IRC logs for #openttd on OFTC at 2020-12-19
⏴ go to previous day
00:06:48 *** nielsm has quit IRC (Ping timeout: 480 seconds)
00:32:50 *** gelignite has quit IRC (Quit: Stay safe! Stay at home! Stop the chain reaction!)
01:44:16 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
01:45:58 *** HerzogDeXtEr has joined #openttd
01:51:49 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
02:24:15 *** tokai|noir has joined #openttd
02:24:15 *** ChanServ sets mode: +v tokai|noir
02:30:50 *** tokai has quit IRC (Ping timeout: 480 seconds)
03:05:52 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
03:06:55 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
03:10:17 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JL0tO
03:14:12 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
03:20:37 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler closed pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
03:20:40 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
03:21:55 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JL0q2
03:26:52 <DorpsGek> [OpenTTD/OpenTTD] btzy commented on pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/JL0mJ
03:37:00 *** Wormnest has quit IRC (Quit: Leaving)
03:40:28 *** debdog has quit IRC (Ping timeout: 480 seconds)
05:37:05 *** Laedek has quit IRC (Read error: Connection reset by peer)
05:53:35 *** gregdek_ has joined #openttd
05:53:45 *** innocenat__ has joined #openttd
05:57:20 *** dvim has quit IRC (Ping timeout: 480 seconds)
05:57:35 *** gregdek has quit IRC (Ping timeout: 480 seconds)
05:57:35 *** innocenat_ has quit IRC (Ping timeout: 480 seconds)
05:57:35 *** gregdek_ is now known as gregdek
05:57:35 *** innocenat__ is now known as innocenat_
07:36:46 *** sla_ro|master has joined #openttd
07:45:20 *** andythenorth has joined #openttd
08:00:20 *** Progman has joined #openttd
08:29:59 * andythenorth looks at JGR's implementation of 'stop expiring'
08:30:37 * andythenorth considering something similar in newgrf
08:30:55 <andythenorth> it will be quite complicated I think to figure out which vehicles should be available in a specific year
08:48:55 *** WormnestAndroid has quit IRC (Remote host closed the connection)
08:49:08 *** WormnestAndroid has joined #openttd
08:50:33 *** jottyfan has joined #openttd
08:56:28 <Eddi|zuHause> isn't "stop expiring" simply "set life to 255"?
09:02:28 <andythenorth> but providing a parameter for the player to choose the year
09:03:23 <andythenorth> opens up the possibility of having 2 generations of vehicle present
09:03:53 <andythenorth> so there needs to be a heuristic to prevent one generation, probably the older one
09:04:40 <Eddi|zuHause> you know exactly what's in the grf
09:14:28 <andythenorth> so the simple rule would to expire the older generation
09:14:44 <andythenorth> even if model life suggests it should still be around
09:15:26 <Eddi|zuHause> that's why "retire early" exists
09:15:52 <andythenorth> does that help here?
09:16:05 <Eddi|zuHause> i use it in CETS exactly for that purpose
09:16:12 <andythenorth> I might be missing something
09:16:37 <Eddi|zuHause> i think you're trying to solve a solved problem
09:17:02 <andythenorth> only Pikka has solved it as far as I know
09:17:06 <andythenorth> or very high daylength values
09:17:19 <andythenorth> but pikka has far fewer vehicles
09:18:29 <andythenorth> the case I forsee is
09:18:50 <andythenorth> choosing to freeze time in 1960 would leave 80 useless vehicles from 1930 still cluttering the buy menu
09:19:05 <andythenorth> (and similar cases when generations transition)
09:20:54 <andythenorth> maybe a var for 'vehicle x is available'
09:20:57 <andythenorth> would solve this
09:21:27 <andythenorth> otherwise there's no way that both cleans up the buy menu AND avoids pathological cases
09:23:05 <andythenorth> ah :D pikka does it by not allowing player to specify precise year
09:23:08 <andythenorth> just a generation
09:24:12 <andythenorth> that would be simpler
09:24:20 <andythenorth> just need to handle vehicles that span generations
09:30:28 * andythenorth wonders if a parameter to just scale intro dates is better
09:31:11 <andythenorth> if normal progression is 30 years...just allow player to put in arbitrary length
09:31:34 * andythenorth looks up what max model life etc is
09:32:16 <andythenorth> and I'd have to scale retire early etc appropriately
09:33:02 <andythenorth> what's a sensible upper level? Daylength 9-10 seems to be well-liked
09:33:10 <andythenorth> so 300 years between new trains?
09:33:44 <Eddi|zuHause> i'm pretty sure 255 is "disable retiring"
09:37:39 <andythenorth> hmm so this strategy wouldn't eliminate daylength
09:37:54 <andythenorth> newgrf spec can't reach high enough
10:04:02 *** Afdal has quit IRC (Ping timeout: 480 seconds)
10:06:17 *** Speedyn has quit IRC (Quit: Lost terminal)
10:09:08 *** Speedy` has joined #openttd
10:28:15 *** HerzogDeXtEr has joined #openttd
10:35:33 <Samu> I don't trust that tiles/day measurement
10:37:03 * andythenorth makes containers
10:37:16 <andythenorth> going to be a lot of declarations :P
10:37:47 <andythenorth> I've already got 33 variants
10:38:03 <andythenorth> each with 1-n random patterns
10:40:22 * andythenorth needs to rethink
10:40:32 <andythenorth> maintaining 100 variants is just going to cause bugs
10:51:04 <andythenorth> oh it's 236 variants :P
11:37:03 <Wolf01> You should write a software to maintain each vehicle
11:45:08 <Wolf01> A simple web page with a form and a nosql db should be enough
11:46:08 <andythenorth> I can just keep it all on an AWS queue
11:46:18 <andythenorth> as long as I remember to copy it to the queue occasionally
11:53:41 *** jalt has quit IRC (Quit: Leaving.)
