IRC logs for #openttd on OFTC at 2023-02-19
⏴ go to previous day
00:00:06 <andythenorth> Hmm can we run grf as wasm in a lambda and just stream the sprites to your screen?
00:00:40 <andythenorth> Bonus points if we can use edge-nodes and ‘computing in the network’
00:16:38 <frosch> the api version is set by the ai
00:16:47 <frosch> if they import an old library, that is the ai's problem
00:17:12 <frosch> i.e. the compatiblity layer applies to the whole vm: ai and libraries
00:22:03 <frosch> hmm, their info.nut says api "1.2" though
00:22:14 <frosch> so maybe the compatibilty script loading is broken 🙂
00:25:13 <frosch> works for me in master, starting their ai loads the compatibilty script
00:25:33 <frosch> in fact their newest tar which adds their "fix" crashes, because it now calls a wrong CreateGroup
01:09:38 <glx[d]> yeah the AIGroup change was in 1.9
02:22:27 *** HerzogDeXtEr1 has quit IRC (Read error: Connection reset by peer)
02:24:04 *** gelignite has quit IRC (Ping timeout: 480 seconds)
02:51:17 *** Tirili has quit IRC (Quit: Leaving)
02:56:43 *** geli has quit IRC (Quit: Stay safe!)
03:06:42 *** tokai|noir has joined #openttd
03:06:42 *** ChanServ sets mode: +v tokai|noir
03:13:29 *** tokai has quit IRC (Ping timeout: 480 seconds)
03:30:55 *** Wormnest has quit IRC (Quit: Leaving)
03:46:39 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:14:04 *** TROILUS4 has joined #openttd
04:18:22 *** TROILUS has quit IRC (Remote host closed the connection)
04:18:23 *** TROILUS4 is now known as TROILUS
05:20:36 *** TROILUS has quit IRC (Read error: Connection reset by peer)
05:20:37 *** TROILUS5 has joined #openttd
05:20:38 *** TROILUS5 is now known as TROILUS
06:09:41 *** Flygon has quit IRC (Remote host closed the connection)
06:54:45 *** HerzogDeXtEr has joined #openttd
07:13:19 *** JohnFranklin[d] has quit IRC (Quit: User went offline on Discord a while ago)
08:54:33 *** sla_ro|master has joined #openttd
09:53:28 <TrueBrain> ugh ... why did frosch force me to write more javascript? 😛
09:55:24 *** gelignite has joined #openttd
10:11:38 <andythenorth> onClick(“truebrain”)
10:11:59 <andythenorth> Interesting conflation there of “me” and co-pilot
10:12:19 <andythenorth> Are we going to start treating AI as part of ourselves?
10:22:01 <dP> Can we be sure TB is not an AI?
10:27:59 <TrueBrain> filters only don't update based on the other filters, as meh, lots of work 😛
10:28:25 <TrueBrain> CoPilot can most likely do it, but .. not feeling it 🙂
10:30:12 <andythenorth> Yeah dynamic combinatorial filters is…meh
10:30:22 <andythenorth> Work, and can be confusing
10:31:37 <andythenorth> Means e.g. selected item in one filter becomes invalid due to another filter. Just say no 🙂
10:47:14 *** TROILUS7 has joined #openttd
10:49:35 *** TROILUS has quit IRC (Read error: Connection reset by peer)
10:49:36 *** TROILUS7 is now known as TROILUS
11:05:23 <FLHerne> oh, so that's what the end-on crash bug is
11:22:32 *** WormnestAndroid has quit IRC (Remote host closed the connection)
11:24:53 <TrueBrain> okay .. next up, bananas-server .. should be relative easy
11:27:05 <Samu> question, how do I get to the head of a vehicle?
11:35:38 <TrueBrain> okay, a tag `standard` and `high` is not a good description 😄
11:35:43 <TrueBrain> the rest all seems fine 🙂
11:50:40 <andythenorth> petern: so is Regions done? 😛
11:50:48 <andythenorth> Or is it cycling.
12:27:48 *** gelignite has quit IRC (Quit: Stay safe!)
13:15:25 *** Flygon has quit IRC (Read error: Connection reset by peer)
13:45:09 <TrueBrain> wow, some AIs really like to write things to the console, lol
13:57:16 <frosch> are you adding a label "chatty"?
