IRC logs for #openttd on OFTC at 2019-06-27
⏴ go to previous day
01:07:23 *** ChanServ sets mode: +v tokai
02:27:43 *** Thedarkb-T60 has joined #openttd
02:54:31 *** AKTheKnight has joined #openttd
04:35:40 *** snail_UES_ has joined #openttd
04:42:18 *** snail_UES_ is now known as Guest6084
04:42:18 *** snail_UES_ has joined #openttd
08:01:14 *** andythenorth has joined #openttd
08:32:43 <andythenorth> it is *not* lunch time
08:37:42 <reldred> you say that but I've been stuffing my face the last half hour
08:38:54 <reldred> do I re-release nabs or not
08:39:41 <reldred> can I even get into my bananas account or not
09:40:48 <DorpsGek_II> [OpenTTD/OpenTTD] JoeCreates commented on issue #7629: Option to upscale game by nearest neighbour https://git.io/fjovH
10:36:05 <nielsm> I might try to PoC it on win32
11:08:54 *** Maarten has joined #openttd
11:23:29 *** Maarten has joined #openttd
11:31:51 <nielsm> I may have forgot something
11:31:58 <nielsm> openttd.cfg: resolution = 5764,2130
11:32:11 <reldred> Well, if the zoom out in the main window then can still do a 1:1 pixel
11:32:50 <reldred> I'm mostly playing on a surface go and my home desktop with a 3440 monitor so I use ui and text scale in ottd atm
11:33:17 <reldred> but yeah, could be useful
11:34:14 <nielsm> 800x600 scaled by factor 3
11:38:46 <milek7> why does ottd have platform-specific drivers? shouldn't SDL be crossplatform by itself? ;P
11:39:29 <nielsm> it easily gets sub-optimal with some of the platform specifics when you try to handle everything via a single cross-platform library
11:39:59 <nielsm> I can't think of any examples right now, but you'll typically end up with several little things that don't behave quite right
11:41:52 <LordAro> milek7: at the very minimum, historical reasons
11:42:39 <LordAro> i'm pretty sure the win32 video driver came first
11:42:55 <LordAro> (because that's what TTD had, of course)
11:44:40 <nielsm> LordAro + peter1138: should I make a PR of that scale parameter?
11:47:09 <nielsm> it's a secret driver parameter
11:47:28 <nielsm> nobody will ever discover it exists :D
11:47:49 <nielsm> (and requires cfg editing or commandline parameter)
11:48:48 <LordAro> no harm in an extra PR :p
11:49:04 <LordAro> really need to find some time to review some...
11:53:57 <reldred> nielsm: I saw that config line you posted above for resolution, how do you set the scale factor?
11:54:16 <nielsm> videodriver = "win32:scale=2"
11:54:20 <reldred> or in those screenshots did you go and do shenanigans to the video driver itself?
11:54:54 <reldred> Cheers, I'm going to have a play with that
11:55:11 <nielsm> i.e. you have to restart the game to change it, though there isn't any reason with my implementation it couldn't be made a regular setting
11:55:40 <nielsm> also, 1024x768x2 is faster than 2048x1536x1
11:56:41 <nielsm> I've also thought about making it part of the resolution specification
11:57:48 <reldred> Hmm, just edited my video driver line and it has no effect on my end
11:58:04 <nielsm> reldred: did you apply the patch I linked earlier?
11:58:18 <reldred> ahh, that would be why
11:58:38 <reldred> I read the convo wrong, thought this was lurking in the backend the whole time but only for the win32 driver
11:59:07 <reldred> I do need to get a build environment going though for ottd, that's the one thing I haven't done yet
11:59:25 <reldred> I never screwed with visual studio before, should I go that route or compile in cygwin the old fashioned way?
11:59:33 <reldred> I mean now that VS is free
11:59:49 <nielsm> I think the VS route is less painful
12:00:33 <reldred> I'm already doing stuff in cygwin for my grf compile framework I put together hence my inclination.
12:01:20 <nielsm> it's just that building for windows with gnu toolchain is not something we test very often
12:02:26 <reldred> Its probably been somewhere between five and ten years since I last compiled ottd from source so I'm a long way behind the curve.
