IRC logs for #openttd on OFTC at 2020-01-12
⏴ go to previous day
00:07:46 *** sla_ro|master2 has quit IRC
00:13:24 <Samu> if (IsTileType(n.GetTile(), MP_WATER)) {
00:13:24 <Samu> c += YAPF_TILE_LENGTH * (WATER_DEPTH_MAX - GetWaterDepth(n.GetTile()));
00:13:38 <Samu> gotta go sleep take care
00:20:00 <glx> I don't think we have a "show reservation" thing for water
00:20:53 <glx> maybe the path is too long to be found without buoys
00:20:59 *** WormnestAndroid has quit IRC
00:21:05 <glx> but yapf should handle that
00:21:38 *** WormnestAndroid has joined #openttd
00:23:11 <FLHerne> nielsm: Was there previously a connection between those long narrow channels?
00:25:33 <glx> to go in this location, it seems it used the direction to dock fallback
00:26:16 <glx> do you get ship is lost notices ?
00:26:33 <nielsm> the ship just circles around in the corner
00:26:42 *** snail_UES_ has joined #openttd
00:27:28 <nielsm> and no, no lost warnings
00:28:19 <nielsm> oh now it did get lost yes
00:28:45 <glx> anyway there's a path it should find it, still a PF bug
00:28:45 <nielsm> maybe it was because I had "advice on players vehicles
00:29:01 <nielsm> when it first got lost, and then it didn't re-post the message
00:29:43 <glx> OPF had a length limit for pathfinding but I think there's no limit for newer PF
00:29:52 <nielsm> but yes it does prefer paths away from the coast with this change
00:30:19 <TrueBrain> glx: I looked why the order of 'ids' can be different. The "ID" it creates are set to "deterministic" ... funny word, as it turns out :P Seems that the global npm package influences it somehow
00:31:10 <glx> I'm currently cleaning stuff, I noticed removing the pr branch wasn't removing all files
00:31:50 <TrueBrain> 'npm run clean' removes everything relevant
00:32:08 <glx> but for some reason it didn't want to run
00:32:17 <TrueBrain> "deterministic - Module names are hashed into small numeric values."
00:32:26 <TrueBrain> glx: it is expecting to run on a linux command line btw :D
00:32:55 <TrueBrain> "clean": "lerna exec rm -- -rf node_modules package-lock.json dist lib",
00:33:11 <TrueBrain> the only command that does that :P
00:33:48 <glx> well rm "partially works" in powershell ;)
00:34:21 <TrueBrain> yeah, the -rf only is not compatible
00:34:32 <TrueBrain> so remove all the "node_modules" folders, and you will be fine :)
00:34:50 <glx> but exec rm doesn't because it's not an executable
00:37:52 <glx> no changes, still the big diffs
00:40:24 <TrueBrain> doesn't matter, really
00:40:38 <TrueBrain> I am more interested in a review atm :)
00:40:43 <glx> yes it's just annoying :)
00:42:20 <TrueBrain> there are more things that are annoying :) I am just more happy if this works, and I can shut down k8s :P
00:50:32 <nielsm> hmm, building canal over river, the river depth should be preserved, right? should there be a max depth to canals? what should happen if the player tries to build canal over a river deeper than the max canal depth?
00:52:31 <nielsm> (error? just lower depth? additional cost?)
00:52:58 <hythlodaeus> if depth means there there has to be a transitioning area, then it might not be a problem at all
00:53:15 <hythlodaeus> because a deeper river would also mean it would be wider
00:54:30 <hythlodaeus> so it's very much like raising or lowering terrain: no tile can be raised/lowered without afecting thhe depth of the surrounding ones
00:55:03 <hythlodaeus> so a 1-tile wide course of water would always have the same depth
01:00:44 <hythlodaeus> alternatively if one still wants the possibility of having 1 tile-wide canals with increased depths: what about using locks for this purpose as IRL?
