IRC logs for #openttd on OFTC at 2020-01-16
⏴ go to previous day
00:10:31 *** ToBeFree has joined #openttd
00:10:54 *** ToBeFree is now known as Guest13857
00:11:04 *** snail_UES_ has joined #openttd
00:34:18 *** HerzogDeXtEr1 has joined #openttd
00:38:50 *** ToBeCloud has joined #openttd
00:39:13 *** ToBeCloud is now known as Guest13858
00:39:18 *** HerzogDeXtEr2 has joined #openttd
00:40:10 *** Wormnest_ has joined #openttd
00:41:20 *** Guest13858 is now known as ToBeFree
01:23:04 *** ToBeFree has joined #openttd
01:41:29 *** hythlodaeus has joined #openttd
02:07:41 *** snail_UES_ has joined #openttd
02:11:59 <DorpsGek_III_> [OpenTTD/OpenTTD] floodious commented on issue #7830: Load font from openttd config file directory, not from working directory. https://git.io/JewMG
02:26:35 <DorpsGek_III_> [OpenTTD/OpenTTD] floodious commented on issue #7830: Load font from openttd config file directory, not from working directory. https://git.io/JewMG
04:08:03 *** Wormnest_ has joined #openttd
04:13:04 <DorpsGek_III_> [OpenTTD/OpenTTD] James103 commented on pull request #7938: Feature: Setting for a year that repeats forever https://git.io/JvUSs
05:11:13 <DorpsGek_III_> [OpenTTD/OpenTTD] Eddi-z commented on pull request #7938: Feature: Setting for a year that repeats forever https://git.io/JvU9i
06:41:57 *** snail_UES_ has joined #openttd
07:41:29 *** crazystacy has joined #openttd
08:29:20 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh commented on pull request #7938: Feature: Setting for a year that repeats forever https://git.io/JvU7Q
08:41:06 *** andythenorth has joined #openttd
08:51:39 *** Smedles has joined #openttd
09:16:22 <andythenorth> what haps Pikka?
09:16:38 <Pikka> is hand-drawing planes a good idea?
09:19:25 <andythenorth> hand drawing ships is...slow
09:21:42 <andythenorth> are ships like planes? o_O
09:32:03 * andythenorth renders all the ships
09:32:10 <andythenorth> no but maybe I should :P
09:54:21 <Pikka> two down, three to go for the basic views of one plane :) then just... 15 to go. and all the other variant sprites...
10:23:00 *** crazystacy2 has joined #openttd
11:20:25 *** innocenat_ has joined #openttd
11:50:48 *** ToBeFree has joined #openttd
12:11:04 <_dp_> tf is sprite sorter doing
12:31:54 <DorpsGek_III_> [OpenTTD/OpenTTD] Eddi-z commented on pull request #7938: Feature: Setting for a year that repeats forever https://git.io/JvUAa
12:35:21 <Eddi|zuHause> _dp_: lesson learned from 20+ years of TTD modding: never question the sprite sorter. if you fix one thing, it breaks in another place
12:36:32 <_dp_> idk, replacing that crap with std::sort pretty much solves 4k :p
12:38:00 <Eddi|zuHause> what about all the corner cases?
12:41:04 <_dp_> Eddi|zuHause, what corner cases?
12:41:35 <Eddi|zuHause> overlapping bounding boxes and stuff?
12:42:14 <_dp_> how's that a corner case, that's only thing it does xD
12:42:29 <_dp_> sorting overlapping stuff
12:45:57 <Eddi|zuHause> yes, but in the translation from 3D-space (bounding boxes) to 3D-space (sprites) there are sometimes cycllic dependencies that cannot be properly resolved
12:46:30 <Eddi|zuHause> seoncd 3D shoudl be 2D
12:46:44 *** WormnestAndroid has quit IRC
12:46:57 *** WormnestAndroid has joined #openttd
12:46:57 <_dp_> Eddi|zuHause, yet it does somehow order them :p
12:47:12 <_dp_> only problem is to define an ordering properly
12:47:42 <Eddi|zuHause> there is no way to order it properly.
12:48:02 <Eddi|zuHause> every order you choose will cause some situation to be resolved the wrong way
12:48:26 <Eddi|zuHause> if you choose a different way, some different situation will be resolved the wrong way
12:50:16 <_dp_> Eddi|zuHause, that means current sorter is broken as well so even more reasons to replace it with something faster :p
12:51:10 <Eddi|zuHause> _dp_: what i meant to say was: you need to make sure the new "faster" method is broken in exactly the same way
12:51:33 <Eddi|zuHause> it must be "bug-for-bug" compatible :p
12:52:55 <_dp_> well, that just means "proper" ordering is the one that current sorter generates xD
12:53:42 <_dp_> should be solvable with at least quadratic algorithm
12:55:58 <_dp_> looking at this sorter though...
12:56:05 <_dp_> does this damn thing even sort? xD
12:57:38 <_dp_> seems bugged even for stuff that's unambiguously ordered
12:58:06 <Eddi|zuHause> if you don't know what it does, how do you know it's cubic?
13:01:47 <_dp_> Eddi|zuHause, coz it lags like crazy :p
13:04:56 <_dp_> also algo itself is clearly cubic, it's just that sprites are mostly unordered and somewhat pre-sorted
13:06:07 <milek7_> there was optimized version somehwere on github
13:08:41 <_dp_> milek7_, is that the one that was reverted? ;)
13:09:40 <milek7_> yes, but reportedly that second commit fixes it
13:10:42 <_dp_> idk, if it uses qsort why not just qsort it all the way
13:10:50 <_dp_> any bug-for-bug thing is gone already anyway
13:21:02 <_dp_> if I understand it correctly problem is that unordered sprites aren't really unordered
13:21:23 <_dp_> otherwise just a center mass qsort already does everything that sprite sorted does
13:23:40 <milek7_> i don't get almost any performance improvement with it anyway
13:24:13 <milek7_> both original and modified cause ~100ms lag when unzooming
13:26:27 <_dp_> even more if sum is precalced
13:27:49 <_dp_> it still lags a bit but it's 150ms vs 1s
13:28:14 <_dp_> and those 150 aren't even in sorter
13:31:13 <milek7_> resolution too low for it to matter?
13:31:42 <_dp_> milek7_, what res are you testing on?
13:31:52 <_dp_> it's not a issue below 4k afaik
13:32:28 <_dp_> yeah, probably too low. you can try removing sorter completely to see what's lagging
13:32:51 *** andythenorth has joined #openttd
13:51:04 *** hythlodaeus has joined #openttd
13:57:52 *** snail_UES_ has joined #openttd
14:39:24 <Samu> whole map screenshot shouldn't exist
14:39:35 <Samu> now im waiting an hour for it to finish
14:48:54 <Samu> anyway, i need to solve the lifetime PR
14:48:59 *** andythenorth has joined #openttd
14:49:10 <andythenorth> why are infrastructure costs?
14:49:13 <Samu> profit too big don't fit the window
14:49:18 <andythenorth> or should I stay out of problems forum? :P
14:52:20 <Eddi|zuHause> it's not like you can just turn them off
15:00:23 <andythenorth> too much fly swallowing
15:00:36 <andythenorth> 'forbid 90 degrees'
15:00:38 <Samu> there's sort by profit this year, last year, and lifetime. it would be odd not to display lifetime profit
15:00:40 <andythenorth> or just don't build them?
15:00:58 <andythenorth> 'airport noise levels' or just don't build lots?
15:01:46 * andythenorth isn't grumpy, just tired :P
15:02:01 <andythenorth> Eddi|zuHause did I explain that I've solved cdist pax problem? o_O
15:02:14 <Pikka> samu, does it still truncate if you stretch the window out? I guess people who want to see the number can just make the window larger? :)
15:03:37 <Samu> it shows the whole thing
15:04:30 <Pikka> andythenorth: FWIW, my recent GRFs all have a "halve costs" parameter which just sets the running and purchase cost factors 1 lower. simples for sandbox play.
