IRC logs for #openttd on OFTC at 2021-02-23
⏴ go to previous day
00:02:51 <milek7_> and now persistent mapping makes practically no difference
00:24:05 *** mirrorbird has quit IRC (Quit: Leaving)
00:24:57 *** hotgirl55 has joined #openttd
00:33:07 *** Progman has quit IRC (Remote host closed the connection)
00:56:40 *** gelignite has quit IRC (Quit: Stay safe!)
01:14:36 *** Tirili has quit IRC (Remote host closed the connection)
01:30:50 *** WormnestAndroid has quit IRC (Remote host closed the connection)
01:32:15 *** WormnestAndroid has joined #openttd
01:37:52 <FLHerne> (this channel is usually a bit quiet during EU nighttime)
01:38:37 <FLHerne> supermop_Home: I looked at it a bit more, the probabilities in switch_color_vehs_a are just too complicated for the number of random bits available
01:39:00 *** WormnestAndroid has quit IRC (Remote host closed the connection)
01:39:12 <FLHerne> If you make them all `1` (i.e. equal probabilities) it works
01:41:19 *** WormnestAndroid has joined #openttd
01:42:06 <FLHerne> As it is, the chances are n/41, so NML needs 6 random bits, and the other one needs 4, while there are only 8 available
01:42:37 <FLHerne> Making it even is a straight 1/16 chance, so it only takes 4 bits and is possible to implement
01:59:13 *** Wormnest has quit IRC (Quit: Leaving)
02:42:15 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on issue #8686: Unable to change townnames in Scenario editor after pr8566 https://git.io/JtMyZ
02:54:32 <supermop_Home> FLHerne thanks!
03:27:00 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:35:54 *** muffindrake2 has joined #openttd
03:37:41 *** muffindrake1 has quit IRC (Ping timeout: 480 seconds)
06:06:53 *** Wuzzy has quit IRC (Remote host closed the connection)
06:57:03 *** andythenorth has joined #openttd
07:03:00 <DorpsGek> [OpenTTD/OpenTTD] EratoNysiad commented on issue #8686: Unable to change townnames in Scenario editor after pr8566 https://git.io/JtMyZ
07:14:26 *** andythenorth has quit IRC (Quit: andythenorth)
07:17:10 *** snail_UES_ has quit IRC (Quit: snail_UES_)
07:19:45 *** andythenorth has joined #openttd
07:23:49 *** sla_ro|master has joined #openttd
07:53:57 *** hotgirl55 has joined #openttd
07:55:27 *** HerzogDeXtEr has joined #openttd
08:38:17 *** hotgirl55_ has joined #openttd
08:38:17 *** hotgirl55 has quit IRC (Read error: Connection reset by peer)
08:41:51 *** roadt__ has quit IRC (Ping timeout: 480 seconds)
08:51:37 *** roadt__ has joined #openttd
08:58:27 *** hotgirl55_ has quit IRC (Quit: Leaving)
10:34:22 <DorpsGek> [OpenTTD/OpenTTD] orudge commented on pull request #8718: Change: [Actions] stop relying on external actions for trivial stuff, and rely on cmake tools to build, test and pack https://git.io/Jt7nA
10:51:21 *** gelignite has joined #openttd
11:16:27 *** andythenorth has quit IRC (Quit: andythenorth)
11:22:11 *** peter1138 has joined #openttd
11:22:11 *** ChanServ sets mode: +o peter1138
11:56:45 *** andythenorth has joined #openttd
12:25:42 *** roadt__ has quit IRC (Ping timeout: 480 seconds)
12:25:59 *** Samu has quit IRC (Quit: Leaving)
12:30:04 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8732: Change: Improve graph period markings https://git.io/Jt70C
12:35:16 *** roadt__ has joined #openttd
12:47:14 *** sla_ro|master has quit IRC ()
12:48:37 <TrueBrain> I like these small steps to change the GUIs :)
12:49:25 <TrueBrain> means we can focus on a few small things, instead on EVERYTHING :D
12:53:33 <andythenorth> big UI change usually fails IMHO
13:05:37 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8732: Change: Improve graph period markings https://git.io/Jt7zt
13:07:59 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8730: Codechange: [OpenGL] Load all OpenGL functions dynamically. https://git.io/Jt7zn
13:09:54 *** andythenorth has quit IRC (Quit: andythenorth)
13:19:05 <TrueBrain> frosch123: assuming you read your backlog, ^^, is that the way to do it with language files? :)
13:19:33 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8732: Change: Improve graph period markings https://git.io/Jt7g3
13:21:19 <TrueBrain> what I kinda miss with GitHub, is a place to work on bigger things, like this UI stuff
13:21:24 <TrueBrain> all these changes are iterations of each other
13:21:32 <_dp_> ui scaling makes these graphs noticeable worse :(
13:21:35 <TrueBrain> now the conversation continues in an PR, which is nice
13:21:40 <TrueBrain> but not the place .. but we don't have another place
13:21:54 <TrueBrain> creating a branch or fork alone is not enough
13:21:59 <TrueBrain> as it is just a continues conversation, basically
13:22:07 <TrueBrain> not sure if there is a solution .. just something I have noticed more often lately
13:22:35 <TrueBrain> guess it is mostly because I am a huge fan of small iterations :D
13:22:52 <TrueBrain> _dp_: I can imagine that breaking these graphs, yes :P
13:23:51 <TrueBrain> hmm, I guess if you could push new commits to an existing PR that you can merge again too, or something .. but I guess that would be rather messy .. hmm
13:23:55 <_dp_> btw, I'm fine with that pr even the way it is
13:24:03 <_dp_> it's definitely an improvement either way
13:24:13 <TrueBrain> yeah, I got that. Yours is a nice addition on top of that
13:24:23 <TrueBrain> I am just a bit annoyed GitHub doesn't really allow that workflow :)
13:24:41 <TrueBrain> I want to merge the current work, while you continue the conversation to further improve it, basically
13:26:16 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8732: Change: Improve graph period markings https://git.io/Jt7gH
13:26:22 <TrueBrain> _dp_: well, I am just annoyed GitHub basically tells us: fuck you, that won't happen :P
13:27:01 <TrueBrain> I just noticed the dots in that graph are VERY oddly placed
13:27:10 <TrueBrain> I have no clue if the dot is the start of the quarter
13:27:28 <TrueBrain> it sits somewhere in the middle of the grid, for no clear reason
13:27:30 <_dp_> btw, I'm starting to think that it would be better to make bg under labels black as well
13:27:40 <_dp_> dot is the whole quarter so it's in the middle
13:28:04 <_dp_> it's a quarterly chart basically, not monthly
13:28:14 <TrueBrain> yeah, but measured at some point
13:28:23 <TrueBrain> now it feels like it is measured half-way through the quarter
13:28:27 <TrueBrain> if you get what I mean :)
13:28:31 <TrueBrain> highcharts does this better :P
13:28:45 <_dp_> well, highcharts one is monthly
13:29:02 <TrueBrain> well, honestly, it shouldn't be a point
13:29:30 <TrueBrain> LordAro: what are we looking at here? :D
13:29:41 *** _2TallTyler has joined #openttd
13:29:43 <LordAro> michi_cc's upscale branch
13:29:47 <_dp_> TrueBrain, you'd have to rename it to C1-C4 or smth to do like highcharts
13:29:49 <LordAro> (with a couple of tweaks to make it compile)
13:30:30 <_dp_> it's quarterly income/profit, it's not rly measured at any point
13:30:38 <TrueBrain> LordAro: now the question, is it an improvement without :)
13:30:51 <TrueBrain> _dp_: exactly why the dots are the problem for me :D
13:30:53 <peter1138> urgh 1) lime pickle, hot burning lips 2) doing an interactive rebase and finding some obsolete code hidden in it
13:30:57 <TrueBrain> it is not their location, but the fact it is a dot :P
13:31:07 <LordAro> i think it might be fun to have as an option, but definitely not default
13:31:29 <TrueBrain> LordAro: well, I do like how it looks :)
13:31:57 <peter1138> LordAro, I did upscaling like that once, this looks a bit tidier though.
13:32:03 <_dp_> TrueBrain, idk, imo having a dot makes no difference
13:32:35 <TrueBrain> _dp_: the problem is with you centering the label, it becomes more fuzzy to me :)
13:32:48 <TrueBrain> but as I said .. it just highlights that the dot is the weird thing
13:33:03 <TrueBrain> peter1138: looks cartoonish :)
13:33:16 <LordAro> cell-shaded OTTD when?
13:34:11 <peter1138> Hmm, that was from November. 20. 11.