01:01:44 <nielsm> yes it would make sense to have a way to build (more expensive) deeper canals
01:04:43 <hythlodaeus> then you could treat them as road types
01:06:43 <hythlodaeus> and deeper canals should not be compatible with aqueducts for instance
01:07:43 <hythlodaeus> or rather they could be linked, but heavier ships would not be able to usr them
01:12:42 <TrueBrain> LordAro: seems my patches are all working; so tomorrow I am going to merge this, and deploy the new DorpsGek_III :D After that we can finally test if your tag_created works :P It does require a .dorpsgek.yml change in OpenTTD before it works btw .. but I will create a PR for that tomorrow too :)
01:14:05 *** ChanServ sets mode: +v tokai
01:20:39 <milek7_> huh, jukebox window crashes when compiled with AddressSanitizer
01:25:51 <LordAro> i've seen similar stacktraces before
01:27:57 <LordAro> none of those strings take other parameters, presumably?
01:37:12 <milek7_> do SetDParamStr copy string?
01:37:35 <milek7_> i think struct returned by _music.GetCurrentSong() goes invalid at the end of statement
01:40:00 <nielsm> no it only copies a string pointer, not the contents
01:40:18 <nielsm> so yeah that would be it I think
01:40:46 <nielsm> when the function returns an empty playlist item because nothing is playing, the object goes out of scope
01:42:52 <milek7_> i don't think it matters whether it is empty or not
01:43:36 <milek7_> as GetCurrentSong returns by copy
01:44:27 <nielsm> this approach won't fix it
01:50:23 *** andythenorth has left #openttd
02:05:00 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh opened pull request #7926: Fix: Use after return for stack allocated object https://git.io/JvfE5
02:18:00 <DorpsGek_III> [OpenTTD/OpenTTD] James103 commented on issue #7925: A certain save-game corrupts the title screen sprites when loaded. https://git.io/Jvf40
02:45:26 <hythlodaeus> nielsm: if eventually you can give me further feedback on how to improve #7870 I would really appreciate it. I'm trying to come up with a solution to solve current UI woes while making it simulatenously more informative, pleasing, and standardized. This changes are important for my future translation plans, but I do not want to alienate others here. If you find some time, I would love to hear your preferences and thoughts o
03:16:48 *** supermop_Home_ has quit IRC
05:20:48 *** WormnestAndroid has quit IRC
05:21:11 *** WormnestAndroid has joined #openttd
05:23:23 *** crazystacy has joined #openttd
05:34:53 *** ToBeCloud_ has joined #openttd
05:34:55 *** ToBeCloud_ is now known as ToBeFree
05:36:01 *** heffer_ has joined #openttd
05:37:48 *** tyteen4a03 has joined #openttd
05:38:59 *** colde_ is now known as colde
05:38:59 *** ToBeFree is now known as ToBeCloud
05:39:55 *** innocenat_ has joined #openttd
05:40:40 *** Sheogorath has joined #openttd
05:40:43 *** daspork has joined #openttd
05:49:17 *** mikegrb_ has joined #openttd
05:49:17 *** super_sp1oky has joined #openttd
05:49:17 *** Extrems has joined #openttd
05:56:39 *** WormnestAndroid has quit IRC
06:36:03 <crazystacy> how about a "skip order" button next to the "reverse direction"?