15:04:31 <Samu> i dont think JGR displays lifetime profit on vehicle lists
15:04:40 <Samu> it only got sort by lifetime profit
15:04:45 <andythenorth> reddit says it does? Or I misunderstood
15:04:51 <andythenorth> anything you want, it's in JGR
15:04:56 <andythenorth> unrelated, it's interesting watching a community switch platform
15:05:03 <andythenorth> reddit is now pretty much 100% JGR
15:05:42 <Samu> i need to install JGR to see what it does
15:05:57 <andythenorth> the added features are quite remarkable
15:06:12 <andythenorth> I find it overwhelming, and my general idea for core OpenTTD is generally 'less'
15:06:33 <Pikka> +1 to removing rivers and canals then? ;)
15:06:39 <andythenorth> there's some gradual loading thing for trains where they drive slowly through stations
15:06:48 <andythenorth> I don't think it got shunting though
15:06:58 <andythenorth> Pikka: +1 to rivers and canals are broken
15:07:08 <andythenorth> I think I diverge on solutions :P
15:07:20 <Pikka> honestly I just can't be bothered drawing them for my landscape set :P
15:09:29 <andythenorth> 5 out 146 trains lost money last year in my game
15:09:44 <andythenorth> but they are all on a route where I have an escape depot at the loading station
15:09:56 <andythenorth> and the problem is I have 4 more trains than I need
15:10:00 * andythenorth deletes 4 trains
15:11:04 <Samu> testing JGR, he's cheating
15:12:02 <Samu> the window is large by default thx to that extra cargo list stuff
15:12:53 <andythenorth> cargo list filter eh
15:13:39 <andythenorth> I did start using JGR to see what would be worth adopting
15:13:53 <andythenorth> but I couldn't see the wood for the trees, and found gameplay very confusing initially
15:20:15 <SpComb> I think that could be said for anyone trying vanilla OpenTTD for the first time :/
15:21:01 <andythenorth> familiar things are familiar
15:21:01 <hythlodaeus> i'm relatively new to openttd and i did not find it all that confusing
15:21:25 <hythlodaeus> the game could use a real manual tho
15:21:51 <SpComb> I suppose the fact that a bunch of useful features are hidden behind ctrl-clicks doesn't really hurt beginners, they just won't use things like shared ordered
15:22:05 <andythenorth> wiki hythlodaeus? o_O
15:22:15 <Pikka> a UI overhaul is well overdue
15:22:21 <hythlodaeus> the wiki is frankly poorly organised
15:22:35 <Pikka> every feature that's ever been added has been stuck in sideways to an interface designed for a much simpler game in 1994
15:22:41 <SpComb> BTW the concept of distinct lines in TpF2 would definitely be worth considering, with vehicles assigned to lines
15:22:46 <andythenorth> I don't like wikis, but I went through our wiki assuming it was crap and needed fixing
15:22:55 <andythenorth> and in some places it was easy to spot what was wrong
15:23:01 <andythenorth> but lots of it kind of 'eh ok'
15:23:08 <SpComb> although please keep the "skip order" thing so that you can control which station individual vehicles are going to
15:23:09 <hythlodaeus> I do not think wikis are replacement for manuals
15:23:21 <andythenorth> you prefer the product of a single mind?
15:23:57 * andythenorth waits for JGRPP to compile :|
15:24:02 <hythlodaeus> no, just something that's organised in a coherent way and doesn't mix gameplay tips and pseudo tutorials with feature listing
15:24:08 <SpComb> although with the more dynamic pathfinding in OpentTD, it would probably be difficult to get the same kind of line overlay as TpF2 has
15:25:03 <SpComb> which is, I guess, the most useful feature for that concept
15:25:18 <hythlodaeus> as for the UI, I do not find it so bad. Some things are indeed poorly sandwiched here and there, but i have seen much much worse
15:25:25 <SpComb> you can zoom in on a piece of track and see what lines are using that track
15:25:28 <hythlodaeus> I like the UI in fact.
15:25:39 <andythenorth> ok first problem in JGR
15:25:42 <andythenorth> I have to make plans?
15:25:55 <SpComb> I haven't run into any of that
15:25:57 <milek7_> automatically creating group when creating shared orders
15:26:06 <andythenorth> I also need to configure my zones
15:26:45 <hythlodaeus> sounds like you're playing a mod that's the fruit of specific individual tastes
15:26:52 <hythlodaeus> does it come with proper documentation?
15:27:11 <milek7_> static platform allocation in tf2 annoyed me
15:27:29 <Eddi|zuHause> documentation? what fantasyworld do you live in?
15:27:32 <FLHerne> andythenorth: Plans are just a way to doodle on the map
15:28:01 <hythlodaeus> Eddi|zuHause: documentation is sort of my thing
15:28:07 <andythenorth> so how do I build the plan?
15:28:12 * hythlodaeus wants to write manual for openttd in the future
15:28:19 <andythenorth> where's the button for 'build this plan'?
15:28:20 <FLHerne> andythenorth: Click 'New', click-and-drag on the world, there'll be a squiggly line on the world
15:28:34 <FLHerne> Under 'Plan list' in the map menu
15:28:38 <Eddi|zuHause> hythlodaeus: there was a guy who wrote a german manual once. he sold about 3 copies...
15:28:44 <hythlodaeus> milek7_: except when they do
15:28:46 <SpComb> yeah, the routing/pathfinding in TpF2 is way too static, being able to better balance trains across multiple platforms would definitely be an improvement
15:29:13 <andythenorth> plans are not on the minimap?
15:29:31 <hythlodaeus> no, openttd needs an ingame manual that functions a bit more like help topics on office software
15:29:39 <Eddi|zuHause> SpComb, milek7_: yeah, that was a problem since train fever. it means you need one platform per line, and makes stations bigger than they need to be
15:30:06 <hythlodaeus> although a little train mascot would be cool
15:30:11 <andythenorth> first thing I did with plans was write my name
15:30:16 <andythenorth> then try and see it on minimap
15:30:16 <Eddi|zuHause> "looks like you're trying to build a train line. can i help you with that?" :p
15:30:18 <FLHerne> andythenorth: Functionally, they're sort of a long sign?
15:30:19 <hythlodaeus> "it seems your train has crashed. woul you like to help with that?"
15:30:26 <andythenorth> what's the % button in vehicle orders?
15:30:31 <andythenorth> it doesn't do anything when I click it
15:30:45 <andythenorth> well it toggles up / down, to be precise
15:30:48 <andythenorth> but nothing else happens
15:31:13 <Eddi|zuHause> hythlodaeus: iirc there is a tutorial scenario/game script, maybe start with that?
15:31:22 <hythlodaeus> but essentially an in game manual with a search function would help a lot
15:31:38 <hythlodaeus> i've seen the tutorial scenario, and i did not like it
15:31:41 <andythenorth> 'load if available' would fix cargodist
15:31:44 <FLHerne> andythenorth: It toggles a column
15:31:50 <Eddi|zuHause> hythlodaeus: and who would keep that ingame manual up to date?
15:31:54 <FLHerne> If there are any actual orders
15:31:57 *** supermop_work has joined #openttd
15:32:09 <andythenorth> 'load if available' / 'load by cargo type' /s
15:32:13 <FLHerne> It shows you what the average load was for each order
15:32:27 <andythenorth> if cargodist loads the wrong cargo, the link appears to then never be destroyed
15:32:39 <andythenorth> always using 'load by cargo type' would solve that
15:32:44 <Eddi|zuHause> andythenorth: in TF2 you can set loading/unloading per cargo type
15:33:02 <Eddi|zuHause> "only load steel at this station"
15:33:08 <hythlodaeus> Eddi|zuHause: the same people who keep the game up to date
15:33:12 <andythenorth> oh but I have to set all the things I don't want to load?
15:33:18 <andythenorth> hythlodaeus: we can't even get a blog post approved
15:33:23 <FLHerne> andythenorth: There's a "set all" at the bottom
15:33:25 <andythenorth> there is no way we'd do a manual
15:33:27 <Eddi|zuHause> hythlodaeus: sorry, i had to actually laugh out loud on that thought :p
15:33:35 <andythenorth> FLHerne: yeah thanks
15:33:58 <FLHerne> I should do jgrpp again
15:34:22 <FLHerne> It does solve a lot of the things I find irritating about trunk still
15:35:02 <hythlodaeus> my point is, the game is made by volunteers, so you get some volunteers to work on the manual too
15:35:11 <andythenorth> they do that in the wiki
15:35:13 <hythlodaeus> why would it be any different from code and graphics
15:35:30 <andythenorth> to be clear, I quite liked the idea of having official docs
15:35:34 <andythenorth> but having done it
15:35:35 <hythlodaeus> great, so they can do the same to a manual
15:35:40 <andythenorth> it's pretty frustrating
15:36:00 <andythenorth> nobody wants to do it, but everybody wants to nitpick it when others do
15:36:04 <andythenorth> and nobody is in charge
15:36:17 <hythlodaeus> i'm saying i'd like to do it
15:36:26 <andythenorth> and only 1 person will have an opinion on infosec issues, but 10 will have an opinion on comma placement
15:36:47 <hythlodaeus> that's why you agree on a style guide beforehand
15:36:49 <Eddi|zuHause> hythlodaeus: we have a wiki
15:37:00 <FLHerne> It's not a very good wiki
15:37:07 <andythenorth> I just fixed the wiki, no bikeshedding, no PRs, no nitpick
15:37:15 <andythenorth> I just deleted and rearranged the entire development section
15:37:27 <andythenorth> I didn't fix all the wiki :P
15:37:30 <andythenorth> just development
15:37:38 <hythlodaeus> guys, first you tell me "no one wants to do, it's frustating"
15:37:58 <hythlodaeus> when i say I want to do it, you're telling me we have a wiki so don't do it
15:38:12 <Eddi|zuHause> that's not what we said
15:38:20 <hythlodaeus> seems like a bit of a vicious cycle
15:38:38 <andythenorth> if you try it, you will find it very frustrating
15:38:45 <andythenorth> and you will likely rage quit
15:39:02 <andythenorth> as this is 100% predictable in advance, I think it's unfair to not warn you
15:39:22 <andythenorth> I came close to rage quitting over 1 blog post
15:39:29 <hythlodaeus> because of writing itself or other people?