13:34:25 <TrueBrain> _dp_: some more looking at it .. I even wonder why it only makes a dot every 3 months, and why not just every month :P
13:34:40 <peter1138> It's good job nobody got paid for working on OpenTTD because, uh, nobody would get paid...
13:35:06 <_dp_> unless you want moving average
13:35:32 <TrueBrain> the more I look at it, the more I release why I never open these graphs
13:35:37 <_dp_> well, not avg I guess but overlapping periods
13:35:41 <TrueBrain> they are just freakinglish weird
13:36:13 <_dp_> or actually avg because quartely income makes even less sense when displayed monthly xD
13:36:26 <TrueBrain> just monthly income
13:36:33 <TrueBrain> instead of quarterly is what I meant, ofc ;)
13:36:36 <FLHerne> peter1138: You're back!
13:36:38 <_dp_> monthly is just too much noise to be useful
13:36:47 <TrueBrain> I don't see how it would be more or less noise than quarterly
13:36:49 <FLHerne> peter1138: I apologise for being inconsiderate and bloody annoying :-(
13:37:25 <_dp_> TrueBrain, monthly jumps average a bit over the quarter, no ?
13:37:44 <TrueBrain> sorry, I don't understand what you wrote there :(
13:38:34 <_2TallTyler> An OpenTTD month is so short I imagine it would be quite jumpy, yes. It's already all over the place quarterly if you have long-distance train routes or ships.
13:38:47 <TrueBrain> that is a valid point
13:38:53 <_2TallTyler> I suspect there's a reason it's quarterly
13:39:06 <TrueBrain> be careful with suspicions there :P
13:39:41 <TrueBrain> but okay, I should just stop looking at that graph, and it annoys me more every minute :D
13:39:58 <TrueBrain> such graphs are more what I would expect
13:41:05 <_2TallTyler> Bar graphs would definitely be a separate PR :P
13:41:18 <TrueBrain> we already agreed your current work is just fine :)
13:41:23 <TrueBrain> the next addition, now that is the question :D
13:41:39 <_2TallTyler> Isn't part of the point of these graphs to compare your company's performance with that of other players? Bar graphs wouldn't work with that.
13:41:42 <_dp_> bar charts would be an interesting change indeed
13:41:53 <_dp_> not so much for the percent increase though imo :p
13:42:01 <TrueBrain> no, not the percentage :)
13:42:07 <TrueBrain> _2TallTyler: I guess ..
13:42:24 <TrueBrain> it is just weird, to have a box of "3 months" and put a single dot somewhere in there to indicate: that month
13:42:54 <TrueBrain> that, in my head, only works if you call it Q1 .. Q4
13:43:10 <TrueBrain> but I agree with your reasoning there, people have different opinions throughout the world what Q1 .. Q4 means :D
13:43:19 <_2TallTyler> I mean, comparing company performance is mostly useless beyond bragging rights, but when I play with my younger brothers (teenagers) they love to rag on me whenever they manage to get a quarter higher than mine.
13:43:19 <TrueBrain> s/that month/that quarter/
13:44:04 <_dp_> not that line chart works perfectly for comparing either...
13:44:10 * _dp_ had some "let's disable dP transport so we can see other companies" moments
13:44:26 <TrueBrain> haha, non-linear scales coming when? :D
13:45:30 <_dp_> idk, I kinda gave up on improving ingame charts too much :p
13:45:34 <TrueBrain> anyway, centered years, absolutely an improvement for me :)
13:45:35 <_dp_> just make them look ok and move on
13:45:49 <TrueBrain> "acceptance", yes :)
13:46:00 <TrueBrain> browser plugin and run highcharts, you say? :D
13:47:32 <_2TallTyler> _dp_ How did you get centered years? Just put the year in the July string instead of Jan?
13:48:00 <_dp_> _2TallTyler, yep, but it would be better to draw year separately
13:51:24 <_dp_> there is a diff in my comment btw
13:51:27 <_2TallTyler> What settings did you use to get clipping? I can't reproduce it.
13:52:06 <_dp_> well, I just changed my ui size to double from quad but I have it all custom
13:52:11 <_dp_> try making window smaller
13:52:49 *** gelignite has quit IRC (Quit: Stay safe!)
13:53:27 <_dp_> oh, and mb up the font scaling
13:53:44 <_dp_> may or may not only happen with custom fonts as well
13:54:16 <_dp_> basically, your year has to be wider than a column
13:54:21 <_2TallTyler> Looks like an issue of text wrapping, not with reserving enough vertical space
13:54:50 <_dp_> it wraps because it's to narrow
13:55:09 <_dp_> and it's too narrow because for months it's correct or they'll overlap
13:56:54 <_2TallTyler> It's not possible with default font. The window can't be made narrow enough.
13:58:32 <_dp_> yeah, looking at it mb it's just my changes that did it
13:58:40 <_dp_> I forgot to change UpdateWidgetSize
13:59:24 <_dp_> though rendering year separately makes more sense anyway
14:01:03 <_2TallTyler> Yes, separately drawn year will be necessary with left-aligned months and center-aligned years. Still don't like centered months, sorry :)
14:02:30 <_dp_> _2TallTyler, oh, looks like you forgot UpdateWidgetSize as well :p
14:03:13 <_dp_> so it's actually your bug :p
14:05:53 <_2TallTyler> Holy duplicate code, Batman
14:09:01 <_dp_> that UpdateWidgetSize is kinda weird though in general
14:09:21 <_dp_> it sums up all the label widgth but then draw divides them equally
14:09:38 <_2TallTyler> I actually find the periods slightly disorienting without vertical grid lines, since the dots don't correspond to a month label. I think I will keep vertical grid lines in this PR.
14:09:48 <_dp_> so it may only work ok-ish because of that year reserving some extra space
14:12:15 <_dp_> ah, no, nvm, it doesn't su
14:43:29 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8718: Change: [Actions] stop relying on external actions for trivial stuff, and rely on cmake tools to build, test and pack https://git.io/Jt7KS
14:44:54 *** snail_UES_ has joined #openttd
14:51:04 *** sla_ro|master has joined #openttd
14:58:44 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8732: Change: Improve graph period markings https://git.io/Jt7iY
15:02:54 <_2TallTyler> "According to accounts from a train forum" is such an unreliable reference :P
15:06:29 *** _2TallTyler has quit IRC (Quit: Page closed)
15:07:08 *** Samu has quit IRC (Ping timeout: 480 seconds)
15:07:11 *** sugarman has joined #openttd
15:07:41 <_dp_> _2TallTyler, how is centered year any more problematic than a left one? especially on the left side
15:08:25 <sugarman> every1 i have questn
15:08:36 <sugarman> how 2 put passengr in2 boat?
15:08:59 <_dp_> also I don't think that's important at all, just having full years show is more than enough to navigate the chart imo
15:09:48 <sugarman> plz help do i need put truck lodin bay or bus stop by?
15:11:05 <sugarman> it says passengr but no can put bus go to bay
15:11:20 *** gelignite has joined #openttd
15:14:48 <Wolf01> sugarman: try with transfer orders or enable passengers destinations
15:15:08 <sugarman> dose that reqiure vehicle?
15:15:11 <sugarman> or take from town without?
15:16:13 <sugarman> i think i misunderstande
15:16:52 <sugarman> do u must put bus up 2 the dock or must not need? can dock take town people no bus?
15:17:43 <sugarman> argh game is 2 hard!!!!!!!
15:18:39 <_dp_> coop players are sure different nowadays...
15:20:06 <sugarman> plz i am newe gamer i play game not good
15:20:24 <sugarman> no hate must only love
15:21:42 <peter1138> When you find a performance problem, spend time fixing it, and then refactor something else which makes that all irrelevant...
15:22:00 <FLHerne> sugarman: Your question is hard to understand
15:22:19 <sugarman> when u get boat dock, how get people in dock?
15:22:27 <sugarman> u need road with bus?
15:22:30 <FLHerne> sugarman: A dock can receive passengers from a town without any other transfers, if it's near enough houses
15:22:36 <FLHerne> You don't need a road
15:22:43 <sugarman> does road help people get?
15:22:49 <sugarman> or no need no whatsoever
15:23:03 <sugarman> and what if i want person to be
15:23:15 <sugarman> moved from 1 town to another town, den i put them on a boat
15:24:53 <FLHerne> sugarman: Notice that "Coverage area highlight" is "On", so some tiles have a blue outline
15:25:04 <FLHerne> Those are the ones that the dock will collect people from
15:25:59 <sugarman> how put oil in a boat?
15:26:03 <sugarman> if oil in middle of big land
15:26:16 <FLHerne> If enough houses are in the area, it'll show "Supplies: Passengers, ..."
