IRC logs for #openttd on OFTC at 2020-03-13
⏴ go to previous day
00:05:19 <Samu> i installed the new microsoft edge
00:17:14 <Samu> this is funny, now I can't reproduce the bug
00:30:03 *** WormnestAndroid has quit IRC
00:31:02 *** WormnestAndroid has joined #openttd
01:44:56 *** WormnestAndroid has quit IRC
01:44:57 *** WormnestAndroid has joined #openttd
04:22:16 *** Decapita1 has joined #openttd
04:23:48 *** snail_UES_ has joined #openttd
04:53:06 *** WormnestAndroid has quit IRC
04:53:31 *** WormnestAndroid has joined #openttd
05:13:51 *** WormnestAndroid has quit IRC
05:14:38 *** WormnestAndroid has joined #openttd
07:30:29 *** D-HUND is now known as debdog
07:46:20 *** sla_ro|master has joined #openttd
08:17:47 *** andythenorth has joined #openttd
08:25:41 *** agentsmith101 has joined #openttd
10:05:34 *** gelignite has joined #openttd
10:25:47 <peter1138> THere are, sadly, cake slices and Gregg's doughnuts in the office as it's somebody's birthday.
10:26:10 <peter1138> I nearly succumbed to a doughnut at 9:15 but... reality came back.
10:26:11 <andythenorth> sadly / not sadly /s
10:26:24 <andythenorth> I will mostly be doing SEO today
10:26:25 <peter1138> I'm still struggling with snacking :(
10:31:28 <peter1138> We've got some really cheap laptops in that don't have caps-lock keys.
10:31:32 <peter1138> Apparently there are quite a few users who use caps-lock to enter individual caps, as they don't know about the shift key...
10:56:19 <Samu> so today I'm going to try reproduce the bug of yesterday
10:56:31 <Samu> so far it's been inconsistent
10:56:45 <Samu> sometimes it happens, sometimes it's behaving correclty
11:07:10 *** Arveen2 has joined #openttd
11:07:20 <Samu> damn, the thing is really inconsistent, I can't find a patern that reliably reproduces it
11:17:48 *** HerzogDeXtEr has joined #openttd
11:24:38 *** arikover has joined #openttd
12:09:59 *** andythenorth has joined #openttd
12:25:00 <Samu> ok, i got somewhere, i found the bug comes from the neighbour tiles
12:25:31 <Samu> the map size is 256x256, yet that line produces this: 0x00007ff73efbaa28 {-64, 1, 64, -1}
12:25:53 <Samu> it should be {-256, 1, 256, -1}
12:34:51 <Samu> what is static doing? seems that static is the error
12:35:31 <Samu> tried removing static, and i can't have it trigger the bug
12:39:04 <Samu> seems static was the problem
12:39:17 <Samu> can't trigger the bug anymore
12:39:35 <Samu> but it's been inconsistent, sometimes it happens, sometimes it doesn't, i need to make sure
12:40:13 *** WormnestAndroid has quit IRC
12:40:26 *** WormnestAndroid has joined #openttd
12:50:26 <dwfreed> static within function scope means that the variable is defined once at the first invocation of the function, and the value of the variable persists between calls
12:53:30 <DorpsGek_III_> [OpenTTD/OpenTTD] SamuXarick opened pull request #8036: Fix: [AI/GS] CanBuildConnectedRoadPartsHere neighbours tiles were at times incorrect https://git.io/Jv6dB
13:08:37 *** tokai|noir has joined #openttd
13:08:37 *** ChanServ sets mode: +v tokai|noir
13:10:57 *** gelignite has joined #openttd
13:41:41 *** agentsmith101 has joined #openttd
13:43:43 *** andythenorth has joined #openttd
13:45:18 *** agentsmith101 has joined #openttd
13:51:03 *** agentsmith101 has joined #openttd
13:52:01 *** agentsmith101 has joined #openttd
14:08:53 *** Lejving has joined #openttd
14:20:32 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 approved pull request #8036: Fix: [AI/GS] CanBuildConnectedRoadPartsHere neighbours tiles were at times incorrect https://git.io/Jv6b2
14:23:53 <DorpsGek_III_> [OpenTTD/OpenTTD] ldpl commented on pull request #8036: Fix: [AI/GS] CanBuildConnectedRoadPartsHere neighbours tiles were at times incorrect https://git.io/Jv6bX
14:28:35 *** snail_UES_ has joined #openttd
14:29:25 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 merged pull request #8028: Codechange: [AzurePipelines] Update MacOS image to 10.14 https://git.io/JvguX