13:57:50 <TrueBrain> seen the new PRs? I didn't do "size" for NewGRFs, as that feels a bit too opinionated in my opinion . we can always add that later
13:59:15 <frosch> yes, i read that i made it in live: forcing something else to write js, while you are asleep 🙂
13:59:31 <TrueBrain> CoPilot is awesome 😛
14:00:41 <TrueBrain> hmm, I want a minimap snapshot every N in-game days, but using the current method is rather expensive .. I just want the delta, basically .. hmm
14:01:41 <TrueBrain> although .. hmm .. initially I wanted an animated gif, but now I am thinking about it .. maybe a slideshow of PNGs is better .. just slower 😛
14:04:00 <TrueBrain> how many years should an AI run to get a bit of an idea if it crashes or not?
14:10:39 <andythenorth> Ok so AdvancedRegions
14:11:22 <andythenorth> Where for a tile range, specific game behaviour is over-ridden by local methods
14:11:47 <andythenorth> e.g industry founding, cargo payment, town expansion etc
14:13:22 <Rubidium> TrueBrain: maybe also log the number of commands or something, as an AI surviving 10 years without building anything isn't useful either, even when it doesn't actually crash
14:16:00 <frosch> you added a class member named "set", isn't that asking for trouble?
14:16:55 <frosch> no idea, maybe just C paranoia 🙂
14:17:17 <TrueBrain> owh, you mean as in a reserved keyword issue
14:17:25 <TrueBrain> we also use fields like `type`, so we should be fine 😛
14:18:12 <frosch> ah, would not have notices that one 🙂 I always use isinstanceof
14:20:58 <dP> whenever I do that my editor annoys me by highlighting them as keywords
14:22:14 <TrueBrain> frosch: I meant the field `type` not the keyword `type` 😛 Hihi, I hate keywords 😄
14:24:09 <TrueBrain> frosch: had to double-check, but yeah, all keys are kept as a dict-key, so a string, so all that should be fine. And in Python `set` is not a keyword, so also not an issue. And in JSON it is also nicely transported, and Javascripts handles it fine. Good, worth the check 😄
14:25:55 <frosch> yaml is the one with the troublesome type magic, just make sure to not add a criterion "on" or "no" :p
14:26:11 <TrueBrain> yeah, you showed .. silly YAML
14:26:21 <TrueBrain> Booleans are sent as YAML booleans, so we should be good 🙂
14:27:52 <frosch> oh dear, i forgot the horrors of validate_packet_size
14:28:04 <TrueBrain> Yeah .. it is better now at least
14:28:21 <TrueBrain> there are so many constraints on the other fields, the chances of it being seen now by any user are very very low
14:28:27 <TrueBrain> (description is capped at 510 chars)
14:31:08 <frosch> you removed tags from the regressions tests, but did not add classification. do you care?
14:31:22 <TrueBrain> no; the regression tests what users can influence and change, mostly
14:31:34 <TrueBrain> I was thinking about adding some regressions for classification, but it is .. ugh .. tricky
14:32:08 <frosch> nah, testing the classification is useless 🙂
14:32:19 <frosch> whenever you change it, you reclassify everything, and then check the diff
14:34:50 <frosch> oh, approved too early :p
14:35:02 <TrueBrain> QUICK, MERGE MERGE MERGE 😛
14:40:52 <TrueBrain> if you insist on making the same remark twice, I insist on making the same comment twice. Hihi, sorry, couldn't refuse 😛
14:42:52 <frosch> copy&paste code results in copy&paste reviews 🙂
14:43:18 <TrueBrain> let me know if you disagree btw; designing for the future can be a pitfall
14:43:36 *** WormnestAndroid has joined #openttd
14:48:36 <frosch> ok, I think some of the "+2" could be "+1", but if it is about leaving space for future externsions, it does not matter
14:49:18 <frosch> however, there is another problem, which was there before. the validation needs to check the *encoded* length of name and description
14:49:35 <frosch> plenty of sets use CJK in their texts
14:49:48 <TrueBrain> frosch: +2 is the length-byte. Bit undocumented where all the `+2`s come from .. took me a while to realise 😛
14:51:10 <TrueBrain> hmm, no, it is even worse
14:51:17 <TrueBrain> that function is just the worst 😛 The need for it alone ..