06:36:17 <crazystacy> or some hotkey. not sure which hotkeys there are, i suspect the openttd wiki hotkey page is out of date
07:09:53 *** snail_UES_ has joined #openttd
08:52:31 *** andythenorth has joined #openttd
08:52:41 *** HerzogDeXtEr has joined #openttd
09:00:59 <andythenorth> deep sea narwhal pie
09:02:39 <andythenorth> Pikka: le question
09:05:15 *** sla_ro|master has joined #openttd
09:30:15 <andythenorth> HEQS and Squid are more fun than Hog and Sam
09:41:18 *** Progman has joined #openttd
10:03:24 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh updated pull request #7926: Fix: Use after return for stack allocated object https://git.io/JvfE5
10:05:53 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #7926: Fix: Use after return for stack allocated object https://git.io/Jvfwc
11:04:12 * andythenorth plays with town growth off
11:17:42 *** hythlodaeus has joined #openttd
11:17:58 *** hythlodaeus_ has joined #openttd
11:19:37 <hythlodaeus_> I'd like to you invite you folks to test my improved english.txt for feedback as p/ issue #7870
11:20:27 <hythlodaeus_> i got some previous positive feedback on how to handle style and capitalization and most people seem to agree. I would like to hear your opinion on the modified tooltips text and general formatting
11:41:32 <TrueBrain> well, here goes nothing .. lets see if it all still works :D
11:41:50 <DorpsGek_III> [OpenTTD/website] TrueBrain merged pull request #140: Add: [Actions] switch to our own actions instead of writing them out https://git.io/JvfcU
11:42:51 <Samu> what newgrf have ships moving slow on canals
11:43:52 <DorpsGek_III> [OpenTTD/website] TrueBrain opened pull request #141: Fix: [Actions] wrongly referenced 'checkout-dispatch' action https://git.io/Jvfrp
11:44:00 <TrueBrain> that was a quick: nope! :D Lalalalaaaa
11:44:37 <TrueBrain> never rename something in the middle of your work :D
11:45:18 <Eddi|zuHause> that's why sou spend the time naming things at the very beginning :p
11:45:32 <TrueBrain> and you NEVER change your mind after that
11:45:40 <TrueBrain> what a silly thing to say :)
11:48:38 <DorpsGek_III> [OpenTTD/website] TrueBrain merged pull request #141: Fix: [Actions] wrongly referenced 'checkout-dispatch' action https://git.io/Jvfrp
11:49:19 <Samu> i dont understand water depths yet
11:51:14 <nielsm> although I want to rework that part really
11:57:48 <TrueBrain> hmm, seems the docker hub webfrontend is broken (again)
11:58:29 <DorpsGek_III> [OpenTTD/actions] TrueBrain opened pull request #2: Fix: 'exec.exec()' also escapes quotes; 'remote-tag' now got an extra quote https://git.io/Jvfoq
11:59:49 <Eddi|zuHause> "Fix: 'exec.exec()' also escapes quotes" is a thing you never want to read in patchnotes
12:00:32 <Eddi|zuHause> because it means there's probably only a thin layer left between you and cthulu...
12:00:44 <DorpsGek_III> [OpenTTD/actions] TrueBrain merged pull request #2: Fix: 'exec.exec()' also escapes quotes; 'remote-tag' now got an extra quote https://git.io/Jvfoq
12:05:12 <DorpsGek_III> [OpenTTD/website] TrueBrain opened pull request #142: Fix: logo in top-right corner wasn't pointing to "Home" https://git.io/JvfoK
12:05:28 <Samu> max depth is only 200 now?
12:05:39 <Samu> ah no, i didn't wait enough
12:05:53 <TrueBrain> I so wish Githubs repository would become public (as in: anonymous access) ... as docker hub really has issues ...
12:07:07 <DorpsGek_III> [OpenTTD/OpenTTD] James103 opened issue #7927: Giant screenshot warning reports incorrect dimensions of screenshot https://git.io/Jvfoi
12:11:00 *** andythenorth_ has joined #openttd
12:11:48 <DorpsGek_III> [OpenTTD/website] TrueBrain closed issue #139: Wrong link target of the link on the OpenTTD-lettering in the page header https://git.io/JveQw
12:11:48 <DorpsGek_III> [OpenTTD/website] TrueBrain merged pull request #142: Fix: logo in top-right corner wasn't pointing to "Home" https://git.io/JvfoK
12:11:51 <andythenorth_> nielsm I suggest 2 sea tile types
12:11:59 <andythenorth_> and a transition type
12:12:42 * andythenorth_ wonders how many water tiles there are
12:13:26 <andythenorth_> we have 2 water cycles in the palette, but really we need to change the colour for deep
12:14:52 <andythenorth_> the transition tile would be a simple dither blend
12:15:03 <nielsm> I wonder if water depth can be handled by a recolour sprite?
