IRC logs for #openttd on OFTC at 2019-03-08
⏴ go to previous day
00:05:24 *** octernion has joined #openttd
00:22:14 <Samu> it's an ai only server :p
00:23:00 <Samu> road vehicle cache soon?
00:23:28 <acklen> Samu: what's the server? I've got a few AIs I could test
00:24:07 <peter1138> Samu, probably not soon.
00:26:04 <peter1138> Everyone is busy I think :)
00:45:08 <Samu> i'm not paying much attention honestly
00:46:45 <Beerbelott> Eddi|zuHause: I do not think what you told me was true about directory precedence was right: Directory where external configuration file sits has no precedence at all
00:47:40 <Eddi|zuHause> Beerbelott: maybe it's due for a clear redesign and redocumentation ;)
00:47:49 <Beerbelott> dbg: [console] /home/berbe/source/src/openttd.cpp:1074 SwitchToMode _file_to_saveload: name = Europe.sav, title = , fop = 1, AbstractFileType = 1, DetailedFileType: 1
00:47:49 <Beerbelott> dbg: [console] /home/berbe/source/src/fileio.cpp:421 FioFOpenFileSp Trying to load Europe.sav
00:47:49 <Beerbelott> dbg: [console] /home/berbe/source/src/fileio.cpp:421 FioFOpenFileSp Trying to load /home/berbe/source/save/Europe.sav
00:47:49 <Beerbelott> dbg: [console] /home/berbe/source/src/fileio.cpp:421 FioFOpenFileSp Trying to load /home/berbe/.openttd/save/Europe.sav
00:48:02 <Beerbelott> (mb I should have used a paste :p )
00:48:22 <Beerbelott> it just goes through a list of predefined dirs
00:48:56 <Eddi|zuHause> Beerbelott: what about the save location?
00:49:47 <Beerbelott> See the 3 fallbacks?
00:50:08 <Beerbelott> 1st it tries to use the provided -g 'as is'
00:52:57 <Beerbelott> then goes throughhttps://github.com/OpenTTD/OpenTTD/blob/master/src/fileio_type.h#L110
00:53:31 <Beerbelott> I have a little bit of headache trying to remember who's who...
00:54:11 *** supermop_work_ has quit IRC
00:54:21 <Beerbelott> because ofc they are build on top of a searchpath... which might not be the same everytime, as I noted in my issue ;)
00:55:17 <Beerbelott> I should start a dedicated debug branch merely tracing directories path build flow
00:55:50 <Beerbelott> TL; DR I do not thing my PR breaks anything
00:57:11 <peter1138> I'm missing a resize event after toggling fullscreen off :/
00:57:21 <peter1138> Seems to depend on the WM.
00:59:09 <Eddi|zuHause> Beerbelott: still, what about the save location?
01:04:47 <DorpsGek_II> [OpenTTD/OpenTTD] Berbe commented on pull request #7339: Fix #7311: External configuration file does not change directories https://git.io/fhpAA
01:05:09 <Beerbelott> Eddi|zuHause: I didn't notice a change of behaviour between trunk & PR version
01:07:06 <Beerbelott> If you or any reviewer spot one, I can help tracing it but heck
01:07:28 <Beerbelott> I've spent sooooo many hours trying to trace/debug/check specific path building throughout the code
01:07:50 <Beerbelott> I'm reaching saturation
01:13:56 <Beerbelott> Eddi|zuHause: It's a real pain as I have no clear picture of the decisions taken and why. Isn't there at least a remnant of a hand drawing summing up decisions about dirs? :(
01:14:55 <nnyby> theres a problem with doing alt+enter on xmonad.. i can debug this
01:15:11 <peter1138> I've changed a lot :-(
01:16:45 <nnyby> hah okay, that's fine. i'll wait to see your changes.
01:19:10 <peter1138> I'm totally confused though.
01:19:33 <peter1138> want_hwpalette still exists, but isn't used for anything.
01:20:43 <peter1138> So the SDL 2 driver never sets SDL_HWPALETTE or SDL_SWSURFACE
01:21:52 <peter1138> fullscreen is now fullscreen-window
01:22:03 <peter1138> and as such the resolution dropdown does nothing in fullscreen mode
01:22:52 <peter1138> I bet you're regretting mentioning that bug now:p
01:22:57 <peter1138> It was about to be merged.
01:23:07 <peter1138> But I can't now, as quite a lot doesn't make sense
01:23:43 <nnyby> yes, looks like there are some code paths that aren't followed here, that we should clean up
01:24:00 <peter1138> Is 8bpp gone in SDL2?
01:24:37 <nnyby> i left the hwpalette stuff in... i wasn't clear if or when _use_hwpalette gets set
01:25:04 <peter1138> Have you tried with -b 8bpp-optimized?
01:25:46 <nnyby> i wasnt aware you can set different blitters like that i'll start messing with that
01:27:24 <nnyby> regarding FULLSCREEN vs FULLSCREEN_DESKTOP.. we can probably change that back to FULLSCREEN. i guess i did fullscreen-desktop because I got into situations where changing the resolution while in fullscreen would permanently change my desktop's resolution, and i didn't like that. that's how sdl1.2 works though i think
01:38:46 <nnyby> okay :3 well -b 8bpp-optimized is obviously broken in the sdl2 branch. and working on sdl1.2. so thats one thing to fix.
01:43:14 <peter1138> Yes, that's why I asked if you'd tried it :p
01:43:22 <peter1138> I knew obviously not, heh.
01:52:31 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
01:52:47 <peter1138> Well, good news is I got fullscreen/window switching "working"
01:53:18 <peter1138> But it doesn't work with SDL_WINDOW_FULLSCREEN :p
01:56:17 *** Thedarkb-X40 has joined #openttd
02:14:43 *** Wormnest has joined #openttd
02:30:13 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
02:43:55 *** Supercheese has joined #openttd
02:51:28 <nnyby> im still trying to get this to work
03:00:30 <nnyby> still using fullscreen-desktop though?
03:01:29 <nnyby> and im assuming 8bpp is still broken..
03:02:02 <nnyby> just curious.. are there cases where this blitter is chosen by default?
03:03:16 <nnyby> well its 2am in the UK so im guessing ur going to bed soon
03:04:57 *** Thedarkb1-T60 has joined #openttd
03:05:25 <peter1138> 8bpp is totally disabled.
03:07:26 <peter1138> nnyby, commit 631e8b45fd7bbd95766ee294304fad38dda946dc
03:07:30 <peter1138> -Change: Disable usage of 8bpp blitters and video modes by default. Many modern OS and hardware cause issues with those.
03:08:04 <peter1138> So basically you only get them by using -b 8bpp-optimized
03:08:11 <peter1138> (Or by having it in your config)
03:08:35 <peter1138> I suspect there'll be players who do have blitter = 8bpp-optimized in their config.
03:09:39 <peter1138> At this point, it might be better to keep the SDL 1.2 driver after all, though.
03:10:50 <nnyby> you suggesting having sdl 1.2 + sdl 2.0 as separate drivers?
03:11:05 <nnyby> still.. maybe possible to get 8bpp working with sdl2
03:11:09 <peter1138> Removing 8bpp support is a big step.
