, honestly :) 22:12:57 <TrueBrain> it is a very weird thing to do :P 22:14:10 <TrueBrain> dbg: [driver] 2147483428 / 1073741434 <- x/y position of the window :D 22:14:48 <glx> feels wrong 22:15:18 <michi_cc> TrueBrain: Cause found, fix is in thinking phase. 22:15:22 <frosch123> ah, uint 22:15:36 <TrueBrain> michi_cc: do tell? Or I will see the change, also fine :) 22:15:47 <DorpsGek> [OpenTTD/OpenTTD] LC-Zorg commented on pull request #8557: Change: darken the background of all graph to increase contrast https://git.io/JtkkL 22:16:33 <michi_cc> TrueBrain: VideoDriver_Win32Base::ChangeResolution might not resize to w x h, but h - something. VideoDriver_Win32OpenGL::ChangeResolution calls AllocateBackingStore with the original w x h after it. 22:17:13 <michi_cc> I'm not really sure anymore why I have that AllocateBackingStore in there anyway. Either I have to fix the w and h params, or maybe it can go at all. 22:19:01 <DorpsGek> [OpenTTD/OpenTTD] frosch123 opened pull request #8581: Fix: Only center the window, when it is smaller than the screen. https://git.io/JtkkG 22:19:04 <TrueBrain> can't say I understand where it does "h - something", but SDL has this code where the height (as I just showed by accident) can be different from what you request, and because of that "w" and "h" are passed everywhere 22:19:07 <frosch123> TrueBrain: does that help? 22:19:10 <TrueBrain> really annoying, honestly :D 22:19:13 <TrueBrain> let me check! 22:19:37 <TrueBrain> bashing windows whenever you get the chance I see :P 22:20:14 <frosch123> TrueBrain: i had 12 years without windows. now i notice windows bugs all days 22:20:15 <TrueBrain> I thin you dont need to include algorithm btw 22:20:20 <TrueBrain> I think LordAro did something magical there 22:20:35 <michi_cc> I think total window size is limit to screen size, but we assume w x h is _client_ size, not window size, i.e title bar. 22:20:37 <frosch123> i don't know whether they also were there 12 years ago, or whether linux fixed them meanwhile 22:20:40 <TrueBrain> yes, that fixes it frosch123 :) 22:21:14 <TrueBrain> but leaving out that include also compiles fine 22:21:20 *** gnu_jj_ has quit IRC 22:21:32 <TrueBrain> ah, included in stdafx.h 22:21:34 <TrueBrain> that explains :D 22:22:49 <DorpsGek> [OpenTTD/OpenTTD] frosch123 updated pull request #8581: Fix: Only center the window, when it is smaller than the screen. https://git.io/JtkkG 22:23:21 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8581: Fix: Only center the window, when it is smaller than the screen. https://git.io/Jtkkr 22:24:16 *** tokai has joined #openttd 22:24:16 *** ChanServ sets mode: +v tokai 22:31:12 *** tokai|noir has quit IRC 22:31:49 <glx> the difference between win32 and win32-opengl is impressive, fast forward in win32 is around 500fps at max, while it's near 1500fps in win32-opengl 22:32:15 <TrueBrain> yeah, I am really happy with it :) 22:32:25 <glx> and that's using everything autoprobed 22:33:01 <glx> one was launched with -ddriver=2 the other with -ddriver=2 -vwin32 22:33:57 <glx> I think I even see a difference in reactivity during newgrf scan 22:34:24 <glx> I have too many newgrf :) 22:36:14 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8557: Change: darken the background of all graph to increase contrast https://git.io/JtkIJ 22:36:19 <DorpsGek> [OpenTTD/OpenTTD] frosch123 merged pull request #8581: Fix: Only center the window, when it is smaller than the screen. https://git.io/JtkkG 22:36:47 <glx> (my config is amd fx-6100 + nvidia 450GTS) 22:37:29 <TrueBrain> I might have been a bit passive aggressive in that comment :P 22:37:39 <TrueBrain> reporting unrelated bugs in merged PRs .. what could possibly go wrong 22:39:02 <TrueBrain> lets see if this more direct response to what he was trying to get done helps :) 22:41:30 <michi_cc> Okay, besides the crash itself, do we want to allow a total window size larger than the screen? OTTD resolution is used as client (inner) size, not total window size. 22:42:13 <TrueBrain> owh, found out what SDL was doing, it was indeed clamping to my screen size 22:42:19 <TrueBrain> I just have 2x 2560x1440 22:42:20 <glx> a window bigger than the screen is almost unusable 22:42:26 <TrueBrain> so yeah .. 3000x1440 fits easily :P 22:42:44 <TrueBrain> michi_cc: I suggest to do the same for Win32 indeed; clamp to max screen size sounds sane :) 22:42:46 <_dp_> hm, somehow I thought that dark thing was rejected... 