IRC logs for #openttd on OFTC at 2021-03-06
⏴ go to previous day
00:01:52 <DorpsGek> [OpenTTD/OpenSFX] orudge approved pull request #28: Add: Polish (pl_PL) and Finnish (fi_FI) translations https://git.io/JqUSE
00:02:12 <DorpsGek> [OpenTTD/OpenSFX] orudge merged pull request #28: Add: Polish (pl_PL) and Finnish (fi_FI) translations https://git.io/JqJp8
00:08:41 *** nielsm has quit IRC (Ping timeout: 480 seconds)
00:24:13 *** Progman has quit IRC (Remote host closed the connection)
00:28:05 *** erle- has quit IRC (Remote host closed the connection)
00:29:30 <DorpsGek> [OpenTTD/OpenTTD] James103 commented on issue #8806: DPI too high makes UI unusable on Mac OS X catalina https://git.io/JqU9a
00:33:59 <DorpsGek> [OpenTTD/OpenTTD] scovit commented on issue #8806: DPI too high makes UI unusable on Mac OS X catalina https://git.io/JqU9a
00:38:03 <DorpsGek> [OpenTTD/OpenTTD] scovit opened issue #8807: Segmentation fault when searching for internet servers https://git.io/JqU9b
00:41:07 <DorpsGek> [OpenTTD/OpenTTD] scovit opened issue #8808: Die with failed assertion on MacOSX with cocoa-opengl driver https://git.io/JqUHv
00:45:03 <DorpsGek> [OpenTTD/OpenTTD] James103 commented on issue #8808: Die with failed assertion on MacOSX with cocoa-opengl driver https://git.io/JqUHv
00:50:41 *** Gustavo6046 has quit IRC (Ping timeout: 480 seconds)
00:54:39 <DorpsGek> [OpenTTD/OpenTTD] scovit commented on issue #8808: Die with failed assertion on MacOSX with cocoa-opengl driver https://git.io/JqUHv
00:58:16 *** Compu has quit IRC (Remote host closed the connection)
01:01:52 *** Gustavo6046 has joined #openttd
01:24:00 *** WormnestAndroid has quit IRC (Remote host closed the connection)
01:26:07 *** WormnestAndroid has joined #openttd
01:26:31 *** WormnestAndroid has quit IRC (Remote host closed the connection)
01:28:43 *** WormnestAndroid has joined #openttd
01:54:12 <DorpsGek> [OpenTTD/OpenTTD] jdeepankur commented on pull request #8535: Fix #8506: Towns shouldn't add junctions to NewGRF roads they cannot build https://git.io/JqU5l
02:27:26 *** Flygon_ has joined #openttd
02:30:26 *** Flygon__ has joined #openttd
02:34:12 *** Flygon has quit IRC (Ping timeout: 480 seconds)
02:37:12 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
02:40:42 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
03:04:49 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
03:12:54 *** Flygon_ has joined #openttd
03:13:17 *** Wormnest has quit IRC (Quit: Leaving)
03:16:54 *** Flygon__ has joined #openttd
03:19:42 *** Flygon has quit IRC (Ping timeout: 480 seconds)
03:23:42 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
03:25:54 *** Flygon_ has joined #openttd
03:32:42 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
03:34:10 *** Wuzzy has quit IRC (Remote host closed the connection)
03:36:42 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
03:36:54 *** Flygon_ has joined #openttd
03:41:25 *** Flygon__ has joined #openttd
03:43:42 *** Flygon has quit IRC (Ping timeout: 480 seconds)
03:48:12 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
03:49:26 *** Flygon_ has joined #openttd
03:56:12 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
04:04:07 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:10:19 <Gustavo6046> What is the best way to merge 3 railway lines into 2 with signals?
04:10:34 <Gustavo6046> Just "X-cross" them and put path signals around the cross?
04:11:13 <Gustavo6046> But then, what if I ever have more than three trains? Wouldn't this be susceptible to lockups?
04:11:40 <Gustavo6046> What if I want a set of railway lines that can support a number of trains that isn't bound by lockups, things like this?
04:15:54 *** Flygon__ has joined #openttd
04:22:42 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
04:41:28 *** jeremy[m] has quit IRC (synthon.oftc.net larich.oftc.net)
04:41:28 *** snail_UES_ has quit IRC (synthon.oftc.net larich.oftc.net)
04:41:28 *** freu[m] has quit IRC (synthon.oftc.net larich.oftc.net)
04:41:28 *** k-man has quit IRC (synthon.oftc.net larich.oftc.net)
04:41:28 *** Extrems has quit IRC (synthon.oftc.net larich.oftc.net)
04:41:33 *** Extrems has joined #openttd
04:41:54 *** snail_UES_ has joined #openttd
04:46:02 *** freu[m] has joined #openttd
04:47:28 *** jeremy[m] has joined #openttd
04:48:23 *** Flygon_ has joined #openttd
04:55:16 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
05:01:54 *** Flygon__ has joined #openttd
05:08:42 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
05:35:39 *** snail_UES_ has quit IRC (Quit: snail_UES_)
05:47:24 *** Flygon_ has joined #openttd
05:54:12 *** Flygon__ has quit IRC (Ping timeout: 480 seconds)
05:55:47 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
05:57:57 *** Smedles has quit IRC (Quit: No Ping reply in 180 seconds.)
05:59:21 *** Smedles has joined #openttd
06:12:53 *** Flygon_ has joined #openttd
06:19:47 *** Flygon has quit IRC (Ping timeout: 480 seconds)
06:33:56 <supermop_Home> Gustavo6046 how do you mean?
06:35:06 <supermop_Home> one line with three tracks reduces to two tracks? or like 3 branches connect to two branches
06:38:14 <supermop_Home> a regular two track line with directional signals will support an arbitrarily large number of trains... basically as many as fit into the length of the line minus 1
06:44:03 <supermop_Home> a 3rd bi-directional track in the center will add a bit more capacity for surges of congestion, but has little sustained effect as it cannot quickly switch direction
06:45:16 <supermop_Home> and if you have a lot of trains heading north that need the extra capacity, eventually you will have a lot of trains returning south
06:50:13 <supermop_Home> the length of the center track and signal spacing on it is something you need to strike a balance with, between favoring many same direction trains following one another on it, or favoring changing direction
06:51:16 <supermop_Home> but if you have just 3 trains going between A and B, a simple 2 track line is plenty. one track north, one track south
06:52:24 *** Flygon__ has joined #openttd
06:53:18 <supermop_Home> keeping at least one track dedicated for each direction prevents track locking up, as Northbound trains can never block southbound trains
06:59:17 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
07:21:12 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #8806: DPI too high makes UI unusable on Mac OS X catalina https://git.io/JqU9a
07:23:05 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #8807: Segmentation fault when searching for internet servers https://git.io/JqU9b
07:23:08 <DorpsGek> [OpenTTD/OpenTTD] LordAro closed issue #8807: Segmentation fault when searching for internet servers https://git.io/JqU9b
08:04:52 *** sla_ro|master has joined #openttd
08:15:32 *** andythenorth has joined #openttd
08:21:04 *** Progman has joined #openttd
08:35:27 *** rptr_ has quit IRC (Quit: Leaving)
08:36:16 <andythenorth> changelogs entries, what's the best case to use?
08:36:34 <andythenorth> I seem to switch between 'Convert' and 'Converted' or 'Fix' and 'Fixed'
08:36:45 <andythenorth> Apple only use 'Converts' and 'Fixes' etc
08:43:00 <Wolf01> I use "[FIX]"/"Fix:" and "Fixed stuff"
08:49:23 <dwfreed> the recommendation for git commit messages is to use present tense verbs
08:49:42 <dwfreed> "fix foo" rather than "fixed foo"
09:00:14 <andythenorth> gets weird when written as a player-oriented changelog about complex changes
09:00:41 <andythenorth> partly because it sounds like an imperative instruction sometimes
09:09:21 <Wolf01> Looks like a roadmap instead of a changelog
09:12:33 <andythenorth> 'Adjust running costs'
09:12:41 <andythenorth> did I adjust them?
09:12:46 <andythenorth> or can the player adjust them?
09:17:11 <Eddi|zuHause> andythenorth: if the player adjusts them, it should have been "allow adjusting running costs"
09:17:52 <andythenorth> that assumes that the reader knows the convention
09:20:37 <andythenorth> because read in isolation 'Adjust running costs' implies the player can adjust them
09:20:53 <andythenorth> 'Create beautiful presentations'
09:21:01 <andythenorth> 'Connect with all your friends on social media'
09:21:06 <andythenorth> 'Share cat videos'
09:21:58 <andythenorth> 'Replace sprites'
09:22:01 <andythenorth> 'Repaint trains'
09:22:07 <andythenorth> 'Make trains faster'
09:22:31 <andythenorth> I am only beating this horse because my changelogs are quite inconsistent
09:23:09 <Eddi|zuHause> that's a feature list, not a changelog
09:23:32 <andythenorth> "but what is a changelog"?