12:04:41 <reldred> I'd imagine VS would have good github support
12:04:58 <reldred> hmmmMMMM, I guess I know what I'm staying up til 1am to do tonight then
12:06:41 <nielsm> actually I don't use any git integration in VS, prefer to do that commandline
12:09:19 <reldred> Hmm, well that gives me something to play with tonight then. I've been meaning to get stuck in and fiddle with some numbers used for terrain generation once I get my head around how some of it operates. The rivers generator for instance. I disagree with just how many it thinks 'many' is.
12:09:25 <LordAro> i don't know about cygwin specifically
12:13:41 <reldred> Yeah I'm not noecessarily committed to cygwin, but I already had it installed on both of my main machines when I started horsing around with the idea of fixing up some of my grf sets, and my old makefile system just needed imagemagick, gcc, and make so yeah, was pretty easy to modify my makefiles to work in cygwin
12:22:43 <reldred> Actually, just thinking about scaling again; it would be nice if the double UI option also made the little miniport window also default to 2x zoom. Would still give the option for a 1:1 100% pixel size zoom level, but otherwise text size and UI size at 2x pretty much solve the rest of scaling issues.
12:23:20 <reldred> Wouldn't be as fast as the driver level 2x scale though I'd imagine but framerates are kinda tied to ticks I;'ve noticed anyway.
12:25:25 <peter1138> We should unify the driver, preferably an OpenGL-based surface, at least.
12:26:42 <peter1138> glfw would be a good place to start.
12:26:51 <peter1138> It's already cross-platform.
12:26:58 <peter1138> Supports Vulkan too.
12:27:37 <peter1138> Basically, make a new driver for glfw instead of trying to modify the existing drivers, then when it's fully tested, remove the others :D
12:28:12 <LordAro> i think michi_cc had an opengl driver somewhere...
12:28:28 <peter1138> Yes he did but it was bolted on to the existing driver backends, and not all of them.
12:28:36 <peter1138> Wrong approach IMHO.
12:29:09 <peter1138> It was also not OpenGL-as-a-surface, it went deeper into the blitters.
12:32:14 <reldred> Right, that's my evening shift at work done. Chat later :). And peter1138; last thing to do to release NABS in its current form is fully port it from the grfmaker sources to that nfo includes system I cooked up nearly a decade ago, because if I'm releasing through bananas I'll probably include the sources in the tarball that goes online. I'm not (and can't) gpl'ing it but I've no problems with sources
12:33:19 <reldred> but yeah, danmack said he sent lifeblood some pm's seven years ago that are still unread and Oz has been gone for ten years now. I had permission once before to develop and release nabs so eh it is what it is.
12:34:30 <peter1138> NML is the current thing :)
12:48:45 <milek7> doesn't SDL2 support hw accelerated surfaces too?
12:52:25 <Eddi|zuHause> milek7: how would SDL2 solve the platform issues?
12:54:25 <milek7> it won't, i just don't think glfw driver is necessary if SDL is already there
12:59:00 <reldred> peter1138: I'll likely move to nml for new projects
12:59:40 <reldred> peter1138: but at the very least I want to remove the dependency from grfmaker from the compile. I mean there's nothing wrong with it, grfmaker is open sourced now and publicly available and it still works just fine
13:00:18 <reldred> any features it doesn't support are easily sidestepped with the custom nfo block I begged them to include way back when
13:07:05 <Eddi|zuHause> reldred: there's also m4nfo, which i've never used, so i can't tell whether it's good or not
13:08:39 <reldred> yeah, had a quick squizz at that. Michael if I remember correctly had been developing/using m4nfo for ages behind closed doors when Patchman first suggested to me to use gcc preprocess+make files when I was complaining about the nabs nfo getting out of control.
13:09:00 <reldred> Turned out he was doing something far more sophisticated the whole damn time (because ofcourse he was)
13:13:54 <reldred> now who was it that I gave my makefiles to back in the day can't remember
13:49:06 <reldred> Hmmm. I definitely want to change the way the building protection works though before I release, I never did actually understand how callback 143 works when I first implemented it.
13:49:45 <reldred> and then restrict building to purely 1900-1950 I think.
13:52:53 *** andythenorth has joined #openttd
14:31:57 <reldred> huh, grfcodec can ingest pngs now
14:32:09 <reldred> what wonders the new decade has wrought
14:32:30 <reldred> thats one whole stage of my makefile that can be deleted now lol
14:32:51 <Eddi|zuHause> modern technology is crazy, right?
14:34:59 <reldred> next minute you'll get me using modern markup instead of writing bytecode...