14:29:30 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 commented on pull request #8028: Codechange: [AzurePipelines] Update MacOS image to 10.14 https://git.io/Jv6bH
14:54:14 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro merged pull request #8036: Fix: [AI/GS] CanBuildConnectedRoadPartsHere neighbours tiles were at times incorrect https://git.io/Jv6dB
14:56:59 <Samu> it really affects how the routes are built
14:57:13 <Samu> there's decisions based on it with pathfinder
14:57:29 <Samu> it's gonna change many other AIs as well
14:57:49 <Samu> hopefully for the better
14:58:20 <LordAro> it would only be broken if you changed save (to a different map size) while using the same instance of OTTD
14:58:28 <LordAro> i can't imagine many people would be doing that
14:58:54 <Samu> how many Ais use the road pathfinder library?
14:59:05 <Samu> all those will be affected
14:59:21 <LordAro> yes, if you play multiple saves with the same OTTD
15:00:44 <Wolf01> Like when I reported a bug in the try..catch construct of the framework we use at work, inside the same function it worked fine, if you use the try..catch in the caller and raised the exception in the called, in the caller you won't get the exception number (always 0). They won't fix it because "it always worked like that and now is too late to fix it"...
15:01:55 <LordAro> "it's not a bug, it 's a feature"
15:02:06 <LordAro> always annoying when that happens
15:02:21 <LordAro> though if there's anything relying on that behaviour...
15:05:28 <Samu> seems to be that case here
15:05:41 <Samu> if the other Ais expect the old behaviour, then it's trouble
15:09:13 <Wolf01> They even made some of the functions work the way they needed and not the way they expected, an example is the collection loop, if you flag as "deleted" one object in the collection then you won't find it in the loop, that's because is how the UI works, they needed to add a "show deleted" property to the collection just to be able to use it in the code
15:09:42 <andythenorth> this is how real software works
15:09:49 <andythenorth> as made by people
15:09:51 <andythenorth> not book authors
15:10:16 <Wolf01> Yes, but people may have other needs
15:10:53 <Wolf01> And if you go "out of standard" because you need it that way, then is toothache after a while
15:11:02 <andythenorth> then they will have regrets
15:11:05 <andythenorth> regrets are common
15:11:14 <LordAro> computers were a mistake
15:11:18 <andythenorth> there is a common patter
15:11:28 <andythenorth> software -> regret -> more software
15:11:42 <Wolf01> Last time I asked for support they answered "you already know these things better than us"
15:11:54 <andythenorth> Wolf01: what are you doing then if you can't work?
15:11:57 <Samu> wondering how the rail got this, probably also bugged?
15:12:28 <Wolf01> andythenorth: I'm on vacation, then netflix and some games, a lot of sleep too
15:13:00 <Wolf01> I don't know if I'm ready for that
15:13:58 <andythenorth> non-track station tiles on corners :P
15:14:50 <Wolf01> It just needs to be ported to master
15:15:10 <Wolf01> BTW, monday I'll have a videocall for a consultation for the base structure of our next software, so I might start to do some homeworks
15:15:31 <Samu> rail doesn't have an equivalent function, just checked :o
15:15:41 <andythenorth> 'page title too short'
15:15:56 <Wolf01> (multi-tenant or single-tenant+catalog application)
15:16:33 <Samu> checked marine, also doesn't have an equivalent function
15:20:22 <Samu> there are 2 more statics in script_road.cpp, but these don't seem to be trouble
15:30:51 <Samu> here's an example how it affects
15:31:03 <Samu> same route, different paths
15:38:10 <Samu> gonna check other AIs than mine, how do they do now
15:46:43 *** Wormnest_ has joined #openttd
15:52:39 <Samu> it's strange, it made no difference for AroAI
15:57:52 <Samu> maybe the road pathfinder doesn't use the function?