09:24:11 * andythenorth reads the changelog for Trees and Tents
09:24:43 <andythenorth> hmm, inconsistent also
09:25:38 <andythenorth> what is the audience for a changelog?
09:25:41 <andythenorth> serious question
09:27:10 <Eddi|zuHause> people who actively seek out to read it. it's not a sales pitch you need to barf on everyone
09:27:31 <andythenorth> and yet a sales pitch is needed, or at least a gameplay guide
09:27:40 <andythenorth> so I think the changelog might get deleted as a concept
09:27:54 <andythenorth> it's utterly redundant in a world containing version control
09:28:00 <Eddi|zuHause> these are very different concepts
09:28:48 <Eddi|zuHause> no. it duplicates information that would also be found in a version control system. but filtered, cleaned up, shortened, more focused
09:29:05 <andythenorth> but for an audience of...? probably nobody?
09:30:43 <Eddi|zuHause> the point is to make the information more accessible. for people who want to know, to not force them to wade through a whole version control system. for people who want to skim a list to find out if anything changed would really bother them.
09:32:46 <andythenorth> my proposition is that in a world of limited interest
09:32:51 <andythenorth> those people don't get what they want
09:33:42 <andythenorth> write a changelog
09:33:46 <andythenorth> write a gameplay guide
09:33:50 <andythenorth> update the first post
09:33:54 <andythenorth> add a release post
09:33:59 <andythenorth> write text for bananas
09:34:08 <andythenorth> possibly update readme
09:34:10 <peter1138> It's missing docks.
09:34:24 <andythenorth> I saw some on your livestream once
09:34:28 <andythenorth> they worked and everything
09:35:00 <andythenorth> all the best demos
09:35:07 <andythenorth> don't click the button that causes asserts
09:37:05 <Eddi|zuHause> i don't know what your problem is.
09:37:42 <Eddi|zuHause> you don't "write" a changelog. you take the existing commit messages, and delete the ones that would be of no interest.
09:37:49 <Eddi|zuHause> possibly reorder them by importance
09:38:56 <andythenorth> that sounds like work
09:39:23 <andythenorth> it's a maintainer expectation that seems to have no value
09:42:48 <Eddi|zuHause> just that you don't see the value, doesn't mean it doesn't exist
09:44:55 <michi_cc> I think it mostly depends on what you want to achieve. If you don't care one bit whether you have 1 or 1000000 "customers", just upload if with a comment of whatever. If you do care about how many people you can reach, these things get value for engaging with people.
09:45:58 <andythenorth> in a world which is mostly reddit 'yolo'
09:46:36 <Eddi|zuHause> spoiler: reddit is not "the world"
09:46:37 <andythenorth> I guess there might be one person who needs a detailed changelog before choosing to never ever upgrade the grf again
09:46:51 <andythenorth> reddit is a pretty good representation of most of the world
09:47:22 <andythenorth> no I caveat, most of the world under the age of about 40 or so
09:48:10 * peter1138 installs Visual Studio Build Tools for no particular purpose.
09:49:11 <Eddi|zuHause> andythenorth: let's assume i'm a user
09:49:23 <Eddi|zuHause> and i want to update your grf
09:49:32 <Eddi|zuHause> then i need a quick overview of changes
09:49:39 <Eddi|zuHause> to see whether i need to adjust my playstyle
09:51:01 <andythenorth> peter1138 what's the gameplay like in those? Is it a good mechanic? Does it have in-game purchases?
09:51:06 <Eddi|zuHause> this changelog needs to be optimized for reading-time.
09:51:39 <Eddi|zuHause> the longer it takes to read, the less people read it, the less value it has
09:51:44 <peter1138> Eddi|zuHause, use a state machine.
09:53:01 <andythenorth> Eddi|zuHause arbitrary claim
09:53:09 <andythenorth> the shortest changelog is
09:53:19 <andythenorth> factually accurate
09:56:39 <_dp_> is Pinocchio a changelog?
09:57:15 <Eddi|zuHause> andythenorth: that's what you get for ignoring competing optimization goals
10:04:38 <peter1138> Er, x86, x86_amd64, amd64, or amd64_x86? what?
10:05:14 <Eddi|zuHause> also, you're missing x64
10:06:59 <peter1138> Oh, amd64/x64 is equivalent here, apparently.
10:07:10 <peter1138> Not Intel's failed 64 bit.
10:07:12 <Sang> Hello, i want to ask about the message "city too close to other city", could the distance between the cities be changed?
10:07:27 <Eddi|zuHause> no, that'd be ia64
10:07:49 <Sang> Because i am building very detailed map of Korea
10:08:25 <Eddi|zuHause> Sang: i think that's hardcoded
10:09:02 <peter1138> Urgh. To install vcpkg, make a git clone of... wtf.
10:09:07 <Eddi|zuHause> peter1138: to me, saying "amd64" acknowleges that "ia64" still exists
10:09:37 <Sang> Can you help me solve it?
10:09:54 <Eddi|zuHause> Sang: other than make a bigger map? probably not
10:10:22 <Sang> I already have the biggest map
10:10:49 <Eddi|zuHause> the game mechanics get a bit wonky if towns are too close to each other
10:11:48 <andythenorth> why am I playing Blitz on a Saturday morning?
10:11:59 <andythenorth> it's like one of the top 7 worst days to play Blitz
10:12:12 <Eddi|zuHause> i mean, you can disable the check when placing towns in scenario editor, and people without that change will be able to load the scenario. but that won't solve any of the gameplay problems
10:13:09 <Sang> well thanks for help I will try anything to solve some things
10:13:15 *** Sang has quit IRC (Quit: Page closed)
10:30:20 <peter1138> Bah, can't get vscode/cmake/vcpkg to play ball.
10:30:35 <peter1138> And then it worked. Damn.
10:31:28 <Eddi|zuHause> i hate when it starts working without giving an indication what was wrong before
10:31:33 *** rptr_ has quit IRC (Quit: Leaving)
10:31:59 <peter1138> It was a magical restart.
10:32:09 <peter1138> Also it appears to be thrashing my CPU :D
10:32:41 <peter1138> Oh no, that's Origin which has broken for some reason.
10:34:10 <peter1138> No language packs. Hmm..
10:35:04 <peter1138> Seems to have put the exe in a subdirectory. How unuseful.
10:39:40 <andythenorth> are you compiling your own version of Trees?
10:40:32 <michi_cc> If you're using the cmake VS project file generator on OTTD, you need to manually set the working dir in the debug options to the build dir.
10:44:30 <FLHerne> andythenorth: yay new CHIPS
10:50:36 *** Gustavo6046 has quit IRC (Ping timeout: 480 seconds)
10:53:42 <FLHerne> I like the new loaders, those will be good
10:54:03 <FLHerne> And mud platforms that actually match :-)
10:54:36 <FLHerne> Huh, somehow I never thought of the arrangement in pic 2
10:54:57 <andythenorth> me neither until I did
10:55:05 <andythenorth> wasn't intentional, just accident
10:55:23 <FLHerne> Those dock ones would look much better with INFRA foundations
10:55:58 <FLHerne> How much of a pain is it for tiles to have custom foundations? I know NewStations does it a bit
10:56:38 <andythenorth> I'm sure it can be done
10:56:47 <andythenorth> I am avoiding learning the station spec
10:57:06 <andythenorth> I could show you a link to the CHIPS tile template, but you'd have to sign a waiver in advance for trauma
10:57:28 <andythenorth> I just edit sprites, cargo-cult the existing nfo, and step away
10:57:34 <FLHerne> well...it might be interesting
10:59:47 <andythenorth> I'm not quite sure what does what
11:00:15 <andythenorth> also there are GRM sprites in play, which I treat as an interesting guessing game (as to intended purpose or how to use them correctly)
11:00:36 <andythenorth> CHIPS is not aided by 'yexo wrote it to learn the newgrf station spec'
11:01:26 <peter1138> Alright, how to diagnose this OpenGL pausing glitch?
11:02:13 <peter1138> Also how to stop the graphs jumping around in scale. Or is that a symptom...
11:03:32 <peter1138> michi_cc, alright... how is that done?
11:04:49 <andythenorth> suppose I should fix all the FIRS 4 beta bugs
11:04:54 <andythenorth> and release it or something
11:06:17 <michi_cc> peter1138: In proper VS, you just right click on the 'openttd' project, select properties and change it in the debugging section. No clue if it is the same in VS Code.