15:39:40 <andythenorth> other people of course
15:39:48 <hythlodaeus> oh don't worry about it
15:40:04 <andythenorth> the project applies a software code review methodology to content
15:40:25 <andythenorth> nobody is prepared to fix it as there's no-one in charge
15:40:39 <andythenorth> but those with the power are comfortable with the software code review methdology applied to all
15:40:46 <hythlodaeus> the last post was written by m
15:40:49 <FLHerne> hythlodaeus: I think the message is "do it, but on the wiki" ?
15:40:54 <andythenorth> the only applicable route is to vote with your feet
15:41:16 <FLHerne> hythlodaeus: I mean, the wiki already has pages describing most of the features in some way
15:42:08 <FLHerne> So a manual would effectively be an improved version of most of the wiki's existing purpose
15:42:18 <andythenorth> FLHerne: do you know how to do the 'trains of any length can load in stations' thing in JGR?
15:42:23 <andythenorth> where they advance as they load?
15:42:35 <FLHerne> andythenorth: No, that's a thing?
15:42:42 <andythenorth> it was in a video in forums
15:42:49 <andythenorth> JGR needs it's own sub-forum
15:43:02 <FLHerne> andythenorth: I do the opposite, and use the programmable routing restrictions to keep my long trains out of short platforms
15:43:02 <andythenorth> it's pretty much 100% of OpenTTD discussion
15:43:16 <FLHerne> (also, make bay platforms work sanely without a million waypoints)
15:43:19 <andythenorth> and the JGR thread is unworkable
15:44:06 <andythenorth> the cdist cargo over-rides in settings eh?
15:45:19 <SpComb> that tooltip PR is a little unfortunate indeed, agreeing on the design/style beforehand would be a better idea IMO than trying to split up the commits, it would be a massive pain to rebase those kinds of changes
15:45:58 <Eddi|zuHause> andythenorth: i think it's in the clickthrough for "stop at start/middle/end of station"
15:46:05 <hythlodaeus> it's impossible to split up the commits
15:46:14 <andythenorth> thanks Eddi|zuHause :)
15:46:18 <hythlodaeus> and we've agreed on a style before
15:46:28 <andythenorth> ha ha the 'through loading' is so funny to watch
15:46:56 <andythenorth> the % load option in vehicle orders absolutely doesn't work though
15:47:02 <Eddi|zuHause> andythenorth: it's another case for "the order GUI is way too overloaded"
15:47:04 <andythenorth> it now says 50% but I have not set that
15:47:10 <hythlodaeus> the rest is just re-writing, which cannot simply be discussedo n every single entry
15:47:16 <andythenorth> it magically changes
15:47:31 <andythenorth> it's supposed to control the amount of cargo refitted to?
15:47:38 <andythenorth> I can refit 50% of vehicles to coal?
15:48:36 <andythenorth> oof JGR won't thank me for bug reports, he gets enough :x
15:48:44 <andythenorth> but this just doesn't work
15:48:49 <Eddi|zuHause> i've not seen this feature
15:49:16 <andythenorth> it's so you can control how the % of the consist refitted to
15:49:27 <andythenorth> so 50% coal, 50% iron ore
15:49:29 <SpComb> hythlodaeus: yeah, I agree that the "split it up into commits so that we can review/decide on each one separately" is a bad idea, the commits would all conflict with eachother if you tried to change them
15:49:31 <andythenorth> but it doesn't work
15:49:54 <andythenorth> the % should be in the 'refit to' menu, it's just broken like this
15:50:04 <Eddi|zuHause> andythenorth: i can't imagine what that is supposed to mean
15:50:35 <hythlodaeus> SpComb: I'll find a way to work it out, but thank you for your support!
15:50:55 <hythlodaeus> if you can drop a line with your opinion on the commit thread, I'd like to read it :)
15:51:17 <hythlodaeus> there's still some stuff I am currently rewriting and some more lines to fix here and there
15:51:25 <FLHerne> andythenorth: I don't understand
15:51:35 <FLHerne> You're looking at the percentages on the right?
15:51:42 <andythenorth> yes, for amount to refit to
15:51:46 <hythlodaeus> but so far I am quite happy with the tooltips I'm working on
15:52:04 <andythenorth> so if I want 50% coal, 25% iron ore, 25% limestone
15:52:45 <FLHerne> andythenorth: That's not what that does
15:53:11 <FLHerne> andythenorth: It displays the average load for that journey segment over the last few things
15:53:29 <FLHerne> Unless I'm still confused about what you're looking at
15:53:39 <andythenorth> in the screenshot
15:53:44 <andythenorth> it shows 100% loading order
15:53:47 <andythenorth> 0% loading order
15:53:53 <andythenorth> and then some 50% I never asked for?
15:53:57 <andythenorth> and I can't type in the box
15:54:19 <FLHerne> The '100% 0% [50%]' column?
15:54:50 <andythenorth> yeah I understand now
15:55:43 <FLHerne> The button shows the overall average, and you can toggle per-segment averages
15:56:32 <andythenorth> it's kind of weird
15:56:55 <FLHerne> andythenorth: There are some more conditional-order types that help with managing weird CDist stuff
15:56:58 <Eddi|zuHause> this window needs to be split into 3 at least: orders, schedule, statistics
15:57:02 <andythenorth> obviously ^^ that's broken, but the concept is fine
15:57:39 <andythenorth> so where's the departure boards?
15:58:28 <Eddi|zuHause> i've not looked at the departure boards since ChillPP
15:58:43 <Eddi|zuHause> which is 5-ish years ago?
15:58:59 <andythenorth> I'm looking at timetables now
15:59:10 <andythenorth> Timetabling full load orders is not recommended?
16:00:24 <FLHerne> Well, what's the point?
16:01:04 * andythenorth tries timetables
16:01:07 <FLHerne> If the vehicle ever has to wait for a full load longer than the timetable, the timetable did nothing and now it's late
16:01:12 <andythenorth> so a train takes 10 days to run a route
16:01:25 <andythenorth> I set it to 20, but the train didn't travel at half the speed?
16:01:37 <FLHerne> If it doesn't, you might as well leave out the 'full load' bit
16:01:55 <andythenorth> yeah I got rid of the full load
16:02:00 <andythenorth> but the train is still taking 10 days to travel
16:02:02 <FLHerne> No, it'll just arrive at the next place earlier
16:02:20 <Eddi|zuHause> if you allow combining full load with timetabling, you immediately run into conflict between the people that want it to leave at the earlier or the later time
16:02:36 <FLHerne> Timetabling never affects the way trains move, just how long they wait in stations
16:02:41 <Eddi|zuHause> andythenorth: you can configure a max speed separately
16:02:51 <FLHerne> I'm not sure why the dwell time can be set separately, tbh
16:03:09 <andythenorth> oh I found the speed limit
16:03:17 <FLHerne> Well, it'll wait once you've set a start date
16:03:28 <FLHerne> Speed limit is in upstream OTTD now, I think?
16:03:43 <andythenorth> I don't understand
16:03:46 <andythenorth> I never use timetables
16:03:48 <FLHerne> ChillPP used to have speed-limiting signals
16:03:55 <FLHerne> I don't know if jgr has them
16:03:59 <andythenorth> I just set the 'wait at station' time for pax and mail trains and ships
16:05:53 <andythenorth> which is the point of it no?
16:06:01 <Eddi|zuHause> setting up train timetables is a pain, which is why almost nobody uses them
16:06:04 <andythenorth> we won't add % full load, because the UI is overloaded
16:06:09 <andythenorth> so it's all solved in JGR
16:06:25 <FLHerne> andythenorth: If you set all the times, and set a start date, you can control when vehicles get anywhere and how long they wait
16:06:27 <andythenorth> and the main reason people are switching to JGR is for this
16:06:34 <FLHerne> andythenorth: I don't think JGR does anything to timetables
16:06:41 <FLHerne> What are you talking about?