15:26:30 <sugarman> wut about oil FLHerne
15:26:36 <sugarman> oil is not passenger
15:27:04 <FLHerne> Either build a canal, or use a truck or train and then transfer orders
15:27:31 <FLHerne> Most things in OpenTTD work similarly for any cargo
15:28:22 <sugarman> THANKS UR SO GOOD AT OPENTTD
15:28:32 <sugarman> FLHerne: wanna play openttd war even though id lose??
15:29:23 <sugarman> its okay you cant be good at everything
15:30:14 *** sugarman has quit IRC (Remote host closed the connection)
15:39:16 <LordAro> "LIBLZMA not found; compiling OpenTTD without LIBLZMA is strongly disencouraged"
15:41:49 <FLHerne> We have the best words
15:45:35 *** Progman has joined #openttd
15:49:30 <FLHerne> Of course, 'disenfranchise' is the correct word
15:49:47 <FLHerne> English is a silly language
16:00:06 *** arikover has joined #openttd
16:06:19 *** Flygon has quit IRC (Read error: Connection reset by peer)
16:10:14 <DorpsGek> [OpenTTD/OpenTTD] nielsmh updated pull request #8716: Fix #8713: Change OTTD2FS and FS2OTTD to return string objects https://git.io/Jt9Ha
16:18:49 <DorpsGek> [OpenTTD/OpenTTD] nielsmh updated pull request #8716: Fix #8713: Change OTTD2FS and FS2OTTD to return string objects https://git.io/Jt9Ha
16:34:05 *** Wormnest has joined #openttd
17:10:20 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8716: Fix #8713: Change OTTD2FS and FS2OTTD to return string objects https://git.io/Jt77E
17:11:58 *** frosch123 has joined #openttd
17:13:08 <frosch123> if you close 1.5 eyes, then "DBaaS" looks very similar to "BaNaNaS"
17:19:28 *** sugarman has joined #openttd
17:20:32 <frosch123> TrueBrain: yes, if you can fix translations with a regex, it's the right thing to do
17:26:45 *** sugarman has quit IRC (Remote host closed the connection)
17:30:56 *** Progman has quit IRC (Remote host closed the connection)
17:33:42 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #8730: Codechange: [OpenGL] Load all OpenGL functions dynamically. https://git.io/Jt7Fk
17:35:34 *** sugarman has joined #openttd
17:43:31 *** sugarman has quit IRC (Remote host closed the connection)
17:43:42 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on issue #8686: Unable to change townnames in Scenario editor after pr8566 https://git.io/JtMyZ
17:46:43 <DorpsGek> [OpenTTD/OpenTTD] nielsmh commented on pull request #8730: Codechange: [OpenGL] Load all OpenGL functions dynamically. https://git.io/Jt7bZ
17:48:57 *** gelignite has quit IRC (Quit: Stay safe!)
17:50:36 <frosch123> michi_cc: your PR works on my machine :)
17:54:28 *** qwebirc85304 has joined #openttd
17:54:41 *** qwebirc85304 has quit IRC ()
17:54:53 *** sugarman has joined #openttd
17:56:45 *** Progman has joined #openttd
18:03:01 <milek7_> shouldn't things in video/cocoa/CMakeLists.txt be guarded with if(NOT OPTION_DEDICATED) like the ones in video/CMakeLists?
18:04:21 *** sugarman has quit IRC (Remote host closed the connection)
18:05:13 <frosch123> now that the CI dockers are defined inside GH actions, you can probably add a CI target for dedicated servers in a container with minimal dependencies
18:05:45 <frosch123> though, probably a waste of cpu time
18:14:53 *** glx is now known as Guest13936
18:16:41 <frosch123> i like how i swapped # and ##, and nielsm then did the same
18:20:43 *** Guest13936 has quit IRC (Ping timeout: 480 seconds)
18:22:14 *** sugarman has joined #openttd
18:22:16 <LordAro> milek7_: i imagine no one's tried to make a dedicated server on macos before
18:23:02 <nielsm> frosch123 I just went by your :D
18:23:05 <nielsm> since I don't remember either
18:23:44 <glx> I guess all the subdirectory can be skipped
18:27:20 *** jottyfan has joined #openttd
18:27:58 <LordAro> glx: it still needs the null video driver, right?
18:28:21 <glx> I mean the add_subdirectory(cocoa)
18:28:50 <frosch123> "dedicated" is a separate video driver
18:29:19 <glx> OPTION_DEDICATED is for dedicated only builds
18:30:43 *** sugarman has quit IRC (Remote host closed the connection)
18:30:48 *** sla_ro|master has quit IRC (Read error: Connection reset by peer)
18:31:05 *** sla_ro|master has joined #openttd
18:31:13 *** andythenorth has joined #openttd
18:31:29 <frosch123> "null" runs ff-only
18:31:59 <frosch123> pff, 3 minute lag spike
18:33:20 <peter1138> Disemcombobulati//@//on1141414411111141141414
18:34:30 <arikover> I am playing with 20210223-master-gcc465efa67, and I just noticed that when restarting a game (command "restart"), the map has everything except industries. Is that intended or is this a bug?
18:37:27 <LordAro> arikover: that seems like a bug
18:38:01 <LordAro> arikover: actually, maybe not
18:38:07 <LordAro> what did you start the game from?
18:39:31 <arikover> I started a fresh game. Not from a savegame, if that's what you mean.
18:39:55 <LordAro> as in, new game -> play a bit -> restart command ?
18:40:00 <frosch123> arikover: "restart" uses the current game settings. if you changed "industry density" to "funding only" in game, then it's expected
18:40:47 <LordAro> `help restart` & `help reload` might explain it
18:43:02 <arikover> Oh I guess I know what is happening. I start with low industries, but also use a GS (Think globally act locally), and this might be changing the industry setting at the start of the game. Is that possible?
18:44:23 *** peter1138 has left #openttd
18:59:55 <DorpsGek> - Update: Translations from eints (by translators)
19:02:25 *** sugarman has joined #openttd
19:08:22 <sugarman> hey sirs can some1 give me download 2 newist build of openttd with gfx tar insid? preferably by mega.nz
19:08:41 <sugarman> openttd.org blocked 4me by org :(
19:09:24 <LordAro> openttd.org is blocked one of the bigger filesharing sites on the internet is not?
19:09:35 <sugarman> cuz is game, openttd
19:09:37 <arikover> frosch123 LordAro: The GS changes the industry density during the game to funding only. So that's it. Thank you for the help!
19:10:11 <sugarman> if u be so kind sir ill make much gratitudinal LordAro
19:10:29 <LordAro> sugarman: i'm feeling generous, win64?
19:11:06 <sugarman> im stuck on archive.org version i culd find and it version mismatch :(((
19:11:47 <sugarman> and preferably portable version but is no problm if installer prob LordAro
19:12:20 <TrueBrain> Don't forget to add the coin miner
19:12:58 <sugarman> also plox keine RAT :)
19:13:17 <michi_cc> TrueBrain: Re #8731, it's not loading 40bpp-anim because a) it would only be loaded after the video driver is loaded (which is not in the debug output snippet), and because b) it is in fact loading 8bpp-optimized, as that is faster.
19:13:30 <LordAro> i figured you'd probably want the portable version :p
19:13:48 <LordAro> if your internet connection is locked down that much, you probably can't install things easily either ;)
19:14:02 <sugarman> thank u very much sir ur veri kind
19:15:26 <TrueBrain> michi_cc: ah, yes, right :)
19:16:04 <michi_cc> So, force a 32bpp with antialias enabled only for OpenGL or always?
19:16:26 <TrueBrain> Most OSes already defaulted to a 32bpp blitter
19:16:37 <TrueBrain> As I found out during testing
19:16:52 <TrueBrain> So doing for OpenGL makes total sense to me
19:17:21 <michi_cc> Yeah, I explicitly changed that as, OpenGL not, only pushing a fifth of the data is always faster.
19:18:17 <TrueBrain> I think we should start looking in deprecating 8bpp blitters in general :) .. more options towards the future with 32bpp :p
19:19:42 <sugarman> can u chat in multi player ?
19:20:08 <nielsm> yes, you just press Enter to open the chat box
19:20:31 <sugarman> nice thanks nielsm ur very cool :)
19:21:04 <Eddi|zuHause> is there magic command line option to turn off opengl?