11:07:00 <peter1138> <OutDir Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">I:\src\OpenTTD\build\Debug\</OutDir>
11:07:35 <peter1138> Hmm, localdebuggerworkingdirectory defaults to bin too, instead of build.
11:08:36 <michi_cc> That's because the VS project generator is the only one that defaults to putting the EXE into a subdir. You might also need to copy the contents of bin/ into the build dir, if you run any AIs that need the compat files.
11:12:55 <peter1138> Changing either of them has no effect. Meh.
11:14:15 <peter1138> Ok, well I give up.
11:16:28 <michi_cc> glx knows a lot more about that stuff, except of course he's not here right now.
11:22:04 *** dihedral has joined #openttd
11:30:10 *** frosch123 has joined #openttd
11:57:26 <andythenorth> what's the number of combinations for the 829 grfs on Bananas?
11:57:36 <andythenorth> turns out my grfs break a lot of other grfs
11:57:48 <andythenorth> going to be a long long long list of incompatible grfs needed
11:58:17 <_dp_> why do uncompressed saves take priority over lzo?
12:00:08 <_dp_> I'd guess lzo is kind of deprecated so it tries to lower the amount of lzo saves in existence
12:00:18 <_dp_> but it's not like it can be removed completely anyway
12:02:25 <_dp_> hm, or it's just because uncompressed was added later and no one bothered with priorities xD
12:10:42 <andythenorth> Termite seems to be catastrophic for other sets
12:10:52 <andythenorth> xUSSR and Japan Set I tested so far
12:10:58 <andythenorth> no vehicles in depots
12:11:28 <andythenorth> FIRS breaks vehicle sets that use cargo subtype refits, so I'm testing all the major train grfs for that
12:11:39 <andythenorth> but along the way I'm discovering I've broken other grfs
12:11:49 <andythenorth> kinda contemplating the OzTrans approach here to grfs
12:11:58 <_dp_> how to call a pair of SaveLoadFormat and compression level?
12:12:06 <_dp_> SaveLoadStuff is the best I could think of %)
12:13:34 <andythenorth> railtype grfs seem to trivially break train grfs
12:13:39 <andythenorth> maybe that's fine
12:13:55 *** HerzogDeXtEr has joined #openttd
12:15:47 *** WormnestAndroid has quit IRC (Remote host closed the connection)
12:16:01 *** WormnestAndroid has joined #openttd
12:16:24 <FLHerne> Yeah, that dep is obvious
12:18:54 <michi_cc> _dp_: SaveLoadPreset maybe?
12:20:34 <_dp_> yeah, a good option I guess
12:20:45 <_dp_> SavePreset though as it only applies to saving
12:25:01 <andythenorth> So what's better?
12:25:09 <andythenorth> learn nfo stations?
12:25:22 <andythenorth> do something else?
12:25:43 <andythenorth> m4nfo stations seems to be the prevailing solution
12:36:06 <peter1138> Freshly 'baked' petit pains
12:36:35 *** dekeract[m] has left #openttd (User left)
13:04:28 <_dp_> error: ‘LoadFilter’ does not name a type; did you mean ‘RoadBits’?
13:04:36 <_dp_> thanks gcc, very helpful :p
13:21:03 *** WormnestAndroid has quit IRC (Remote host closed the connection)
13:24:20 *** WormnestAndroid has joined #openttd
13:27:57 <andythenorth> maybe I make a grf that uses NRT?
13:28:30 <DorpsGek> [OpenTTD/OpenTTD] JGRennison opened issue #8809: Crash/assertion failure when removing airport when hangar window open https://git.io/JqTCW
13:30:20 <andythenorth> omelette, eggs :)
14:04:02 <DorpsGek> [OpenTTD/OpenTTD] James103 opened issue #8810: "Aircraft ran out of fuel" messages point to tile 0 instead of where the aircraft is. https://git.io/JqTlw
14:20:15 *** tokai|noir has joined #openttd
14:20:15 *** ChanServ sets mode: +v tokai|noir
14:25:52 <frosch123> widelands added libcurl for downloading add-ons
14:26:48 <TrueBrain> PR for OpenTTD when?
14:27:02 *** tokai has quit IRC (Ping timeout: 480 seconds)
14:27:51 <frosch123> they also use github as add-on database :)
14:28:02 <frosch123> some ideas seem to be successful
14:31:50 <frosch123> "sync_safe: Optional: A boolean indicating that this add-on will not desync in replays and network games. This should not be set by the add-on author." <- hmm
14:32:36 <frosch123> i wonder who will set it... community poll? how often did this add-on desync?
14:34:57 <frosch123> hmm, their add-on repo has a lot of "translator bot" commits
14:35:21 <frosch123> sounds like the andy idea: bananas adds translations to contents
14:35:46 <TrueBrain> Honestly not a bad idea
14:35:56 <frosch123> would need grf container 3 :p
14:36:04 <frosch123> so that the texts are not part of the md5sum
14:36:17 <TrueBrain> Ideally translations are a new package next to the current
14:36:20 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: [WIP] Change: add support for next/previous railtype global hotkeys https://git.io/JqT4v
14:36:28 <andythenorth> means I never have to pull untrusted commits
14:36:52 <frosch123> TrueBrain: that sounds like compatbility hell
14:37:05 <TrueBrain> So updating translations doesn't require downloading the whole newgrf again
14:37:07 <frosch123> we do not even offer a strgen download, for just that reason
14:37:28 <TrueBrain> I was thinking about zbase
14:37:31 <andythenorth> even with newgrf eints, I didn't ship some translations for a year or more
14:37:34 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: [WIP] Change: add support for next/previous railtype global hotkeys https://git.io/JqT4k
14:37:38 <andythenorth> just because I didn't know they were there
14:37:42 <TrueBrain> Offering a new download every day because someone translated a string
14:37:49 <TrueBrain> Would be .. evil :D
14:38:12 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: [WIP] Change: add support for next/previous railtype global hotkeys https://git.io/JqT4I
14:38:52 <frosch123> i think we had that problem with yeti
14:39:22 <TrueBrain> So it is not only grfv3, but also a way to sideload translations :D
14:39:31 <supermop_Home> TrueBrain you'd need some kind of algorithm like Instagram that knows when the user last updated and offers them a new version only when enough has changed that you think it meets their threshold for wanting one
14:39:37 <TrueBrain> And know when to update
14:39:38 <supermop_Home> and im sure people will love that
14:39:54 <frosch123> TrueBrain: i guess it does not need grfv3 then
14:40:15 <TrueBrain> Can strings be sideloaded or anything?
14:40:15 <frosch123> grf would just contain baselang then, and the translations would use the basetexts as key
14:40:17 <supermop_Home> though it would be hilarious to start feeding people newgrfs based on ML of their behavior
14:40:24 <TrueBrain> Merged client-side?
14:40:41 <supermop_Home> access the microphone
14:40:48 <frosch123> newgrf have no stable id for texts, that would identify them accross different grf versions
14:40:59 <frosch123> so using the english text as key is the closest we have
14:41:03 <supermop_Home> "oh this guy was talking about trolley buses the other day, show him this NewGRF"
14:41:10 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: [WIP] Change: add support for next/previous railtype global hotkeys https://git.io/JqT4O
14:41:39 <TrueBrain> We could link it to the md5sum, translations ..
14:42:03 <TrueBrain> Dunno .. sounds it is doable, but needs work :D
14:42:18 <TrueBrain> And a heavy modification to eints, I guess :p
14:42:28 <TrueBrain> And link BaNaNaS to eints somehow ..
14:42:55 <Timberwolf> Most of those algorithms it'd be something like, "oh you downloaded a trolleybus GRF, now you will only be shown trolleybuses until the end of time, trains and boats and industry GRFs do not exist for you".
14:43:11 <supermop_Home> andythenorth drawing a 6px puff of steam is hard
14:43:30 <supermop_Home> about as hard as a 7px old-timey flywheel
14:44:07 <supermop_Home> not sure if i need the big brass dome thing
14:44:24 <supermop_Home> without it this looks more like a compressor than a steam engine
14:45:45 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: [WIP] Change: add support for next/previous railtype global hotkeys https://git.io/JqT4l
14:47:05 <frosch123> i do not have enough cores to compile widelands
14:47:25 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: [WIP] Change: add support for next/previous railtype global hotkeys https://git.io/JqT40
14:49:07 <DorpsGek> [OpenTTD/OpenTTD] grossws commented on pull request #7851: [WIP] Change: add support for next/previous railtype global hotkeys https://git.io/JqT4a
14:59:19 <supermop_Home> stubby steam engine better?