14:35:30 <reldred> to be fair bytecode doesn't bother me too much, I spend plenty of time analysing hex in my dayjob anyway
14:36:32 <Eddi|zuHause> wait, you don't program your set using a spreadsheet and feeding that into a code generator?
14:39:03 <reldred> ooohhh thats for trains
14:39:09 <reldred> yeah I see the benefit of that for sure
14:39:51 <reldred> Jeeze why do I always find the really clever stuff on openttdcoop?
14:40:57 <LordAro> that's mostly where it happens :p
14:43:43 <Eddi|zuHause> reldred: the basic framework should be easy enough to adapt to things other than trains
14:47:15 <reldred> I still need to figure out what the next project will be after I re-release nabs anyway. Does nml do stations yet?
14:47:31 <reldred> I mostly did stations, buildings and newobjects
14:48:47 *** snail_UES_ has joined #openttd
14:50:05 <andythenorth> nml doesn't do stations yet...
14:50:38 <Eddi|zuHause> stations might need some serious reworking on the generator end
14:51:08 <Eddi|zuHause> i mean, there's no reason why you couldn't generate NFO instead of NML, but the structures of stations is probably a bit different
15:25:13 *** andythenorth has joined #openttd
16:07:05 *** Viv796 is now known as Guest6114
16:11:06 *** sla_ro|master has joined #openttd
16:15:10 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjoqT
16:15:45 *** Guest6114 has left #openttd
16:28:43 *** Wormnest has joined #openttd
16:30:15 <DorpsGek_II> [OpenTTD/OpenTTD] auge8472 commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjoqW
16:42:10 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjoqV
18:47:49 *** HerzogDeXtEr has joined #openttd
19:00:16 *** argoneus has joined #openttd
19:06:19 *** Thedarkb has joined #openttd
19:20:14 *** frosch123 has joined #openttd
19:42:10 *** andythenorth has joined #openttd
19:45:27 * andythenorth looking at _saveload_crash_with_missing_newgrfs
19:47:17 *** Progman has joined #openttd
19:57:10 *** Thedarkb has joined #openttd
20:00:19 *** tokai|noir has joined #openttd
20:00:19 *** ChanServ sets mode: +v tokai|noir
20:01:50 <frosch123> savegame is made with newer version?
20:01:56 <frosch123> andythenorth: what fork are you using? :p
20:02:34 <andythenorth> the issue occurs when OpenTTD loses a newgrf
20:05:59 <frosch123> "Station::Get(order->GetDestination())->ship_station.tile != INVALID_TILE > 0" <- what compiler does peter use, that does not complain about that :p
20:10:06 <frosch123> did you move road-hog to git?
20:10:15 *** gelignite has joined #openttd
20:10:40 <frosch123> first thing to notice is that version detection for firs-git is broken, so it identifies as version 1
20:10:54 <frosch123> but apparently you also just bumped grfid or something?
20:11:19 <andythenorth> I'm assuming the broken is Hog though
20:11:44 <frosch123> anyway, my ottd refuses to consider the road-hog from the .zip even remotely compatible with the save
20:12:01 <andythenorth> it was fine whilst I was playing the game
20:12:10 <andythenorth> it was only after save-load that the issue appeared
20:12:40 <andythenorth> as per the ticket, if the grf is too broken for OpenTTD to detect as compatible, that's expected
20:13:24 <andythenorth> it just makes newgrf dev quite tricky
20:13:53 *** Speedy` has joined #openttd
20:17:14 <frosch123> the savegame wants road-hog version 0
20:17:35 <frosch123> the grf i have is minimum compaitble 1176
20:17:46 <andythenorth> yes there's no reason I would have set 0
20:17:54 * andythenorth wonders if there's a bug in the grf compile somewhere
20:17:55 <frosch123> so, somehow the version got lost
20:18:00 <andythenorth> I quite often lose versions
20:18:12 <andythenorth> or rather, OpenTTD often can't find a compatible version
20:18:27 <frosch123> btw. i uses bananas2 code to read the required grf from the savegame :p
20:19:08 <andythenorth> rocket science :)
20:20:48 <frosch123> so, what is the exact issue? that ottd crashes when loading the savegame without roadhog? (i can reproduce that)
20:21:01 <frosch123> or does it also crash when you somehow force it to use the "incompaitble" hog?