16:06:51 <FLHerne> Oh, no, the automate thing
16:07:21 <Eddi|zuHause> autoseparation is horrible if you ever want to deviate from it. that's why it never made it to trunk
16:07:30 <andythenorth> it's literally impossible to play OpenTTD without JGR timetables
16:07:47 <andythenorth> I thought we had auto-separation in trunk, but only for trains?
16:07:57 <andythenorth> there's some magical ctrl-click a button to make timetables work?
16:08:04 <andythenorth> I tried it for ships, it doesn't work
16:08:18 <FLHerne> andythenorth: There's auto-start-date, but it's basically useless
16:08:28 <Eddi|zuHause> there's nothing "auto" abut that ctrl+click
16:08:51 <FLHerne> (because you have to get the timetable worked out in advance, and then it breaks as soon as you add any more trains or there are significant delays)
16:09:05 <Eddi|zuHause> andythenorth: it needs a fully timetabled order list to work
16:09:13 <FLHerne> Eddi|zuHause: It automatically calculates evenly-spread start dates, no?
16:09:22 <FLHerne> Which is marginally more auto than before
16:09:37 <FLHerne> But still doesn't make it worth doing
16:09:44 <Eddi|zuHause> FLHerne: pretty much
16:10:57 <FLHerne> andythenorth: Hm, this 'scheduled dispatch' thing is new to me
16:11:44 <FLHerne> andythenorth: The 'automate' and 'auto separation' things dynamically modify the timetable so it matches running times but avoids bunching
16:12:38 <FLHerne> There always used to be a problem that if some major delay/jam happened, the delay got automated into the timetable, and then it would remain a self-fulfilling prophecy
16:13:05 <FLHerne> (because the trains would wait, and block trains on other lines, and then everything cascaded)
16:13:23 <FLHerne> Maybe it's been improved since
16:19:12 <andythenorth> does it break when vehicles upgrade?
16:19:49 <FLHerne> Not that I noticed, anyway
16:19:54 <andythenorth> what else is in JGR
16:19:57 * andythenorth looks in the repo
16:20:10 <FLHerne> andythenorth: Ship collision avoidance is nice
16:20:21 <andythenorth> I think I tried the patch for that
16:20:21 <FLHerne> And multi-track level crossings that don't kill all your buses
16:20:43 <Eddi|zuHause> i think i wrote a patch for that
16:21:14 <andythenorth> I already tried the polylines track tool last time I played JGR
16:21:16 <FLHerne> I think Eddi did write the patch for that :P
16:21:34 <FLHerne> There must be /some/ way to solve the savegame-upgrade issue?
16:22:06 <FLHerne> (temporary third state?)
16:22:40 <Eddi|zuHause> FLHerne: it just needs a saved setting, which i didn't include back then because updating it would have been a pain
16:22:48 <Eddi|zuHause> FLHerne: should be easier with the new system
16:23:28 <Eddi|zuHause> FLHerne: so old games would just load with the setting turned off, and it would "only" crash all your vehicles if you enabled it mid-game
16:25:48 <andythenorth> building rivers!
16:34:16 <andythenorth> how do I custom bridge head?
16:35:24 <Eddi|zuHause> build a track onto it?
16:35:48 <andythenorth> I was trying to build the bridge onto track
16:37:07 <hythlodaeus> in regards to the previous topic: my idea is that timetabling is primarily meant to be used on passenger vehicles, especially buses (and now trams too)
16:37:20 <andythenorth> more town growth speeds!
16:37:33 <andythenorth> custom bridge heads?
16:38:12 <FLHerne> hythlodaeus: The problem is that timetabling as it actually exists is clearly not "meant" to do anything sane :P
16:38:24 <andythenorth> I thought it was for use with BROS?
16:38:30 <andythenorth> for those realistic UK maps etc?
16:38:38 <andythenorth> seriously there are people in reddit doing that
16:38:53 <hythlodaeus> i don't see how one could harmonize it will full load orders either
16:39:10 <hythlodaeus> but I can imagine it working very well on bus and a tram routes
16:39:21 *** sla_ro|master has joined #openttd
16:39:25 <FLHerne> It seems very much to have been designed and implemented without assessing the user-experience for any use case
16:39:25 <hythlodaeus> because it can prevent traffic jams
16:39:46 <FLHerne> hythlodaeus: The version in JGR handles that alright
16:39:47 <hythlodaeus> agreed, but how could we improve on that?
16:39:57 <FLHerne> Because of the automate/auto-separation thing
16:40:09 <andythenorth> it's for model trains
16:40:12 <hythlodaeus> what does it do, if I may ask?
16:40:17 <andythenorth> but it's sold and used for vehicle separation
16:40:33 <FLHerne> Upstream doesn't have that, so it needs impractical levels of micro-managing to get timetables set up usefully and then keep them updated
16:41:38 <FLHerne> hythlodaeus: In upstream, you manually set the time for each journey segment and station (there's an 'Autofill' tool, but it's useless because it leaves no margin for delays)
16:42:18 <FLHerne> (especially if used when the route is first set up, because then there's no waiting cargo and it doesn't account for loading times either)
16:42:30 <FLHerne> Then you set the start times for each vehicle
16:43:16 <FLHerne> There's a hidden ctrl-click that sets them all at once for a set of shared-orders vehicles with <overall time>/<number of vehicles> spacing
16:43:51 <FLHerne> If anything changes later (there are delays, or a new stop, or more vehicles) you have to manually redo most of the above
16:44:30 <FLHerne> JGR has a thing that automatically updates the timetable to match actual running times
16:44:35 <hythlodaeus> yeah I usually just trash vehicles and do it all over, when i need to readjust
16:44:45 <FLHerne> And, optionally, manipulates it to maintain even vehicle spacing
16:44:47 <hythlodaeus> it's easier to just do one vehicle and clone all over
16:44:59 <FLHerne> When it works, it's great
16:45:09 <hythlodaeus> when does it not work?
16:45:51 <FLHerne> It can get stuck in feedback loops, particularly when multiple routes interact
16:46:12 <hythlodaeus> how do you think it can be improved?
16:46:41 <FLHerne> Some delays happen, the timetable gets updated to allow for them, vehicles wait to keep the service spread out but block other vehicles
16:47:14 <FLHerne> ...which get delayed, so /their/ timetable gets lengthened, and they wait and block the first ones
16:47:36 <FLHerne> I'm not really sure :P
16:47:42 <hythlodaeus> i see, but still why not suggesting an improved version of this to upstream?
16:47:59 <hythlodaeus> certainly someone will be able to come up with a solution
16:48:01 <FLHerne> tbh, I think it works about as well as an automated non-super-intelligent thing could do
16:48:13 <andythenorth> they block, not like the reddit post
16:48:20 <FLHerne> hythlodaeus: I mean, that patch has been around for about a decade
16:48:43 <FLHerne> No-one actually /has/ come up with a solution that fixes the various perceived problems :P
16:48:54 <hythlodaeus> has someone ever tried?
16:49:08 <FLHerne> It's definitely an improvement
16:49:14 <hythlodaeus> also have we had a look to how other games do it?
16:49:35 <hythlodaeus> i suppose transport fever might have timetabling
16:49:42 <FLHerne> Which is a much bigger rework of how timetables as a concept should exist
16:50:51 <hythlodaeus> well it's never to late to re-assess things
16:51:12 <FLHerne> The problem there is that a complete change from how they currently work is confusing
16:51:20 <hythlodaeus> just right now they're working on to implement water depth, which will actually make ship gameplay interesting
16:51:23 <FLHerne> (and might cause compat problems)
16:51:34 <hythlodaeus> everything always does
16:51:46 <FLHerne> And also that it's still very complicated to use
16:51:51 <andythenorth> honestly, I think we'd have cover to throw out timetables :P
16:51:51 <hythlodaeus> and again, water depth will likely break a ton of savegames
16:51:58 <FLHerne> The difference is that the complexity actually achieves something :P
16:52:26 <FLHerne> But unlike the auto-separation hack, it's not a "one-click" solution to vehicle bunching
16:52:37 <FLHerne> Which is probably the main thing people actually want from timetables...