15:02:14 <andythenorth> "very small pixels"
15:03:59 <supermop_Home> i'm torn as i think the flywheel really needs to be red, but so does the gas can
15:04:47 <supermop_Home> i'd prefer some kind of color coding of grey or black = steam, red = gasoline/diesel, blue = electric
15:07:40 <supermop_Home> Timberwolf that would somewhat funny
15:08:18 <supermop_Home> i avoid looking at any 'discover' or suggested / reels stuff on instagram
15:09:12 <supermop_Home> but some early viewing i made has convinced it that i primarily want to see Yugoslavian brutalist architecture and Spanish language Simpsons memes
15:09:28 <supermop_Home> so that's all that is ever there in that tab
15:11:41 <supermop_Home> i think i'd like to make a station grf next
15:16:24 <supermop_Home> andythenorth chips is 10 years old?
15:17:12 <_dp_> hm... is static_assert(x == 5 && x <= 8) too weird to use?
15:18:29 <nielsm> unless it's generated code yes that's just weird to write
15:19:32 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8535: Fix #8506: Towns shouldn't add junctions to NewGRF roads they cannot build https://git.io/JqTBS
15:20:12 <_dp_> well, they're for completely different reasons... == 5 to make sure that it stays the same regardless of ifdefs
15:20:24 <_dp_> and <= 8 because of some implementation details down the line
15:53:11 <_dp_> how do I pass struct to std::thread? it keeps demanding I std::ref() it and it works but doesn't look right
15:53:17 <_dp_> I want to copy that struct, not ref it
15:54:25 <frosch123> usually you pass a lambda to std::thread
15:55:40 <_dp_> for lambda arguments I mean
15:56:20 <_dp_> though I guess it's StartNewThread that may be at fault here, not std::thread itself
15:59:38 *** Flygon__ has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:00:37 <nielsm> you only get a pointer sized variable to pass to the thread yeah
16:00:44 <frosch123> it forwards everything, so SaveFileToDisk defines the parameter type
16:00:49 <nielsm> if you want to pass something bigger you need to pass a pointer to it
16:01:07 <frosch123> nielsm: are you looking at ottd 0.6 source?
16:01:48 <nielsm> I'm just commenting on a conversation I haven't been following at all
16:02:09 <frosch123> StartNewThread is a variadic template that perfect-forwards all arugments to the functor
16:03:19 <frosch123> oh, wtf. why does it do it like that...
16:04:48 <frosch123> _dp_: it does a decay_copy
16:05:19 <frosch123> so your compressionconfig may lack some copy-constructors or similar
16:10:55 <LordAro> TrueBrain: right, so where are these TICC/TOCCs likely to need to go?
16:14:34 <TrueBrain> Getting dinner now, but basically to start in the videodriver around the game tick, draw tick and sleep :)
16:14:47 <TrueBrain> Can do more details later tonight ;)
16:15:54 <DorpsGek> [OpenTTD/OpenTTD] LordAro dismissed a review for pull request #7851: [WIP] Change: add support for next/previous railtype global hotkeys https://git.io/Jtyqq
16:20:50 <LordAro> openttd: /home/lordaro/dev/OTTD/openttd/src/pathfinder/yapf/yapf_destrail.hpp:201: bool CYapfDestinationTileOrStationRailT<Types>::PfCalcEstimate(CYapfDestinationTileOrStationRailT<Types>::Node&) [with Types = CYapfRail_TypesT<CYapfRail1, CFollowTrackT<TRANSPORT_BEGIN, Train, true>, CNodeList_HashTableT<CYapfRailNodeT<CYapfNodeKeyTrackDir>, 8, 10>, CYapfDestinationTileOrStationRailT, CYapfFollowRailT>;
16:20:56 <LordAro> CYapfDestinationTileOrStationRailT<Types>::Node = CYapfRailNodeT<CYapfNodeKeyTrackDir>]: Assertion `n.m_estimate >= n.m_parent->m_estimate' failed.
16:20:59 <LordAro> that's... not the crash i was going for
16:22:10 <LordAro> and it looks suspiciously like an overflow
16:24:44 <_dp_> frosch123, I didn't rly get what you were trying to say there but std::move looks like a correct solution
16:29:15 <TrueBrain> LordAro: the trick btw is to set the count low, like 10, and watch which increases when you game stutters .. it should be either of those 3 :p
16:29:39 <TrueBrain> If none, try TIC/TOC instead
16:32:01 *** Gustavo6046 has joined #openttd
16:32:12 *** Gustavo6046 has quit IRC ()
16:33:03 *** snail_UES_ has joined #openttd
16:34:23 *** Wormnest has joined #openttd
16:36:02 *** Gustavo6046 has joined #openttd
16:37:48 <supermop_Home> i need to find a freight forwarder in germany
17:01:34 <Gustavo6046> Is there no NewGRF that adds FIRS-compatible cargo choppers/airplanes?
17:01:49 <Eddi|zuHause> Gustavo6046: av8?
17:02:42 <Gustavo6046> Oh, that looks liek it could be it!
17:03:20 <Gustavo6046> "Full range of airplanes and helicopters" sounsd very promising, and kind of like how I'd describe FISH or 2cc in their respective categories.
17:04:01 <frosch123> there are very different versions of av8
17:04:08 <frosch123> i recommend to use av8, and avoid av9
17:04:56 <Gustavo6046> I'll use av8 2.1, I think that was the version I saw?
17:05:14 <Gustavo6046> frosch123: any particular reason to avoid av9? Is it like an unfinished/beta version, like FIRS 4?
17:05:49 <frosch123> no, both are "finished" but they have very different "visions" about what an aircraft set should be
17:05:57 <frosch123> av8 has lots of different aircraft
17:06:35 <frosch123> av9 is an intentionally trimmed down version. not sure how many, about 10 aircraft
17:06:44 <frosch123> so it has almost no variety
17:08:07 <DorpsGek> [OpenTTD/OpenTTD] ldpl updated pull request #8805: Feature: Allow client and server to negotiate on compression to use for the savegame https://git.io/Jqf7C
17:13:11 <Gustavo6046> Also, I think I finally figured a solution to that passenger/two-way full load disproportion dilemma.
17:13:44 <Gustavo6046> Let me explain. I'd try to set up passenger lines, and have vehicles 'wait for any full load' at both sides.
17:14:24 *** ChanServ sets mode: +v tokai
17:14:57 <Gustavo6046> This was a problem, because they'd always wait too much in the side with more passenger influx, while there were few vehicles to pick up passengers from the other side, resulting in one side with too many waiting vehicles while the other had increasing amounts of people waiting (= hurt ratings).
17:15:19 <Gustavo6046> I solved this by making *two* sets of shared orders; one would wait at a side but not at the other, and the other set would wait at the *other* side.
17:15:56 <Gustavo6046> Is there a way to solve this problem with just a single set of shared orders while still waiting for a full load when possible?
17:16:23 <Gustavo6046> Preferably a simple solution - 'simple' here means not using that confusing timetable thingamajig thingy.
17:18:57 <Eddi|zuHause> you can just set the wait times at the stations to something longer than a full unload and load would take, then there is no clumping up due to different load times anymore
17:21:11 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
17:21:21 <Eddi|zuHause> there's unfortunately no way to set a load time as "don't wait longer than this for a full load"
17:25:25 <LordAro> TrueBrain: trouble is, i don't really know where those places are
17:25:34 <LordAro> i'm usre i can find them, given time, but...
17:37:19 <andythenorth> Eddi|zuHause conditional orders of course
17:37:26 <andythenorth> the intuitive and often-suggested solution
17:37:34 <andythenorth> whereby a vehicle loops via another station
17:37:38 <andythenorth> completely sensible
17:37:39 <LordAro> TrueBrain: actually, i'm having no luck reproducing this at all
17:40:43 <Gustavo6046> andythenorth: hmm?
17:41:12 <Gustavo6046> like, take all cargo without waiting, then go to dummy station right next, and come back to load more if conditional jump succeeds?
17:45:14 *** gelignite has joined #openttd
17:47:47 <Eddi|zuHause> yeah, pretty much
17:49:01 <TrueBrain> LordAro: I would say those 2 problems are completely unrelated :D :D
17:49:07 <TrueBrain> but yeah if you cannot reproduce it, you cannot debug it
17:49:46 <LordAro> yeah, i found the things eventually
17:50:00 <TrueBrain> maybe peter1138 can add some TICC / TOCCs and see if he can pin down where the slowdowns happen
17:50:05 <LordAro> but yeah, no luck (reliably) reproducing it
17:50:16 <LordAro> very occasional jumps, but nothing consistent
17:50:20 <TrueBrain> I mainly need to know what is slowing down
17:50:31 <LordAro> (i tried debug & release builds)
17:50:46 <TrueBrain> if you look at the TICC/TOCC macro you see they are really simple macros; sometimes I use them to only print values that are above a limit I set
17:51:30 <Gustavo6046> I keep finding myself missing a vehicle set of some type.