16:04:04 <Samu> ah, it uses it incorrectly
16:04:23 <Samu> AIRoad.CanBuildConnectedRoadPartsHere(cur_node, path.GetParent().GetTile(), next_tile)
16:04:33 <Samu> if the result is -1, it's still true
16:12:48 <Samu> looks like i found another bug
16:15:22 <Samu> it still gets the correct value of -1
16:21:36 <Samu> it's converting from uint to int
16:21:45 <Samu> and it seems to be doing it correctly
16:26:48 <Samu> very few AIs use this function correctly
16:27:07 <Samu> the majority accepts -1 as true
16:30:32 <Samu> Pathfinder.Road-3 is for some weird reason is testing with start tile being the same as the end tile
16:31:05 <Samu> this returns -1, which ends up as true
16:32:30 <Samu> many many AIs use Pathfinder.Road-3 / Pathfinder.Road-4 libraries
16:32:48 <Samu> the library itself is bugged
16:36:10 <Samu> ah, i see why the pathfinder gives start tile the same as the end tile
16:36:29 <Samu> it's looking at all neighours of the current tile
16:37:01 <Samu> one of the neighours is the same as the previous tile
16:38:02 <Samu> CanBuildConnectedRoadsHere returns -1, in which it's accepted as true... no real harm here, but whatever...
16:38:57 <Samu> it's an invalid neighbour, shouldn't be taken into consideration
16:39:22 <Samu> it's a matter of a road vehicle walk backwards
16:39:38 <Samu> it doesn't do that, so the neighbour shouldn't be considered
17:10:01 *** daspork has joined #openttd
17:22:52 <Samu> the library is just accepting every tile as valid
17:45:52 *** sla_ro|master has joined #openttd
17:52:11 *** gelignite has joined #openttd
18:40:52 <andythenorth> I should make an RV set
18:42:26 *** frosch123 has joined #openttd
18:49:10 <andythenorth> trucks in that video so nice :)
18:52:05 <frosch123> is it advantageous when the trailer is longer than the engine?
18:52:07 <LordAro> that's actually a really nice set of guidelines
19:02:44 <andythenorth> based on planetmaker's 1.9 I believe
19:06:58 *** WormnestAndroid has quit IRC
19:07:09 *** WormnestAndroid has joined #openttd
19:23:38 *** Progman has joined #openttd
19:49:36 *** ChanServ sets mode: +v tokai
19:52:54 <andythenorth> so if you were a reddit user
19:53:07 <andythenorth> and someone said 'get Iron Horse from content downloader'
19:53:15 <andythenorth> which grf would you download?
19:55:36 <frosch123> do other games also have addons for addons?
19:56:28 <glx> I guess openttd manages dependencies :)
19:59:44 <andythenorth> same if you want FIRS
20:00:37 *** tokai|noir has joined #openttd
20:00:37 *** ChanServ sets mode: +v tokai|noir
20:02:31 <nielsm> nice, the danish borders are closed from tomorrow and one month forward
20:02:42 * andythenorth wonders about all this
20:02:52 <TrueBrain> I wonder if they will put up walls again :P
20:03:13 <frosch123> i also joined homeoffice club :)
20:03:29 <andythenorth> curious if UK will manage to hold the current line
20:03:35 <andythenorth> or if Sunday newspapers will force a change
20:03:57 <nielsm> we're all gonna get infected anyway might as well make it quick
20:04:11 <TrueBrain> working from home is really boring, I noticed ...