11:55:10 <jalt> Hi, is there a way to preview NewGRFs from the ingame browser?
11:57:32 <jalt> A good start would be to see what vehicles, if any, were included, and whether it included anything other than vehicles (stations, industries, etc.). Is there any viewer/tool?
12:13:34 *** Timberwolf has quit IRC (Remote host closed the connection)
12:17:14 <Wolf01> Use the website link if any, and hope the author put some images there :P
12:18:33 *** ekangmonyet_ has joined #openttd
12:19:55 *** ekangmonyet has quit IRC (Read error: Connection reset by peer)
12:23:54 <jalt> Alas, that puts me back where I started, using only the "major" NewGRFs mostly by themselves. It's not realistic to start a new game and try to examine what changed for each of the NewGRFs on bananas.
12:26:24 *** ChanServ sets mode: +v tokai
12:27:15 <andythenorth> this is probably why there are 'recommend me newgrf' threads every day
12:28:58 <andythenorth> and it's mostly dead
12:29:50 <andythenorth> what are you looking for?
12:33:12 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
12:48:05 <FLHerne> And some notes on what things are/do
12:48:26 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/JLEuN
12:55:44 * andythenorth won't be using that
12:55:49 <andythenorth> I should sort out blacken
12:56:03 <andythenorth> pep8 formatting of multiline dicts is stupid and offensive
12:56:06 <andythenorth> blacken's is rad
12:57:19 <andythenorth> that's a load of manual formatting I won't be doing now
12:57:27 <TrueBrain> you are aware black applies PEP-8 on your source, right? :)
12:57:49 <andythenorth> well the results are quite different to using autopep8
12:57:59 <TrueBrain> it is PEP-8 nevertheless :P
12:58:10 <andythenorth> ok I retract my earlier comment
12:58:31 <TrueBrain> if we stop caring about the details!
13:01:19 * andythenorth also likes the cake that blacken returns to stdout
13:01:56 <andythenorth> although my shell has black background, and blacken has black text....so I literally only see cake
13:07:55 <TrueBrain> my shell also has a black background, but shows everything fine :)
13:08:17 <andythenorth> I only want cake
13:12:22 <TrueBrain> it seems ..... discussions don't trigger events yet on GitHub
13:13:46 <Wolf01> Mmmh, I only find guides to remap controls for no man's sky in VR mode... but I can't even move the cursor or open the steam overlay, I would like to know if it's actually playable in VR :/
13:33:26 <andythenorth> discussions trigger me!
14:01:45 *** Progman has quit IRC (Remote host closed the connection)
14:11:53 <Samu> multiplication with large numbers can be scary
14:19:39 <Wolf01> It already shipped, it's called mashinki
14:19:56 <Wolf01> You can even drive vehicles manually now
14:21:39 <Wolf01> The only thin I coundn't get int is the different pace, is faster on some aspects and slower in some others compared to OpenTTD
14:27:51 <andythenorth> openttd has a very natural pace
14:50:19 <Samu> I'm trying to create the worst case scenario for MoveGoodsToStation, testing for overflows, if such is possible
14:51:55 <Samu> and... heliports is the station with the most coverage
14:52:07 <Samu> that can be placed in high numbers
14:57:00 <Samu> 1 house can be covered by 80 heliports
15:09:09 <Samu> 80 heliports + 36 docks + some larger airports with big coverage
15:16:26 <Samu> + 4 city airports + 4 international airports
15:17:49 <Samu> and it would be 4 more intercontinentals if it had 1 more tile coverage, but alas, it's not
15:19:27 <Samu> total, 124 stations for 1 tile
15:22:15 <_dp_> irrc newgrfs can increase station catchment to 16 or smth
15:24:25 <Samu> best_sum max value is 255*15
15:25:34 <Samu> company_sum[owner] is 255*(124-14)
15:30:56 *** Timberwolf has joined #openttd
15:31:12 <Samu> hmm can a cargo have a rating of 0?
15:31:26 <Samu> if that is possible, we can have a division by 0
15:41:32 <Samu> it isn't possible, just checked
15:49:02 *** DasPoseidon has joined #openttd
15:49:34 *** DasPoseidon has quit IRC ()
15:49:40 <Samu> question: Is 0xFFFFll ugly?
15:50:23 <Samu> or is (uint64)0xFFFF less ugly?