17:51:37 <Gustavo6046> But I think now I've caught 'em all!
17:51:38 <TrueBrain> but in reality, we should add some PFEs to monitor the game/draw tick better
17:53:16 <LordAro> as long as it doesn't slow things down too much
17:58:45 <LordAro> i fixed the pathfinder crash
17:58:52 <LordAro> ...by converting a load of ints into int64s
18:15:07 <TrueBrain> is that really fixing it? :D
18:16:03 <TrueBrain> and adding PFEs: every tick it already does PFEs, just in a place that is not useful to measure these kind of problems .. so I want to either move up the current ones, or add a new parent, not sure what is best yet
18:16:20 <TrueBrain> basically, draw-tick is not measured, but UpdateWindows is, which is always called in the draw-tick
18:16:24 <TrueBrain> just a tiny bit more too :)
18:16:38 <LordAro> seems to... just not sure it's the correct solution
18:16:47 <LordAro> as best as i can tell, it's always triggered this assert
18:17:02 <LordAro> (savegame from #7670)
18:18:02 <TrueBrain> well, by adding 32bits, you make the chance of it happening smaller, I guess :P
18:18:17 <TrueBrain> you could also make it an OverflowSafe<int> I guess :P
18:18:26 <TrueBrain> I guess a lot, I see :D
18:19:00 <TrueBrain> I really wonder why an estimation would be more than 2^31
18:19:07 <TrueBrain> sounds really weird :P
18:19:38 <LordAro> all the estimations i was printing out were of the same magnitude
18:19:44 <LordAro> just that this one pushed it over the limit
18:20:02 <TrueBrain> sounds like there a bug hiding below this
18:20:21 <LordAro> i was afraid you were going to say that :p
18:20:32 <TrueBrain> I know you don't like that answer, don't worry :)
18:20:45 <TrueBrain> but there are only 16M tiles in a maxed-out game
18:20:50 <TrueBrain> even at a penalty of 100 per tile
18:20:55 <DorpsGek> TrueBrain: 2147483648
18:21:06 <TrueBrain> @calc pow(2,31)/100/16000000
18:21:06 <DorpsGek> TrueBrain: 1.34217728
18:21:12 <TrueBrain> well, it would get really close to that
18:21:26 <TrueBrain> "really" being relative here :D
18:21:41 <TrueBrain> so what the hell is the pathfinder doing?
18:21:53 <LordAro> is the thing it's doing that's pushing it over the limit 1735715127 + 419000300
18:22:24 <TrueBrain> those numbers are already insane, given we use ~100 for scoring per tile :P
18:22:31 <TrueBrain> but, I have not read at all how estimations are made
18:24:37 <TrueBrain> I love how the comments suggest something the code is NOT doing
18:25:53 <TrueBrain> that "d" value, how is that calculated? Why doesn't it have a comment explaining? :P
18:26:09 <Eddi|zuHause> @calc pow(2,31) <-- how have you lived in IT and not memorized this number? :p
18:26:09 <DorpsGek> Eddi|zuHause: Error: invalid syntax (<unknown>, line 1)
18:26:45 <TrueBrain> x and y are multiplied by 2
18:26:59 <TrueBrain> and x by YAPF_TILE_CORNER_LENGTH, and y by YAPF_TILE_LENGTH / 2
18:27:05 <TrueBrain> yes, that makes absolutely sense to me (?!?!)
18:27:17 <LordAro> it works fine in release mode, of course
18:27:20 <Gustavo6046> Also, how do I prevent road congestion when I use many road vehicles?
18:27:24 <LordAro> well, i imagine the pathfinder's going a little nuts
18:27:46 <TrueBrain> LordAro: mostly I wonder if it explains some absurd train-tick times, honestly :)
18:28:15 <LordAro> release builds that have asserts disabled :p
18:28:23 <TrueBrain> those will work just fine :P
18:28:33 <LordAro> what are the rules for that, these days?
18:28:44 <TrueBrain> non-beta non-RC only
18:28:56 <TrueBrain> for the extra SSPPPPEEEEDDDDDDD
18:29:20 <TrueBrain> never was part of that choice, still no clue if that is a good idea btw LordAro :P
18:29:40 <TrueBrain> Gustavo6046: disable breakdowns :D :D :D
18:30:05 <TrueBrain> so the cost is 1700M and the estimation is 420M
18:30:47 <TrueBrain> do I want to know what route the PF is trying to plot?
18:31:02 <TrueBrain> it is like it starts in a town in the top left, and goes allllll the way to the bottom right to go allllll the way back
18:31:14 <TrueBrain> I guess that is one way to make it go nuts :D
18:31:33 <Gustavo6046> supermop_Home: also sorry I forgot to respond that day
18:31:34 <LordAro> the game is full of AIs, but even so, seems unlikely
18:31:35 <Gustavo6046> Thanks for the tips! :)
18:31:48 <TrueBrain> but I hope the PF says after a while: I had enough of this, I will just go close to the destination and see what happens
18:32:02 <Gustavo6046> I thought breakdowns were an important part of the balance
18:32:06 <Gustavo6046> like, the balance of the gameplay
18:32:07 <LordAro> there are various protections in place for that, but none of them seem to be getting triggered
18:32:17 <LordAro> Gustavo6046: almost every player disables them immediately
18:32:19 <TrueBrain> Gustavo6046: you are asking for a solution, I am presenting one :) Without any further context, of course
18:32:37 <Gustavo6046> But more importantly, they often lock the roads when they enter the stops/stations to load and unload.
18:32:43 <Gustavo6046> Even when there are free stations
18:32:59 <Gustavo6046> They don't seem to do it now
18:35:08 <TrueBrain> okay, what the "d" value basically is: shortest distance from current position to destination
18:35:21 <TrueBrain> with a penalty of 100, give or take
18:35:32 <TrueBrain> means there is a train that wants to go 4M tiles away from where it is
18:35:36 <TrueBrain> if I read this right :P
18:35:42 <TrueBrain> well, it could be 144
18:35:53 <TrueBrain> @calc 419000300 / 144
18:35:53 <DorpsGek> TrueBrain: 2909724.3055555555
18:36:16 <TrueBrain> that is a long long track :D
18:36:28 <LordAro> it's built some fairly lengthy routes, judging from the smallmap
18:36:33 <LordAro> but probably not that long
18:36:58 <DorpsGek> [OpenTTD/OpenTTD] JGRennison commented on issue #7670: Road vehicle pathing cache does not always pick up changes in road network https://git.io/fjyVR
18:37:19 <LordAro> i think he's watching us
18:37:22 <TrueBrain> JGR is spoiling the fun again :P
18:37:59 <TrueBrain> so yeah, I keep forgetting people can change the penalties .. this should really have been a developers-only setting
18:38:05 <TrueBrain> to tune in the first year of the release of YAPF :P
18:38:09 <TrueBrain> and come with very big !!!!!!!!!!!!!
18:38:27 <Gustavo6046> Is there a real benefit to lining two roads (in high-congestion road vehicle lines), or is it not a substantial difference?
18:38:35 <LordAro> who's the guy who keeps trying to convince everyone that changing those settings is a good idea?
18:39:04 <TrueBrain> anyway, the penalties for "d" are fixed
18:39:11 <TrueBrain> so this is still an insanely long track, I would have to guess
18:40:18 <TrueBrain> anyway, I am just reading code and making guesses .. I could also open the savegame and make sure what I think is true :P
18:41:13 <TrueBrain> waaayyyy to lazy for that :)
18:41:33 <TrueBrain> just keep being surprised why YAPF doesn't have some basic comments in the F / G / H value parts
18:41:37 <TrueBrain> and why it is not called F / G / H
18:41:40 <TrueBrain> while it clearly is :P
18:44:21 <TrueBrain> anyway, for 1.12, UI redesign + removing of settings people shouldn't touch?
18:44:30 <TrueBrain> both will cause drama on forums anyway, so we better can get it over with :P
18:45:00 <TrueBrain> "you remove my ability to break the game by setting stupid PF settings!!!!!!!"
18:45:17 <TrueBrain> You don't allow me to play this game how I want it: broken and without any purpose!