03:11:45 <nnyby> but yeah.. we have some options. Keep sdl 1.2 around.. or get this working with sdl 2
03:29:48 *** Beerbelott has left #openttd
04:14:58 <Eddi|zuHause> if this is an option, then keeping sdl1 and 2 as separate drivers seems like the preferable solution
04:52:43 *** tokai|noir has joined #openttd
04:52:43 *** ChanServ sets mode: +v tokai|noir
06:12:12 *** Supercheese has joined #openttd
06:28:21 <nnyby> i think i just got 8bpp blitter working, using an 8bit sdl shadow surface. and then blitting to a 32bit surface when displaying to screen. this is how the 1.2 driver works. i'll add a new commit to my branch
06:30:37 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
06:32:38 <nnyby> really a small change, and things are already in place to make this work
06:42:42 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas commented on pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhpjB
08:16:09 *** andythenorth has joined #openttd
09:29:32 *** Speedy` has joined #openttd
10:10:19 *** andythenorth has joined #openttd
10:11:44 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhhf0
10:16:07 <andythenorth> eh that's an interesting idea
10:16:15 <andythenorth> port-type industries that build next to rivers
10:16:22 * andythenorth might need to redesign these industries
10:17:39 <DorpsGek> andythenorth: pikka was last seen in #openttd 2 days, 16 hours, 32 minutes, and 16 seconds ago: <Pikka> eek
10:17:57 <andythenorth> so 1.9 is done, what's next?
10:22:58 <peter1138> Approval and merging.
10:23:15 <andythenorth> I pulled the SLD 2 PR
10:23:36 <peter1138> I was about to merge it, then I found it was very unfinished.
10:25:30 <andythenorth> subjectively assessment of SDL 2 performance on Mac OS
10:25:33 <andythenorth> "about the same"
10:25:56 <andythenorth> peter1138: ^^^ winner, winner, chicken dinner
10:26:41 <peter1138> Yes but that doesn't mean it's good yet :p
10:27:39 <andythenorth> in a galaxy far far away
10:27:52 <andythenorth> it was hoped that SDL 2 would allow 'full animation' to be used once more
10:28:05 <andythenorth> but it probably doesn't, yet
10:28:26 <peter1138> Why can you use full animation?
10:28:45 <andythenorth> been nerfed for years
10:28:57 <peter1138> It's been working for years.
10:29:14 <andythenorth> there was a change either in Apple's graphics drivers, or in Intel hardware about 8 years ago
10:29:20 <andythenorth> palette animation is fucked
10:29:26 <peter1138> Palette animation totally works.
10:29:37 <andythenorth> I've had two credible explanations
10:29:43 <peter1138> OpenTTD hasn't use 8bpp modes by default for 5 years.
10:29:46 <andythenorth> 1. Apple or Intel removed drivers
10:29:54 <peter1138> Palette animation is done internally by OpenTTD.
10:29:59 <andythenorth> 2. there isn't enough bandwidth between graphics cards and CPU
10:30:02 <peter1138> So what are you talking about?
10:30:55 <andythenorth> with 'full animation' off graphics rendering averages 0.4ms
10:31:14 <andythenorth> with 'full animation' on, it averages 0.6ms
10:31:43 <andythenorth> this is on an empty map, and scales horribly as the map gets busier
10:33:09 <andythenorth> with 'full animation' off, a month takes 3s on FFWD
10:33:21 <andythenorth> with 'full animation' on, a month takes about 27s on FFWD
10:33:57 <peter1138> Because it's redrawing the screen.
10:34:11 <andythenorth> welll on Mac OS it was a performance regression about 8 years ago
10:34:21 <andythenorth> it used to be substantially faster on FFWD with full animation on
10:34:25 <peter1138> 8 years ago FFWD with full animation off wouldn't've taken 3s
10:34:31 <peter1138> It's only since that change I made.
10:34:53 <peter1138> 8bpp is faster for palette animation, yes.
10:35:05 <andythenorth> hang on what's the command to show bllitter?
10:35:05 <peter1138> But absolutely no OS supports 8bpp modes these days.
10:35:16 <peter1138> So we do it in software.
10:35:26 <peter1138> This is slower, but that's obvious
10:35:49 <peter1138> When you say "palette animation is broken", that would be somethhing like the colours are fucked up or it just plain doesn't animate.
10:35:58 <andythenorth> no that got fixed
10:36:17 <peter1138> That was broken *because* it was trying to use unsupported 8bpp mode.
10:37:31 <andythenorth> let's try some blitters
10:37:46 <peter1138> 3s vs 27s is normal.
10:37:59 <peter1138> 3s is actually a bug.
10:38:08 <peter1138> It wasn't really intended to skip viewport frames being rendered.
10:38:33 <peter1138> With FFWD, it is basically animating the palette as fast as it can.
10:38:59 <andythenorth> yeah don't fix that bug :)
10:39:10 <andythenorth> or give me a flag I can compile with to keep it ;P
10:40:43 <andythenorth> yeah all blitters perform about the same
10:41:56 * andythenorth thinks of evil option
10:42:02 <andythenorth> suppress animation in FFWD
10:43:22 <andythenorth> basically, I really like animated water
10:43:31 <andythenorth> and I really like being able to FFWD very fast to test newgrf
11:23:53 <Eddi|zuHause> ... or we could do palette animation with shaders
11:24:21 <peter1138> Yes if you rewrite the rendering system.
11:24:39 <Eddi|zuHause> "how hard could it be"
11:29:51 <peter1138> Maybe I'll get it another go.
11:30:48 <peter1138> Something something
11:33:56 <michi_cc> I might have to rebase it tough, still using old commit message syntax :p
11:34:15 <Eddi|zuHause> in which world do i ever "finish" anything?
11:35:47 <michi_cc> It is working fine on my Windows system. Last time somebody tried it it didn't work there, but I don't have their system to test, so...
11:40:05 <peter1138> I think I had a crash but didn't investigate much.
11:41:41 <peter1138> michi_cc, that uses textures for surfaces but not individual sprites like my old patch, right?
11:42:11 <michi_cc> Yeah, it's mostly a new window driver, not a complete blitter replacement.
11:42:46 <michi_cc> I did start a bit on individual sprites as you can see with the accelerated mouse cursor stuff.
11:43:50 <peter1138> That's a big job though. You'd need to create spritesheets on the fly, and then use VBOs for tiles, and maybe chunks of tiles.
11:47:18 <peter1138> OpenGL started to go mad with individual textures for each sprite :)
11:47:43 <peter1138> And creating a sprite sheet needs some kinda of sprite pre-loading.
11:47:51 <peter1138> So the sprite cache system needs to... change.
11:53:27 *** andythenorth has joined #openttd
11:56:45 <michi_cc> peter1138: There's a reason why the branch is where it is :p
11:56:59 <peter1138> michi_cc, yeah, but that isn't necessary to see an improvement, right?
11:57:35 <peter1138> It needs SDL of course.
11:59:10 <michi_cc> Results vary with graphic hardware, but it should help those system very challenged with palette animation, especially if used with the "40bpp" blitter.
12:00:07 <michi_cc> peter1138: OSX is not using SDL by default.
12:01:55 <peter1138> michi_cc, I'm aware, hence why I specified :)
12:17:35 <Eddi|zuHause> vehicles at 4x zoom need smoother movement and animation...
12:17:46 <Eddi|zuHause> but "someone" rejected #5300
12:24:47 <Eddi|zuHause> maybe i should include something like that in #7000 (yay feature creep!)
12:28:10 <peter1138> Yeah progress gets some funny values sometimes.
12:28:28 <Eddi|zuHause> i think we fixed those
12:32:19 <planetmaker> I think that was reverted
12:33:39 <Eddi|zuHause> i don't see anything
12:35:41 <Eddi|zuHause> planetmaker: as far as i can tell, those lines are still there in src/train_cmd.cpp
12:36:46 <peter1138> It wasn't reverted, a suitable fix was found.