20:04:19 <TrueBrain> nielsm: euh .. no, please lets not do that ;)
20:04:19 <andythenorth> there are all kinds of guides for it
20:04:35 <andythenorth> Friday has been a bonza day for content marketing around covid
20:04:46 <andythenorth> twitter had all kinds of 'guide to x' from bloggers
20:04:57 <andythenorth> I had to stop reading
20:05:04 <andythenorth> but then that applies to twitter every day also
20:09:32 <TrueBrain> we got it shared too :P
20:09:53 <TrueBrain> it is a very nice article .. I liked the "chart 7" the most :D
20:10:07 <TrueBrain> just never realised that
20:10:18 <andythenorth> that's the one I mentioned yesterday
20:10:22 <andythenorth> it went round a lot of the UK
20:10:31 <andythenorth> people will die anyway
20:10:42 <andythenorth> the guy is a professional speech writer / blogger
20:11:05 <andythenorth> not a virologist, epidemiologist, public health professional, or even elected politician
20:11:17 <andythenorth> put charts in anything, people believe it's science
20:11:34 <andythenorth> but he's done well on ambulance chasing
20:12:39 <frosch123> well, the key message is, slow the spreading to reduce the peek impact
20:12:46 <frosch123> which sounds plausible
20:13:12 <andythenorth> he has no skin in the game
20:13:17 <andythenorth> the real issue is winter 2020
20:13:34 <andythenorth> a certain number of people with underlying health conditions are going to die anywya
20:13:39 <andythenorth> whether we panic or not
20:14:18 <frosch123> are you trying to predict the influenca of the next season?
20:14:30 * andythenorth tends to agree with the UK's public health approach, which seems to be an outlier
20:15:52 <frosch123> what does uk health staff say?
20:16:05 <andythenorth> the UK appears to be modelling that this is unstoppable fundamentally, sans vaccine
20:16:15 <andythenorth> it will keep recurring until there's population immunity
20:16:26 <andythenorth> there's a concept in epidemiology around r
20:16:37 <andythenorth> can't remember the details, we used it for a study on viral games / films
20:16:49 <andythenorth> and what goes linear or exponential
20:17:07 <frosch123> yes, it levels at around 70%
20:17:12 <frosch123> noone is disputing that
20:17:31 <frosch123> but it still makes a difference how fast you arrive at the 70%
20:17:50 <andythenorth> the healthcare system is overwhelmed at much lower numbers
20:18:28 <andythenorth> based on what doctors I know report, I'm guessing maybe even single digits overwhelms current healthcare
20:19:12 <andythenorth> so the UK approach appears to be, save the social distancing for the height of the epidemic
20:19:26 <andythenorth> instead of confining everyone indoors when ~no-one has it
20:19:45 <andythenorth> population won't comply with 10-14 weeks of quarantine, then again from September through March 2021
20:19:58 <andythenorth> and we don't have the law enforcement capacity to enforce it
20:21:14 <andythenorth> anyway, it will be interesting to live inside an experiment
20:21:21 <frosch123> ok, so the uk idea is to reach a defined spread speed, and it considers current speed too slow
20:21:55 <frosch123> andythenorth: you are already living in that other experiment for 3 years
20:23:35 <andythenorth> I think UK is just realistic about how much public will comply
20:23:38 <TrueBrain> lol .. I rather be part of a system that at least tries to do something about everyone dying, than a system that says: fuck it, they are dead already :P
20:24:15 <andythenorth> TrueBrain: based on social media here (until I closed the browser windows), a lot of UK public agrees
20:24:37 <andythenorth> so if Sunday newspapers create a lot of fuss, scientific view will probably be pushed aside
20:24:49 <andythenorth> (agrees with you)