19:22:28 <Eddi|zuHause> also, the problem with LTO is that it's essentially like a full recompile even though you changed only 1 compile unit
19:22:30 *** Tirili has quit IRC (Ping timeout: 480 seconds)
19:22:41 <LordAro> sudo apt-get remove nvidia-driver
19:22:57 * Eddi|zuHause slaps LordAro around a bit with a large trout
19:23:15 <FLHerne> Some of the corner angles look much more aligned than others
19:24:07 <FLHerne> (this is 'Passenger Carriage' in 1.3.5)
19:24:22 *** gelignite has joined #openttd
19:25:14 *** supermop_Home has quit IRC (Remote host closed the connection)
19:26:50 <Eddi|zuHause> michi_cc: i see no real difference between commenting and uncommenting the define
19:28:03 <michi_cc> Okay, in that case I think I'll leave the code as is. Better being safe, especially if it isn't doing much.
19:28:04 *** sugarman has quit IRC (Remote host closed the connection)
19:28:57 <Eddi|zuHause> well, i haven't done too much testing... loading a large savegame, zooming out, and scrolling around a bit while paused
19:29:16 <Eddi|zuHause> fps drops from 30 to around 18 in both cases
19:29:51 <Eddi|zuHause> in the commented case it seemed rather constant, in the uncommented case a bit more jumpy, but seemed to average out
19:30:36 <Eddi|zuHause> also, i'm starting to get annoyed that the game doesn't start on display 0
19:30:53 *** sugarman has joined #openttd
19:32:07 <TrueBrain> have your mouse on display 0 when starting :)
19:32:22 <TrueBrain> and you can start without OpenGL by using "-v sdl" :)
19:32:56 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler opened issue #8734: Vehicles not rendered in 2CC when dragged inside a depot https://git.io/Jt5JC
19:33:09 <Wolf01> Hmmm, I think I passed out for a hour or 2
19:33:29 <Eddi|zuHause> TrueBrain: i know that. but no other game does it like that, and that's the annoying part
19:34:04 <frosch123> you can also pass the display via command line
19:34:14 <andythenorth> Wolf01 you alive?
19:34:27 <TrueBrain> Wolf01: should we be worried, call 112, or?
19:34:45 <TrueBrain> would be fun "giving your address" :D
19:34:50 <andythenorth> I had a (probably not Covid, probably another mystery virus with all the same symptoms) Jan 2019
19:34:55 <Eddi|zuHause> ok, so in the non-opengl case, fps is more like 12
19:34:59 <andythenorth> then I spent a lot of 2019 randomly falling asleep
19:35:39 <andythenorth> don't only Americans get mono?
19:35:55 <andythenorth> it's a thing in US sitcoms, but I know nobody in Europe who has ever mentioned it
19:36:31 <andythenorth> oh we call it something else
19:36:36 <TrueBrain> andythenorth: that is either local bias, or just not wanting to hear :P
19:36:54 <TrueBrain> but many people get it really young here, which is a good thing
19:37:06 <andythenorth> oh we call it glandular fever
19:37:08 *** sugarman has quit IRC (Remote host closed the connection)
19:38:03 <TrueBrain> "Ziekte van Pfeiffer" it is called here
19:39:12 <LordAro> mm, 31s fresh compile time
19:39:22 <LordAro> 40 cores is quite nice
19:39:23 <TrueBrain> Debug build, I assume :P
19:39:28 *** _2TallTyler has joined #openttd
19:39:34 *** Wolf01 is now known as Guest13947
19:39:52 <LordAro> and yes, also debug build
19:39:54 <Wolf01> Woop, windows crashed with netflix
19:41:25 <DorpsGek> [OpenTTD/OpenTTD] James103 commented on issue #8735: Transparency of houses, etc., rendered too dark https://git.io/Jt5Jp
19:41:58 <TrueBrain> how does he do that ...
19:42:41 <LordAro> both issues are presumably opengl?
19:42:43 <Wolf01> <andythenorth> don't only Americans get mono? <- I've got mono, but the development SDK, maybe they get the runtime :P
19:43:00 <TrueBrain> does OpenTTD has something you can click that puts something in your clipboard or something about what OpenTTD you are running, with all the settings etc?
19:43:10 <TrueBrain> so debugging what is being used exactly is easier?
19:43:12 <LordAro> TrueBrain: 2:20 for release build :(
19:43:25 <TrueBrain> LordAro: yeah ... I stopped trying Release builds because of that
19:43:26 *** Guest13947 has quit IRC (Ping timeout: 480 seconds)
19:43:35 <TrueBrain> happy Debug builds now run at 9999 fps simulation for me
19:44:05 <_2TallTyler> I'm here if I can answer any of those questions for your TrueBrain
19:44:17 <TrueBrain> _2TallTyler: nah, in your case I am pretty sure they are caused by OpenGL :)
19:44:30 <TrueBrain> but I notice often we have to ask basic information of users
19:44:34 <TrueBrain> most which they wouldn't know how to answer
19:44:58 <_2TallTyler> I figured as much. I certainly don't know if I'm running OpenGL...just whatever it launched with automatically
19:45:00 <TrueBrain> the current reports just show how bad of a tester I am :D
19:46:31 <michi_cc> I'd tend for it to be a bug in the 40bpp-anim blitter, not the OpenGL driver, as transparency is handled by the blitter.
19:47:25 <FLHerne> Why is there a 40bpp blitter?
19:47:40 <FLHerne> (macOS does high colour depth natively now?)
19:48:04 <michi_cc> Because letting the GPU do OTTD palette animation/CC overlay is better than letting your CPU do it.
19:48:25 *** sugarman has joined #openttd
19:48:35 <andythenorth> oof I need to get the build working eh :P
19:50:29 <FLHerne> I assumed it was RGBA with 10 bits per channel, like Photoshop etc.
19:51:34 <FLHerne> Google says RGBAM is + 'Mint', but only for very niche lighting things
19:53:35 *** _2TallTyler has quit IRC (Quit: Page closed)
19:55:35 <michi_cc> M == mask in GRF speak.
19:59:52 <TrueBrain> LordAro: gave certbot any more attention?
20:00:14 <LordAro> completely forgot about it
20:00:20 <TrueBrain> letsencrypt reminded me :D
20:00:54 <TrueBrain> michi_cc: as a FYI, the above 2 tickets also happen with the 8bpp-optimized blitter
20:01:04 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
20:01:19 *** WormnestAndroid has joined #openttd
20:02:39 <michi_cc> That almost rules out OpenGL, as the 8bpp case is really only applying our palette and nothing more. Maybe one of the video driver refactorings broke something with palette animation.
20:02:48 <TrueBrain> the drag&drop is OpenGL only
20:02:54 <TrueBrain> the transparency happens also on win32
20:03:56 <TrueBrain> transparency only happens in 8bpp, so yeah, I will bisect what causes that
20:04:12 <TrueBrain> drag&drop also happens with 40bpp-anim
20:04:56 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on issue #8734: Vehicles not rendered in 2CC when dragged inside a depot https://git.io/Jt5JC
20:05:22 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on issue #8735: Transparency of houses, etc., rendered too dark https://git.io/Jt5Jp
20:05:47 <TrueBrain> right, lets see what broke transparency .. finding a known-good commit first :P
20:06:33 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #8736: Fix #8731: Always use a 32bpp blitter if font anti-aliasing is enabled. https://git.io/Jt5kG
20:06:35 <sugarman> are signals important 4 makin trainz
20:09:23 <TrueBrain> Bisecting: 41 revisions left to test after this (roughly 5 steps)
20:09:44 <LordAro> sugarman: depends how complicated you want to make things
20:10:07 <sugarman> if i make a big ring of tracks aroun a buncha townz
20:10:13 <sugarman> n put a buncha trains on it
20:11:07 <LordAro> TrueBrain: why on earth did you decide on red(ish) for links on the wiki?
20:11:20 <TrueBrain> that is orange darling
20:11:24 <TrueBrain> like all openttd.org links are
20:11:31 <TrueBrain> for like .. 15 years now :)
20:11:37 <LordAro> yeah... but it's the wiki
20:11:46 <LordAro> every wiki ever uses red links for "page does not exist"
20:11:58 <TrueBrain> that is why it is not red
20:12:08 <TrueBrain> I just took the website CSS :)
20:12:44 <TrueBrain> deadlinks are grey btw
20:12:56 <LordAro> also, it is #DD6000, orange is rather lighter :p
20:13:22 <TrueBrain> pretty sure with that much green you cannot consider it red either
20:13:27 <TrueBrain> it left red like 0x003000 ago :P
20:13:44 <LordAro> "mediumblue" is much better, imo
20:14:06 <TrueBrain> pretty sure if we ask 10 people we get 10 different answers :D
20:14:13 <TrueBrain> but the why is simple, I just copied the website CSS :)
20:15:22 <TrueBrain> don't you hate that you are bisecting on a 32bpp blitter, where you wanted to test a 8bpp?