20:24:19 <andythenorth> the first one: it crashes immediately on loading the savegame
20:25:22 <LordAro> frosch123: bananas2 code exists? :o
20:25:40 <frosch123> ah, i see, it does not crash in the ordinary afterload, it's in afterloadvehicle
20:26:46 <frosch123> LordAro: some, database tables; stuff to read metadata from savegames, newgrfs, basesets and heightmaps; and locally/unpublished also some layer2 code
20:29:39 <frosch123> we should have deleted afterloadvehicles 10 years ago
20:29:50 <frosch123> those separate afterload functions have always been crap
20:30:40 <LordAro> it would be nice if afterload wasn't quite so big though
20:30:45 <LordAro> it's a pretty massive beast
20:31:37 <frosch123> ideally we would have some Desc structure which descript the bitmapping in the map array, similaer as for structures in the pool
20:31:56 <frosch123> then there could be some algorithm to move them into the right place instead of incrementally moving them around
20:32:12 <frosch123> then at least all the mapaccessors would work for the more advanced migrations
20:35:36 <LordAro> would be good to get some automated testing of some sort as well
20:36:07 <andythenorth> I am mechanical turk?
20:36:57 <frosch123> i considered starting with the coop archive, and then loading every savegame with every savegame-bumping ottd version and resaving
20:37:19 <frosch123> then compaing the result of migrating version 1->2->3 vs 1->3
20:37:28 <frosch123> but most old ottd conversions do not compile :p
20:38:20 <LordAro> that'd explode with the number of saves pretty quickly
20:38:41 <LordAro> and would not be quick, either
20:38:59 <frosch123> 50000 saves or so :)
20:39:13 <frosch123> smatz used to compile every revision to plot the compilation time
20:39:20 <frosch123> but he was also a gentoo user :p
20:40:02 <andythenorth> don't we have jenkins for that? :P
20:40:58 <frosch123> i think the peek is the noai merge
20:45:16 <michi_cc> peter1138: My opengl branch so far is mostly surface-only. I started dabbling with (texture) blitters, but not pushed anywhere. And except for the platform-specific initialization code, most things are already in a separate file/class.
21:03:42 <andythenorth> make OpenTTD great again!
21:21:41 *** HerzogDeXtEr has joined #openttd
22:06:19 *** Wormnest has joined #openttd
22:11:14 *** Supercheese has joined #openttd
22:14:11 *** Supercheese has joined #openttd
22:14:11 *** Wormnest has joined #openttd
22:14:11 *** tokai|noir has joined #openttd
22:14:11 *** Maarten has joined #openttd
22:14:11 *** AKTheKnight has joined #openttd
22:14:11 *** Extrems has joined #openttd
22:14:11 *** dwfreed has joined #openttd
22:14:11 *** rocky11384497 has joined #openttd
22:14:11 *** Smedles has joined #openttd
22:14:11 *** y2kboy23 has joined #openttd
22:14:11 *** heffer_ has joined #openttd
22:14:11 *** ToBeFree has joined #openttd
22:14:11 *** reldred has joined #openttd
22:14:11 *** Antheus has joined #openttd
22:14:11 *** rubenwardy has joined #openttd
22:14:11 *** techmagus has joined #openttd
22:14:11 *** daspork_ has joined #openttd
22:14:11 *** acklen_ has joined #openttd
22:14:11 *** greeter has joined #openttd
22:14:11 *** mikegrb has joined #openttd
22:14:11 *** innocenat_ has joined #openttd
22:14:11 *** synthon.oftc.net sets mode: +v tokai|noir
22:16:34 *** Smedles has joined #openttd
22:16:34 *** daspork_ has joined #openttd
22:19:56 *** heffer_ has joined #openttd
22:19:56 *** ToBeFree has joined #openttd
22:19:56 *** rubenwardy has joined #openttd
22:19:56 *** techmagus has joined #openttd
22:19:56 *** acklen_ has joined #openttd
22:19:56 *** greeter has joined #openttd
22:19:56 *** mikegrb has joined #openttd
22:19:56 *** innocenat_ has joined #openttd
22:57:51 <Samu> tell me a number between 1 and 5000, someone
23:12:36 <Samu> one more week and hopefully it's over
23:21:01 *** Eddi|zuHause has joined #openttd
23:37:15 <Eddi|zuHause> hm, something's wrong with firefox, i can't open any menus...
23:54:30 *** Thedarkb has joined #openttd
continue to next day ⏵