20:25:35 <TrueBrain> I am happy I can do my part by working from home :) And happy I have a boss and manager that makes that possible too :)
20:25:52 <TrueBrain> (especially given my job is visiting clients :P)
20:26:01 <andythenorth> we might stop visiting clients
20:26:15 <andythenorth> I think it's overkill, but realistically they will start cancelling soon
20:26:19 <TrueBrain> it is currently up to our clients .. I am guessing that will change soon too :)
20:26:49 <andythenorth> wife didn't go to Amsterdam, everything closed :P
20:27:34 *** Xaroth6 has joined #openttd
20:27:37 <andythenorth> apparently European flights are very cheap right now though
20:28:14 <TrueBrain> yeah .. would be a surprise if airliners survive :)
20:28:30 <andythenorth> that hotel in Tenerife that got quarantined, it's good, I've stayed next to it
20:28:35 <andythenorth> probably very cheap right now
20:29:22 <frosch123> TrueBrain: politics always saves big companies on the first occassion
20:29:38 <andythenorth> not sure it will for airlines
20:29:39 <frosch123> and for unknown reasons airlines count as big companies
20:30:01 <andythenorth> on the plus side, it gives us delay of 6 months carbon output
20:30:34 <milek7> there they bail out national airline every few years
20:30:57 <andythenorth> frosch123: so is germany locked down?
20:31:05 <andythenorth> 'locked down' such a broad term :P
20:31:12 <andythenorth> so widely used last 2 days
20:31:26 <frosch123> andythenorth: it's a federal country, every state does what they want
20:31:38 <frosch123> one announced to close schools next week for 4 weeks
20:31:44 <frosch123> that caused others to follow
20:31:55 <andythenorth> it becomes a race to look active
20:32:04 <andythenorth> closing schools kills grandparents apparently
20:32:13 <andythenorth> not sure if the evidence supports that, but it's an assertion made
20:32:48 <andythenorth> kids are an asympomatic vector, parents have to work, grandparents do childcare, grandparents die
20:32:50 <frosch123> my company canceled most events 2 weeks ago, and now they told everyone using public transport to go to home office
20:33:03 <frosch123> which leaves people not using public transport upset :p
20:33:18 <frosch123> but not sure what stops them claiming they would use public transport
20:33:41 <andythenorth> it's funny how big a deal home working is still
20:33:51 <andythenorth> Twitter was full of smug people like me how are set up for home working
20:34:12 <andythenorth> then both my broadband and 4G went slow while I was updating a production site
20:35:01 <andythenorth> also we had to assess our vendors :(
20:35:09 <andythenorth> because we can remotely work, but can they?
20:35:13 <andythenorth> what if a rack goes down?
20:35:22 <frosch123> andythenorth: anyway, public opinion tends toward: all neighbouring countries have harsher measures, why not we
20:35:24 <andythenorth> the apocalypse will come with a lot of paperwork
20:35:56 <andythenorth> frosch123: in the UK, I think we are used to voting for lower social well being in favour of some small number of people getting richer
20:36:03 <andythenorth> we've been doing it for century or more
20:36:15 <andythenorth> it's a tradition
20:37:39 <andythenorth> 'what is your SLA response time in the event of societal collapse'
20:39:12 <andythenorth> 'please describe how do you notify customers if you become aware of civilisational threat'
20:39:31 <milek7> I doubt closing borders have much sense now
20:39:42 <milek7> at this point it just spreads locally
20:39:43 <andythenorth> 'do you give service credits in the event that the contract cannot be delivered due to apocalypse'
20:39:56 <andythenorth> milek7: +1, but it appeals to people eh
20:40:14 <andythenorth> much easier to blame it on external actors, simple political gambit
20:40:25 <andythenorth> it's Wolf01 and all the other italians!