20:15:44 <TrueBrain> LordAro: I kinda gave up debating "what colour should this link be" for openttd.org like 10 years ago
20:16:29 <TrueBrain> for years people have been saying they will fix our CSS
20:16:43 <TrueBrain> I think you put the most effort in it
20:16:48 <TrueBrain> gives you a bit of an idea :P
20:17:00 <glx> hey some did fix parts of the CSS
20:17:00 <andythenorth> nothing wrong with it
20:17:33 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8736: Fix #8731: Always use a 32bpp blitter if font anti-aliasing is enabled. https://git.io/Jt5If
20:17:46 <TrueBrain> andythenorth: clearly that is a matter of opinion
20:19:43 <andythenorth> mine is more valid of course
20:19:52 <andythenorth> just not everyone recognises that yet
20:21:15 <TrueBrain> now I come to think of it, the transparency .. is that not just 32bpp vs 8bpp
20:21:46 <TrueBrain> lets compile 1.10.3 and check ..
20:23:07 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on issue #8735: Transparency of houses, etc., rendered too dark https://git.io/Jt5Jp
20:23:19 <TrueBrain> michi_cc: with ^^, I kinda vote on making 32bpp the default for OpenGL on all platforms too
20:24:26 <TrueBrain> michi_cc: I think there will be more of those regressions for a lot of players, as they were used to how 32bpp does it
20:24:32 <TrueBrain> 8bpp is just .. subpar :P
20:31:33 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on issue #8670: Crash on macOS Big Sur after downloading graphics to 200% https://git.io/JtiPZ
20:41:03 <frosch123> oh, i thought that was already closed
20:41:32 *** Gustavo6046 has joined #openttd
20:41:37 <TrueBrain> :D You can have the pleasure of closing an issue :D
20:42:13 <frosch123> i had to look up the pr
20:42:22 <TrueBrain> it was already 3 days ago
20:42:37 <frosch123> it was the osfx issue that was closed
20:43:32 <TrueBrain> E_TOO_MANY_TICKETS :D
20:46:10 <TrueBrain> I like how our NewGRF Settings window shows "Palette"
20:46:18 <TrueBrain> as that is the important piece of information to show our users :D
20:46:46 <frosch123> it's important for newgrf older than 10 years
20:47:03 <TrueBrain> as user, I have no clue what it means or what I have to do with the info, honestly :)
20:47:06 <frosch123> like that vehicle grf that is almost as old as ottd, but people still talk abuot it
20:47:25 <TrueBrain> I use to have 2 GRFs, that gave me a game with trams
20:47:27 <frosch123> TrueBrain: rename the button to "if things look pink, click here"
20:47:33 <TrueBrain> I now have 100 GRFs .. and I hav eno clue which gives me trams :D
20:47:47 <TrueBrain> frosch123: no, not a button or anything .. "Detailed NewGRF information" show that piece of info
20:47:47 <frosch123> use the text fitler for "tram"
20:48:09 <TrueBrain> frosch123: tnx, that worked surprisingly well
20:48:21 <glx> haha only 100 GRFs, the scan should be fast for you
20:48:45 <frosch123> TrueBrain: at the bottom there is a "toggle palette" button
20:48:55 <frosch123> it toggles the thing in the details
20:49:04 <TrueBrain> button is disabled :P
20:49:19 <frosch123> yes, in-game with no dev-settings
20:49:33 <TrueBrain> I am at the intro menu
20:49:36 <glx> opengfx uses windows palette only IIRC
20:49:48 <TrueBrain> with a clean config
20:49:49 <DorpsGek> [OpenTTD/OpenTTD] orudge commented on pull request #8718: Change: [Actions] stop relying on external actions for trivial stuff, and rely on cmake tools to build, test and pack https://git.io/Jt5tW
20:49:51 <frosch123> we even allow toggling the palette when the newgrf sets it properly
20:50:06 <frosch123> oh, maybe that is a dev-setting then
20:50:20 <TrueBrain> so maybe we shouldn't show it for non-dev or something, dunno :)
20:50:52 <TrueBrain> anyway, it just made me giggle, like: what do I have to do with this piece of intel :D
20:51:09 <frosch123> newgrf-dev-tool allows you to toggle the palette unconditionally
20:51:24 <frosch123> without them, you can only toggle them for old newgrf without palette information
20:51:36 <frosch123> TrueBrain: what do you to with grfid and md5sum?
20:51:43 <TrueBrain> frosch123: I was wondering the same, honestly
20:51:50 <TrueBrain> but I could imagine the case where someone asks you for that, or what-ever
20:51:53 <TrueBrain> but it is really thin too
20:51:55 <TrueBrain> same with Parameters
20:52:02 <TrueBrain> I get these bytes .. like .. euh .. okay, tnx?
20:52:02 <frosch123> imo, there is the space, so show something
20:52:32 <frosch123> i think md5sum is more useful than the "tags" on bananas
20:52:50 <frosch123> sometimes you need the md5sum. the tags are always useless :p
20:52:53 <TrueBrain> "View Details" button to show these things would work well too, I guess :
20:53:19 <TrueBrain> but okay, that NewGRF interface is not for the faint of heart anyway :)
20:53:20 <frosch123> and then? when i open the newgrf window, i have a lot of empty space on the right
20:53:36 <FLHerne> frosch123: Could the palette be determined by heuristics?
20:53:46 <TrueBrain> so the reason to add information most users don't know what to do with is because otherwise it is too empty? :D
20:53:56 <FLHerne> "Everything is pink, we should use the other one" seems like something the game can know on its own :p
20:54:06 <frosch123> FLHerne: not worth the effort
20:54:15 <frosch123> every newgrf since 2012 sets it
20:54:34 <frosch123> those people who use newgrf older than that, also know what palette means
20:54:39 <andythenorth> what are 'parameters'
20:54:45 <andythenorth> as we're playing this meta game :)
20:55:14 <andythenorth> I should stop playing Blitz
20:55:18 <andythenorth> it just funds Putin anyway
20:55:24 <frosch123> maybe ottd should ask "in what year did you play ottd first?" on start-up?
20:55:30 *** jottyfan has quit IRC (Quit: jottyfan)
20:55:38 <frosch123> then use that to disable/enable old silly features
20:55:48 <TrueBrain> frosch123: well, it is funny to walk through the interface after having for years :)
20:55:55 <andythenorth> or move 'toggle palette' to parameters
20:56:00 *** sugarman has quit IRC (Remote host closed the connection)
20:56:02 <andythenorth> rename parameters to settings
20:56:08 <andythenorth> rename grfs to mods
20:56:23 <TrueBrain> sometimes the audience indeed seems to be another group than me :)
20:56:37 <andythenorth> you are not a group :)
20:56:45 <TrueBrain> have it your way :P
20:56:45 <andythenorth> oh dear andythenorth
20:56:54 <TrueBrain> want more salt with that?
20:57:09 <Timberwolf> frosch123: copy protection style quiz! List a feature and multiple choice answers of "this feature was first introduced by [ ] TTO [ ] TTD [ ] TTDPatch [ ] OpenTTD"
20:57:20 <andythenorth> salty salty tears
20:57:26 <andythenorth> oh wait, wrong game
20:57:28 <andythenorth> this isn't Blitz
20:57:36 <TrueBrain> or make a configuration setting: "I am old, give me back all my information nobody ever looked at" :P :P :P
20:57:51 <TrueBrain> would work for the mapgen too!
20:57:56 <andythenorth> can we rename another button pls?
20:58:01 <andythenorth> View unreadableme
20:58:19 <frosch123> TrueBrain: anyway, you can tie tha "show palette info" to the "enable palette toggle button"
20:58:26 <frosch123> there is no point in showing it, if you can't toggle it
20:58:45 <frosch123> you can probably also tie md5sum to some developer setting
20:58:48 <TrueBrain> honestly, that confused me most, the distance between that text and the button
20:58:51 <TrueBrain> I did not link those together :D
20:58:55 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8718: Change: [Actions] stop relying on external actions for trivial stuff, and rely on cmake tools to build, test and pack https://git.io/Jt5qI
20:59:15 <TrueBrain> now you mentioned the button
20:59:17 <TrueBrain> it makes more sense :)
20:59:22 <frosch123> it's a classical gui :) display at the top, radio buttons at the bottom
20:59:31 <TrueBrain> yeah .. it fits the style of the game :)
20:59:40 <frosch123> a lot of ancient software follows the same pattern as ancient hardware
20:59:45 <TrueBrain> but okay, just random observation while trying to load a tram NewGRF :)
21:00:22 <TrueBrain> now lets look at the lovely bug: the tram is in front of the depot, going towards the depot .. yet, pressing the button: goto depot, says: unable to find depot :D
21:01:29 <frosch123> that only applies to trams? not to other road vehicles?