15:51:20 <Samu> p.second.remainder = 0xFFFFll * (a % b) / b;
15:51:33 <Samu> p.second.remainder = (uint64)0xFFFF * (a % b) / b;
15:58:12 <LordAro> Samu: i'd go ULL, personally
15:59:26 <Samu> you're right, it's unsigned
16:13:18 *** Wormnest has joined #openttd
16:13:20 <andythenorth> black is my new crack
16:13:25 <andythenorth> still running it manually but eh
16:17:15 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:23:06 <Samu> no more possible overflows
16:24:11 <Samu> @calc 0xFFFFFFFFFFFFFFFF
16:24:11 <DorpsGek> Samu: 18446744073709551616
16:37:42 *** andythenorth has quit IRC (Quit: andythenorth)
17:00:36 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JLuWm
17:06:40 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8378: Remove undefined behaviour from save/load code https://git.io/JLuWj
17:11:37 *** jottyfan has joined #openttd
17:22:28 <DorpsGek> [OpenTTD/OpenTTD] btzy updated pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/fhGxd
17:22:54 <DorpsGek> [OpenTTD/OpenTTD] glx22 approved pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JLu4r
17:24:08 <DorpsGek> [OpenTTD/OpenTTD] btzy updated pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/fhGxd
17:24:12 <TrueBrain> don't forget to review the related workflows PR too; otherwise we still don't have nightlies :D
17:24:59 <DorpsGek> [OpenTTD/OpenTTD] btzy commented on pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/JLu4y
17:26:07 <DorpsGek> [OpenTTD/workflows] glx22 approved pull request #10: Add: improve workflows for upcoming OpenTTD release workflow https://git.io/JLu4b
17:26:37 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain closed issue #8309: Docker Hub new policy means nightlies / pull-request-checks will start to fail 1st of November https://git.io/JURqa
17:26:40 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIiDr
17:26:52 <DorpsGek> [OpenTTD/workflows] TrueBrain merged pull request #10: Add: improve workflows for upcoming OpenTTD release workflow https://git.io/JIhRs
17:27:09 <TrueBrain> now we wait for 90 minutes to see if it really works :P
17:27:25 <TrueBrain> or rather, play the game: what did I forget :D
17:30:56 <glx> well first test will be in 15 minutes :)
17:31:29 *** Progman has joined #openttd
17:31:38 <Eddi|zuHause> what? we actually PLAY the game here?
17:32:14 <TrueBrain> what the hell gave you that impression?!
17:34:09 <Eddi|zuHause> sorry, i misread :)
17:34:51 <Eddi|zuHause> well, you said those words. in that order. but it meant something different :p
17:36:56 <Eddi|zuHause> hm, sometimes games need 2 attempts to start. and i don't quite understand why
17:38:50 *** andythenorth has joined #openttd
17:42:43 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain requested changes for pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/JLuEq
17:43:04 <TrueBrain> This made me giggle: Note: This can be optimized so that merely changing the "group by" dropdown value does not need to call GenerateVehicleSortList again.
17:43:13 <TrueBrain> like ... yes ...... and so can many more places ... what is your point? :P
17:43:18 <TrueBrain> silly comments, in my opinion :)
17:45:58 *** jottyfan has joined #openttd
17:46:01 <Samu> im trying to explain line 4083
17:49:02 <Samu> yesterday someone told me "it is an integer numerator in the fraction x/65535 that is closest to (a % b) / b", but I'm still unsure how to translate that into a comment
17:49:15 <Eddi|zuHause> i remember a "TODO: doesn't work -- WHAT doesn't work?!?" :p
17:52:53 <andythenorth> I definitely leave comments like that around in my code TB :P
17:52:59 <andythenorth> usually as an apology to others
17:55:45 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8391: RFC: Feature: Public key company authorization https://git.io/JLuzG
18:01:55 <glx> TrueBrain: first test passed ;)
18:02:17 <TrueBrain> milek7: saw your question last night; I thought, lets answer in public, as it might stimulate others too :D
18:03:41 <glx> but timing seems off by 15 minutes
18:04:29 <TrueBrain> schedule is drifting like a mofo on GHA
18:05:12 <TrueBrain> might need to set eints a bit earlier now .. this was rather close to the deadline :D
18:05:49 <TrueBrain> but so far I have seen a drift from up to 14 minutes with schedules
18:06:04 <TrueBrain> I am very tempted to just use DorpsGek to trigger flows:P
18:08:05 <Samu> /* Convert the remainder to a numerator in the fraction x/65535 that is closest to (a % b) / b */
18:10:17 <TrueBrain> disabled all projects on Azure DevOps :)
18:14:09 <glx> if all schedule have the same drift it should be ok
18:14:25 <TrueBrain> well, eints has to be in 18:NN :)
18:14:32 <TrueBrain> otherwise my CE(S)T code will fail :P
18:14:41 <TrueBrain> so if the 18:45 hits at 19:01 .. it will run :)
18:14:45 <TrueBrain> not the worst thing btw
18:16:12 <glx> yeah to be safe the margin should be in hours, not in minutes
18:16:37 <TrueBrain> well, this one was 16 minutes late
18:18:36 <glx> I guess it depends on how many workflows are set to start at the same time
18:19:45 <TrueBrain> it is something at GitHub at least .. but it is a bit annoying
18:19:56 <TrueBrain> maybe someone could experiment a bit with different times
18:20:00 <TrueBrain> like, is 18:01 working better?