12:37:45 <peter1138> Apparently I can delete andythenorth's branch for that PR. I wonder why he hasn't? :p
12:38:16 <planetmaker> peter1138, did you manage to import grfcodec? Today it succeeded for me
12:38:20 <peter1138> Basically the signal timers were said to be "days" but actually weren't.
12:38:24 <peter1138> We made them days :-)
12:38:32 <peter1138> planetmaker, yes I did, I posted the URL in that other place.
12:38:58 <peter1138> But really it needs to be in OpenTTD :-)
12:39:39 <peter1138> We've got nml already, makes sense to take over grfcodec from openttdcoop.
12:39:53 <peter1138> Means we can update it with all the new NewGRF features.
12:40:39 <planetmaker> of course it should be in OpenTTD
12:43:33 <peter1138> I don't have permission to do that.
12:43:35 <planetmaker> I guess I can just as well try to clone it to OpenTTD
12:44:53 <planetmaker> hm... [x] Azure Pipelines? I guess, yes
13:21:20 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
13:27:30 <planetmaker> hm, we also should clone OpenGFX, OpenSFX and OpenMSX to the openttd account...
13:29:18 <peter1138> I've been telling everyone we don't maintain them :p
13:32:32 <planetmaker> you cannot play OpenTTD w/o base graphics. And we have some which we can offer. We do offer them. So we should put that under OpenTTD umbrella, too
13:33:50 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
13:38:35 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
13:39:12 *** luaduck has joined #openttd
13:39:47 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas commented on pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhhkQ
13:40:02 <andythenorth> planetmaker I wondered about moving them under OpenTTD
13:40:15 <andythenorth> I think we do need to recruit maintainers
13:40:59 <planetmaker> however without base sets OpenTTD is back to <1.0 time
13:41:13 <planetmaker> they're as essential as the game itself
13:41:41 <peter1138> I have the DOS and Windows originals, they're not essential ;)
13:41:57 <planetmaker> I know you all do not like it. But that's not how you can offer a game
13:42:11 <planetmaker> "yeah, i got some nice code. But please go buy here this game so you can play mine"
13:42:33 <peter1138> Btw, I never said we shouldn't add them to OpenTTD's github, just that we haven't.
13:42:41 <planetmaker> and drooling on the misery of the state of graphics and sound etc to make this game a true game on its own right
13:43:01 <peter1138> If we do, I will update my posts that say we don't maintain them.
13:43:12 <peter1138> I see you have imported grfcodec \o/
13:43:24 <planetmaker> surprisingly worked nicely today
13:43:37 <planetmaker> yesterday was not grfcodec day it seemed :P or not github day. :P
13:43:41 * peter1138 deletes his import.
13:43:49 <andythenorth> github hg imports from devzone are very flakey
13:44:01 <andythenorth> not sure why, the errors are quite opaque
13:44:11 <peter1138> I never saw errors, it just took a long time.
13:44:18 <peter1138> Like, around 40 minutes.
13:44:44 <andythenorth> btw, afaict, the main blocker to coop -> devzone is replicating bundles
13:44:46 <peter1138> Ok, let's import OpenGFX and the rest.
13:45:15 <andythenorth> coop -> github :P
13:45:30 <planetmaker> well. bundles is not dead nor is there need to take it down
13:45:32 <peter1138> planetmaker, is it importing yet? :D
13:45:41 <andythenorth> from Azure Pipelines
13:45:53 <andythenorth> Github releases can't do that
13:46:01 <planetmaker> github realeases... no?
13:46:02 <peter1138> If bundles can pull from git instead of hg, then it's almost there...
13:46:04 <andythenorth> Github pages could do that, but none of you would like it
13:46:05 <planetmaker> I never looked tbh
13:46:15 <planetmaker> bundles is just an artefact storage
13:46:16 <peter1138> Github releases are just tarballs of the git tree.
13:46:24 <andythenorth> they don't publish docs etc
13:46:43 <planetmaker> Honestly, I really see no problem in keeping the coop bundles storage
13:46:44 <andythenorth> github pages can publish, but only one version, unless we build something specific to do it
13:46:54 <andythenorth> planetmaker well then that's a key / secrets problem
13:47:04 <planetmaker> It's a simple file storage - and that is rather easy to maintain
13:47:08 *** mikegrb has joined #openttd
13:47:17 <peter1138> andythenorth, bundles can poll github to check for changes, no?
13:47:26 <peter1138> And that can be done anonymously.
13:47:26 <planetmaker> peter1138, bundles doesn't poll anything
13:47:27 <andythenorth> that just gets source
13:47:36 <andythenorth> bundles is just publishing
13:47:39 <planetmaker> it gets artefacts pushed by build service
13:47:43 <andythenorth> I was actually going to write a ticket for this
13:47:50 <andythenorth> I spent last week banging my head on Azure
13:47:58 <andythenorth> and got both progress, and depressed about it
13:48:00 <peter1138> Okay, so what build service does it use at the moment?
13:48:09 <andythenorth> which is replaceable trivially by AP
13:48:13 <peter1138> jenkins can pull from github
13:48:14 <planetmaker> everything NewGRF build on devzone
13:48:15 <andythenorth> I've got FIRS building on AP
13:48:24 <andythenorth> TB is pretty adamant we should use AP
13:48:33 <planetmaker> but providing ssh access to any servce... no issue either
13:49:01 <andythenorth> apart from "the AP UI is fucking awful", it's really good
13:49:02 <peter1138> Yes I'm sure, but it's trivial to switch jenkins to pull from git, rather than reimplementing it all in AP.
13:49:27 <planetmaker> jenkins already pulls from git for current nml repo
13:49:29 <andythenorth> yes, that could be a migration step, but it's probably not a destination
13:49:39 <andythenorth> I'll make a paste
13:49:45 <peter1138> But it's better to do that than not have anything
13:49:49 <peter1138> No, don't doa Samu.
13:50:10 <planetmaker> does anyone have a server available which one could use to update all of devzone?
13:50:15 <planetmaker> for a month or so?
13:50:35 <planetmaker> currently it has all repos - they could be moved to github before
13:50:58 <planetmaker> it has jenkins. it has bundles.
13:52:11 <planetmaker> eints is a critical one. The translator for everything
13:52:23 <planetmaker> everything newgrf and baseset
13:53:00 <peter1138> So just the devzone stuff, or everything?
13:54:24 <planetmaker> idealy I'd shortly move everything. Update hypervisor and immediately move back all stuff
13:54:39 <planetmaker> and then conveniently update the VMs on my own server one by one
13:54:59 <peter1138> Switching to a modern hypervisor?
13:55:03 <planetmaker> though... I could probably try to update the HV from centos6 to 7 without prior migration
13:55:22 <planetmaker> currently openvz. But yes, that would need updating too. centos7 would allow docker. I'd use that then
13:55:38 <planetmaker> at least that's my plan
13:55:40 <peter1138> Ah, switching to docker containers rather than VMs
13:55:45 <andythenorth> ^^ I'd make a ticket, but I'm not sure where I'd make it right now :P
13:55:46 <peter1138> Sensible for resources.
13:55:55 <planetmaker> would suffice for most things. I don't have anything windows
13:56:06 <andythenorth> planetmaker: have I missed anything in that paste?
13:56:07 <planetmaker> andythenorth, ticket for what?