16:53:20 <andythenorth> I think it's back to the goals tbh
16:54:13 <andythenorth> we have people actually using timetables to build model railways, which is a thing people do in model railway world
16:54:17 <FLHerne> In general, TIP is a much better way to achieve what the (few) people already using timetables use them for
16:54:28 <andythenorth> and we have people trying to use timetables for a better game, and failing
16:55:26 <FLHerne> Whereas auto-sep is a hack that solves a commonly-requested basic usecase, but does nothing for the existing users
16:55:38 <andythenorth> simplifying timetables probably ruins life for the people who are creating spreadsheets to configure their timetables
16:56:00 <hythlodaeus> apparently Transport Fever has no timetables at all ahah
16:58:49 <FLHerne> Timetable Improvement Patch
16:58:55 <FLHerne> the thread I linked above
16:59:23 <hythlodaeus> what is the concept behind that one? and what are the shortcomings?
17:00:30 <hythlodaeus> hmm it uses route nodes
17:00:38 <hythlodaeus> isn't it sort of what waypoints are for?
17:15:27 <nielsm> argh the baseset translation building is so terribly slow in the windows cscript implementation
17:19:15 <nielsm> (will the cmake build fix/improve that?)
17:23:45 <DorpsGek_III_> [OpenTTD/OpenTTD] lcd047 opened pull request #7942: Fix: [SDL2] support pasting from clipboard on Linux https://git.io/JvTUV
17:25:53 <Samu> jesus, visual studio is really making me wanna buy an SSD
17:26:12 <Samu> they no longer care optimizing software for HDDs
17:27:45 <Samu> finally it cancelled build
17:27:52 <Samu> only took 10 minutes ffs
17:29:08 <supermop_work> people still buy HDDs?
17:56:28 <peter1138> nielsm, that may have been my fault? :( I think there's an md5sum in there, heh.
17:58:53 <supermop_work> andythenorth: i use timetables for 'better game'
17:58:57 <supermop_work> seems to work fine
17:59:10 <supermop_work> ofc could be less tedious
17:59:15 *** Progman has joined #openttd
18:00:00 *** HerzogDeXtEr has joined #openttd
18:00:18 <supermop_work> 24 clock helped, now i generally just schedule everything at like 30 day period, try to have services sharing a line 10 or 15 days apart
18:00:43 <supermop_work> its generally fine and not too hard for simple things
18:00:52 <supermop_work> like a tram running in a loop in a city
18:01:25 <andythenorth> what do I click?
18:01:29 <supermop_work> autofill timetable, add a bit a slack to get a round number
18:01:42 <andythenorth> I never use them, i just use the 'wait at station' tool
18:01:47 <supermop_work> the buy that many trams and ctrl click start date
18:02:13 <supermop_work> works fine, seems better than any attempt at 'autoseparation'
18:03:16 <supermop_work> "oh it takes tram 1 27 days to complete its run, pad it out to 30 days, buy two more trams"
18:04:10 <supermop_work> next time tram 1 completes the first item in orders, ctrl click set start date
18:04:28 <supermop_work> set to the expected date, then start the new trams
18:12:17 <_dp_> surprisingly seems correct though
18:24:12 <andythenorth> deep and shallow oil rigs :P
18:24:23 <andythenorth> deep sea ones have a very slow animation cycle, they bob up and down :P
18:24:27 <andythenorth> also ships that roll and pitch
18:25:26 <nielsm> deep and shallow fishing grouns producing different cargo
18:26:14 <nielsm> freshwater fishing grounds that only exist on deeplakes at high elevation
18:27:24 <hythlodaeus> tiles look fine, but hard to tell the difference without an ingame screenshot
18:28:30 <nielsm> deep water? I consider the current efforts at making tiles for it very much experimental stage
18:34:34 <hythlodaeus> but good work so far
18:34:49 <supermop_work> bridges that have max depth in addition to length
18:34:54 <hythlodaeus> this will be one hell of an update for ships, that's for sure
18:35:06 <Samu> bridges max height and max depth
18:35:09 <nielsm> well it's not approved yet :P
18:35:34 <hythlodaeus> people seem excited about it, so it's good
18:36:01 <nielsm> eugh I can't find a free cb variable to put water depth info in for industries
18:36:35 <Samu> question: how many digits is considered too much money?
18:37:01 <Samu> I put 10 for resizing purposes, but I don't know openttd approach to that
18:37:13 <supermop_work> various long span types of bridges could be set to have a higher basic cost than those with shorter span, but be better / cheaper at tall heights or deep water
18:37:25 <Samu> like finances window, how do you resize money displays?
18:38:26 <hythlodaeus> also as this will affect the positioning of ports, I really hope they get revised as structures. I'd personally would like port building to be a bit more like train stations, with differing platform lengths and such
18:39:09 <supermop_work> so building a wood trestle across a shallow valley is cheaper than spanning it with a suspension bridge, but a trestle over a deep fjord is prohibitively expensive
18:39:11 <nielsm> Samu: if you're okay with rounded values you can use {CURRENCY_SHORT} to "shorten" the values to millions/billions/etc instead
18:39:55 <supermop_work> but the suspension bridge is reasonable
18:40:29 <supermop_work> as it doesn't care how far down the dirt is under the main span
18:40:30 <Samu> lifetime profit displaying CURRENCY_SHORT, hmm
18:41:43 <supermop_work> but not over the open ocean
18:42:21 <nielsm> supermop_work: first would be adding bridges with longer spans than 3 tiles
18:42:31 <nielsm> (I think that's the max right now?)
18:43:11 <supermop_work> well its more a limit of the period of repeating sprites for bridges
18:43:15 <nielsm> and having some kind of dynamic way to select span lengths, either for the player or automatic cost-minimising
18:43:25 <supermop_work> there are only a few patterns hardcoded to use
18:43:36 <supermop_work> better newgrf bride support would help
18:44:58 <supermop_work> you can make a bridge with a huge 'clear span' now, but it is only in the form of : [A-B-B-B-B...-B-A]
18:46:03 <supermop_work> and the others are like A-B-C-B-A
18:47:13 <supermop_work> or A-B-C-B-D-B-C-B-A
18:47:31 <supermop_work> SC4 was interesting for building bridges
18:53:06 <andythenorth> we need a decent ship set
18:58:07 <Samu> Money CompanyFinancesWindow::max_money = INT32_MAX;
19:00:31 <DorpsGek> Samu: Error: There's really no reason why you should have underscores or brackets in your mathematical expression. Please remove them.
19:02:16 <nielsm> it's 2 billion and something
19:02:21 <Samu> #define INT32_MAX 2147483647i32
19:02:53 <Samu> heh, what I had, but with a different code
19:03:00 <nielsm> it's very easy to get more than that also
19:03:47 <Samu> resize->height = GetVehicleListHeight(this->vli.vtype, 1);
19:03:47 <Samu> for (uint i = 0; i < 4; i++) SetDParamMaxValue(i, INT32_MAX);
19:04:03 <Samu> *size = maxdim(*size, GetStringBoundingBox(STR_VEHICLE_LIST_PROFIT_THIS_YEAR_LAST_YEAR_LIFETIME));
19:04:45 <Samu> but i guess this is more elegant
19:08:36 <nielsm> huh also why are grf bridges different from most other grf features, it does not use action 1/2 to indicate graphics
19:12:37 <Eddi|zuHause> nielsm: because everybody was too scared to specify/implement it
19:13:03 <Samu> what are the chances a single vehicle can make 10 digits profits this year + last year + lifetime?
19:14:02 <nielsm> aaaa no this won't work, I have to make it a 60+x variable to be able to query tiles at other locations
19:19:16 <Eddi|zuHause> nielsm: if we ever attempt action1/2/3 for bridges, it must be able to have custom layouts, not just the builtin tile configurations
19:20:14 <nielsm> and support making tiles below impassable (occupied by pylons)
19:21:58 <Eddi|zuHause> ... and possibly wider bridges (i.e. connect to neighbouring bridge)
19:25:04 <Eddi|zuHause> i always thought bridges should be built more like stations, so you could have some buttons for suspension bridge span and total length
19:25:17 *** andythenorth has joined #openttd
19:26:38 <Eddi|zuHause> the GRF part could then inherit much of the object spec
19:31:48 <Samu> I'm not too good at gui resizes
19:32:12 <Samu> tell me if I broke the display with your choice of fonts
19:33:30 <Samu> probably should have done max(size->width, bodingbox.width)
19:36:06 <nielsm> i'm so good at run-on sentences
19:37:17 <supermop_work> Eddi|zuHause: would a 2 tile wide bridge be able to have road on one side and rail on the other?
19:38:20 <Eddi|zuHause> supermop_work: since they would still be technically 2 separate bridges, all that needs is the right detection code
19:39:17 <supermop_work> manhattan bridge would be tough to model
19:40:19 <Eddi|zuHause> is that a dual layer bridge?
19:40:27 <Samu> plz test #7919 by opening vehicle list windows
19:40:53 <Samu> tell me if the width is fine, or broken with your choice of fonts
19:41:16 <Samu> the width about displaying tiny text of profits
19:42:46 <Eddi|zuHause> long url is long
19:43:12 <Eddi|zuHause> that's like 3 nested urls?