18:27:49 *** HerzogDeXtEr1 has joined #openttd
18:34:03 *** HerzogDeXtEr has quit IRC (Ping timeout: 480 seconds)
18:41:49 <milek7> TrueBrain: well, the plan is
18:41:56 <milek7> the paragraph you just quoted :P
18:42:17 <TrueBrain> that lacks depth :)
18:43:22 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick opened pull request #8398: Change: Distribute left over cargo to stations according to the highest remainder https://git.io/JLuoN
18:43:36 <TrueBrain> can you limit the amount of contributions per user on GitHub? :(
18:50:31 <milek7> TrueBrain: maybe I'll just implement it later, and then have discussion
18:50:40 <milek7> I basically wanted to know if my proposed method is fine, or you have other ideas
18:50:54 <TrueBrain> and that is exactly what I replied: someone needs to draw up where we want to go with this :)
18:51:09 <TrueBrain> it is fine if you don't want to btw; but it would be nice if someone set some dot on the horizon for this
18:51:17 <TrueBrain> very hard to evaluate ideas without it
18:54:52 <TrueBrain> so if your proposed idea is fine strongly depends on where we want to take this
18:55:26 <TrueBrain> if = whether, english hard
18:57:00 <TrueBrain> and if you don't get what I am trying to say, that is also fine :) Basically I ask if you can sit on the chair of an architect ;) Not everyone enjoys that seat, which is perfectly fine :)
18:58:16 <milek7> well yes, I don't know want you want, it's there :P
18:58:19 <milek7> you want other possibilities described?
18:58:56 <TrueBrain> no :) Basically, we need to know how far we want to carry this idea. You mention friendlist, which means sharing of information. Do we want to do that in a decentralized way? Or more centralized? Per cluster of server?
18:59:06 <TrueBrain> do we want authentication servers to process friendlist? Lookup via email?
18:59:13 <TrueBrain> so many ideas and possibilties
18:59:19 <TrueBrain> and a lot of them restrict others
18:59:46 <TrueBrain> so basically I need someone to go wild with the ideas and draw up something that is consistent and that others are like: WOW, I WANT THAT
19:00:26 <TrueBrain> so it is not so much about possibilities, more about the end of a track, which station we sort-of want to end up in
19:00:44 <TrueBrain> do we want friendlists? do we want that between more than 1 server? do we want that "globally"?
19:01:12 <TrueBrain> to give an example, say, we want friendlists. Everyone agrees, that is awesome. That means the default should not be a key per server, as that would make that impossible
19:01:16 <milek7> fine, so you want other ways described
19:01:17 <TrueBrain> it could still be an opt-in
19:01:27 <milek7> I just don't know why you replied "no :)"
19:01:30 <TrueBrain> I am just looking where you want to go with this :)
19:01:40 <TrueBrain> it is not about other possibilities; it is about a consistent combined idea :)
19:02:11 <TrueBrain> sometimes it can be as simple as: we want a place where creators can store their NewGRF and users can download it
19:02:16 <TrueBrain> (that was all it took for BaNaNaS)
19:02:49 <TrueBrain> and we need to ask users what they want, is kinda what I try to imply :)
19:03:02 <TrueBrain> do they want a friendslists, or do they want full privacy
19:03:11 <TrueBrain> this is btw not days or weeks worth
19:03:17 <TrueBrain> just a simple paragraph :P
19:04:02 <milek7> but I described: global pubkeys shared with other clients, clients store friendslist locally (pubkey+last known name mapping), optional privacy mode (which breaks friendslist but keeps company authorization working)
19:05:11 <TrueBrain> (not related to your work btw; was for an MS replacement)
19:05:18 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler reopened pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
19:05:21 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
19:10:26 <TrueBrain> milek7: what you describe there, in other words, are: I am going to take a left at the end of the street, walk for 3 blocks, take a right, and there I am going to sit down. What I am trying to ask is: where are you going to? :D
19:10:50 <TrueBrain> but if you like I can draft something I think we should be heading .. and if there is some consensus with the bigger server-admins, we can see if your ideas fit there
19:11:03 <TrueBrain> but the current answer is: I do not know if they are good ideas or not
19:11:08 <TrueBrain> I lack information :)
19:12:00 <TrueBrain> (and I am not trying to be difficult, to be clear; it really is not possible for me to know if those ideas are a good idea or not. We need community input)
19:16:30 <milek7> we're talking in circles:P
19:16:35 <milek7> I described where I want to be (normally public global pubkeys, friendslits stored locally at clients, maybe masterserver offering query service where given pubkeys are currently playing, optional privacy mode)
19:16:36 <milek7> and you seemingly want me to describe other possiblities I have not chosen
19:16:52 <milek7> and this is fine, but just tell me that :)
19:17:04 <TrueBrain> I am telling you what I want in many different ways
19:17:13 <TrueBrain> as it goes with a language barrier and completely different backgrounds, that is not easy
19:17:26 <TrueBrain> if I knew what to tell you for you to understand what I try to say, I would, no worries :)
19:17:30 <TrueBrain> but I am trying here ;)
19:18:04 <TrueBrain> and I am not interested in other possibilities; I want to know where you see this going :) You keep telling me the next step, I want to know the end station (sort-of .. doesn't have to be black and white)
19:18:36 <TrueBrain> short-term vs long-term
19:18:40 <TrueBrain> point-solution vs vision
19:18:49 <TrueBrain> what-ever they call it in your own day-to-day life :)
19:20:06 <TrueBrain> bah, why did they remove "re-run failed jobs"
19:20:11 <TrueBrain> now I have to rerun them all :'(
19:20:22 <TrueBrain> milek7: honestly, these things work better over voice in 90% of the cases :)
19:26:03 <milek7> well, working authorization and friendlist is the end station, isn't it?