13:56:15 <andythenorth> explaining the problem
13:56:34 <andythenorth> as always, people try to help, but without knowledge of the problem, suggested solutions aren't always viable
13:56:41 <andythenorth> and we have to make the problem clear :)
13:56:55 <planetmaker> hm... where. Good point
13:57:08 <andythenorth> we're in limbo currently
13:57:15 <andythenorth> transitions are hard :)
13:57:39 <planetmaker> There is IMHO no big issues to keep eints running either
13:58:07 <planetmaker> both bundles and eints are pretty small applications with little cross-dependency
13:58:35 <planetmaker> my real main concern is: I need to update the HV... which I consider risky without contingency
13:58:40 <andythenorth> or at least as sane as TB gets
13:59:51 <andythenorth> I was hoping we'd make progress on this today :)
14:00:00 <andythenorth> first thing is to know what 'this' is :)
14:00:26 <planetmaker> Upgrade DevZone / Migration of repos to github
14:00:59 <andythenorth> ok, it has a title :)
14:01:17 <planetmaker> I didn't yet see anything missing in your past
14:01:42 <planetmaker> not sure really about azure though
14:01:47 <planetmaker> but yes... let's try
14:01:56 <planetmaker> but that would be kinda the last step :)
14:02:20 <andythenorth> do I miss anything? where does newgrf wiki live?
14:08:39 <peter1138> OpenGFX importing? Woo
14:09:03 <peter1138> Shall I update my posts now? :D
14:09:22 <planetmaker> first try gave me a 500 on github. 2nd is running :P
14:09:37 <peter1138> I'll wait til it's actually done it, I guess.
14:10:23 <peter1138> ^ That's going to be somewhat nerfed with non-rect-catchment.
14:11:10 <peter1138> Well, can also extend with extra tiles, heh.
14:11:55 <peter1138> Might be an argument for massive platforms and loading extra to have a large catchment.
14:12:08 <peter1138> But that's separate.
14:13:24 <planetmaker> un-nerfing that is not difficult. I don't see a big issue, peter1138
14:13:43 <peter1138> Not really an issue :)
14:13:47 <planetmaker> and it's a definitive win on grasping the situation more intuitively
14:14:21 <peter1138> I'd be surprised if 3.3 actually compiled us...
14:14:50 <planetmaker> indeed. And indeed
14:15:08 <andythenorth> hmm, wonder if nmlc builds with pypy
14:16:10 <planetmaker> I didn't have success
14:16:36 <andythenorth> did frosch write a custom cpython module for part of nmlc?
14:16:46 <planetmaker> for sprite encoding
14:16:46 <andythenorth> "Several studies even suggest that PyPy is about 7.5 times faster than CPython."
14:17:04 <planetmaker> which was a definitive speed-up
14:17:46 <andythenorth> when it shipped, it cut FIRS compile times below 1 min
14:18:06 <andythenorth> anyway, nml is a digression :)
14:18:18 <andythenorth> we need to pin down this coop infra / service plan
14:18:34 <andythenorth> planetmaker: have you had a go with the Azure Pipelines interface?
14:19:38 <LordAro> 4.8+ is C++11 support, iirc
14:19:46 <LordAro> might need 5 for C++14
14:40:09 <planetmaker> andythenorth, not much yet. First thing for that is to create a build docker container. Otherwise azure is pointless
14:47:11 <andythenorth> there are positives and negatives
14:47:26 <andythenorth> the positives are, once I understood what was needed, building FIRS was easy
14:47:39 <andythenorth> the negatives are
14:47:50 <andythenorth> - learning the AP web interface is about one day of rage
14:48:03 <planetmaker> so you build it using azure?
14:48:16 <andythenorth> - nearly all the docs are useless, they are accurate, but they link off to other Azure things all the time
14:48:31 <andythenorth> - and reference services that may or may not be available
14:48:39 <planetmaker> then please implement that in a general manner to make it useful
14:49:01 <andythenorth> - there is *absolutely no way* we could offer this as a general service to newgrf authors
14:49:09 <andythenorth> it's a total brain fuck at first
14:49:21 <andythenorth> but TB thinks we can template everything
14:49:42 <andythenorth> I honestly think that even moving authors to Github is going to be very hard
14:49:49 <planetmaker> tbh I don't feel like I will be doing that with azure docker any time soon.
14:50:03 <andythenorth> devzone isn't broken
14:50:07 <andythenorth> it's just not the future
14:50:18 <planetmaker> it becomes 50% less useful since you are not using it any longer
14:50:36 <andythenorth> for me, the nice thing about devzone is that it was an easy to user service that did everything
14:50:36 <planetmaker> similar motivation to fix anything ;)
14:51:00 <andythenorth> telling authors, 'first create a github account'
14:51:13 <andythenorth> 'then create a microsoft account'
14:51:18 <planetmaker> yes, I think that will be a big hurdle.
14:51:26 <planetmaker> Similar, they will have to relearn from mercurial to git
14:51:36 <planetmaker> Dear, I tell you, that's a stony unpleasant path
14:51:48 <andythenorth> 'then authorise an Azure Pipelines -> Github service connection"
14:51:54 <andythenorth> "then create API keys"
14:52:01 <andythenorth> "then create github pages account"
14:52:09 <andythenorth> "then create a special branch for github pages"
14:52:19 * andythenorth is over-exagerating, but tried all of this last week
14:52:30 <planetmaker> well. That's what *I* did for every author who asked. Just for the services on my server
14:52:32 <andythenorth> users are gonna be like, 'fuck all that, I'll use dropbox'
14:52:51 <andythenorth> or just post code in forums
14:53:04 <planetmaker> create repo, install key, explain procedure, setup jenkins, adopt build scripts, if needed, ...
14:53:08 <planetmaker> that simply doesn't scale
14:53:17 <planetmaker> yes, they will. I'm sure, too
14:54:01 <planetmaker> there's now one user of DevZone left really. Maybe two: George and the WAS people. And myself
14:54:17 <planetmaker> Everything else is about... abandoned
14:55:29 <Eddi|zuHause> well, you can migrate all repos to github, but don't ask people to do that manually...
14:55:58 <andythenorth> planetmaker: Ammler was confusing PITA sometimes, but loss of Ammler = loss of some of the enthusiasm
14:56:06 <andythenorth> and loss of yexo was mega :(
14:56:18 <andythenorth> I don't like bothering spike when the infra breaks
14:56:31 <planetmaker> I've been doing that a few times already
14:56:36 <planetmaker> I have his mobile number :P
14:56:52 <planetmaker> but I'm not fond of doing that either
14:56:53 <andythenorth> oof, based on activity, it really is just me and George
14:57:06 <andythenorth> I was thinking this was a bigger problem
14:57:21 <andythenorth> my assumption was that we need to build a web service for non-technical grf authors
14:57:24 <andythenorth> which is easy to use
14:57:29 <andythenorth> and I don't know how to do that
14:57:52 <andythenorth> but there really aren't any active
14:58:00 <planetmaker> I have a jenkins project which somewhat works like "put your zip file with code and graphics here and we build your newgrf"
14:58:25 <Eddi|zuHause> that sounds a bit over the top
14:59:00 <andythenorth> there are plenty of active grf projects in forums, but not many that are building long-term software projects
14:59:05 <andythenorth> that's probably fine
14:59:36 <andythenorth> some stuff is disposable, we don't need to worry about the longevity of it
14:59:44 <planetmaker> Well. My main motivation back then was to offer a place to collect NewGRF projects. So that it becomes easy to find them
14:59:58 <planetmaker> the rest then evolved from that
15:00:45 <planetmaker> well... so we can conclude: let's scrap it altogether. No-one needs it
15:00:53 <planetmaker> no one will shed a tear
15:01:40 <planetmaker> I'm sure the russian Guys will miss it though... sometimes I have the feeling they rely on the build service for building
15:02:21 <Eddi|zuHause> well, the build service not building was a major setback for CETS
15:02:51 <andythenorth> what's minimum viable?