21:02:34 <andythenorth> isn't this just pathfinder caching though?
21:02:34 <TrueBrain> its a fun little bug :D
21:03:00 <frosch123> sometimes the pathfinder starts searching on the next tile, maybe it has issues if the next tile is already the depot
21:03:14 <TrueBrain> no, it has to do with half-tiles
21:03:17 <TrueBrain> where it can turn around
21:03:21 <TrueBrain> it seems the PF doesn't understand those
21:03:47 <TrueBrain> while on that half-tile, it cannot find the depot
21:03:51 <TrueBrain> when it leaves, it can
21:04:00 <TrueBrain> busses etc never enter the half-tile
21:04:06 <TrueBrain> it turns around before the half-tile happens
21:04:09 <TrueBrain> I guess that is related
21:04:58 <glx> maybe exit direction is not properly detected for these
21:05:18 <TrueBrain> the fun part: having to open follow_track.hpp :D
21:06:44 <TrueBrain> if you don't hear from me in the next 2 days, call help
21:06:53 <glx> I think I opened it not too long ago for buses
21:08:24 <michi_cc> Okay, there's a very easy solution for #8734 and a quite hard one.
21:09:31 *** _2TallTyler has joined #openttd
21:11:24 <michi_cc> TrueBrain: You prefer to basically always use 32bpp I guess?
21:11:38 <TrueBrain> michi_cc: I think it avoids a lot of annoyance by users, yes
21:11:53 <TrueBrain> I do like your PR btw, so maybe "both" is the correct answer :)
21:12:20 <TrueBrain> but really, if you are used to 32bpp transparency, you won't get used to 8bpp transparency anymore :P
21:14:11 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on issue #8735: Transparency of houses, etc., rendered too dark https://git.io/Jt5Jp
21:14:52 <michi_cc> Okay, lets merge #8736 then. Removing the prefer 8bpp altogether can be another commit.
21:15:06 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #8736: Fix #8731: Always use a 32bpp blitter if font anti-aliasing is enabled. https://git.io/Jt5kG
21:15:18 <TrueBrain> michi_cc: exactly :)
21:15:43 <TrueBrain> okay, GetTileTrackStatus() returns 0 on half-tiles
21:16:56 <TrueBrain> and there is a "hack" there, which makes the PF think the tile is a full tile straight
21:17:17 <TrueBrain> so that explains why the PF is like: DUDE WTF :D
21:17:37 <michi_cc> TrueBrain: Okay, so for #8734, I completely missed that the mouse cursor can have a company colour remap applied. Easy fix: just remove separate cursor drawing. It was mostly intended to allow fast mouse updates than the game loop, but you've made that obsolete :)
21:17:57 <michi_cc> The hard fix is to improve the GL shader code to cater for all the remaps.
21:18:08 <glx> oh but the hack was for buses I guess
21:18:27 <TrueBrain> glx: no, the hack is for trams specific; and it is in the PF btw, not in GetTileTrackStatus, to be clear :D
21:18:41 <TrueBrain> what I like about the OpenGL mouse handling
21:18:47 <TrueBrain> that it doesn't snapshot the buffer for bytes
21:18:53 <TrueBrain> end copy them back to "undraw"
21:19:01 <glx> ah then the hack feels wrong has half tile exit should be entrance in this case
21:20:03 <_2TallTyler> Question about strings: When game_creation.ending_year is set to 0, the setting reads "Never". I see the string in english.txt, but can't find where that string is specified when the setting is 0. Any ideas?
21:20:05 <TrueBrain> michi_cc: how "hard" is the hard fix?
21:20:45 <michi_cc> However hard it is to do Blitter_32bppOptimized::Draw as shader :)
21:20:46 <glx> searched for the string id _2TallTyler ?
21:20:50 <andythenorth> can we have drive through stations from jgrpp? /me wonders
21:22:06 <TrueBrain> michi_cc: :D Well, I think you can best estimate what solution is better .. having the mouse done by OpenGL is just better than the software solution OpenTTD uses, I think, but having to write complex shaders can break easily .. so yeah, nice problem to balance :D
21:23:01 *** Tirili has quit IRC (Remote host closed the connection)
21:23:04 <frosch123> michi_cc: the mouse cursor is also broken when if consists of multiple sprites
21:23:14 <frosch123> do you need a screenshot, or do you know what i mean?
21:23:20 <glx> _2TallTyler: SettingEntry::SetValueDParams() I guess
21:24:08 <_2TallTyler> Aha, figured it out. Thanks.
21:24:09 <michi_cc> frosch123: That is supposed to work as it is iterating over _cursor.sprite_count
21:24:53 <frosch123> hmm, i think it draws both sprites, but the x-offset is off, they are overlapping
21:25:38 <frosch123> oh, the mouse-cursor clipping does not account for gui zoom
21:26:50 <michi_cc> Do you have some example cursor what is multi-part?
21:27:28 <frosch123> articulated vehicles
21:27:32 *** sla_ro|master has quit IRC ()
21:27:35 <frosch123> all the big engines in nuts
21:30:47 <frosch123> maybe it misses the sprite-offsets for the secondary sprites
21:36:40 <TrueBrain> I STRANDED A TRAM \o/
21:37:01 <TrueBrain> it derailed and got stuck
21:37:20 <_2TallTyler> Easy, just pull the pole off the wire ;)
21:38:28 <TrueBrain> so yeah, we don't have trackbits for halftiles
21:38:38 <TrueBrain> they are a bit centered around rails :D
21:38:55 <frosch123> no, trains and ships have the same issue
21:39:00 <frosch123> depots, tunnels, bridges
21:39:14 <frosch123> all of them are enterable from one side only, and do not traverse
21:39:23 <frosch123> they are all special-cased in the pathfinder
21:39:39 <TrueBrain> but rails don't have half-tiles, do they?
21:39:42 <TrueBrain> sorry, I am a bit confused :D
21:39:48 <frosch123> depot are the same thing
21:39:57 <TrueBrain> rails in depot, yeah, fair
21:39:59 <frosch123> you can enter from one side, but only exit on the same one
21:40:11 <frosch123> same for tunnels and bridges
21:40:27 <TrueBrain> roads and trams just have it in many more places
21:40:33 <TrueBrain> and for trams a single case is patched
21:40:43 <TrueBrain> so that annoyed someone enough :D
21:41:01 <frosch123> try grepping for depot in the pathfinder code :p then add the tram-special case in the same places
21:41:21 <frosch123> actually, this sounds vaguely familar. i think i did that 12 years ago
21:41:36 <TrueBrain> I really wonder if we shouldn't just add trackbits, but that might be a bit more work :D
21:41:59 <frosch123> that does not solve anything
21:42:09 <frosch123> then the pathfinder thinks it can enter from the other side
21:42:20 <TrueBrain> if we add half-trackbits
21:42:47 <frosch123> gettiletrackstatus also already has a "enterdir" parameters
21:42:56 <frosch123> i think bridges/tunnels/depots depend on that
21:43:08 <frosch123> so, if anything, make trams more similar to depot/tunnel/bridges, they work well
21:43:23 <TrueBrain> rvs have a similar issue
21:43:32 <TrueBrain> so it is a lot of "exception" code
21:43:50 <frosch123> rvs just ignore those single-bit tiles, they pretend they don't exist :p
21:43:56 <glx> so any half-road (tram or not) should be handled like a depot
21:43:57 <TrueBrain> yeah, which is a bug on its own
21:44:12 <TrueBrain> but for now, I am trying to understand what this hack for trams is suppose to do
21:44:17 <frosch123> you can bulldoze single roadbits of competitors, but not single trambits
21:44:41 <glx> but RVs can't turn around on half tile, they would move on grass ;)
21:44:45 <TrueBrain> if there is a half-tile, it acts like it is a full tile over either axis
21:45:05 <TrueBrain> which feels like it tried to address this bug, but kinda failed or something
21:45:17 <TrueBrain> glx: why wouldn't they be able to turn around earlier? :)
21:46:14 <TrueBrain> ah, wait, another place corrects the tram hack
21:46:36 <TrueBrain> it is a two-for-one-hack
21:47:33 <TrueBrain> single-bit trams act like they are x-axis or y-axis, but you cannot exit on the other side
21:47:38 <TrueBrain> so .. half-tiles :D
21:48:36 <TrueBrain> Some comments: /* sometimes the roadveh is not on the road (it resides on non-existing track) how should we handle that situation? */
21:53:27 *** hotgirl55 has joined #openttd
21:55:15 *** Samu has quit IRC (Quit: Leaving)
21:58:30 *** _2TallTyler has quit IRC (Quit: Page closed)
21:59:43 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #8738: Fix #8123: trams on half-tiles couldn't find depots https://git.io/Jt5Yr
21:59:56 <TrueBrain> well, instead of fighting the system, I just extended the already "fix halftile issue"
22:00:01 <TrueBrain> not sure what to think of it, honestly
22:01:55 <LordAro> TrueBrain: have you tried pointing 2 half-tiles at each other?