14:52:41 <TrueBrain> but yes, they are missing `.encode()`
14:53:26 <frosch> in good news: andy no longer has to fix the trailing | in the tags-list on the website
14:53:50 <TrueBrain> the md5sum is also funny .. that `+2` is overshooting it by far
14:54:26 <TrueBrain> but yeah, as description is capped to 510 bytes, not chars
14:54:34 <TrueBrain> nothing is really broken .. just silly. Will fix that in another PR soon
14:55:23 <frosch> haha, i have no idea about js, but that "use strict"; is really funny to me 🙂
14:55:41 <TrueBrain> 😄 And it actually does something useful
14:56:04 <TrueBrain> it forces you do declare a var before using it
14:56:09 <TrueBrain> which prevents a lot of common mistakes 🙂
14:56:47 <frosch> using nops to encode future information is a long tradition in computer science
14:57:12 <TrueBrain> yeah ... "it had to be backwards compatible!!"
15:10:23 <frosch> not really qualified to review that, but the things i understood looked good 🙂
15:10:37 <TrueBrain> I wasn't qualified to write it, but yeah ... if it aint broken .. 😛
15:12:44 <frosch> surprisingly many sets had no tags at all
15:13:49 <TrueBrain> okay, so now .. merge and bring to production? Or add classifications to other content first .. hmm
15:14:12 <TrueBrain> guess the latter, as the first will give some QQ on some of the content that no longer has tags 😛
15:16:49 <frosch> i would have been fine with the former 🙂
15:17:12 <TrueBrain> well, approve bananas-server then, and I will think about it over dinner
15:17:38 <frosch> oh, i thought i already did 🙂
15:17:40 <TrueBrain> Tnx for the reviews 🙂
15:22:16 <TrueBrain> funny you mention that ... means another library isn't working 😛
15:26:11 <nnyby> Does anyone have any thoughts on automatically upgrading trains? The process of selling old trains and buying new ones for monorail tracks feels like it could be improved somehow. I saw this brief discussion from a few years ago: https://github.com/OpenTTD/OpenTTD/issues/6617
15:28:58 <nnyby> in particular, it's annoying to have to manually replicate the train setup: which cars to buy and re-creating the train's route
15:29:05 <Xarick> TallTyler: I'm experimenting with vehicle lengths and testing if a train is on tile to avoid sudoku drivers
15:29:30 <Xarick> the bottom right is my tests, resulting in even less crashes
15:29:45 <Xarick> top right is current 13.0
15:30:34 <andythenorth> nnyby: JGRPP has template train replacement
15:30:37 <Samu> is it kamikaze or sudoku ? or harakiri?
15:30:51 <nnyby> andythenorth: thanks, i'll look into that
15:31:31 <andythenorth> there's a jgr-specific channel on the discord server, but not irc
15:34:21 <frosch> even docs and examples 🙂
15:34:31 <TrueBrain> yeah .. not even lazy, I know, shocker 😛
15:36:01 <TrueBrain> let's hope that resolves most, if not all, of the sentry issues .. as I really don't understand them otherwise 🙂
15:37:38 <Samu> oh snap, on another different test, it gets worse results, im so sad I can't improve this
15:41:43 *** gelignite has joined #openttd
15:45:01 <TrueBrain> time to disconnect some connections!
15:45:03 <Samu> master + my changes: 60 crashes
15:45:22 <Samu> master + my other changes: 70 crashes
15:47:32 <Samu> i get very mixed results :o
15:48:08 <TrueBrain> bah, strong-referenced tasks don't solve all issues for sure .. meh
15:52:51 *** Wormnest has joined #openttd
16:05:07 *** herms has quit IRC (Quit: bye)
16:16:31 *** TROILUS3 has joined #openttd
16:17:29 <Xarick> this is a very subjective matter
16:18:08 *** TROILUS has quit IRC (Remote host closed the connection)
16:18:08 *** TROILUS3 is now known as TROILUS
16:30:09 *** TROILUS has quit IRC (Read error: Connection reset by peer)
16:30:20 *** TROILUS has joined #openttd
17:37:44 <Samu> I'm testing 5 different approaches to level crossings
18:03:37 <Samu> that moment you realize you made a code mistake after doing tests... :(
18:04:17 <petern> That moment you realise you made a code mistake after merging the PR
18:15:22 <Samu> the PBS savegame has some hidden nasty surprises after 3 years
18:15:46 <Samu> the first 2 years has very low crashes
18:16:03 <Samu> suddenly on the 3rd year, it skyrockets
18:16:48 <andythenorth> petern: you mean...every time...for me?