15:03:20 <andythenorth> actually talk CETS first, I'll ask my questions in a bit
15:03:36 <Eddi|zuHause> planetmaker: i've never understood why it's not working. it was relying on custom nml, and one or more steps in that failed
15:04:11 <Eddi|zuHause> planetmaker: building the custom nml binary, or using it, or both
15:04:24 <planetmaker> Eddi|zuHause, yes... I didn't quite work it out either... it *should*... but somehow doesn't
15:04:37 <planetmaker> you want to investigate yourself?
15:04:51 <planetmaker> give me your ssh key and... look through
15:04:56 <Eddi|zuHause> i have no hope of finding anything out that way...
15:05:39 <planetmaker> how would you have hope?
15:06:04 <Eddi|zuHause> i'm hopeless with figuring out build systems
15:06:26 <planetmaker> being on the system, it's not much more than debugging some shell scripts
15:06:29 <andythenorth> they're usually just ill-documented and the errors are opaque
15:06:34 <planetmaker> jenkins does not more than executing shell scripts
15:06:37 <andythenorth> but build systems are just pipelines and triggers
15:06:53 <Eddi|zuHause> it's nothing to do with the specific build system used
15:06:59 <Eddi|zuHause> it was always this way
15:07:08 <andythenorth> funny how brains work
15:07:13 <andythenorth> I can't think easily in trees :P
15:07:19 <andythenorth> but build systems...they're just pipes
15:07:29 <Eddi|zuHause> i need someone to tell me "put your code here" and a button "build it"
15:08:20 <andythenorth> takes all sorts to make a world ;)
15:08:51 <Eddi|zuHause> i have a similar problem with using GUI frameworks
15:09:06 <andythenorth> they're just configuration
15:09:12 <andythenorth> and event handling
15:09:59 <andythenorth> Eddi|zuHause ok you do the data structures and recursive trees, and we'll do the pipelines :P
15:10:41 *** octernion has joined #openttd
15:10:58 <planetmaker> I actually would like to scratch everything. And then try to build a completely new stuff. But when I'm honest with myself: I totally lack the time and knowledge to get that going in any forseeable timeframe
15:11:23 <planetmaker> ammler was ... insane enough to just do that on live systems
15:11:29 <andythenorth> well, my shopping list is this
15:11:30 <planetmaker> s/live/production/
15:11:49 <andythenorth> 2. continuous integration
15:12:01 <andythenorth> 3. artefact publishing (including, optionally, html docs)
15:12:13 <Eddi|zuHause> sounds like a plan, do it :)
15:12:27 <andythenorth> there are a couple of extras that are nice to have, I have no real need
15:12:34 <planetmaker> maybe I should really do it
15:12:41 <planetmaker> 1) move everything to github
15:12:44 <andythenorth> 4. continuous deployment (tagged releases -> bananas via musa)
15:12:56 <planetmaker> then everything else is not so important anymore as the code is the single most important thing
15:12:57 <andythenorth> 5. issue tracker (I don't really need them)
15:13:29 <andythenorth> I also find the redmine 'files' tab incredibly useful, although I could just post to forums
15:13:32 <andythenorth> but I use it a lot
15:13:53 <planetmaker> I'll start migrating all repos to the openttdcoop 'team' on github
15:14:03 <andythenorth> I also use coop pastebin a lot, but there is just pastebin.com
15:14:22 <andythenorth> we don't use it often, but it's incredibly useful when we do
15:14:36 <andythenorth> mostly making outlines of specs
15:14:41 <andythenorth> or refining problem definition
15:15:14 <planetmaker> I think we once had briefly etherpad.o.o :P
15:15:24 <andythenorth> repos -> github, with current jenkins
15:15:53 <andythenorth> but anyone relying on mercurial for action 14 version number, will have a bad time
15:16:02 <andythenorth> alberth has a possible fix for that
15:18:09 <Eddi|zuHause> really, just count revs since tree root
15:19:12 <planetmaker> yes... that would work for 98% of the cases... probably good enough
15:19:28 <planetmaker> been thinking about that, too
15:19:38 <andythenorth> yes that might also be a solution Eddi|zuHause
15:19:55 <planetmaker> currently some use 'days since 2000'. That's possible with git as well
15:20:22 <Eddi|zuHause> that works, except when you use build on push..
15:20:25 <andythenorth> so if I publish multiple branches of FIRS, which one gets which rev?
15:20:38 <andythenorth> it is acceptable to get broken results btw
15:20:48 <andythenorth> as long as they're predictable in advance when designing the system
15:21:08 <andythenorth> using nightly/push builds of newgrfs can come with health warnings
15:21:10 <Eddi|zuHause> andythenorth: each branch needs its own linear build number, possibly with changed GRFID
15:21:13 <andythenorth> but not clown shoes surprises
15:21:24 <andythenorth> changing GRFID has some problems
15:21:41 <andythenorth> there are multiple moving parts
15:21:51 <Eddi|zuHause> andythenorth: i see no sane way to synchronize build numbers between branches
15:21:58 <peter1138> What happened to Ammler & Yexo?
15:22:02 <andythenorth> I think we have to choose what breaks
15:22:03 <Eddi|zuHause> andythenorth: so just live with it
15:22:09 <planetmaker> ah, it doesn't matter if two branches in a newgrf repo have the same indicated version, does it?
15:22:19 <planetmaker> peter1138, they simply disappeared
15:22:31 <andythenorth> planetmaker probably not, unless something else makes the wrong assumptions
15:22:38 <Eddi|zuHause> planetmaker: imho the case is rare enough to not worry about
15:22:47 <andythenorth> yexo is working at Google as an SRE
15:22:51 <peter1138> Drifted off, much like I did, I guess.
15:22:52 <andythenorth> and Ammler just disappeared
15:23:02 *** Thedarkb-T60 has joined #openttd
15:23:12 <andythenorth> site reliability engineer
15:23:30 <planetmaker> ah... yes, google tells me :P ty
15:24:04 <planetmaker> last time I stalked him he was still at DekTec
15:24:34 <peter1138> Basically people grew up and got real jobs, eh?
15:24:50 <peter1138> OpenGFX still importing, eh?
15:24:56 *** synchris has joined #openttd
15:25:05 <planetmaker> no. Failing consistently. I opened a service ticket with github
15:25:21 <planetmaker> I might do it by hand... if needed
15:25:26 <planetmaker> but that's dreadful :|
15:25:36 <peter1138> Have you tried the other repos?
15:25:48 <planetmaker> so waiting for an easy solution. No, I haven't yet
15:26:06 <planetmaker> I've started to import all coop repos from top to bottom to the github coop account
15:26:27 <planetmaker> fun stuff like 2cctrainset
15:26:41 <peter1138> Is there a repo limit?
15:27:29 <andythenorth> sorry about LinkedIn, it's worse than Hitler or something :(
15:27:33 <planetmaker> I will find out about a repo limit :)
15:27:50 <planetmaker> if there is... there'll be an openttdcoop2 and openttdcoop3 and ... :P
15:28:38 <planetmaker> does github offer the option to archive repos?
15:28:45 <andythenorth> if the repo has eints commits, it seems to make broken commits
15:28:51 <andythenorth> which github chokes on
15:28:58 *** Thedarkb1-T60 has joined #openttd
15:28:59 <andythenorth> frosch knows how to fix
15:29:27 <peter1138> Scrap the history and import afresh from the last commit? :p
15:29:46 <peter1138> andythenorth, did that branch work? :p
15:30:06 <peter1138> Dunno if that's compileable on OS X.