19:43:29 <Eddi|zuHause> i herd u liek urls, so we put some urls in ur url
19:45:31 <andythenorth> oh I was looking at JGR also eh
19:45:37 <supermop_work> basically 8 tracks and 3 road lanes when opened
19:47:44 <andythenorth> JGR fixes RV overtaking
19:47:52 * andythenorth sees why everybody is switching
19:48:24 <Samu> andythenorth, test my stuff too :p
19:49:11 <Samu> there's now a square between "Aircraft Available" and "Manage List" buttons
19:49:21 <supermop_work> now is 4 tracks and 3 lanes on bottom with 4 lanes on top
19:49:21 <Samu> unclickable, but it looks ugly
19:49:58 <andythenorth> hmm, JGR fixes depot flipping as well, not sure about that
19:50:14 <andythenorth> articulated RV overtaking just works as far as I can tell
19:50:50 <Samu> what if the vehicle is extra-extra-extra-long?
19:51:01 <Samu> with many articulated pieces
19:51:37 <andythenorth> didn't test that
19:51:49 <andythenorth> give JGRPP a test
19:52:33 <andythenorth> I might switch to developing for JGR
19:53:14 <Eddi|zuHause> ... or you could start backporting stuff?
19:53:59 <andythenorth> it's a different newgrf spec
19:54:09 <andythenorth> and player base is switching
19:54:25 * andythenorth trying to work out what programmable signals is for
19:54:34 <andythenorth> no, OpenTTD is in the best shape for years
19:54:44 <andythenorth> and JGRPP is important for that
19:54:57 <andythenorth> it reduces social pressure to incorporate dubious fixes
19:55:12 <andythenorth> and reduces the amount of 'devs are horrid horrid people' stuff in forums
19:55:19 <andythenorth> which is project death
19:55:47 <Eddi|zuHause> still, some things are worth backporting
19:55:53 <andythenorth> I'm tending to agree
19:56:02 <andythenorth> but whether we'll agreee on what those are.... :D
19:56:21 <andythenorth> I can make a signal that is programmable with 'deny'
19:56:29 <andythenorth> but I don't see the point, it just blocks a route
19:57:27 <andythenorth> oh there's supposed to be a conditional before it
19:59:39 <andythenorth> anyone know what the new cdist modes do?
19:59:50 <andythenorth> asymmetric (nearest), asymmetric (equal)
20:00:02 <andythenorth> I didn't find any docs apart from the giant forum thread
20:00:33 <Eddi|zuHause> andythenorth: i think it's for your problem that the distance influence should be high for passengers, but low for cargo
20:00:57 <andythenorth> there is also a cdist setting for every cargo in JGR
20:00:58 <Eddi|zuHause> so "equal" wouldn't take into account distance
20:01:39 <supermop_work> is setting cdist individually per cargo 'fun'?
20:02:17 <andythenorth> only got one train :P
20:02:48 <andythenorth> it's not optimal, for FIRS Steeltown
20:03:12 <andythenorth> they're over-rides on defaults though
20:04:19 <andythenorth> oh wait, there's 2 kinds of programmable signal? :o
20:04:35 <Eddi|zuHause> yeah, there's different kinds
20:04:44 <supermop_work> the grf that defines the cargos should define how CDist-y they are
20:04:57 <Eddi|zuHause> one is "turn X green when Y is red" which only makes sense for block signals, so is useless
20:05:13 <Eddi|zuHause> and one is "only allow trains with speed > blah"
20:05:31 <supermop_work> same what that it essentially sets what the payment rate and decay are
20:06:43 <andythenorth> nah, it's too specific to a map and play style supermop_work :)
20:06:58 *** Hythlodaeus_ has joined #openttd
20:07:08 <supermop_work> then the user should also set all the payment rates per map....
20:08:05 <supermop_work> i mean that the user should still have a general setting for cargo distance demand, but how coal might be different to grain would scale with that
20:08:39 <andythenorth> it would work if demand was economic
20:08:45 <andythenorth> which I would like to try, but eh
20:08:59 <supermop_work> having coal prefer to go far and grain near on map X, but other way around on map Y makes no sense
20:09:04 <andythenorth> oh vanilla can turn off the signal GUI :o
20:09:14 <andythenorth> amy ignorance never ends :)
20:10:23 <supermop_work> maps also shoulding put a power plant in a far corner from coal in the middle of nowhere though
20:12:22 <andythenorth> signals on bridges and tunnels eh
20:14:00 <supermop_work> funnily enough the manhattan bridge only allows one train per track at a time
20:14:08 <supermop_work> as it bounces too much otherwise
20:19:11 <andythenorth> TEMPLATE BASED TRAIN REPLACEMENT :o
20:19:18 <andythenorth> oh, but how do I use it :(
20:19:56 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh commented on pull request #7942: Fix: [SDL2] support pasting from clipboard on Linux https://git.io/JvTqm
20:20:19 <andythenorth> there are buttons, but I can make the window small enough to fit on my screen
20:20:40 <andythenorth> so template-based train replacement is only for users with > 13" screens
20:21:05 <nielsm> iirc when TBTR was submitted as a PR for master it also had big trouble with UI scaling here
20:21:14 <andythenorth> well I can't test it :P
20:21:18 <andythenorth> the window doesn't scale
20:21:39 <andythenorth> I need to manage slots in vehicle groups
20:21:46 <andythenorth> but I don't know why
20:23:52 <Eddi|zuHause> supermop_work: it would make sense to have the grf be able to set a cargodist category, instead of the way too generic "passengers, express, everything else" that we have now
20:25:16 <andythenorth> I suspect players know better than authors
20:25:17 <supermop_work> Eddi|zuHause: exactly
20:26:11 <Eddi|zuHause> supermop_work: there are two problems with that: 1) which categories should there be, and 2) what to do with older grfs?
20:26:46 <supermop_work> old grfs, all of the 'cargo' cargo is in one category, same as now
20:27:01 <andythenorth> show train length in vehicle window?
20:27:03 <andythenorth> seems quite nice
20:27:06 <andythenorth> I often want to know that
20:27:14 <andythenorth> I have to send train to depot to find out
20:27:16 <supermop_work> can't you just look at the train?
20:27:54 <andythenorth> to figure the length?
20:27:57 <andythenorth> surprisingly not
20:27:57 <Eddi|zuHause> i don't think that is a problem i ever came across
20:28:07 <andythenorth> how to build correct length station?
20:28:24 <andythenorth> or passing loops
20:29:24 <DorpsGek_III_> [OpenTTD/OpenTTD] lcd047 updated pull request #7942: Fix: [SDL2] support pasting from clipboard on Linux https://git.io/JvTUV
20:29:41 <Eddi|zuHause> dunno, maybe i'm too good at remembering what train lengths i built
20:30:33 <andythenorth> lots of additions there
20:30:43 <supermop_work> yeah i've never had this problem andy
20:30:49 <andythenorth> although the 'running' is a bug, those are el engines on non-el track
20:32:11 <andythenorth> 'Group' is an addition
20:32:25 <andythenorth> and all the extra info per engine
20:35:08 <Eddi|zuHause> first impression? i don't like the multi-line entries
20:35:49 <andythenorth> I think they're tied to an improved breakdowns implementation
20:35:54 <andythenorth> I should turn breakdowns on or something
20:40:24 <TrueBrain> that one nightly build I want ... gets stuck on MacOS build for some reason :( BOOOOO
20:40:38 <TrueBrain> cancel + rerun, see if that helps ...
20:41:04 <TrueBrain> now the step that took 36+ minutes took 2 seconds
20:41:06 <TrueBrain> guess that is better :D
20:44:00 <DorpsGek_III_> [OpenTTD/OpenTTD] lcd047 updated pull request #7942: Fix: [SDL2] support pasting from clipboard on Linux https://git.io/JvTUV
20:46:24 <TrueBrain> "xz: (stdin): File format not recognized"
20:47:01 <andythenorth> Eddi|zuHause: that breakdown patch thread escalated quickly :o
20:47:06 <TrueBrain> who doesn't install xz these days :(
20:47:13 <andythenorth> bring back Dalestan
20:47:19 <andythenorth> capitalisation eh
20:47:54 <Eddi|zuHause> andythenorth: i haven't read any threads in ages, no clue what you mean
20:48:26 <Eddi|zuHause> i have no intention of clicking on any links
20:48:48 <DorpsGek_III_> [OpenTTD/OpenTTD] lcd047 commented on pull request #7942: Fix: [SDL2] support pasting from clipboard on Linux https://git.io/JvTmf
20:48:48 <DorpsGek_III_> [OpenTTD/OpenTTD] lcd047 closed pull request #7942: Fix: [SDL2] support pasting from clipboard on Linux https://git.io/JvTUV
20:48:56 <TrueBrain> HE WHO CANNOT BE PHISHED
20:49:23 <TrueBrain> the security dude in me applauds you Eddi|zuHause :)
20:49:29 <TrueBrain> and lol @ PR ... that is ... a new one :P
20:55:23 *** frosch123 has joined #openttd
20:56:07 *** Thedarkb has joined #openttd
20:57:05 <Eddi|zuHause> bot-induced ragequit?