19:26:08 <milek7> you want other features?
19:32:11 <TrueBrain> (I am not ignoring you, I am checking to see if the nightly is behaving :D)
19:32:33 <TrueBrain> there is this huge chain of events that is happening, which is pretty to see
19:32:50 <TrueBrain> schedule -> build -> publish CDN -> publish docs -> republish website
19:42:45 <TrueBrain> (it is way too technical, but I am wayyyy to tired to dumb this down :P)
19:42:57 <TrueBrain> (not dumbing down for you, but for, say, Discord people to read :D)
19:43:40 <TrueBrain> I deduced this information from the snippets you have been given btw; possibly I misunderstood or linked information incorrectly ;)
19:45:23 <TrueBrain> owh, we clearly forgot another PR .. oops
19:48:21 *** gelignite has joined #openttd
19:50:52 <milek7> you didn't describe interaction with passwords, guess we want to remove them?
19:51:04 <TrueBrain> exactly the things I want to know ;)
19:51:21 <TrueBrain> which is a choice; I have not given it any thoughts :)
19:52:19 <TrueBrain> we need to rewrite this without naming pubkey ... we need that for ingame too .. no clue how yet :D
19:53:21 <TrueBrain> "based on crypto, an unique ID is generated", or some horseshit
19:53:35 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLu9y
19:53:39 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler closed pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
19:55:18 <milek7> company access list should probably have 2 levels: member, and owner (owner can manage access list)
19:55:30 <TrueBrain> now you are getting what I am after ;)
19:55:44 <TrueBrain> so copy my gist, and fill in all these things you are thinking about
19:56:04 <TrueBrain> and it really can be ambitious; it is easy enough to remove stuff (or "postpone" the implementation)
19:56:54 <TrueBrain> I like that idea btw, access policies of some sorts (can be REALLY simple, but does help)
19:57:55 <TrueBrain> tomorrow I will write a similar one with a central entity, just so we can compare the pros and cons of them .. I honestly don't know what works
19:58:14 <TrueBrain> possibly a hybrid is the best solution
19:58:23 <milek7> if we're replacing game passwords too, invitations?
19:58:53 <TrueBrain> would also solve openttd:// implementation, for multiple OpenTTD versions
19:59:09 <TrueBrain> my head works weird btw .. as now I am thinking: auto-updater? Download correct OpenTTD version on join?
19:59:19 <milek7> time-expirable secret that adds you to access list as member?
19:59:19 <TrueBrain> that should NOT be part of this proposal btw :P
19:59:41 <TrueBrain> invitation-code, basically, sounds like a good replacement of passwords to me
20:00:20 <TrueBrain> how about GameScript access?
20:00:46 <TrueBrain> (for tournaments or quests-servers or what-ever)
20:05:47 <TrueBrain> really curious what server-owners are going to think about it, really :)
20:06:10 <TrueBrain> what I like about the decentralized approach you take here, is that it is privacy first, basically
20:06:40 <TrueBrain> so maybe we can make a service that allows you to store your secret identity based on a login, as an opt-in or something .. dunno
20:07:06 <TrueBrain> feels wrong to store a secret externally, but on the other hand .. we are going to get tickets with: I can no longer join my company after reinstall, and shit :D
20:08:27 <milek7> well, it could store secret encrypted... back to passwords :D
20:08:48 <TrueBrain> but a private password that is never shared with servers ;)
20:08:53 <TrueBrain> AES encrypted blob, that is not a bad idea
20:09:05 <TrueBrain> implementation is a few lines of code too
20:10:31 <TrueBrain> also allows an identifier for cloud-saves :D
20:11:25 <TrueBrain> clean install: full privacy mode -> opt-in friendlist -> opt-in store secret remotely -> opt-in cloudsaves :D
20:12:54 <andythenorth> opt-in DRM for grfs :P
20:13:03 * andythenorth straight from a nice place to a wrong place
20:13:06 <andythenorth> 100 miles an hour
20:13:39 <TrueBrain> hmm, this also works for my crazy MMO idea :P
20:14:58 <Samu> everything is by SamuXarick
20:14:59 <andythenorth> one giant scrolling world
20:15:13 <andythenorth> everybody on a single server
20:15:14 <TrueBrain> andythenorth: yup, shards
20:15:33 <andythenorth> I use words as though I know what they mean :P
20:15:37 <andythenorth> key management skill
20:15:39 <TrueBrain> initially I was thinking of having tickets to do handoffs between servers, but this is really tricky
20:15:44 <TrueBrain> as it depends on the position of your viewport etc
20:16:18 <TrueBrain> but if we have a way to unique identify a user over a cluster of servers
20:16:24 <TrueBrain> that becomes a lot easier :P
20:16:58 <TrueBrain> milek7: I guess what might be nice for a way for a server to indicate that it is part of a cluster, and even if you are in privacy mode, asks you if it is okay to use the same pubkey for that cluster of servers
20:17:51 <Samu> I didn't push those commits, I don't understand what's happening
20:17:59 <TrueBrain> say I run some tournament servers, battling for the biggest profit in 10 years, or something
20:18:06 <TrueBrain> one day I might be on server A
20:18:09 <Samu> I have a workflow on my fork?