18:47:24 <andythenorth> just do a patchpack
18:48:51 * andythenorth is somewhat demotivated to make grf currently
18:49:24 <andythenorth> it's the tedious yak-shaving to finish FIRS v4 I think
18:50:25 <LordAro> JGR's patch doesn't fix the crash either
18:51:25 <andythenorth> particularly trying to find all the train grfs that FIRS 4 breaks
18:51:28 <andythenorth> is really boring
18:57:09 <frosch123> just compare that to skipping cargo slot 31 :p
18:57:37 <frosch123> andy picks really weird fights
19:00:42 *** J0anJosep has joined #openttd
19:02:05 <DorpsGek> - Update: Translations from eints (by translators)
19:02:29 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8688: Feature: Hide block signal GUI by default https://git.io/JqToF
19:05:54 <andythenorth> frosch123 can't rely on them using cargo slot 31
19:06:11 <andythenorth> the issue seems to be relying on GOOD for livery
19:06:19 <andythenorth> but FIRS doesn't always define GOOD
19:06:58 <frosch123> ok, fine. GOOD makes no sense
19:07:14 <andythenorth> yeah, I have spent a good part of the day trying to test the major train sets
19:07:15 <frosch123> they could just as well use MAIL
19:07:32 <andythenorth> I ran into extra lolz issues that Iron Horse and Termite break some of the train sets
19:14:38 <peter1138> Ok, got one out the frezer to defrost.
19:15:11 <peter1138> Seems to have been in there a while...
19:15:21 <TrueBrain> peter1138: if you have the time, did you pick up on the places to start debugging the stalls you are getting earlier? :D
19:16:10 <peter1138> No, I gave up trying to debug as it wants to build/run from build/Debug/ instead of build/, and so can't find any data.
19:16:50 <TrueBrain> yeah, in the build folder, run Debug/openttd.exe, and you are fine
19:16:52 <TrueBrain> but it needs that :D
19:17:38 <peter1138> LocalDebuggerWorkingDirectory gets set to bin instead of build.
19:17:46 <peter1138> Dunno why or how to fix it.
19:18:03 <TrueBrain> it is why I use MSYS to start OpenTTD on Windows
19:18:12 <TrueBrain> it is more in the way I can digest :D
19:22:50 <peter1138> Dunno what sets it to bin, but "bin" is a pretty vague thing to search for.
19:23:33 <peter1138> Hmm, or is the cmake build supposed to be done in bin/?
19:23:45 <peter1138> message(FATAL_ERROR "In-source builds not allowed. Please run \"cmake ..\" from the bin directory")
19:23:50 <DorpsGek> [OpenTTD/OpenTTD] ldpl updated pull request #8805: Feature: Allow client and server to negotiate on compression to use for the savegame https://git.io/Jqf7C
19:24:32 <peter1138> set_target_properties(openttd PROPERTIES VS_DEBUGGER_WORKING_DIRECTORY "${CMAKE_SOURCE_DIR}/bin")
19:26:40 <DorpsGek> [OpenTTD/OpenTTD] ldpl commented on pull request #8805: Feature: Allow client and server to negotiate on compression to use for the savegame https://git.io/JqTKx
19:27:18 <peter1138> Balls, still doesn't work :/
19:29:23 <peter1138> set_target_properties(openttd PROPERTIES VS_DEBUGGER_WORKING_DIRECTORY "${CMAKE_BINARY_DIR}/")
19:29:27 <peter1138> set_target_properties(openttd PROPERTIES VS_DEBUGGER_WORKING_DIRECTORY "${CMAKE_BINARY_DIR}/")
19:30:04 <peter1138> Seems like it might be more correct, but I guess my CMake integration is not doing anything with that anyway.
19:30:27 <TrueBrain> you should ask glx :D
19:31:19 <peter1138> I'm open to other ways to get this going. I just did VS Code + CMake + VS2019 Build Tools
19:32:22 <peter1138> Hmm, there's no traditional vs code run/debug targets, just whatever this Cmake extension does.
19:32:57 <peter1138> Yeah, it starts it from Debug. Hmm.
19:37:55 <peter1138> "cmake.debugConfig": {
19:37:55 <peter1138> "cwd": "${workspaceFolder}/build"
19:39:07 <peter1138> Of course, because it's a debug build, it's kinda slower anyway.
19:39:50 <peter1138> Erm, now it's running, I have zero idea where to start.
19:39:59 <peter1138> I mean, it gliches,but...
19:40:10 <TrueBrain> does the stuttering still happen? See earlier, I pointed out 3 places to TICC/TOCC
19:44:30 *** jottyfan has joined #openttd
19:44:49 *** J0anJosep has quit IRC (Quit: Konversation terminated!)
19:48:29 <peter1138> Got lots of numbers but zero no idea how to interpret it.
19:48:44 <TrueBrain> you want to know which one spikes .. game-tick or draw-tick
19:48:52 <TrueBrain> or that sleep has a value of ..
19:48:55 <DorpsGek> TrueBrain: 16.666666666666668
19:49:18 <TrueBrain> so a count of 10-ish for TOCC, give them a good name, wait for a stutter, and see which one is all of asudden high
19:49:22 <peter1138> Hmm. gametick 32000, 32000, 3200, 177000, 32000, 30000
19:50:50 <TrueBrain> that is rather interesting ..
19:51:03 <TrueBrain> I did not expect that, honestly
19:51:27 <TrueBrain> well, you can go in GameLoop, and TICC/TOCC functions there
19:51:30 <TrueBrain> see if you can trace the function
19:51:34 <TrueBrain> or maybe make a flamegraph
19:51:37 <TrueBrain> see if that shows what is going on
19:56:03 <peter1138> How did I ever develop with just vim?
19:56:35 <michi_cc> The other function to time would be UnlockVideoBuffer() (i.e. video_driver.cpp:31, because that is where the OpenGL driver might do work).
19:57:09 *** _2TallTyler has joined #openttd
19:57:31 <TrueBrain> peter1138: a true vim believer will tell you vim can do that too .. but don't believe them
19:57:52 <peter1138> Yes, I never got as far as ever making ctags and all that gubbins work.
19:58:03 <peter1138> And then it was obscure keypresses.
19:58:47 <peter1138> Hmm, see, if it's the gameloop, them the performance ui stuff should pick it up...
19:59:28 <TrueBrain> not everything is measured, sadly enough
19:59:33 <TrueBrain> there are some odd functions that are outside it
20:03:49 <peter1138> I should remove the performance measurer and see if it's that ;-)
20:04:55 * LordAro uses a large amount of greps
20:05:25 <TrueBrain> so do I, but for different things I am sure :D
20:05:39 <TrueBrain> I should make "grep -n" default ..
20:19:57 <Wuzzy> And after this PR is done, I think OpenSFX is release-ready (don't forget to manually update the date in changelog.txt first ;-) )
20:20:26 <Wuzzy> oh wow, i can't wait to see a new OpenSFX release!! 😀
20:21:00 <peter1138> Barking up the wrong tree, don't think it's the gameloop.
20:21:44 <peter1138> this->UnlockVideoBuffer();
20:22:02 <peter1138> ^ normally ~ 300-800us
20:22:09 <TrueBrain> peter1138: guess there is a reason michi_cc asked :D
20:22:14 <peter1138> sometimes gets 165000us
20:22:16 <Wuzzy> ↑ this includes a couple of new sounds, now 100% libre (no more CC Sampling Plus nonsense, woohoo!)
20:22:31 <TrueBrain> not sure how to debug that issue further .. michi_cc might have ideas
20:23:48 <peter1138> Is your threading change still a PR?
20:24:09 <TrueBrain> and there is no threading with OpenGL
20:24:17 <TrueBrain> (there is with my PR)
20:25:46 <TrueBrain> but it is more interesting to understand why OpenGL stutters for a bit, honestly :)
20:28:02 <michi_cc> peter1138: Which graphics card are you running on?
20:29:25 <michi_cc> Okay, two tests: 1) in src/video/opengl.cpp, uncomment the define at line 15 (NO_GL_BUFFER_SYNC); 2) Before line 592, add "this->persistent_mapping_supported = false;"
20:31:23 *** tokai has quit IRC (Quit: c('~' )o)
20:35:27 <peter1138> If 1 is done and 2 is not done, the lag seems to improve.
20:35:37 <peter1138> if 2 is done, regardless of 1, the lag seems to stay.
20:37:42 <michi_cc> Okay, then I have no real idea either. 1) simply causes the opengl backend to not wait for the previous frame to be done, which will lead to visible glitches. So this means the driver (or the GPU) occasionally spend time doing something.
20:38:25 <michi_cc> Going further done would need stuff like Nvidia Nsight Graphics and other vendor tools.
20:42:34 <peter1138> I'm going to try pr 8741
20:45:06 <peter1138> As predicted, nothing changes.