20:57:20 <Eddi|zuHause> to be fair, the CI still does not propagate error messages properly
20:59:46 <TrueBrain> on that I agree .. someone should port it to GitHub Actions :D
21:03:08 <TrueBrain> if it aint a pull-request .... :P
21:03:43 <TrueBrain> I guess adding the commit checker alone would help a lot
21:03:50 <TrueBrain> you can even look into annotation
21:04:24 <glx> only way to test actions "locally"
21:08:55 <TrueBrain> I push stuff to master in my fork
21:08:57 <TrueBrain> that works well too
21:10:14 * andythenorth donates to forums
21:10:28 <andythenorth> TrueBrain: OpenTTD funds are still 'fine'?
21:10:32 <andythenorth> or AWS ate it all? :D
21:11:01 <TrueBrain> I did not receive any PANIC message from mister o :)
21:11:13 <andythenorth> planetmaker: does coop need any funds?
21:12:16 <frosch123> very likely, it needed funds when it was active
21:13:35 <TrueBrain> I believe andythenorth is in the wrong xmas month or something :P
21:13:54 <andythenorth> I didn't pay my tax bill yet :P
21:14:02 <andythenorth> I have a lot of money in my account until end of month
21:15:07 <frosch123> TrueBrain: maybe he sold all unwanted xmas gifts on ebay
21:15:34 <TrueBrain> andythenorth .. that makes no sense at all :P
21:15:53 <andythenorth> sense of spending is related to money in account NOW
21:16:14 <andythenorth> currently I feel rich and can donate
21:16:28 <andythenorth> Feb I will feel poorer :P
21:16:40 <andythenorth> life is contextual :P
21:16:45 <Eddi|zuHause> just make sure your tax report shows the donation as tax-deductable
21:17:06 <andythenorth> I never bother :P
21:17:22 <Eddi|zuHause> what's the point of donating, then?
21:17:32 <Samu> I have a coding style issue
21:17:43 * andythenorth is not very tax efficient
21:18:47 <TrueBrain> andythenorth: just leave out the word "tax" :P
21:21:50 *** grossing has joined #openttd
21:25:44 <nielsm> 3 if it has to change, but it smells like a wrong approach to whatever it is
21:28:07 <Samu> im trying to deconstruct
21:28:19 <andythenorth> why is google SERPS awful now?
21:28:24 <andythenorth> it's showing favicons and crap
21:29:07 <Samu> trying to have my patch with the least amount of changes
21:29:15 <Samu> when compared with original
21:32:22 <Samu> patching backwards sucks
21:36:58 *** grossing has joined #openttd
21:39:12 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 opened pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTY7
21:39:50 <glx> and of course this PR will escape commit check ;)
21:40:28 <TrueBrain> hmm .. that is fixable .. how did I do that ..
21:40:34 <TrueBrain> basically, it needs to see a workflow once
21:40:38 <TrueBrain> and it will run your PR workflow after that
21:40:46 <TrueBrain> not sure it matters, tbh
21:40:52 <TrueBrain> for other repos, we simply did [Actions] btw
21:40:55 <TrueBrain> not really important
21:41:31 <glx> github prevents running actions from "external" PR
21:41:47 <glx> it's a safety measure for paid plans
21:43:25 <DorpsGek_III_> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTOf
21:43:34 <TrueBrain> silly enough, that is a false statement :)
21:43:43 <TrueBrain> forks cannot use secrets
21:44:05 <TrueBrain> but I noticed last week that it runs the workflow from my PRs, once the master has any workflow
21:44:37 *** grossing has joined #openttd
21:45:07 <TrueBrain> glx: can you run your code in your fork, so we can see it works? :D
21:46:16 <Samu> Money base_cost = IsCanal(tile) ? _price[PR_CLEAR_CANAL] : IsRiver(tile) && _game_mode == GM_NORMAL && !_settings_game.construction.dynamite_river && !_cheats.magic_bulldozer.value ? (Money)0 : _price[PR_CLEAR_WATER];
21:47:02 <glx> ternary should be a last resort choice
21:47:20 <glx> and use parenthesis with imbricated ternarys
21:47:53 <glx> or at least around complex && args
21:48:38 *** Hythlodaeus_ has joined #openttd
21:49:00 <nielsm> ternaries: simple condition (no && or ||), simple values (no chaining or internal ternaries)
21:49:18 <nielsm> if it gets any more complex than that, split it into multiple lines
21:49:32 <nielsm> or see if it might make sense to make a function out of it
21:49:38 <glx> don't try to be smarter than the compiler ;)
21:49:59 <Samu> linux builders complained about the 0
21:50:48 <nielsm> Money{ 0 } would be better
21:51:29 <nielsm> I considered making that a ternary, but figured I may as well make a function of it
21:52:05 <nielsm> it would have ended up as a horribly long chained one otherwise
21:52:15 <Samu> funny, this is the ClearTile_Water function
21:53:37 <Samu> so it's ok to create a function for a DoCommand
21:54:51 <nielsm> simple, static helper functions like that can make code easier to read, and the compiler might often inline them anyway
21:55:46 <nielsm> (also reminder that the "inline" keyword in C++ does not mean the function must or even just should be inlined, it just means it's legal to define the function multiple times and you promise every definition of it across all translations units are identical)
21:56:22 <nielsm> what a clear front page
21:56:42 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 updated pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTY7
21:57:48 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 commented on pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTOD
21:58:19 <TrueBrain> all that is left is to trigger this when a build is done :)
21:58:41 <TrueBrain> glx: how ironic, that your example failed :P
21:58:42 <glx> weird it ran fine before the rebase and forced push
21:59:04 <TrueBrain> glx: checkout@v2 changed some rules
21:59:14 <TrueBrain> (it checks out a lot less code)
21:59:14 <frosch123> how does the nightly work? does the nightly trigger a rebuild of the whole jekyll?
21:59:50 <TrueBrain> glx: yes. Most likely because @v2 changed stuff :) It seems origin/master doesn't exist ;)
22:00:06 <TrueBrain> it never fetches 'master' in @v2
22:00:26 <glx> ha yes, I'll need the openttd checkout I guess
22:00:51 <TrueBrain> no, you need to checkout 'master' too :)
22:01:09 <TrueBrain> "fetch all branches", but only for master, basically
22:01:17 <TrueBrain> @v1 did a full git clone
22:01:21 <TrueBrain> which is TERRIBLY slow
22:01:28 <TrueBrain> @v2 only picks up what is absolutely needed
22:01:33 <TrueBrain> the rest you have to get yourself
22:01:36 <TrueBrain> like .. the master branch :)
22:02:12 <TrueBrain> "git fetch --no-tags --prune --depth=1 origin +refs/heads/master:refs/remotes/origin/master" or something?
22:03:06 <TrueBrain> @v1 took 40 seconds for OpenTTD .. @v2 only 7
22:03:14 <TrueBrain> just to highlight how much better it got :)
22:03:49 <Samu> the way I have rivers being permanent is hacky
22:04:12 <Samu> it actually demolishes the river at cost 0 and rebuilds it right away
22:04:26 <Samu> with the same random bits
22:05:06 <Samu> saying "you can't do that"
22:05:06 <TrueBrain> frosch123: why did you ask?
22:06:15 <frosch123> i just assumed it would be cron/time triggered
22:06:35 <TrueBrain> nah, it is a huge chain :)
22:06:49 <TrueBrain> otherwise new tags are annoying :D
22:06:55 <LordAro> Samu: yup, that's terrible
22:07:09 <TrueBrain> hello mister awesome
22:07:31 <Samu> it gives the idea nothing is done
22:07:42 <LordAro> Samu: you said it yourself, it's really really hacky
22:07:59 <LordAro> TrueBrain: that's lord awesome to you
22:08:22 <TrueBrain> my apologizes lord sinky
22:08:31 <TrueBrain> sinky? Well, I guess I am going with this now
22:09:27 <LordAro> Samu: ok, why do you think it's hacky?