20:18:21 <TrueBrain> from the cluster perspective, it would be nice if you identify yourself as the same user
20:18:24 <TrueBrain> so it can track your stats
20:19:01 <milek7> I would say.. disable privacy mode
20:19:19 <TrueBrain> yeah, but that makes it a 0 / 1 nob
20:19:28 <TrueBrain> I was wondering if adding a 0.1 would be nice too :)
20:19:32 <TrueBrain> just spitting out ideas :)
20:21:11 <Samu> someone can explain why do I have github actions on my openttd fork?
20:23:59 <Samu> seems to only work for the master branch
20:24:30 <Samu> that's the least interesting branch as it's equal to the upstream/master one
20:31:38 <LordAro> Samu: you, or possibly GH, has enabled them
20:31:53 <LordAro> you should be able to disable them in the repo settings
20:36:37 <DorpsGek> [OpenTTD/website] glx22 commented on pull request #177: Add: upcoming release has new filename descriptions https://git.io/JLubZ
20:45:22 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #8399: Fix: [OSX] Warning about ambiguous method (-Wobjc-multiple-method-names). https://git.io/JLuNO
20:49:55 <andythenorth> hi-cube intermodal containers? o_O
20:49:57 <DorpsGek> [OpenTTD/website] glx22 approved pull request #177: Add: upcoming release has new filename descriptions https://git.io/JLuNA
20:49:59 <andythenorth> to make my life harder?
20:50:39 <DorpsGek> [OpenTTD/OpenTTD] mattkimber opened pull request #8400: Fix c02ef3e4: [NewGRF] Variable 0x44 is always HZB_TOWN_EDGE for roadstops https://git.io/JLuAU
20:56:04 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #8399: Fix: [OSX] Warning about ambiguous method (-Wobjc-multiple-method-names). https://git.io/JLuA9
20:59:31 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #8399: Fix: [OSX] Warning about ambiguous method (-Wobjc-multiple-method-names). https://git.io/JLuNO
21:01:31 <DorpsGek> LordAro: I have not seen Yexo.
21:01:45 <LordAro> TrueBrain: i thought the seen db was kept?
21:11:08 <DorpsGek> TrueBrain: Bjarni was last seen in #openttd 9 years, 10 weeks, 6 days, 20 hours, 52 minutes, and 1 second ago: <Bjarni> heh
21:11:32 <TrueBrain> So that is interesting ...
21:11:43 <DorpsGek> TrueBrain: frosch123 was last seen in #openttd 1 day, 0 hours, 29 minutes, and 5 seconds ago: <frosch123> i guess you want the full sf experience
21:11:50 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #8400: Fix c02ef3e4: [NewGRF] Variable 0x44 is always HZB_TOWN_EDGE for roadstops https://git.io/JLupy
21:12:00 <DorpsGek> TrueBrain: I have not seen yexo.
21:12:26 <TrueBrain> No clue! Seems it decided to forget about him :(
21:13:06 <TrueBrain> But clearly the DB is in working order ..
21:18:12 <TrueBrain> I did some manual filtering to clear up old channels .. maybe that went wrong? Dunno .. nothing we can do about it now :)
21:20:12 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8398: Change: Distribute left over cargo to stations according to the highest remainder https://git.io/JLuh2
21:23:06 <TrueBrain> glx: would you mind merging and tagging the website PR? Tnx :)
21:24:38 <DorpsGek> [OpenTTD/website] glx22 merged pull request #177: Add: upcoming release has new filename descriptions https://git.io/JIhBu
21:28:12 <DorpsGek> [OpenTTD/OpenTTD] mattkimber commented on pull request #8400: Fix c02ef3e4: [NewGRF] Variable 0x44 is always HZB_TOWN_EDGE for roadstops https://git.io/JLujH
21:28:51 <TrueBrain> Tnx! Not logged in on mobile, so can't do it myself :)
21:40:42 <DorpsGek> [OpenTTD/OpenTTD] mattkimber updated pull request #8400: Fix c02ef3e4: [NewGRF] Variable 0x44 is always HZB_TOWN_EDGE for roadstops https://git.io/JLuAU
21:44:05 <Timberwolf> I hate trying to do anything related to squashing commits in git :(
21:46:35 <glx> it's not that hard with interactive mode
21:47:21 <TrueBrain> We can always squash from the GitHub interface just before merging
21:48:05 <DorpsGek> [OpenTTD/OpenTTD] mattkimber updated pull request #8400: Fix c02ef3e4: [NewGRF] Variable 0x44 is always HZB_TOWN_EDGE for roadstops https://git.io/JLuAU
21:48:24 <glx> yeah, but sometimes squashing is needed for the commit checker
21:48:46 <Timberwolf> Finally got there. And yes, interactive mode helped :)