20:47:09 <TrueBrain> seems your GPU needs to warm up or something
20:47:45 <LordAro> 8741 should make the pauses less obvious though
20:48:41 <TrueBrain> that is true; unlocking happens outside the game-state-lock
20:49:00 <TrueBrain> so the game can just do another tick
20:49:08 <TrueBrain> so the stutter is ~15ms less
20:49:17 <TrueBrain> well, no, the game can continue
20:49:22 <TrueBrain> the next drawing will make it block
20:50:17 <TrueBrain> we also need someone with AMD to track down the slowness ..
20:50:43 <TrueBrain> if I had a job, I would buy one :P
20:51:21 <peter1138> That's a thing, after a while it does actually stop glitching, but not in a reproducible manner...
20:52:25 <milek7> what with AMD? on windows or linux?
20:52:56 *** _2TallTyler has quit IRC (Quit: Page closed)
20:53:31 <peter1138> Intel CPU, Nvidia GPU, Windows.
20:53:40 <frosch123> TrueBrain: random fact. when you pause widelands for a while, unpausing will immediately trigger a autosave
20:53:54 <TrueBrain> ALMOST right, I would say ;)
20:56:02 <TrueBrain> better than us, at least :)
20:59:51 <Venemo> hey guys. I've set up port forwarding for port 3979 between my VPS and my PC. I have verified that I can connect to this port using telnet, but openttd just says "server offline"
21:00:10 <DorpsGek> LordAro: OpenTTD uses TCP and UDP port 3979 for server <-> client communication, UDP port 3978 for masterserver (advertise) communication (outbound), and TCP port 3978 for content service, a.k.a. BaNaNaS (outbound)
21:01:41 <_dp_> udp is a tough nut to forward
21:02:14 <Venemo> errr... I used this command: 'ssh -R 0.0.0.0:3979:localhost:3979 user@vps -N'
21:02:49 <_dp_> I tried to forward udp once but I think I gave up
21:02:50 <LordAro> are you sure you need to port forward?
21:03:00 <michi_cc> Didn't somebody post a patch/PR for proper scaling of window padding and PIP here, or am I imagining that?
21:03:03 <LordAro> i would be very surprised if VPS is blocking ports
21:05:57 <Venemo> I'd like to use port forwarding becasue the PC sits behind multiple routers
21:06:34 <LordAro> that's usually only an issue when the server is behind multiple (home) routers
21:06:52 <milek7> Venemo: why don't you run server on vps itself?
21:07:01 <Venemo> the server is on my PC
21:08:02 <Venemo> seemed easier this way :)
21:08:50 <LordAro> i think you have found that it's probably not the case
21:09:07 <LordAro> unless your VPS is seriously low powered
21:10:30 <Venemo> well, the VPS runs centos 6 so I didn't want to get into the trouble of building the game for this old os
21:11:05 <LordAro> luckily for you, that's well supported by our generic-linux binaries
21:11:11 <LordAro> but you should probably think about upgrading anyway
21:13:11 <Venemo> where is the generic binary? I only found ubuntu and debian ones
21:13:31 <LordAro> oh, only for 1.11 betas
21:13:59 <LordAro> but i wouldn't think you'd have issues building 1.10.3 for COS6 anyway
21:14:57 <Venemo> I guess it's worth a try
21:16:45 <Venemo> it has gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44)
21:16:52 <LordAro> yeah, but that has enough C++11 support afaik
21:25:09 <Gustavo6046> Trains drive on the right :)
21:53:46 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #8812: Codechange: Allow driver factories to decide at runtime if the driver is usable. https://git.io/JqTSV
21:56:08 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8812: Add: Option to (dis-)allow accelerated video drivers. https://git.io/JqTSD
22:04:29 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8812: Add: Option to (dis-)allow accelerated video drivers. https://git.io/JqT9q
22:06:40 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
22:12:48 *** andythenorth has quit IRC (Quit: andythenorth)
22:15:11 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8812: Add: Option to (dis-)allow accelerated video drivers. https://git.io/JqT97
22:15:57 *** jottyfan has joined #openttd
22:21:03 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8812: Add: Option to (dis-)allow accelerated video drivers. https://git.io/JqTHU
22:25:31 *** gelignite has quit IRC (charon.oftc.net liquid.oftc.net)
22:25:31 *** dihedral has quit IRC (charon.oftc.net liquid.oftc.net)
22:25:31 *** freu[m] has quit IRC (charon.oftc.net liquid.oftc.net)
22:25:31 *** D-HUND has quit IRC (charon.oftc.net liquid.oftc.net)
22:25:39 *** gelignite has joined #openttd
22:27:34 *** dihedral has joined #openttd
22:29:18 *** freu[m] has joined #openttd
22:32:01 <DorpsGek> [OpenTTD/OpenTTD] spnda commented on pull request #8812: Add: Option to (dis-)allow accelerated video drivers. https://git.io/JqTHF
22:40:17 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:40:25 <TrueBrain> michi_cc: I guess mostly I hope this triggers someone to make a Vulkan driver or something :P
22:41:17 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:41:46 <TrueBrain> case and point, thank you :
22:41:55 <michi_cc> TrueBrain: Would you make driver names translatable?
22:42:50 <TrueBrain> Allegro .. SDL (X11) .. SDL (OpenGL) .. Win32 (GDI) .. Win32 (OpenGL) .. Cocoa .. no, I do not think they are translatable pieces in there?
22:43:53 <TrueBrain> SDL can be a bit tricky ... X11 is only known after the backend is initialized
22:43:57 <TrueBrain> so it would be just SDL maybe
22:45:37 <TrueBrain> if (and that is still an if for me) we do this, we should also have a dropdown for music and sound driver
22:45:55 <TrueBrain> doesn't all have to be done at once or for 1.11 even
22:45:57 <spnda> thing is, one could technically implement SDL for windows as an option. But that would be besides the point, that would just be the same effectively
22:46:19 <LordAro> is there ever more than one option (beyond GL/non-GL) on any normal system?
22:46:23 <TrueBrain> I debug a lot with SDL for Windows, yes. But we do not officially support that :D
22:46:32 <TrueBrain> LordAro: currently, no
22:46:43 <TrueBrain> that is why I mentioned Vulkan, as possible future :)
22:46:54 <spnda> Anything made by intel, amd or nvidia after 2015/16 would support Vulkan
22:47:17 <TrueBrain> and we have Metal on macOS
22:47:19 <spnda> but honestly, the benefits are tiny for a game like this and idk if I want to spend the few thousand lines implementing a vulkan driver for this game
22:47:28 <spnda> i'd rather spend the time extending my vulkan raytracer lol
22:47:35 <spnda> which is what vulkan is really made for
22:48:14 <TrueBrain> but again, I am not saying it is a good idea to make a dropdown for the video driver; just mentioning that I had that in my head, for what-ever that is worth :D
22:48:50 <spnda> Well I had the idea today that the Game Options window could be a tabbed window, fusing the options and game options together into different tabs.
22:49:07 <spnda> That would also allow a tab for video settings, possibly configuring any driver, resolution or blitters
22:49:15 <TrueBrain> for my UI redesign drafts those windows look completely different :D
22:49:17 <spnda> And maybe even a audio tab with audio volume or something
22:49:26 <TrueBrain> and yes, exactly that
22:49:30 <TrueBrain> be more like other games, basically
22:49:31 <spnda> TrueBrain: is that inside a pr or branch of yours?
22:49:38 <TrueBrain> no, nothing that fancy yet
22:50:35 <TrueBrain> but I haven't done Options yet in that format; hope to do that next week
22:51:05 <spnda> but thats just something from gimp? or is that actually coded?
22:51:21 <TrueBrain> I made a NWidget -> HTML converter
22:51:36 <TrueBrain> so it is HTML of which I took an image :)
22:51:53 <TrueBrain> but written in a language I could in theory copy/paste in OpenTTD C++ code
22:52:08 <spnda> oh what? that's a bit over the top is it not lol
22:52:16 <TrueBrain> I needed something to do quick mockups with
22:52:24 <TrueBrain> while still being limited by NWidget system
22:52:32 <TrueBrain> GIMP slowed me down a lot
22:52:41 <TrueBrain> C++ was too darn slow too
22:52:54 <TrueBrain> quick iterations in mockups is rather important :)
22:53:37 <TrueBrain> colours are wrong, editbox is not implemented, etc .. its a mockup :P
22:53:48 <spnda> it's very noticable, yeah
22:56:22 <spnda> there is no NWidget for tabs just now is there?