15:30:09 <andythenorth> peter1138 that PR?
15:30:21 <peter1138> Not a PR yet, just a branch.
15:30:43 <andythenorth> that compiled fine
15:31:42 <andythenorth> 16 people listed on contact page, 6 are active contributors, 7 are gone and 3 are occasional
15:31:51 <peter1138> Fuck it, I'm eating the last Biscoff Lotus biscuit...
15:32:13 <peter1138> To be fair, it's just a twin pack anyway, but "last" sounds more dramatic.
15:32:14 * andythenorth wonders if that contact page is pointless or not
15:32:50 <peter1138> Contact should be "info@openttd.org" only :p
15:33:06 <peter1138> Yeah, just the top 3.
15:33:10 <andythenorth> when I rebuild the website, we decided not to confuse that with changing content
15:33:20 <andythenorth> but we could change it now
15:33:20 <planetmaker> that is totally fine
15:33:43 <peter1138> Btw, I think the credits window in game needs to scroll faster.
15:33:46 <planetmaker> though doesn't hurt to have the active people e-mail listed there... but is it useful? Dunno
15:33:51 <peter1138> It used to speed up with FFWD but that doesn't happen now :p
15:34:05 <andythenorth> peter1138: credits window, what where?
15:34:34 <peter1138> Question Mark menu - About OpenTTD
15:35:09 <planetmaker> does it already mention some of the newer contributors?
15:35:22 <andythenorth> yeah that's....slower than star wars
15:35:45 <andythenorth> oh is this why samu asks for 'expert in x'? :P
15:35:51 <peter1138> I wonder if 60ms or even 30ms is better.
15:35:58 <peter1138> Nah, I doubt he's looked at it.
15:36:05 <andythenorth> I am pretty certain Alberth never ever wants to be referred to as 'GUI expert'
15:36:46 <andythenorth> I like inactive section
15:36:52 <andythenorth> "TrueBrain.... (active)"
15:37:20 <andythenorth> didn't Ludde go on to invent Spotify or something?
15:37:48 <peter1138> 60ms is okay, that's 1 pixel every 2 game ticks.
15:37:56 <peter1138> Because that's a relevant figure...
15:38:37 <Eddi|zuHause> is it scaled with font size?
15:38:56 <peter1138> Thanks for reminding me to check.
15:39:00 <Eddi|zuHause> it probably should :)
15:39:54 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
15:39:54 *** Thedarkb-T60 has joined #openttd
15:40:14 <Eddi|zuHause> as in, measure line height, divide that by constant, and scroll by that amount
15:41:09 *** Thedarkb1-T60 has joined #openttd
15:47:56 <peter1138> All important stuff, eh.
15:51:04 <michi_cc> peter1138: Updated my opengl branch to current master.
15:53:35 <Eddi|zuHause> michi_cc: make it a draft PR?
15:54:36 <michi_cc> It is Windows only for now and didn't work for the one I got feedback from; seems less-than-draft to me.
15:56:07 *** Beerbelott has joined #openttd
15:56:49 <Eddi|zuHause> michi_cc: might make it easier to find next time :)
15:57:02 <peter1138> It's not PRs get merged ;)
15:57:17 *** andythenorth has left #openttd
15:57:22 *** andythenorth has joined #openttd
15:57:38 *** supermop_work has joined #openttd
15:57:45 <peter1138> It's not like PRs get merged ;)
15:57:56 <planetmaker> yeah... they start to pile-up...
15:58:09 <peter1138> I should just make less of them.
15:58:16 <peter1138> Quality of quantity.
15:58:30 <Eddi|zuHause> so much for that :p
15:58:39 <planetmaker> anyway... RL calls... see ya tomorrow or sunday
15:58:58 *** supermop_work_ has joined #openttd
16:00:33 <andythenorth> how do you find quality without quantity? :P
16:00:42 <andythenorth> quality is just a sieve
16:01:20 <andythenorth> but also cull some, not necessarily yours either :P
16:01:28 <peter1138> I've got about 2 hours tonight before I go to watch Captain Marvel.
16:03:05 <peter1138> 24 commits of replacing SmallVector with std::vector
16:04:08 <Eddi|zuHause> tgp is weird... "perlin noise with large primes", and then it uses "small primes" (53, 101, 167)
16:13:10 <andythenorth> shall I put a novelty PR in for fixing tropic?
16:13:14 <andythenorth> then someone can reject it
16:13:39 <peter1138> You have a patch for it?
16:13:58 <peter1138> I wish my own variant was faster.
16:14:15 <Eddi|zuHause> i was looking into using perlin noise with the rivers, instead of pure randomness
16:14:48 <peter1138> Perlin works nicely when you use it in multiple layers, but TGP doesn't work for that.
16:15:25 <peter1138> Needs to use off-map storage, and then you're talking lots of RAM.
16:15:27 *** Gustavo6046 has joined #openttd
16:15:31 <peter1138> Although, it's only world gen, so maybe it's okay.
16:16:20 <Eddi|zuHause> i wasn't about to rewrite tgp (yet), was just looking for a function i can reuse as river priority function
16:17:06 <peter1138> Variety works much better as a perlin-overlay.
16:18:26 <peter1138> And there's some techniques to make nicer valleys.
16:24:50 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth opened pull request #7340: Change: replace tropic landscape map gen rules https://git.io/fhhmJ
16:25:06 <andythenorth> I fully expect that to get closed, but eh, it makes my maps better
16:26:13 <Eddi|zuHause> peter1138: tgp also has this weird sine transform, that kinda conflicts with variety i feel
16:27:06 <andythenorth> it flattens everything down
16:27:24 <andythenorth> I did try just deleting bits of TGP before, there are quite "interesting" results from that
16:33:08 <DorpsGek_II> [OpenTTD/OpenTTD] michicc approved pull request #7231: Add: Add flag and railtype sprite type to draw pre-combined ground sprites. https://git.io/fhhm0
16:35:43 <DorpsGek_II> [OpenTTD/OpenTTD] michicc approved pull request #7289: Add: Configurable ship curve penalties (YAPF) https://git.io/fhhmg
16:36:11 <Eddi|zuHause> as far as i got in this document, it says here you need a reproducible mapping from integers into real [-1,1] interval, which you can either do with a hash function, or storing random values into an array
16:36:51 <Eddi|zuHause> the bits of tgp i read is that he uses a hash function here
16:37:18 <DorpsGek_II> [OpenTTD/OpenTTD] michicc approved pull request #7335: Fix #7334: Ship lost after crossing bridge due to path cache not being consumed while on final bridge end. https://git.io/fhhmr
16:39:43 <DorpsGek_II> [OpenTTD/OpenTTD] michicc approved pull request #7234: Feature: Game setting to define how industries with neutral stations accept and supply cargo from/to surrounding stations. https://git.io/fhhmi
16:40:17 <andythenorth> eh the climate selection buttons at game start have inconsistent padding
16:42:30 <Eddi|zuHause> missing scale factor somewhere?