12:15:04 * andythenorth_ wondering if it can be automated
12:15:04 <Samu> ooops, nielsm got an assert
12:15:18 <andythenorth_> the blend could be masked
12:15:45 <TrueBrain> wtf, the ' is still there .. this should not be happening ... hmmmm ... sometimes automation is a bitch :(
12:16:02 <Samu> an oil rig was spawning, and crashed
12:16:27 <hythlodaeus_> nielsm: it's more of a matter of how depth will be handled by scenarios and the map gen
12:17:06 <hythlodaeus_> you're gonna need new sprites for mid transitioning depth lines, that's for sure
12:17:15 <Samu> ah no, actually an industry was being removed from the map
12:18:03 <hythlodaeus_> but mostly I think depth will be a river vs ocean matter, with most transitioning happening on shorelines
12:18:08 <andythenorth_> nielsm is it trivial to recolour base set sprites? I’m assuming yes because CC, etc
12:18:11 <Samu> i suspect just using magic bulldozer and demolishing an industry should be enough to trigger the crash
12:18:32 <andythenorth_> nielsm how deep are we going?
12:18:45 <nielsm> right now the max depth is presented as 300 m
12:19:14 <nielsm> with 0 presented as 10 m, and the remaining presented as value * 20 m
12:19:18 <Samu> industry must be close to coast
12:19:28 <andythenorth_> transition at 8?
12:19:45 <andythenorth_> or configurable?
12:20:04 <nielsm> I was thinking transition at 2-3 and 10-11
12:20:30 * andythenorth_ wonders about non-navigable water
12:20:37 <andythenorth_> and sprites for it
12:20:49 <nielsm> that's pretty much parks in towns
12:21:00 <nielsm> non-navigable water isn't represnted on the map
12:21:18 <nielsm> (unless you have water and put newobjects in it to make it impassable)
12:21:26 <andythenorth_> I’m thinking about ships always requiring at least depth 2
12:21:38 <andythenorth_> think it’s too big a change maybe
12:22:23 <andythenorth_> but if rivers had depth 1 or 2 we could make the 1 sprites narrower
12:22:45 <andythenorth_> and depth 1 coasts get rocks
12:22:47 <Samu> nielsm, try demolishing industries with magic bulldozer that were built on land, near coast tiles, you get a crash
12:22:58 <nielsm> I think the normal draught of ships isn't much more than 2-5 metres?
12:23:10 <andythenorth_> but I think it leads to all shores being non-navigable
12:24:57 <crazystacy> now we'll need viking ships
12:25:31 <andythenorth_> we could use river mouths to create depth 2 bays and estuaries in sea
12:25:53 <nielsm> the suez canal is 20-30 metres deep I think
12:26:03 <andythenorth_> just pathfind to deeper water and expand a bit
12:26:42 <nielsm> so yeah depth 2 in my current units is enough to support any ship
12:26:50 <nielsm> and depth 1 will support most non-gigantic
12:27:22 <andythenorth_> a bit for dredgeable?
12:27:55 <andythenorth_> dredgeable tiles are imvalid for pathfinding but otherwise depth 1
12:27:57 <nielsm> I'd just make newgrf able to query the depth of water tiles, so industries can check it before construction
12:28:09 <DorpsGek_III> [OpenTTD/website] TrueBrain opened pull request #143: Update: bump all Python dependencies to latest version https://git.io/JvfoN
12:28:11 <nielsm> depth 0 is also valid, it's "very shallow"
12:28:57 <andythenorth_> yeah I’m just obiwan with my depths :)
12:31:57 <andythenorth_> dredgeable = player can make tile navigable
12:32:45 <nielsm> hm, the port of esbjerg in denmark, which faces the north sea, has the deepest docks at 10 m
12:32:54 <Samu> how do i propose a fix to a pr?
12:33:06 <nielsm> I don't really think it's worth making "too shallow to navigate" water
12:33:17 <nielsm> just make it newobjects if you want that
12:33:38 <andythenorth_> inplementation could be objects
12:33:59 <andythenorth_> but adding rocks manually to just remove them immediately is weird gameplay
12:34:27 <nielsm> can newobjects specify they want to be placed automatically by worldgen?