18:18:03 <andythenorth> hmm shall I find my glasses
18:18:06 <andythenorth> you're all blurry
18:25:28 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
18:43:44 <DorpsGek> - Update: Translations from eints (by translators)
19:13:17 <andythenorth> everything is sharp again
19:18:11 <andythenorth> also how did we live without chunky bevels petern ?
19:18:19 <FLHerne> nnyby: "recreating the train's route" <- when you build the first new train, you can copy the orders from one of the existing ones
19:18:26 <FLHerne> if you haven't sold them all already
19:19:12 <andythenorth> meh, am I the only person still using the sprite font?
19:19:18 <FLHerne> (in the orders window, with an empty order list select 'go to' and click on the train you want to copy orders of)
19:19:24 <FLHerne> or ctrl-click to share orders as usual
19:19:35 <FLHerne> sprite font best font
19:19:38 <petern> andythenorth: Everything was fields
19:19:46 <FLHerne> (and the less said about the tiny sprite font the better)
19:19:59 <andythenorth> goes it throw out sprite font?
19:20:16 <andythenorth> because fancy typographic quote mark, but wrongly
19:22:12 <andythenorth> unrelated, but this is just how my brain works when I try playing a game 😦
19:22:33 <andythenorth> I am writing town GS stuff, but really, 1 whole second to set growth rates?
19:22:48 <andythenorth> "real time computing"
19:23:45 <andythenorth> also totally unrelated, but town-type label schema?
19:23:54 <andythenorth> reldred: TallTyler ^ town labels?
19:24:40 <TrueBrain> frosch: : okay, let's merge this what we have, and slowly add other classifiers; but not today, as I have to redo the BaNaNaS PR, as people made new commits 😛 Silly people .. WHY DO THEY ADD STUFF! 😛
19:24:47 <andythenorth> ^ all houses and inhabitants are identical in BORG town 😛
19:25:23 <andythenorth> MILK for a town with lots of roundabouts (Milton Keynes for non-English channel viewers)
19:25:30 <frosch> TrueBrain: how did your ai crash tests progress?
19:25:33 <andythenorth> SWIN for the town with 1 magic roundabout
19:25:45 <frosch> my plan was to scan scripts for usage of specific api functions
19:25:51 <TrueBrain> frosch: not much, wrote down what I want it to do. Now to figure out how to do that
19:25:58 <frosch> but ofc, running and checking what it does is another otion
19:26:07 <andythenorth> NYNY for New York (named it twice)
19:26:18 <andythenorth> PATX for Paris, Texas
19:26:40 <andythenorth> oh wait, town _types_ not actual townsa
19:26:50 <TrueBrain> frosch: no, I think we should have two separate systems there
19:27:02 <andythenorth> FAIL for most of the north of england 1970-2010
19:27:34 <TrueBrain> so one that just looks at the nuts files and goes: owh, this is what you are doing
19:27:42 <andythenorth> oh are we categorising AIs? 😮
19:27:48 <andythenorth> buses, planes etc?
19:28:13 <TrueBrain> as testing of an AI actually doesn't crash etc is CPU intensive shit 😄
19:28:38 <frosch> the nut classification is about as fishy as the newgrf classification
19:28:56 <frosch> if there is an ai-library which does the actual api calls, it fails
19:29:01 <andythenorth> you will probably find that most AIs are about "GSList"
19:29:07 <TrueBrain> so we also need to classify libraries 😄
19:29:09 <TrueBrain> but yeah, I get what you mean
19:29:13 <frosch> we can check whether it builds engines (AIEngine), but not what type of engine
19:29:16 <andythenorth> "I heard you like lists"
19:29:21 <frosch> but we can detect whether it builds roads or rails
19:29:35 <andythenorth> how many AIs do we have now?
19:29:35 <frosch> so technically we can also do a "rondje" label: builds vehicles, but no infra
19:30:33 <frosch> plenty, want to classify "idle" and "idlemore"?
19:32:11 <frosch> so many ais, which are no competitors
19:32:36 <frosch> oh, there is even "idlemoremore"
19:34:31 <andythenorth> "An AI that does ABSOLUTELY nothing. Period. See the entries 'Idle' and 'IdleMore' for more information."
19:51:28 <TrueBrain> but yeah, I Was thinking making a site where people can see how AIs are oding
19:51:47 <TrueBrain> both in terms of economy
19:52:04 <TrueBrain> but to also use it as classifier .. it is not the worst idea, just .. how ..