21:48:54 <TrueBrain> Lately I have been doing that for people .. lowers the curve of learning git by a lot :)
21:49:46 <glx> touching other people's PR is ok if they don't have to work more on it
21:50:14 <glx> else it can be a pain if they don't know git enough
21:50:18 *** Samu has quit IRC (Quit: Leaving)
21:51:07 <TrueBrain> Basically, turns out GitHub doesnt have very good guides on our rebase strategy
21:51:21 <TrueBrain> And it makes making PRs more difficult for people
21:51:50 <TrueBrain> Lot of new stuff to learn at once
21:52:27 <glx> worse part being the default commit message when you update a file via git
21:54:15 <Timberwolf> The documentation of "stuff you need to do" was good, imo. (Assuming I haven't made some dumbass mistake I haven't noticed)
22:08:30 <andythenorth> it was a serious learning curve for me to git :P
22:08:50 <andythenorth> 5 years of doing it differently, with minimal training, did not help
22:09:55 <andythenorth> learning git where 'force push is banned' and 'interactive rebase looks like cheating' does not help when committing to openttd :P
22:11:28 <andythenorth> oof ran out of nml switch IDs
22:12:04 <glx> well force push is banned except for PRs
22:12:35 <glx> usually force push is a bad idea if other people work on the same branch
22:12:46 <andythenorth> anyway now I know
22:12:59 <andythenorth> now I hide all my sins with rebase -i
22:13:16 <andythenorth> hmm so I think I ran out of spriteset IDs
22:13:20 <andythenorth> this is...interesting
22:13:38 <glx> like recently I needed to force push to master (luckily it didn't affect any people)
22:14:31 <andythenorth> I didn't know spriteset IDs were finite :P
22:15:34 <glx> and it was not a too aggressive rebase, just a removal of latest commits, but still it's not something recommended to do :)
22:16:55 <andythenorth> hmm concurrent spritesets 1/255
22:17:15 <milek7> tbh, do we need unprotected companies.. at all?
22:17:28 <andythenorth> Concurrent spritegroups: 230/256
22:17:31 <andythenorth> looks more plausible
22:18:21 <glx> milek7: with an "authentication" system I think they can be protected by default
22:18:36 <andythenorth> hmm I don't use spritegroups in the place this is failing
22:19:17 *** sla_ro|master has quit IRC ()
22:19:20 <milek7> yeah, but I wonder if it needs to support unprotecting at all
22:21:21 <andythenorth> running out of IDs in nfo wasn't a thing
22:21:31 * andythenorth wonders why nml runs out
22:23:03 <andythenorth> the message is imprecise, I'll paste
22:23:17 <andythenorth> nmlc ERROR: "generated/iron-horse.nml", line 17830: Unable to allocate ID for [random]switch, sprite set/layout/group or produce-block. Try reducing the number of such blocks.
22:23:56 <glx> in NFO you handle the ID manually, while nml may not detect a now free ID
22:25:29 <andythenorth> in NFO there was stupid tricks to just keep chaining through same repeating IDs
22:25:57 <andythenorth> this is a HUGE block of generated code, it probably just consumes too many IDs
22:28:59 <andythenorth> it would be 2k lines if reversed non-8/8 vehicles actually worked in openttd :)
22:29:51 <andythenorth> I can 99% definitely restructure the code generator to not trip this
22:29:57 <andythenorth> I'd just like to know why it happens :D
22:30:37 <glx> I think the first random switch use at least 6
22:36:43 <glx> basically a switch needs at least [number of chains] ids
22:38:24 <glx> switch_spritelayer_cargos_intermodal_cars_default_16px eats a lot of IDs ;) (well some are duplicated I think but still)
22:38:37 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:40:13 <glx> is switch_spritelayer_cargos_intermodal_cars_random_default_16px_stake_flatrack_ZINC marked as optimised ?
22:42:43 <glx> (anyway optimised or not, that should not change the number of used IDs)
22:47:19 <glx> hmm are 16px, 24px, 32px chosen via a switch ?
22:49:25 <andythenorth> vehicles chain directly into the one they need
22:49:58 <andythenorth> I intend to get rid of the single-option random switches, and chain direct to ss
22:50:17 *** gelignite has quit IRC (Quit: Stay safe! Stay at home! Stop the chain reaction!)