16:47:27 *** m3henry has joined #openttd
16:48:51 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth opened pull request #7341: Change: add windowshade to rail station construction window https://git.io/fhhmy
16:49:03 <DorpsGek_II> [OpenTTD/OpenTTD] michicc commented on pull request #7301: Feature: When filtering purchase list by cargo type, make buy button perform a refit if required. https://git.io/fhhm9
16:50:38 <DorpsGek_II> [OpenTTD/OpenTTD] michicc commented on pull request #7336: Add: AI API for vehicle group colours https://git.io/fhhmQ
16:51:27 <DorpsGek_II> [OpenTTD/OpenTTD] michicc merged pull request #7335: Fix #7334: Ship lost after crossing bridge due to path cache not being consumed while on final bridge end. https://git.io/fhp4Y
16:51:55 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #7341: Change: add windowshade to rail station construction window https://git.io/fhhm5
16:52:11 <DorpsGek_II> [OpenTTD/OpenTTD] michicc merged pull request #7289: Add: Configurable ship curve penalties (YAPF) https://git.io/fhN1H
16:52:36 <DorpsGek_II> [OpenTTD/OpenTTD] michicc commented on pull request #7234: Feature: Game setting to define how industries with neutral stations accept and supply cargo from/to surrounding stations. https://git.io/fhhmF
16:56:30 <Eddi|zuHause> max height[mapsize]: { 10, 11, 17, 19, 49, 63, 73 }, ///< Mountainous <<-- who came up with these values? there seems to be quite a bit of a jump there from 19 to 49
16:58:34 <michi_cc> Someone from the whole more heightlevels stuff?
17:03:25 <Eddi|zuHause> i think we should completely revisit this approach at some point :)
17:06:52 *** Wormnest has joined #openttd
17:07:38 <Eddi|zuHause> also, why is tgp bothering with these fixed point types, when all the maths casts it to double anyway?
17:10:31 <andythenorth> how about we do something good? :D
17:11:06 <peter1138> Eddi|zuHause, because map-array.
17:12:30 <Eddi|zuHause> also, i think tgp does the interpolation wrong, it seems
17:12:47 <Eddi|zuHause> well "wrong" is relative here, since it's all arbitrary anyway
17:12:55 <peter1138> Yes, it's designed for speed, not 'correctness'
17:13:32 <peter1138> I remember dealing with that one :)
17:14:16 <peter1138> I'll be able to rebase my PRs in about 1hr30m.
17:22:46 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7341: Change: add windowshade to rail station construction window https://git.io/fhhYC
17:23:31 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7341: Change: add windowshade to rail station construction window https://git.io/fhhYl
17:24:08 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7341: Change: add windowshade to rail station construction window https://git.io/fhhY8
17:24:30 <peter1138> I've just realised there's an online file editor... o_O
17:24:53 <peter1138> Ew, not nice to use.
17:25:35 <peter1138> andythenorth, regarding padding, padding is never scaled by zoom level. This is why it sometimes looks a bit odd.
17:26:01 <andythenorth> I didn't notice until today
17:27:57 <Eddi|zuHause> we used the online editor for the news posts, it was nice for that
17:28:11 <peter1138> Also some sprites are slightly different sizes and use offsets to position them, which can mess up the centring code.
17:28:21 *** supermop_work_ has quit IRC
17:28:30 <peter1138> Eddi|zuHause, yeah, not so much a 1000 source file :-)
17:28:34 <peter1138> andythenorth, fix your PR ;)
17:28:39 <Eddi|zuHause> i can imagine :)
17:29:28 <peter1138> "It'd be nice if the "jungle" was more coherent to a "region", like an entire mountain, instead of a blob of the mountain, but I really have no idea how that could be achieved."
17:29:37 <peter1138> That ... can be acheived with perlin layers ;)
17:30:54 <DorpsGek_II> [OpenTTD/OpenTTD] michicc commented on pull request #7341: Change: Add windowshade to rail station construction window https://git.io/fhhYz
17:32:21 <Eddi|zuHause> while at it, we should also find a nicer method to make water borders
17:32:30 <Eddi|zuHause> instead of this jagged shore cut off
17:32:47 <peter1138> Sure. You can guess...
17:33:04 <Eddi|zuHause> do i have 3 guesses?
17:33:14 <Eddi|zuHause> can i wish for more wishes?
17:34:13 *** Progman has joined #openttd
17:34:34 <peter1138> Hmm, wonder if I still have my old railtypes newgrfs.
17:34:45 <peter1138> There were some I created when developing the feature.
17:35:11 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7231: Add: Add flag and railtype sprite type to draw pre-combined ground sprites. https://git.io/fhhYw
17:36:27 <Gwyd> hi is planetmaker around?
17:37:24 <Gwyd> okay, you can probably help though
17:39:07 <Gwyd> I've drawn the 4 missing sprites for OpenGFX and written the code for them. I pm'd andythenorth on tt-f, which he forwarded to planetmaker, who asked me if I have a patch for the source code. I don't know how to go about writing/generating this.
17:39:39 *** HerzogDeXtEr has joined #openttd
17:40:12 <peter1138> Did you modify the existing OpenGFX code?
17:40:34 <Gwyd> Yes, from OpenTTDcoop Dev zone
17:41:02 <peter1138> Did you use mercurial or just downloaded the source as a zip file type thing?
17:41:42 <peter1138> Ok, then the answer is you don't have a patch :D
17:42:55 <Gwyd> so it would be more helpful to post an issue to the repository, with the files and their appropriate locations attached?
17:43:15 <peter1138> I would have suggested trying to use mercurial, but we are in the process of moving it off that.
17:43:27 <peter1138> Yeah, I think there's an issue open already?
17:44:21 <peter1138> If you post the files to that then they are at least there for someone else to deal with.
17:44:39 <Gwyd> I'll update the current issue then
17:51:04 <peter1138> Gwyd, nice design :)
17:51:28 <peter1138> Is the position really correct, though? All the way over to the left?
17:52:29 <Gwyd> yes, I checked against the current sprites
17:53:59 <Gwyd> I initially thought they would be centred but I checked /sprites/png/gui/gui04.png
18:03:11 *** supermop_work has joined #openttd
18:06:38 *** supermop_work_ has joined #openttd
18:08:15 *** supermop_work__ has joined #openttd
18:09:28 *** supermop_work___ has joined #openttd
18:13:05 *** m3henry has joined #openttd
18:15:39 *** supermop_work_ has quit IRC
18:16:16 *** supermop_work__ has quit IRC
18:33:21 *** m3henry has joined #openttd
18:44:38 <peter1138> Mushrooms, onion, chicken...
18:45:02 <peter1138> Oooh, sweet potato sag aloo
18:45:18 <Eddi|zuHause> i had fried potatos with onions, pork and eggs
18:45:25 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN updated pull request #7234: Feature: Game setting to define how industries with neutral stations accept and supply cargo from/to surrounding stations. https://git.io/fh5qg
18:45:26 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN dismissed a review for pull request #7234: Feature: Game setting to define how industries with neutral stations accept and supply cargo from/to surrounding stations. https://git.io/fhhmi
18:46:05 <peter1138> andythenorth, fixy fixy?
18:47:15 <peter1138> Hmmm, need to find my 3D glasses
18:50:40 <Samu> because they tend to pick the same stuff
18:51:16 <nielsm> it would not "improve AI's", it would make their choice of routes less unviable
18:53:41 <peter1138> I'm not opposed to it per se.
18:54:07 *** Thedarkb-T60 has joined #openttd
18:54:38 <peter1138> Which AI uses groups sensibly?
18:54:45 <nielsm> IMO it removes some of the element of competition and could make multiplayer more bland
18:56:08 <peter1138> Distribution is still dependent on rating, no?
18:56:41 <nielsm> yes, but when everyone gets a chance there's much less incitement to be the very best
18:57:52 <nielsm> if you have three companies each with one station, ratings 80%, 78%, and 76%, they'll get almost equal amounts, and there's not really any reason for the 76% company to do any better
18:58:13 <nielsm> while the current system where only the top two stations get any at all, it forces the third company to compete harder
18:58:31 <nielsm> and that makes for more interesting gameplay, I think :)
19:01:20 <peter1138> How does the 3rd company compete if they don't get anything?