20:41:28 <andythenorth> 'does your data center include resilience arrangements for the following: fire, flood, earthquake, loss of power, emergent pandemic'
20:41:33 * andythenorth could do this all day
20:42:16 <andythenorth> 'we noted that all the staff named on the tender are now dead, please advise us which staff you are nominating as replacements, and their qualifications, experience, and vetting procedure'
20:42:30 <andythenorth> TrueBrain: ^ tell me your life doesn't contain these questions :P
20:43:17 <TrueBrain> its funny how we laughed about a scenario like this 6 months ago ;)
20:44:03 <Wolf01> <andythenorth> it's Wolf01 and all the other italians! <- yup, I don't know if news outside Italy reports how many people gets fined or even arrested for breaking the rules (or just ignoring common sense at all)
20:45:34 <Wolf01> They found a lot of old people at pubs playing cards, youngs organizing rave parties on abandoned buildings because they can't go to disco pubs, people drunk-driving at night
20:45:56 <Wolf01> At least 20-30 of these at day
20:47:37 <TrueBrain> not surprising, but yeah .. like .. come on ..
20:47:39 <Wolf01> And not to mention the riots in prisons
20:48:27 <andythenorth> this is all policing problems eh
20:48:46 <frosch123> as long as the number of corona "experts" grows faster than the virus, we are fine :p
20:49:02 <andythenorth> it is a growth area for blogging
20:49:11 <andythenorth> partly because the official sites are very very cautious
20:49:26 <andythenorth> some publish base statistics and very little else
20:50:03 <frosch123> probably because it involves a lot of paperwork
20:51:01 <frosch123> hr at my company was also funny. they send an email at 10am, they they will write an email about how to procede later. that email came at 17:30
20:51:14 <frosch123> so, 7:30 hours of paper work on what to actually communicate
20:52:32 <andythenorth> I have days like that
21:26:07 <andythenorth> Wolf01: so 11 prisoners died in riots?
21:33:41 *** sla_ro|master has joined #openttd
21:38:03 *** ChanServ sets mode: +v tokai
22:01:17 * LordAro is going to Rotterdam this weekend
22:03:15 <frosch123> yeah, there is a certain thrill whether you are allowed to go home afterwards
22:04:34 *** Progman has joined #openttd
22:05:03 <andythenorth> totally forgotten my own codebase :P
22:05:06 <andythenorth> I only know horses
22:05:48 <LordAro> lumpy milk at the bottom of this cup of tea
22:05:53 <LordAro> guess that's gone off then
22:06:03 <andythenorth> probably good for you
22:06:09 * andythenorth drinks a lot of kefir
22:07:14 <andythenorth> I feel I've seen it twice today
22:08:03 <andythenorth> extend base vehicles with classes
22:08:10 <andythenorth> or reimplement them as newgrf
22:10:41 <spnda> Any setting in nmlc where it just goes through an entire file and checks for each single error? Maybe like nmlc --verbose?
22:11:21 <glx> so it stops at first error
22:31:29 *** Progman has joined #openttd
22:44:55 *** tokai|noir has joined #openttd
22:44:55 *** ChanServ sets mode: +v tokai|noir
23:34:21 *** andythenorth has left #openttd
23:38:50 *** ChanServ sets mode: +v tokai
23:47:39 <Samu> i dont have an account on that website
23:49:08 <Samu> AIRoad.CanBuildConnectedRoadPartsHere(cur_node, path.GetParent().GetTile(), next_tile)
23:49:13 <Samu> AIRoad.CanBuildConnectedRoadPartsHere(cur_node, path.GetParent().GetTile(), next_tile) > 0
23:50:04 <Samu> because it can return -1
23:50:39 <Samu> -1 is true but should be false
23:57:11 *** tokai|noir has joined #openttd
23:57:11 *** ChanServ sets mode: +v tokai|noir
continue to next day ⏵
![https://i.imgur.com/iEHJ25F.png[/img]](https://i.imgur.com/iEHJ25F.png%5B/img%5D){kind=link}
![https://i.imgur.com/mZDXtbz.png[/img]](https://i.imgur.com/mZDXtbz.png%5B/img%5D){kind=link}