22:56:46 <TrueBrain> hence me experimenting a bit with it
22:56:52 <spnda> Also, yes, that mockup looks good, just need to actually indicate that they're tabs and not just some buttons (which is what it looks like)
22:57:16 <TrueBrain> which is rather difficult with the limited colours and styles currently available :)
22:58:47 <TrueBrain> but I first want to mock the content, what should be in windows etc
22:58:59 <TrueBrain> and from there zoom out to how it should look, and after that what colours
22:59:39 <TrueBrain> maybe also, and this will be a shocker, have a window to configure hotkeys
22:59:47 <TrueBrain> as it really surprised me that is only possible by editing a file :P
23:02:21 <spnda> TrueBrain: Btw, how would one access the currently used video driver from global scope?
23:02:41 <TrueBrain> there is a function for that, something something VideoDriver Instance
23:02:45 <TrueBrain> can't remember the exact name
23:03:10 <spnda> for #8813, I want to check all the monitor's refresh rates and add all of them to the dropdown list (if available ofc)
23:03:58 <TrueBrain> VideoDriver::GetInstance()
23:04:08 <TrueBrain> and yeah, it should also auto-select them on first run
23:04:29 <spnda> not just first, users could connect new displays even while the game is running
23:04:44 <TrueBrain> and there it becomes a bit tricky :P
23:04:57 <TrueBrain> you could also adjust it if people change where OpenTTD is on which monitor
23:04:57 <spnda> My answer: Update everytime one updates the game options window
23:05:16 <spnda> And only on game launch detect the refresh rate of the current monitor
23:05:23 <TrueBrain> we solved it for fonts etc by having a top item called "(autodetect)" or what is the name
23:05:43 <TrueBrain> although that is currently lacking what that auto-detected value is :D
23:05:48 <TrueBrain> but that is a small improvement
23:06:10 <TrueBrain> and I was talking about auto-select btw, not detect
23:06:17 <TrueBrain> so on first run, it should pick a refresh-rate that fits your machine
23:06:24 <TrueBrain> instead of "60" or what-ever default
23:07:12 <TrueBrain> or every run, if we go the "(auto-detect)" route, I guess
23:07:18 <TrueBrain> well, make something pretty, I would say :D
23:07:35 <TrueBrain> but as I wrote, I would prefer we keep that for 1.12, but I would love to get some input from others devs on that too
23:07:41 <TrueBrain> I might be too scared of users doing stupid shit :P
23:08:30 <spnda> Yeah, I think 1.12 is the better one for the auto detect stuff. However, the refresh option could be very good for 1.11. Or maybe 1.11.1?
23:08:48 <TrueBrain> as I wrote, above 60 it becomes a bit touch&go what the game does
23:08:55 <TrueBrain> a semi-big game really suffers
23:09:01 <spnda> personally 144 has been fine but I'll test again
23:09:02 <TrueBrain> we need to make some more changes to make that a lot better
23:09:25 <LordAro> agreed, not enough time to get it in 1.11 sanely, imo
23:09:26 <spnda> yeah as you commented in my pr (thx btw)
23:09:52 <TrueBrain> but, that said, I really wouldn't mind if 1.12 doesn't take another year :P
23:09:58 <TrueBrain> as honestly .. backporting sucks balls
23:10:12 <spnda> atleast keep to the yearly april 1st schedule :)
23:10:51 <TrueBrain> its so fun on Steam Discussions
23:10:54 <TrueBrain> people claiming it is a joke
23:11:02 <TrueBrain> and not like: is this a joke?
23:11:05 <TrueBrain> but: THIS CLEARLY IS A JOKE
23:11:21 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #8812: Add: Option to (dis-)allow accelerated video drivers. https://git.io/JqTSV
23:11:23 <TrueBrain> can't believe a 10 year old joke still works :P
23:12:06 <TrueBrain> reminds me, I still don't know why michi_cc asked about translations :D
23:17:33 <spnda> Also from brief testing from my last multiplayer world (isn't large, just medium) and I get fluid 240fps constantly
23:17:52 <TrueBrain> just zoom out once :P
23:18:07 <spnda> zooming in and out really fast gets it down to 210 no further
23:18:17 <spnda> zooming out doesnt change a thing
23:18:27 <TrueBrain> when you have a decent (NewGRF) game, being zoomed out takes easily 10ms to render the viewports
23:18:45 <TrueBrain> even on a 8700K I have
23:19:31 <spnda> As this is a single thread game I guess my 9600K @ 5.2GHz would be quite a bit faster
23:19:51 <TrueBrain> mine run at 4.5GHz :P
23:19:57 <TrueBrain> but so yeah, my friend, you are not our average user :D
23:20:28 <spnda> 8700k default core clock is 4.7ghz?
23:20:37 <TrueBrain> I haven't anything running
23:20:41 <TrueBrain> so it downscaled, I guess :)
23:20:50 <TrueBrain> but default is 3.7GHz
23:21:48 <TrueBrain> even without NewGRFs, fully zoomed out on a 256x256 empty map is 10ms for the viewport :P
23:22:35 <TrueBrain> I do have full animation and full detail on, to be clear :)
23:22:59 <spnda> wait are you rendering on CPU or GPU?
23:23:13 <TrueBrain> irrelevant; this is the viewport drawing
23:23:15 <TrueBrain> that happens in software
23:23:26 <TrueBrain> sorting of sprites etc
23:23:33 <TrueBrain> just .. the bigger the monitor ...
23:23:36 <TrueBrain> the worse the problem gets :D
23:23:44 <spnda> well either way, 1440p, 1k map, fully zoomed out, 0.x to max of 2ms for viewport rendering
23:24:06 <TrueBrain> I have to say, this is with a Debug build .. Release would do a bit better
23:24:19 <spnda> yeah that would definitely be the difference, im using release
23:24:58 <TrueBrain> real problem starts when you load TWs NewGRFs :D
23:25:01 <TrueBrain> they are the worst :P
23:25:12 <spnda> did.... openttd just crash updating a newgrf
23:25:25 <TrueBrain> backtrace or it didn't happen
23:26:00 <spnda> it just closed so thats kinda hard
23:26:15 <TrueBrain> should be a crash.log?
23:26:27 <spnda> no no crash, just instant exit
23:26:42 <TrueBrain> but that won't get your FPS above 30
23:27:19 <spnda> and it just crashed again downloading newgrfs
23:27:31 <TrueBrain> attach a debugger and find out why!
23:27:32 <spnda> i have to debug this be right back
23:27:36 <TrueBrain> but for now, I am off to bed :)
23:35:28 *** NGC3982 has joined #openttd
23:37:15 <spnda> there, issue for that crash opened. idk seems to be some multithreading issue idk
23:38:48 <spnda> oh also translations bot made the 25000th commit today
23:44:41 <LordAro> that's not our assertion...
23:44:48 <LordAro> can you post more detail?
23:44:55 <spnda> thats it, vs doesnt even catch it
23:45:02 <LordAro> stacktrace or anything?
23:45:10 <spnda> literally nothing but ill try again
23:45:31 <LordAro> can you reproduce it with a clean config? (-c foodir/foo.cfg)
23:45:43 <LordAro> and of course, a completely clean build
23:45:56 <LordAro> (James will ask you all these questions shortly, i'm just getting ahead of him :p )
23:46:13 <TrueBrain> I have the exact same error when making screenshot when I compile myself
23:46:31 <TrueBrain> Couldn't figure out why
23:46:35 <glx> compiling master to check here
23:46:51 <TrueBrain> For sure a local issue with my setup
23:47:07 <TrueBrain> Seems to be mixing vcpkg with msvc libraries
23:47:17 <TrueBrain> Then gave up, because .. MSVC
23:48:36 <TrueBrain> Debugger had issues tracing the problem, not triggers on the error basically
23:48:48 <glx> I think I have latest VS2019, not sure if my vcpkg is up to date
23:48:54 <TrueBrain> Had to step by step find the line triggering it
23:49:11 <TrueBrain> Deep inside libpng I gave up :p
23:49:38 <glx> and of course it has to download symbols, since I'm now on 20H2
23:52:06 <TrueBrain> So issues writing the file, basically
23:53:44 <TrueBrain> Would fit my issues, just no clue why it fails to write :D
23:53:55 <TrueBrain> I start the game via MSYS
23:55:16 <TrueBrain> Second solutions could be the cause too
23:55:35 <TrueBrain> Owh well, really sleep time for me now :D
23:56:10 <spnda> oh btw, it might be unrelated but I get "dbg: [misc] String too long for destination buffer" when I launch the game
23:58:29 <spnda> I've updated my issue with some info from where the issue comes
23:59:14 *** erle- has quit IRC (Quit: Leaving)
continue to next day ⏵