19:02:01 <peter1138> It's also worthwhile posting these sorts of thoughts to the PR, so others can consider them.
19:03:52 <Samu> i could change. maybe 50% for the top
19:04:03 <Samu> and the rest is divided between the other stations
19:04:17 <nielsm> I'd suggest trying to distribute by the ratio of the squares of the ratings
19:04:36 <Samu> ratio of the squares of the ratings... uhmm help
19:04:50 <nielsm> I'll write a comment on the PR detailing it
19:17:38 <peter1138> Yup, I actually tested an AI commit ;)
19:18:03 <Samu> hmm colours is something that doesn't attract me much
19:18:17 <peter1138> Does not matter, other people requested it, and players can do it.
19:18:56 <Samu> Content Update 3.6.0 -- Path of Exile: Synthesis
19:18:57 <peter1138> Testing #7231 will be tricky.
19:19:37 <peter1138> Ready to launch in approximately 8 hours. Hmm.
19:19:49 <peter1138> 9.1 GB update, nice.
19:19:55 <peter1138> And I don't even play it.
19:20:24 *** octernion has joined #openttd
19:21:20 <Samu> the league is starting in 1h40m if i'm not mistaken
19:21:39 <peter1138> I don't know what that means.
19:21:50 <Samu> The Synthesis Challenge League
19:22:02 <peter1138> Oh, I found some 2017 diffs.
19:23:50 <Samu> no one here plays poe i guess
19:24:18 <peter1138> I have it installed, I loaded it once, that's about it.
19:24:44 <peter1138> Well, that's encouraging.
19:24:55 *** gelignite has joined #openttd
19:25:31 <peter1138> "stuff.diff" similar
19:25:51 <peter1138> That contains an old patch to remove the group name uniqueness 'feature'
19:25:56 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7184: Feature: Distribute cargo to multiple stations or industries https://git.io/fhhs3
19:28:08 *** Thedarkb1-T60 has joined #openttd
19:29:07 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh approved pull request #7234: Feature: Game setting to define how industries with neutral stations accept and supply cargo from/to surrounding stations. https://git.io/fhhsC
19:30:47 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN merged pull request #7234: Feature: Game setting to define how industries with neutral stations accept and supply cargo from/to surrounding stations. https://git.io/fh5qg
19:33:50 <Samu> gonna experiment with nielsm suggestions
19:37:33 *** andythenorth has joined #openttd
19:39:57 <andythenorth> daylength strikes again
19:40:04 <andythenorth> also what do I need to fixy fix?
19:41:51 *** supermop_work___ has quit IRC
19:46:33 *** supermop_work has joined #openttd
19:47:21 *** supermop_work_ has joined #openttd
19:52:07 *** DoppelD27 has joined #openttd
19:52:20 <Samu> uint cur_worst = amount * pow(st->goods[type].rating, 2) / pow(ratings_sum, 2);
19:53:43 <nielsm> also, it's sum of squares, not square of sum
19:54:10 <nielsm> x*x + y*y + z*z is not equal to (x+y+z) * (x+y+z)
19:54:39 <nielsm> if you take the ratio of squares to ratio of square of sum you get a useless value
19:54:59 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 commented on pull request #7301: Feature: When filtering purchase list by cargo type, make buy button perform a refit if required. https://git.io/fhhsH
19:56:02 <LordAro> nielsm: not that it *really* matters, but i'd just use a single PR for backporting purposes
19:56:26 <nielsm> can do that next time :P
19:56:50 <nielsm> not like issue numbers cost money :)
19:57:51 *** frosch123 has joined #openttd
19:58:40 <Samu> uint cur_worst = amount * st->goods[type].rating * st->goods[type].rating / ratings_square_sum;
20:01:17 <peter1138> Yeah, that's not all that useful :/
20:01:19 <nielsm> what I did before writing up that comment on the PR was put some numbers into Excel and try out a couple formulas
20:01:32 <nielsm> to see how the distribution ended up
20:02:06 <nielsm> (putting some numbers into excel and experimenting with formulas is a common game balancing technique, by the way)
20:04:42 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
20:13:34 <nielsm> "PfnChooseRoadTrack pfnChooseRoadTrack = &CYapfRoad2::stChooseRoadTrack; // default: ExitDir, allow 90-deg" --- I can't imagine a road pathfinder that would disallow 90 degree turns :P
20:14:23 <peter1138> nielsm, there was actually a template for it there, just never used!
20:18:20 <nielsm> peter1138: did you test road path caching with bridges and tunnels?
20:18:38 <DoppelD27> Good evening. Is there a way to find Wiki Pages which are not translated to german yet?
20:18:51 <peter1138> nielsm, indeed, no I haven't.
20:19:19 <peter1138> nielsm, due to the tile being involved, it will appear to work fine.
20:19:29 <peter1138> So it needs a deeper investigation than for ships.
20:27:15 <nielsm> okay it handles the local road reconstruction harrassment fine (as well as you could expect)
20:29:16 <peter1138> It'll handle bridges and tunnels fine, but may unnecessarily invalidate the cache.
20:29:43 <peter1138> Although actually, in the RV cache it explicitly doesn't add tiles that are not junction tiles, so it may actually work perfectly fine.
20:31:00 <nielsm> seems to go fine if I remove the tile right after tunnel exit, while the bus is in tunnel
20:47:30 <Samu> got a real test case here nielsm
20:48:02 <Samu> does seem to do what you ask
21:01:46 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 opened pull request #7345: Fix 13962a847, 00d28a500: forgotten squirrel_export run https://git.io/fhhZl
21:03:25 *** octernion has joined #openttd
21:10:02 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh approved pull request #7345: Fix 13962a847, 00d28a500: forgotten squirrel_export run https://git.io/fhhZi
21:15:03 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 merged pull request #7345: Fix 13962a847, 00d28a500: forgotten squirrel_export run https://git.io/fhhZl
21:18:38 <Eddi|zuHause> and somehow my river widening works only in one diagdir, and not the other
21:19:21 <Eddi|zuHause> i need to play with this a bit more to understand what happens
21:19:42 <andythenorth> it's certainly interesting
21:37:08 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7346: Fix: rebase to 1.9 branch for backport PRs https://git.io/fhhn3
21:39:06 <nielsm> andythenorth: why does Horse not have a nuclear steam-electric locomotive?
21:39:41 <glx> ERROR: This commit is not on top of master
21:39:55 <andythenorth> nielsm: what effect sprite would it use?
21:40:08 <andythenorth> green glowing steam?
21:40:25 <nielsm> just to placate the silly players who insist on using steam locos forever
21:40:31 <nielsm> (becauyse of aethetics)
21:40:42 <andythenorth> they need daylength!!!!!
21:44:26 <glx> will open an issue on the CF
21:46:32 <DorpsGek_II> [OpenTTD/CompileFarm] glx22 opened issue #29: Commit checker always expect origin/master, even for non master PRs https://git.io/fhhn4
22:48:51 *** octernion has joined #openttd
22:50:33 *** Progman_ has joined #openttd
22:56:09 *** Progman_ is now known as Progman
23:12:49 *** supermop_work_ has quit IRC
23:16:45 *** Thedarkb-T60 has joined #openttd
23:31:46 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6993: Enhancements: Allow the scenario (.scn) to run by using the parameter in the openttd.cfg file. https://git.io/fhhC0
23:34:24 *** supermop_work has joined #openttd
23:35:37 *** supermop_work_ has joined #openttd
23:43:44 *** andythenorth has joined #openttd
continue to next day ⏵