22:03:19 <DorpsGek> [OpenTTD/OpenTTD] frosch123 opened pull request #8739: Fix: vehicle-cursor size-limit did not account for the interface zoom level. https://git.io/Jt5Y1
22:03:33 <TrueBrain> good test-case LordAro :D
22:04:07 <LordAro> how about if there's a more favourable depot on the other side?
22:04:14 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8739: Fix: vehicle-cursor size-limit did not account for the interface zoom level. https://git.io/Jt5YM
22:04:21 <TrueBrain> LordAro: exactly what I tried :D
22:04:46 <TrueBrain> the PF does that well :D
22:05:05 <TrueBrain> it is just sad that this is needed to fix it ..
22:05:15 <TrueBrain> but by the looks, fixing it properly takes ... more effort
22:05:19 *** peter1138 has joined #openttd
22:05:19 *** ChanServ sets mode: +o peter1138
22:05:38 <frosch123> TrueBrain: is it possible to deduplicate that code with QueryNewTileTrackStatus ? or is that impossible?
22:05:52 <TrueBrain> hmm .. good question
22:06:13 <TrueBrain> the C++ magic in follow_track is high
22:06:34 <LordAro> follow_track is one of the more readable parts of the PF, imo
22:06:42 <TrueBrain> the code inside, yes
22:06:46 <LordAro> it's all the yapf stuff that is weird
22:06:47 <TrueBrain> how it is constructed .. euh ..
22:06:54 <TrueBrain> many template parameters
22:06:58 <TrueBrain> and helpers functions that read them
22:07:15 <LordAro> like the costrail.hpp(?) has a base class which isn't used anywhere else
22:07:28 <LordAro> true, but they're well enough named
22:07:30 *** andythenorth has quit IRC (Quit: andythenorth)
22:07:39 <TrueBrain> yeah, just my C++ knowledge lacks here greatly :)
22:07:48 <LordAro> you're not just calling proc() and hoping for the best :p
22:08:00 <TrueBrain> so I think it could be deduplicated, I just don't know how :D
22:08:19 <TrueBrain> I would have to create an instance of CFollowTrackT I think
22:09:11 <TrueBrain> if you follow that template, boy ...
22:09:15 <TrueBrain> so many redirects to other names
22:09:38 <frosch123> maybe pull out a c-style function that does the shared part, and call that one from both
22:09:58 <TrueBrain> I do wonder what that if does in general, in that depot function
22:10:04 <TrueBrain> like .. it only fails if the tram is not on any tram tracks
22:10:07 <TrueBrain> is that .. realistic?
22:10:25 <TrueBrain> so I could also just remove that check
22:11:42 <frosch123> TrueBrain: i think that "sometimes the roadveh is not on the road" refers to the animation states when a rv is reversing
22:11:56 <TrueBrain> no, the function, YapfRoadVehicleFindNearestDepot
22:12:03 <frosch123> road-vehicle/tram reversing and their frame is really weird
22:12:06 <TrueBrain> first checks if the RV is on any track
22:12:15 <TrueBrain> before it does the pathfinding
22:12:31 <frosch123> disable_node_optimization ?
22:13:10 <TrueBrain> looking through the history
22:13:19 <TrueBrain> added to check if the road below is compatible with vehicle
22:13:24 <TrueBrain> otherwise refuse to do depot-pathfinding
22:13:38 <frosch123> TrueBrain: that test it the same as in SetOriginFromVehiclePos
22:13:46 <frosch123> pretty sure it refers to rv while reversing
22:14:14 <frosch123> trains reverse instantly, roadvehicles have some weird animation states, that do not refer to trackbits
22:14:54 <frosch123> so: reverse a road vehicle, wait till it is orthogonal to the road direction, then click find depot
22:16:39 <frosch123> RoadVehicle::GetVehicleTrackdir() contains quite some magic
22:16:45 <frosch123> also, standard roadstops
22:16:52 <TrueBrain> was what introduced the if
22:16:53 <frosch123> what trackbits do they have :p
22:17:35 <frosch123> so, what happens when a roadvehicle is loading in a drive-in roadstop, and then you find a depot?
22:17:48 *** Progman has quit IRC (Remote host closed the connection)
22:19:21 <TrueBrain> is what I am using to test
22:19:54 <TrueBrain> but okay, lets not create more bugs by removing something I don't understand .. lets see if I can deduplicate it :)
22:20:26 <TrueBrain> yeah, for sure that function can return wrong values :)
22:20:33 <frosch123> that added SetOriginFromVehiclePos, YapfRoadVehicleFindNearestDepot copied it from there
22:20:55 <TrueBrain> can you remove a half-tile while a bus is on the tile ..
22:21:50 <TrueBrain> too bad they didn't leave a comment about when that happens
22:21:54 <TrueBrain> just that it happens
22:22:05 <frosch123> r5033 is about multistop handling
22:23:05 <TrueBrain> but okay, I would love to clean up this code, but I guess fixing the problem is more useful .. so deduplicate, and use in a few more places, I guess? :)
22:25:23 *** andythenorth has joined #openttd
22:28:44 <frosch123> hmm, no idea who yanek is
22:28:51 <frosch123> dorpsgek logs start in 2007
22:29:04 <frosch123> tt-forums has only unrelated reports on that day
22:30:09 <frosch123> but when something is introduced at 11:00 and fixed at 17:00, i think irc
22:30:29 <frosch123> there was not even a nightly inbetween
22:31:29 *** hotgirl55 has joined #openttd
22:36:07 <frosch123> night, need more sleep
22:36:10 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
22:36:32 *** Beerbelott has joined #openttd
22:37:07 <TrueBrain> now I wonder what happens if I apply the same logic for RVs ..
22:37:19 <Beerbelott> Hello! How do you configure AI script_settings on a headless instance?
22:37:31 *** mirrorbird has joined #openttd
22:42:30 *** mirrorbird has joined #openttd
22:43:20 *** mirrorbird has quit IRC (Remote host closed the connection)
22:43:38 <TrueBrain> possibly via openttd.cfg? Dunno .. you can always prepare a savegame and load that
22:45:48 <TrueBrain> hmm, no, for RVs it is not as simple as duplicating the same tram trick .. guess there is more to it
22:47:32 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:55:11 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8738: Fix #8123: trams on half-tiles couldn't find depots https://git.io/Jt5Yr
22:55:15 <TrueBrain> okay, enough fiddling with trying to make RVs turn on half-tiles :P
22:55:23 <TrueBrain> lets keep it at fixing trams :)
22:56:06 <TrueBrain> ugh, I added the issue in the commit message again .... fucking spam in the issue now :P
22:56:22 <Beerbelott> TrueBrain I set start_date for an AI to a specific value, but what AI::GetStartNextTime() returns is different
22:56:34 <TrueBrain> it is always randomized a bit, I believe
22:56:36 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
22:56:56 <Beerbelott> I wonder if the value I set is being loaded of if any of the AI::START_NEXT_* are being used
22:57:00 *** andythenorth has quit IRC (Quit: andythenorth)
22:59:05 <TrueBrain> yeah, it is always +/- 60 days from what you entered in start_date
22:59:14 <Beerbelott> Hmmm I don't get where the value is being randomised when it's being read from GetSetting (either in AIConfig or ScriptConfig)
22:59:19 <TrueBrain> well, [0, 60), I believe
22:59:32 <TrueBrain> _start_date_config defines how the config behaves
22:59:32 <Beerbelott> So I trust it's randomised on load from config files?
22:59:42 <TrueBrain> which has a value int random_deviation;
23:00:01 <TrueBrain> which is being added by AddRandomDeviation
23:00:26 <TrueBrain> which is called on any config change
23:00:32 <TrueBrain> so on load is the most likely situation, yes
23:00:37 <TrueBrain> well, load of a game
23:00:50 <TrueBrain> wouldn't be much random if it was not per game :)
23:01:34 <Beerbelott> OK thx for all this
23:01:39 <Beerbelott> thought I was crazy
23:02:11 <Beerbelott> Also, AI load doesnot work if you set it on the 1st line of the [ai_players] config block
23:02:22 <Beerbelott> it *has* to be at least on the 2nd line
23:02:31 <Beerbelott> It's a bit confusing
23:02:32 <TrueBrain> that sounds like a worthy bug report :)
23:02:48 <Beerbelott> Oh I know that place (:
23:03:39 <glx> but for dedicated servers it should work
23:03:50 <TrueBrain> so the entries also indicate which company slot?