19:52:19 <TrueBrain> means after upload we need to send it somewhere to be evaluated, and then come back to update the classification
19:52:28 <TrueBrain> as it will take, what, 5 minutes to do all the simulation
19:52:53 <TrueBrain> take for example one of those AIs that only builds helicopters
19:52:58 <TrueBrain> when you start in 1950, it seems to do nothing
19:53:02 <TrueBrain> as there are no helicopters yet 😛
19:57:05 <TrueBrain> figuring out what configuration to test against will be the hardest .. also with NewGRFs etc
19:58:58 <dP> there was some website at some point that did ai competitions
20:02:29 <dP> I don't think there is much point to rank ais on anything but default settings though
20:02:41 <dP> there are just too many combinations of settings
20:04:24 <andythenorth> can we afford to just leave AIs running?
20:05:08 <TrueBrain> for how long did you have in mind? 😛
20:05:18 <andythenorth> on some resetting cycle
20:05:28 <andythenorth> but now I have a better idea
20:05:33 <andythenorth> can we simulate the AIs?
20:05:36 <Samu> I had some high hopes for master + length + train on tile || adjacent
20:05:45 <Samu> turns out it's not as good as expected
20:06:34 <andythenorth> if we know roughly how an AI performs statistically, for a range of conditions
20:06:41 <andythenorth> then we can just run the simulation of them
20:06:50 <TrueBrain> yes, that is called running the game
20:06:54 <TrueBrain> that is one big simulation
20:07:03 <andythenorth> no we need to simulate the simulation
20:07:53 <andythenorth> michi_cc[d]: what's the expected behaviour of vehicle name callback when player renames vehicle? 🙂
20:09:26 <Samu> the length versions tests vehicle_length, if it's small enough to fit between rail gaps, it advances the multi-level crossing
20:10:14 <Samu> the longer vehicles would however just try to get by as fast as possible, even collide vs trains if so
20:11:49 <Samu> the shorter vehicles would stop instead of colliding with train, if they still can stop
20:12:05 <Samu> if a train then passes behind them, the vehicle is safe
20:12:48 <Samu> but then another vehicle of the same size could be following the vehicle in front and it would queue, waiting on the rail tracks :(
20:14:04 <Samu> there's really no easy solution, or a wonder solution
20:14:09 <TrueBrain> frosch: strictly seen possible, to do classification async from the upload, and update the upload when the results arrive 😛 But do we want to go that way ..
20:14:59 <Samu> the current solution in 13.0 is really weird
20:15:44 <Samu> if the road vehicle is already walking on the level crossings, it won't stop moving, even if it means going head on with a train on the side
20:19:38 <frosch> TrueBrain: we can also go the phone-home path. when an ai crashes, users get a "report bug" button. this sends the script error+backtrace to dorpsgek. dorpsgek creates issues in a "content feedback" repository while checking for duplicates, and assigns the bugs to the content owner, who already has a gh account :p
20:20:21 <TrueBrain> so rebuilding Sentry? 😛
20:21:51 <frosch> i just like the fact: since we already tricked content authors to register at gh, we can also assign bugs to them :p
20:24:48 <frosch> 47% if ais and 75% of gs are migrated to gh accounts
20:27:13 <frosch> lol, i was just wondering why there are so many music sets
20:27:21 <frosch> some weird spammed the list in 2017
20:28:17 <frosch> traditional irish/scottish/english/welsh/canadian tunes volumes 1 to 5
20:33:36 <glx[d]> andythenorth: so they cloned our internal "dummy" AI ?
20:35:03 <frosch> glx[d]: there are "idle", "idlemore" and "idlemoremore" ais, and they are having an argument what "doing nothing" means exactly
20:35:14 <frosch> pretty sure they would disagree with "dummy ai"
20:35:20 <TrueBrain> frosch: but even if we collect crash statistics, still says little of how an AI is doing 🙂
20:35:34 <glx[d]> dummy ai really does nothing, but it crashes too
20:35:51 <glx[d]> because it doesn't even have a loop
20:36:04 <frosch> TrueBrain: my impression was that "crashiness" is so dominant, that anything else is not measureable anymore
20:36:20 <TrueBrain> owh, the AIs I tried ran fine for the first few game-years
20:36:45 <TallTyler> frosch: These are bangers. I particularly like the Canadian albums.