22:09:59 <Samu> because in the settings, i say that rivers are "indestructible"
22:10:15 <Samu> they really aren't, they're destroyed and rebuilt at cost 0
22:10:33 <LordAro> right, so something that says it is one thing but is actually another
22:10:36 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 updated pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTY7
22:10:41 <LordAro> do you think that's a good thing?
22:10:47 <Samu> gives the effect that it's indestructible, so it's all good
22:11:00 <LordAro> just because it works, doesn't mean it's maintainable, understandable, or just straight up good code
22:11:09 <LordAro> by the sound of it, it's all 3
22:11:37 <LordAro> hacky is fine when you need to demonstrate something to a customer in 3 hours time and you need to get something working
22:11:45 <milek7_> TrueBrain: content-type is text/html
22:11:47 <LordAro> there are no time or space pressures here, so take the time to do it properly
22:12:10 <milek7_> while content looks rather like xhtml
22:12:23 <DorpsGek_III_> [OpenTTD/workflows] TrueBrain opened pull request #1: Add: workflow to automate publishing of documentation https://git.io/JvT3q
22:12:35 <TrueBrain> milek7_: I hope you understand your comment is fully out-of-context to me :)
22:12:37 *** Hythlodaeus2 has joined #openttd
22:12:56 <Samu> I don't know of another way :(
22:13:30 <TrueBrain> glx: it did do the fetch, I guess
22:13:37 <Samu> i don't want to pass an error to the command
22:13:51 <Samu> would break area clearing :(
22:13:52 <TrueBrain> glx: debugging becomes a bit easier if you add "set -x" at the start of the script
22:14:27 <TrueBrain> glx: did you rebase locally btw?
22:14:30 <Samu> i actually want it to skip clearing the tile
22:15:19 <glx> yes locally master and the branch are in sync with openttd master
22:15:38 <TrueBrain> so why does the rebase fail .. odd
22:16:16 <TrueBrain> replace "master" with "${GITHUB_BASE_REF}" btw, before I forget to mention :D
22:16:56 <milek7_> TrueBrain: docs.dev.openttd.org serves xhtml files but Content-Type in header is text/html
22:17:04 <milek7_> it should be application/xhtml+xml
22:18:03 <TrueBrain> nah, you are solving the wrong problem there
22:18:09 <Samu> gonna try return CommandCost(); see what happens, instead of clearing at 0
22:18:17 <TrueBrain> let me convert those pages to html5
22:19:08 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 updated pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTY7
22:19:17 <Samu> * Creates a command cost return with no cost and no error
22:19:28 <DorpsGek_III_> [OpenTTD/workflows] TrueBrain updated pull request #1: Add: workflow to automate publishing of documentation https://git.io/JvT3q
22:22:49 <TrueBrain> right, time to provision this in production
22:23:20 <milek7_> doxygen output is still xhtml
22:24:02 <TrueBrain> ah, you only mention the root, which was xhtml for some silly reason
22:24:24 <glx> hmm trailing whitespace errors during the rebase it seems
22:25:06 <Samu> wow, i can return CommandCost(); just like that!
22:25:43 <TrueBrain> glx: still weird, as it is a clean rebase .. so that should also be an issue locally?
22:26:22 <glx> locally it works if I type the commands
22:29:44 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 updated pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTY7
22:31:02 <TrueBrain> glx: seems the PR checkout is already a merge
22:31:15 <TrueBrain> (..) origin +739695c458412cf2760f17d2d6d8dd687298cec1:refs/remotes/pull/2/merge
22:31:23 <TrueBrain> still, rebase should work tbh
22:32:21 <TrueBrain> "HEAD is now at 739695c Merge d734ced058d4d28781665b29f54eaee3da8a5021 into 3b177af8263df5212bf3ea3bcd048846f704a868"
22:32:25 <TrueBrain> so something is funky
22:33:52 <TrueBrain> the @v1 also did that
22:35:07 <TrueBrain> glx: I would guess it has to do with fetch-depth
22:35:15 <TrueBrain> so I guess you just have to set fetch-depth to 0
22:36:09 <TrueBrain> but it is what it is :)
22:36:18 <TrueBrain> uses: actions/checkout@v2
22:36:38 <glx> I guess default works fine for PRs with only 1 commit
22:37:06 <TrueBrain> fetch-depth by default is 1
22:37:09 <TrueBrain> so possibly it only gets the merge commit
22:37:13 <TrueBrain> and not the commits below it
22:37:15 <TrueBrain> so yeah, that is possible
22:38:40 <TrueBrain> would even make sense, I guess
22:38:45 <glx> oh the checkout HEAD^ scenario looks better
22:39:28 <TrueBrain> that removes the merge commit :P
22:39:51 <TrueBrain> yeah, fetch-depth of 2
22:39:57 <TrueBrain> so that would be the merge commit + 1 more from the PR
22:40:02 <TrueBrain> but this is always an issue for us
22:40:08 <TrueBrain> as you don't know how many commits are in the PR
22:40:20 <TrueBrain> us wanting to rebase to master is the issue here
22:40:29 <TrueBrain> why did we do that again?
22:41:05 <TrueBrain> fetch-depth of 2 and checkout HEAD^ removes the merge-commit. Commit-checker should "just work" after that I guess
22:42:02 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 updated pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTY7
22:43:04 <TrueBrain> and I am waiting for ACM to provision me a certificate
22:45:16 <Samu> bah.. water_cmd.cpp is still a large diff
22:45:34 <Samu> I don't think I can make it shorter without ruining code style
22:45:49 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 updated pull request #7943: Move commit checker from Azure Pipelines to Github Actions https://git.io/JvTY7
22:46:23 <LordAro> Samu: lines of code is not a bad thing
22:46:42 <LordAro> Samu: incidentally, you're aware that commit should really be split up much more
22:46:50 <LordAro> probably into 3 separate PRs, given it's doing 3 different things
22:48:22 <glx> hmm ok HEAD^ is wrong, it's head of master
22:49:04 <LordAro> bit weird that the checkout is already a merge? can't that be "fixed" ?
22:49:37 <Samu> split into more commits hmm
22:49:39 <glx> I guess I can set the rigth ref
22:49:40 <DorpsGek_III_> [OpenTTD/workflows] TrueBrain updated pull request #1: Add: workflow to automate publishing of documentation https://git.io/JvT3q
22:49:49 <Samu> deconstructing and patching backwards, you mean
22:50:52 <Samu> thing 1 and thing 3 can't be separated that easily
22:52:42 <TrueBrain> glx: pretty sure you are on the wrong track there :)
22:52:51 <TrueBrain> basically, GitHub pulls in the pr/merge, which is fine
22:52:58 <TrueBrain> but for some reason, which I cannot remember, we want to rebase to master
22:53:05 <TrueBrain> that is only going to work if you do fetch-depth of 0
22:53:12 <TrueBrain> as you don't know how many commits are in a PR
22:53:16 <TrueBrain> if we don't do the rebase
22:53:38 <TrueBrain> other solution arrise
22:53:46 <TrueBrain> but as long as you want the rebase .. you are a bit out of luck :)
22:54:05 <glx> I don't remember why we rebase :)
22:55:55 <TrueBrain> so in that case do a fetch-depth of 2, do a "checkout HEAD^" (that removes the merge commit)
22:56:22 <TrueBrain> hmm, no, I am not sure which side of the merge it picks
22:56:31 <glx> but doing that it picks master head :)
22:56:38 <TrueBrain> so the left side, makes sense
22:56:50 <TrueBrain> so yeah, checkout@v2, ref: ${{ github.ref }}
22:57:00 <TrueBrain> so avoids the merge I would guess
22:57:25 <TrueBrain> also interesting: my certificate is still not there :(
22:57:42 <glx> but I'm sure there's the correct ref somewhere
22:58:23 <TrueBrain> dump the env in an action and see
22:58:36 <TrueBrain> (maybe not in this branch, as you are making Azure Pipelines going maddddd :P)
22:58:52 <TrueBrain> I work on Actions in 2 branches: 1 to trigger actions in my fork, and one for the PR :)
22:59:14 <glx> I can do the tests on my other branch indeed
23:00:22 <TrueBrain> hmm, the pr/merge is done, to make sure things like workflows are as they are in master
23:00:37 <TrueBrain> so not doing that might be a bit bad :D
23:00:49 <TrueBrain> so I guess we have to change the commit-checker instead :D
23:26:23 *** andythenorth has left #openttd
23:49:33 <Samu> crap, i broke a feature :(
23:50:54 <Samu> stupid me, how do I undo a rebase now :(
23:51:49 <Samu> nevermind, i still have my very old patch file here to see what I had
continue to next day ⏵
![https://i.imgur.com/sWTcRZA.png[/img]](https://i.imgur.com/sWTcRZA.png%5B/img%5D){kind=link}