23:03:59 <TrueBrain> not exactly what I expected, but I guess that makes sense
23:04:06 <TrueBrain> so not a bug Beerbelott ! :D
23:04:52 <Beerbelott> glx I confirm it doesnot work on dedicated server
23:05:07 <Beerbelott> I set on the 1st line, GetStartNextTime() returns 730
23:05:25 <Beerbelott> Which is 2 years fro mthe defaults settings somewhere
23:05:52 <TrueBrain> okay, I am going to close pathfinders and RVs ... so many places trams have code to allow them to reverse on half-tiles, I give up :D
23:06:13 <TrueBrain> (not only in the PF, but also in the cmd of trams)
23:06:33 <Beerbelott> I notice when the server runs, I notice there is always one "Server" client: I suppose it takes the 1st spot
23:06:43 <TrueBrain> it should be a spectator
23:06:48 <TrueBrain> there should be no company
23:07:13 <Beerbelott> No, there is none indeed
23:07:44 <Beerbelott> But the AI load is only triggered when the game runs... which occurs on my server only when there is at least one player connected, due to a setting
23:08:17 <TrueBrain> yes, AIs start after a few game days
23:08:24 <TrueBrain> so a paused game never starts an AI :)
23:08:25 <Beerbelott> Hence the AI load seems to only be triggered after the 1st player company is created... SO I suppose that's where the problem comes from? COllision between real players and the AI?
23:08:51 <TrueBrain> if company #1 exists, no AI can start on that company, no
23:09:03 <Beerbelott> So if I set an AI on sport, say #3, and 3 players exist when the server tries to load the AI: it will never happen?
23:10:36 <Beerbelott> I thought this list was a dynamic one of AIs, but it actually is a fixed list of players spots
23:10:44 <TrueBrain> company spots, it seems
23:11:00 <TrueBrain> so you can just define all spots with the same AI if you like
23:11:15 <TrueBrain> AIs start 1 after the other, with their delays
23:11:25 <TrueBrain> just you would have more AIs :D
23:11:36 <TrueBrain> well, the GUI works like that
23:11:39 <TrueBrain> so I guess it is intended
23:11:43 <TrueBrain> might not be very flexible, but ..
23:11:55 <Beerbelott> Well I certainly don't wanna fill aserver with AI up. SOunds even less useful than cryptomining on that box :D
23:11:55 <TrueBrain> btw, Beerbelott , setting start_date to 0 should start it from the moment a new game is created
23:12:16 <LordAro> sounds like something in one of Samu's (closed) PRs
23:12:26 <Beerbelott> Imagine a server full with AI. Do people watch that for fun? :D
23:12:41 <TrueBrain> Beerbelott: there are tournaments, yes :)
23:12:44 <glx> GUI is also defines max AI and disable extra lines
23:13:04 <TrueBrain> it is not really user friendly, the approach taken, honestly
23:13:08 <TrueBrain> as in, singleplayer was considered
23:13:19 <Beerbelott> glx There indeed is a max_no_competitors setting
23:13:29 <glx> yeah originaly AIs were SP only
23:13:35 <Beerbelott> So the idea would be to spam the AI in all the spots with that setting?
23:13:43 <TrueBrain> and the GUIs are build around the SP experience
23:13:53 <TrueBrain> and the ini file is a representation of the GUI :D
23:14:03 <glx> then someone allowed them on the server
23:14:07 <Beerbelott> Sounds awfully like a bypass
23:14:27 <TrueBrain> another approach on a multiplayer server is to use the admin port to start AIs
23:15:15 <Beerbelott> SOunds like a hack too: server confs are believed to be deterministic and self-sufficient
23:15:39 <glx> well AIs in MP are kind of a hack
23:15:48 <TrueBrain> are we looking for solutions or principles? :D
23:16:31 <Beerbelott> I think I got my hack, but a config solution would be nice (:
23:16:47 <TrueBrain> requires GUI changes too
23:16:50 <TrueBrain> so not an easy task ;)
23:17:03 <glx> you also have the option to use start script
23:17:16 <TrueBrain> but that is the same as using start_date 0, I guess
23:18:34 <glx> AI management is not optimal (even in SP)
23:19:00 <TrueBrain> 1.12 the big UI overhaul? :D
23:19:21 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:19:27 <glx> start_date setting itself is implemented as a hack in the AI config
23:19:41 <TrueBrain> it has more exceptions than I dare to count :)
23:20:49 <TrueBrain> but would be nice to change the whole UI to be more for ... players, instead of developers / NewGRF authors / ... :D
23:20:52 <TrueBrain> remove all the historical bla
23:21:35 <LordAro> needs one of LC's new game guis :)
23:21:53 <TrueBrain> link or it didn't happen
23:22:13 <LordAro> you're gonna make me search for it, aren't you?
23:22:21 <TrueBrain> you brought it up :)
23:22:24 <TrueBrain> you can't half-ass that :P
23:24:25 <glx> there was a suggestion with AI/GS and newgrf GUI "merge"
23:25:02 <TrueBrain> frosch123 went a bit further, and I agree with him: merge it together into presets
23:25:54 <TrueBrain> also works well for settings etc
23:26:21 <TrueBrain> tabs for online content is a bit annoying for when you are searching
23:26:34 <LordAro> the idea is to reduce scrolling
23:26:42 <LordAro> headers don't really help with that
23:26:45 <TrueBrain> hence those kind of tabs, and collapse them by default
23:26:55 <TrueBrain> works nearly the same as tabs :)
23:26:58 <glx> grouping newgrf by category seems nice, but load order matters for newgrf
23:27:21 <TrueBrain> and what I would do, is always keep the categories visible
23:27:32 *** supermop_Home has joined #openttd
23:29:12 <TrueBrain> so, reasons to remove 8bpp: allow PNG loading, allow nicer shadows, allow (future) smoother news transitions
23:29:22 <TrueBrain> well, they are all possible in 32bpp only
23:29:25 <glx> I think categories are ok for the available list, but not for the active config
23:29:29 <TrueBrain> so no need to remove 8bpp, just you can't have it with 8bpp :D
23:30:03 <TrueBrain> in general, most of those UIs are too busy
23:30:06 <TrueBrain> too many things in there
23:30:11 <TrueBrain> that doesn't work, not really
23:31:01 <TrueBrain> my idea would be to integrate online content with the screen you are at, as in: when you are looking at AIs, integrate the online content for AIs there
23:31:10 <TrueBrain> having a single window with all online content, feels ... weird
23:31:21 <TrueBrain> it was fun when there were like 100 things
23:31:46 <glx> opening online content by mistake is not fun ;)
23:31:55 <TrueBrain> and sure, tabs solve that issue
23:32:14 <TrueBrain> but .. there are other solutions too honestly, that just make the window not needed :P
23:32:38 <TrueBrain> a single window with "Installed mods" would be sufficient
23:33:00 <TrueBrain> server-side filtering
23:33:12 <TrueBrain> that would also help a lot with the lag etc
23:33:29 <glx> yeah newgrf, ais, GS in mods config
23:34:10 <TrueBrain> I will see if I can work that out a bit tomorrow
23:34:28 <TrueBrain> but just fluent integration of local and remote content
23:34:35 <TrueBrain> also for savegames for when cloud-saves :P
23:35:06 <glx> yeah and get scenario from bananas when you want to start a scenario
23:35:26 <TrueBrain> just show the ones you have locally, and the first N best rated from remote
23:35:37 <TrueBrain> when you type in a filter, allow fetching from remote when you click the "fetch from remote"
23:36:38 <TrueBrain> rating system ... another problem :D
23:36:50 <TrueBrain> giving stars has shown to not really work in this world
23:39:58 <TrueBrain> but yeah, if/when we make presets, we can redo AI settings etc too
23:41:36 <TrueBrain> what is nice about people being able to upload presets, that we can see a bit what is popular :D
23:41:45 <TrueBrain> gives a bit of telemetry at least :)
23:42:36 <TrueBrain> right, time to get some sleep
23:42:38 <Beerbelott> Alright, thx for your help (:
23:48:19 <LordAro> #5147 contains some interesting ideas
23:48:29 <LordAro> still haven't found what i've got in my head though
23:48:35 *** arikover has left #openttd (ERC (IRC client for Emacs 27.1))
continue to next day ⏵