20:37:14 <frosch> "no music" is my favorite set 🙂
20:37:36 <andythenorth> could we have subtypes for objects?
20:37:40 <andythenorth> so each object has an ID
20:37:52 <andythenorth> then it has another set of subtypes, all show in object build menu?
20:37:58 <andythenorth> maybe a word for those
20:38:02 <Brickblock1> aren't object ids limited
20:38:10 <andythenorth> object ids are limited
20:38:16 <andythenorth> that's why I invented subtypes
20:38:20 <frosch> we already have object views
20:38:29 <andythenorth> yeah, only 4 though and they're weird
20:38:36 <frosch> mostly meant for orientations, but not sure whether actually used like that
20:39:20 <andythenorth> I could use one ID per FIRS industry
20:39:26 <andythenorth> then subtypes for each tile
20:39:52 *** herms has quit IRC (Quit: bye)
20:41:47 <frosch> will you add a puzzle gs? points if you recreate the exact industry layout with objects without looking at an industry?
20:41:53 <TrueBrain> ugh, okay, let's update BaNaNaS and run the reclassifier ...
20:42:10 <TrueBrain> hoping nobody posts an update in between 😛
20:42:20 <frosch> shut down the frontend container :p
20:42:43 <frosch> but yeah, no activity today
20:42:55 <TrueBrain> they all should be backwards compatible .. at least, my tests showed we are .. about the find out!
20:44:09 <Samu> "master - mask" should have the same results as "master + length" on the first 2 savegames, why didn't it?
20:44:50 <michi_cc[d]> andythenorth: Player name is player name and always wins.
20:47:30 <Samu> omg no... :( found another issue again
20:48:13 <TrueBrain> frosch: oops, totally forgot .. we need a better name for "resolution high/standard" 😄
20:48:22 <TrueBrain> it now will show up in-game as "high" or "standard", as tags
20:48:24 <TrueBrain> which makes no sense 😄
20:48:50 <Samu> passing DiagDirection to GetTileTrackStatus for some reason results in different game states
20:48:53 <frosch> ah, i guess there should be no label for "standard"
20:49:13 <TrueBrain> that is a bit weird in drop-downs
20:49:19 <TrueBrain> "All" or "High-res"
20:49:22 <TrueBrain> that feels incomplete 🙂
20:49:39 <TrueBrain> `has-high-resolution` ?
20:50:20 <frosch> well, it's a bool, like "has-sound", "has-32bpp", "has-high-res", "builds-trains"
20:50:32 <TrueBrain> yes, that is not how you wrote it, and not how I wrote it 🙂
20:50:39 <TrueBrain> 32bpp/8bpp isn't a bool now, which I kinda like
20:50:43 <TrueBrain> but high-res will become a bool now 😄
20:51:29 <frosch> fair, some people hate 32bpp, so they want to tag for 8bpp 🙂
20:52:09 <Brickblock1> Same for high res
20:53:04 <reldred> frosch: Jokes on them then because a lot of 8bpp sets are 32bpp under the hood 😈
20:53:30 <frosch> Brickblock1: but you can disable that in the settings
20:53:40 <frosch> while some 32bpp sets only have placeholder 8bpp sprites
20:53:41 <reldred> Can’t be bothered fixing palette? 32bpp it 😎
20:53:48 <frosch> so disabling 32bpp does not really work
20:53:50 <TrueBrain> I can be convinced to do the same for 32bpp btw .. has-32bpp, instead of 32bpp/8bpp
20:54:20 <frosch> no, now i like the 8bpp and 32bpp 🙂
20:54:29 <TrueBrain> `set=rail-station,palette=8bpp,has-high-res=False,has-sound-effects=False`
20:54:35 <TrueBrain> looks good-enough to me!
20:54:44 <Brickblock1> I would still prefer because down scaling does not look that good
20:57:07 <TrueBrain> I really like that the front-end has no knowledge of the classifiers
20:57:11 <TrueBrain> makes changing them really easy
20:59:42 <TrueBrain> even the linters said: sure!
20:59:45 <TrueBrain> okay, let's bring this to production
21:00:36 <dP> imo zoom filter would be more practical instead of hires bool
21:01:15 <Brickblock1> would it not be harder to implement tho
21:01:54 <dP> zoom level is in the sprite
21:06:58 <TrueBrain> okay .. let's see how this goes ...