22:43:14 <frosch123> michi_cc: dragging a window across two screen works fine in windowed mode 22:43:33 <michi_cc> TrueBrain: Well, that is clamping is exactly why it is crashing right now. 22:43:59 <michi_cc> frosch123: I don't have a second monitor here, but I am assuming the Windows clamping is actually to desktop size, not really monitor size. 22:44:34 <TrueBrain> michi_cc: in SDL, it reads back the real size it got, and sets w/h based on that; something like that possible, to set the _window thingies to the right size? 22:45:49 <michi_cc> Win32 code does that all as well, there's just a second call to AllocateBackingStore that I think is completely useless. But apparently I must have though of something with it, even if I can't see it right now. 22:46:41 <glx> https://docs.microsoft.com/en-us/windows/win32/gdi/multiple-monitor-system-metrics 22:47:11 <TrueBrain> michi_cc: yeah ... one thing I learnt today: there is always this one flow that makes a line of code that is already there have a reason :D 22:49:02 <andythenorth> which FIRS author :P 22:49:13 <glx> shh 22:49:41 <TrueBrain> is there more than one? :P 22:50:14 <glx> there's more than one person who understands FIRS generator ? 22:50:22 <frosch123> andythenorth: i dbout you can proof that you did not say something 22:50:24 <andythenorth> well there are 8 forks 22:50:34 <andythenorth> are they authors? :P 22:50:58 <frosch123> does anyone of them fork the firs source? or do they all take the generated nml and go from there? 22:51:08 <andythenorth> mixed 22:51:17 <andythenorth> about half and half I think 22:51:35 <andythenorth> it's definitely much easier to edit nml 80 times accurately 22:51:37 <frosch123> impressive, i did not expect anyone to understand pynml :) 22:51:41 <andythenorth> compared to editing a value in python 22:51:56 <michi_cc> glx: The limiting is all down internally to Windows. Apparently the default message handler for WM_GETMINMAXINFO does this. It is possible to handle the message yourself to override it, but we don't right now. 22:52:44 <glx> anyway I have only one monitor :) 22:53:14 <glx> and if I had a second one it would be a different resolution 22:54:13 <glx> so not very easy to expand a window on both 22:56:57 <glx> https://docs.microsoft.com/en-us/windows/win32/gdi/the-virtual-screen <-- oh not sure it would be fun to support that 23:00:19 <DorpsGek> [OpenTTD/OpenTTD] LC-Zorg commented on issue #8498: Add some way to find and recover crashed vehicles https://git.io/JLbdS 23:00:21 <glx> but maybe we already do 23:04:59 *** gelignite has quit IRC 23:12:03 *** snail_UES_ has joined #openttd 23:13:12 <andythenorth> how about a notification centre instead of news? :P 23:17:01 *** sla_ro|master has quit IRC 23:19:45 <michi_cc> TrueBrain: Okay, I think I have figured it out, including working switch to fullscreen. Now carefully putting each but into the proper commit. 23:20:29 <TrueBrain> :D 23:20:34 <TrueBrain> he said butt :) 23:20:34 <michi_cc> During this, I've unfortunately noticed that persistent buffer mapping somehow glitches in fullscreen mode. Seems I get synchronisation wrong somewhere. 23:21:22 <michi_cc> Going to push the fixes first though, so you can look at them. 23:29:57 *** frosch123 has quit IRC 23:33:33 *** Wolf01 has quit IRC 23:45:05 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #7744: Draft Feature: OpenGL video driver https://git.io/JeO8B 23:45:57 <michi_cc> TrueBrain: I hope I haven't messed up something. I've shuffled things around, so look carefully. 23:48:30 <TrueBrain> will do! 23:49:16 <michi_cc> Looking at the force-push diff, the part interesting to you is actually really small. I think you can ignore the OpengL stuff. 23:50:34 <TrueBrain> I can, and indeed :) 23:52:44 <TrueBrain> ah, this is why SDL didn't crash .. I already didn't implement ChangeResolution :P 23:52:52 <TrueBrain> that was already taken care of in the Base class for SDL :D 23:53:59 <michi_cc> The important thing to note is that OpenGLBackend::Resize will not get you a video buffer pointer now. 23:54:57 <TrueBrain> my OpenGL implementation is a near-identical copy of the Win32 :P So I could just copy/paste your changes :D 23:55:58 <TrueBrain> going out of fullscreen doesn't restore last window size btw, for Win32 23:56:06 <TrueBrain> otherwise, it works now :D :D 23:56:19 <TrueBrain> wait, it does restore res 23:56:21 <TrueBrain> not location 23:56:24 <TrueBrain> that is the difference, sorry :D 23:56:39 <TrueBrain> not sure that is relevant btw .. on SDL it restores to the last location too 23:56:43 <TrueBrain> that is why I noticed :) 23:56:47 <michi_cc> Never watched for that, but I'm quite sure this is in master as well. 23:56:53 <TrueBrain> most likely 23:58:08 <TrueBrain> right, this all works fine now, sweet :) 23:58:25 <TrueBrain> tomorrow I figure out what to do with palette for SDL .. bit messy, must be a clean way I am sure 23:58:31 <TrueBrain> and make this work I have now in some better commits :D 23:58:52 <TrueBrain> I btw put SDL-OpenGL in another file, that really helped 23:58:57 <TrueBrain> might want to do the same for Win32