22:51:29 <andythenorth> nml does optimise them out though
22:52:13 <glx> then maybe nml doesn't optimise the order of declarations (I really don't know how the generation is done), but I hope it does at least group actions 2 and actions 3 by usage
22:53:14 <andythenorth> I will try restructuring the generator and find out :P
22:53:37 <andythenorth> current structure is just one long list of all possible intermodal containers ever
22:53:43 <andythenorth> very simple to write :P
22:54:28 <glx> maybe just remove some stuff, so we can see the nfo
22:54:39 <glx> and how well nml organised things
22:55:30 <andythenorth> removing is easy :D
22:55:48 <andythenorth> reading raw nfo...is not
22:56:06 <glx> but I think ID generation is done once the reference tree is complete
22:56:35 <glx> (I didn't check the source)
22:57:04 <andythenorth> these are in a procedure, so they're not linked to just one action 3
22:57:08 * andythenorth wonders how that all works
22:57:42 <glx> procedures are actions 2, they should be handle as other references
22:57:52 <andythenorth> no I'm wrong, they're not strictly a procedure, they're just used globally
22:59:18 <Timberwolf> Yes, you have almost as much code as I have sprites :p
22:59:58 * andythenorth searches 'intermodal'
23:02:06 *** jottyfan has joined #openttd
23:02:25 * glx will download the file, browser doesn't like to display it :)
23:03:17 <andythenorth> I could strip the other vehicles out, but eh
23:06:27 <glx> 10924 results for intermodal
23:06:51 <glx> many are just the sprites
23:08:36 <andythenorth> near the end of the file looks like the switches
23:12:17 * andythenorth restructured the generator
23:12:21 <andythenorth> seems to work now
23:12:33 <andythenorth> concurrent spritegroups has dropped
23:12:45 <andythenorth> does nml implicitly add spritegroups?
23:13:08 <andythenorth> at some point action 3 has to go to an actual action 2 :P
23:14:05 <andythenorth> before I restructed, last working compile was 236/256
23:14:17 <andythenorth> after restructure 69/256
23:14:24 <glx> restructuring reorders the nml file ?
23:14:56 <glx> so maybe nml is not smart enough :)
23:16:37 <andythenorth> I don't know what the magic does, but this works
23:17:09 <andythenorth> I interleaved the cargo_type_in_veh switches with the spritesets
23:17:33 <andythenorth> presumably it figures out now that no action 3 resolves to ALL the spritesets
23:18:16 <glx> but it should have figured out in the other case too
23:20:20 <andythenorth> I am about to commit my fixed version, but is the old code wanted for a test case?
23:21:04 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler opened pull request #8401: Change: Towns don't build dead-end road bridges https://git.io/JLzst
23:21:23 <glx> old code is current version, or never commited ?
23:21:59 <andythenorth> I never committed a version that trips the limit
23:22:23 <glx> hmm maybe put it in a branch then
23:24:02 <glx> but if output nml is the exact same code, just in a different order, there's an issue in nmlc I'd say :)
23:25:27 <andythenorth> it's the same logic
23:25:31 <andythenorth> the flow might be different
23:25:41 <andythenorth> master has the working version, branch the failing
23:25:57 <andythenorth> if you build locally, you might want
23:25:58 <andythenorth> make install PW=16 2>&1 | grep -v "returns a constant"
23:26:22 <andythenorth> should fail on line 17761
23:26:34 <glx> haha I can disable optimisation message ;)
23:28:41 <andythenorth> I should cherrypick your commit :P
23:31:33 <glx> I told you, just comment the lines in switch.py
23:32:02 <andythenorth> oof :) I missed that :)
23:32:05 <andythenorth> silly andythenorth
23:32:09 <glx> all the "optimising" are in this file
23:41:34 <andythenorth> I used pypy, it's a shitload faster for me
23:42:01 <glx> wow there are many unreferenced things
23:44:12 *** nielsm has quit IRC (Ping timeout: 480 seconds)
23:45:16 <andythenorth> I'll fix the unreferenced soon
23:45:17 <milek7> TrueBrain: we could also do external identities, like 3PID in matirx
23:45:41 <glx> hey but at least they are real unreferenced ;)
23:46:34 <andythenorth> they're not unreferenced in the successful (fixed) version
23:46:45 <andythenorth> that's interesting?
23:47:27 <milek7> extend access lists to store alternatively also 3PIDs
23:47:29 <glx> well if I can have fail and not fail for the same data it's better I think
23:47:53 <milek7> servers could query chosen identity server about given pubkey
23:48:04 <glx> oh even master has unreferenced :)
23:49:56 <milek7> and if not present on given identity server, pass challenge-response from them to the client along with URL which is asked for client to open, which contains some oauth flow or whatever
23:50:39 <glx> but it looks like master and the branch are identical, minus the reorder so it should be enough to try to trace events
23:51:37 <milek7> so you could do something like: add 3PID:tt-forums.net:Milek7 as company member, and then ottd would ask me to open browser to associate my pubkey with tt-forums account on identity server
23:51:44 <glx> oh of course make failed because I don't have grfcodec in path, but that's ok :)
23:51:48 <milek7> but that's a lot of complexity, and I'm not sure if this is worth it
23:51:57 <glx> I now have 2 nml to test things
23:52:39 <andythenorth> I should say...I have a fix, so if this is some hard nmlc thing
23:52:44 <DorpsGek> [OpenTTD/OpenTTD] stormcone commented on pull request #8401: Change: Towns don't build dead-end road bridges https://git.io/JLzcY
23:52:56 <andythenorth> how many other people will ever trigger it?
23:53:01 <andythenorth> just seems a bit...weird :)
23:53:02 *** Laedek has quit IRC (Read error: Connection reset by peer)
23:53:37 <glx> it's something to check at least, if nmlc can be smarter it's better for everyone
23:54:34 <andythenorth> ok sleeping time
23:54:36 *** andythenorth has quit IRC (Quit: andythenorth)
23:57:27 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8401: Change: Towns don't build dead-end road bridges https://git.io/JLzcd
continue to next day ⏵