21:09:00 <frosch> the js filters do not show up
21:09:21 <frosch> oh wait, they are still unclassified, so some caching
21:11:27 <glx[d]> high-res is extra zoom ?
21:11:48 <frosch> it is >50% 2x or 4x sprites
21:11:55 <TrueBrain> and the API did not appreciate reloading 5k entries
21:12:28 <TrueBrain> there you go, filters
21:12:33 <frosch> oh, i assumed it always reloads everything
21:13:26 <frosch> there are a surprising amount of grfs matching has-high-res and has-sound
21:13:41 <frosch> ah, yes, giant cats... how could i forget
21:14:00 <TrueBrain> okay, in-game listing is showing the same information
21:14:04 <TrueBrain> this seems to have worked 😄
21:14:45 <frosch> yay, the tags are dead, long live the tags
21:14:56 <TrueBrain> finally useful shit 😄
21:15:44 <frosch> i can search for "high-res 32bpp landscape" and it works 😮
21:16:02 <TrueBrain> yup; and it is actually useful to search like that now 🙂
21:17:18 <TrueBrain> I am actually happy that you can now filter on NewGRF content .. curious what the user will think of it
21:17:40 <frosch> i am curious whether anyone will notice, or whether it needs a news post :p
21:18:02 <TrueBrain> yeah, I am considering making a post in Discord channel #developer-updates 🙂
21:18:57 <frosch> just post a screenshot of the in-game content window with some useful search result?
21:19:17 <frosch> "ai powered newgrf content search"
21:22:40 <TrueBrain> small improvements, with hopefully a big impact 🙂
21:22:52 <TrueBrain> we could also add a search-bar on the website btw .. super trivial now
21:23:43 <frosch> the dropdowns are good enough, noone is using the website anyway
21:24:07 <frosch> oh, wow, i did not expect the filter to add to the url
21:24:17 <TrueBrain> reload even works 🙂
21:24:19 <frosch> you can actually share links to reddit
21:24:48 <TrueBrain> initially I wanted to do that server-side, but that was weird and sluggish
21:24:57 <TrueBrain> this requires that you enable Javascript, but .. yeah .. small price to pay
21:25:13 <TrueBrain> everything works fine if you don't have javascript enabled btw; you just can't filter
21:25:17 <frosch> are there actually people who disable js?
21:25:27 <TrueBrain> strictly seen I could make this CSS-only
21:25:37 <TrueBrain> but that is a tiny bit tricky with CSP
21:27:02 <TrueBrain> 10k unique visitors per day .. still amazed how busy our wiki is 🙂
21:28:02 <TrueBrain> the adoption of HTTP/2 is insane 🙂
21:29:02 <SpComb> there's only a handful of browsers, and they agressively auto-update
21:32:37 <frosch> i still can't believe how well this tagging actually works
21:44:24 *** nielsm has quit IRC (Remote host closed the connection)
21:55:44 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
22:23:27 <Samu> how do i pause a server when there's nobody playing?
22:32:31 <glx[d]> min_player setting, or something like that
22:32:58 <glx[d]> type "list_settings min" it should show
22:38:02 <Samu> im the server and spectating
22:39:27 <glx[d]> anything above 0 will pause it if you spectate
22:40:08 <Samu> i created a game and it didn't start paused :(
22:40:13 <glx[d]> but maybe it's only for dedicated servers
22:40:34 <Samu> ah, nevermind, i was playing, dumb me
22:40:41 <Samu> i forget this starts in company 1
22:40:55 <glx[d]> ah yes when you start you are playing
22:41:12 <Samu> wait it still didn't pause :(
22:45:44 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:47:12 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:48:36 <Samu> nobody wants to play with Disasters, Recessions, Breakdowns, Plane crashes, Inflation, Infrastructure mai
23:09:42 <Samu> I'm almost done with these level crossing tests
23:15:26 <Samu> final results, won't test anymore
23:16:34 <Samu> i wanted to use ||, not &&
23:22:13 *** Samu has quit IRC (Quit: Leaving)
23:49:44 *** dale_ has quit IRC (Remote host closed the connection)
23:50:15 *** Eddi|zuHause2 has quit IRC (Read error: Connection reset by peer)
23:51:04 *** Eddi|zuHause2 has joined #openttd
23:54:16 *** sla_ro|master has quit IRC ()
continue to next day ⏵