12:34:37 <andythenorth_> objects are weird anyway, what are they for?
12:34:48 <nielsm> decoration and being in the way
12:35:12 <andythenorth_> they’re immovable afaik
12:35:18 <andythenorth_> unless player built?
12:35:58 <andythenorth_> anyway having rivers lower landscape - above and below waterline both
12:36:51 <andythenorth_> then we can complain about slow map gen :D
12:38:33 <DorpsGek_III> [OpenTTD/website] TrueBrain updated pull request #143: Update: bump all Python dependencies to latest version https://git.io/JvfoN
12:38:41 <TrueBrain> these fucking chores :(
12:39:51 <andythenorth_> nielsm river seed should be at coast, running up also :p
12:40:32 <TrueBrain> hmm .. can see if code works on 3.8, good point
12:41:41 <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick requested changes for pull request #7924: Feature: Water tiles have a depth https://git.io/JvfKI
12:41:59 <Samu> oh it looks like a bad review
12:42:02 <DorpsGek_III> [OpenTTD/website] TrueBrain updated pull request #143: Update: bump all Python dependencies to latest version https://git.io/JvfoN
12:42:03 <TrueBrain> 3.8 it is :) Tnx LordAro :)
12:42:05 <Samu> i still dunno how to use this
12:43:13 <DorpsGek_III> [OpenTTD/OpenTTD] Milek7 commented on pull request #7926: Fix: Use after return for stack allocated object https://git.io/JvfKq
12:43:38 <DorpsGek_III> [OpenTTD/actions] TrueBrain opened pull request #3: Fix: default-values are set by "action.yml", and inputs are always strings https://git.io/JvfKm
12:46:56 <DorpsGek_III> [OpenTTD/actions] TrueBrain updated pull request #3: Fix: default-values are set by "action.yml", and inputs are always strings https://git.io/JvfKm
12:47:38 <DorpsGek_III> [OpenTTD/website] TrueBrain updated pull request #143: Update: bump all Python dependencies to latest version https://git.io/JvfoN
12:51:26 <DorpsGek_III> [OpenTTD/actions] TrueBrain updated pull request #3: Fix: default-values are set by "action.yml", and inputs are always strings https://git.io/JvfKm
12:52:08 <DorpsGek_III> [OpenTTD/actions] TrueBrain updated pull request #3: Fix: default-values are set by "action.yml", and inputs are always strings https://git.io/JvfKm
12:52:10 <TrueBrain> I keep forgetting to run 'npm run all', ffs
12:52:50 <DorpsGek_III> [OpenTTD/website] TrueBrain merged pull request #143: Update: bump all Python dependencies to latest version https://git.io/JvfoN
12:54:00 <DorpsGek_III> [OpenTTD/actions] TrueBrain merged pull request #3: Fix: default-values are set by "action.yml", and inputs are always strings https://git.io/JvfKm
12:55:03 <TrueBrain> seems that bumping the versions made the 'fetch-downloads' a lot quicker
12:55:14 <TrueBrain> but that might just be me-wanting-it-to-be-quicker :D
12:58:26 <TrueBrain> okay, that should be it
13:01:45 *** andythenorth has joined #openttd
13:02:02 <TrueBrain> LordAro: if you have some cycles, I would love a review on the DorpsGek one; it should be trivial (the github workflow is a copy from website, so that is fine. I mean the Python part :D)
13:02:25 <andythenorth> currently river is seeded on a spring, and pathfinds to a coast (or dead-ends in a lake)
13:02:47 <andythenorth> this causes (1) the unsatisfying lakes (side issue)
13:03:13 <andythenorth> (2) means that lowering terrain along river path is hard, because no guarantee that we'll be able to lower the next level
13:03:40 <andythenorth> starting rivers at coast means we just walk 'along' or 'up', lowering land as we go
13:03:43 <nielsm> the entire river generation code is rather poor
13:03:53 <andythenorth> rubidium put it in as a nice quick hack
13:03:58 <andythenorth> it was always "I made something work"
13:04:07 <andythenorth> and it's deliberately not elaborate :)
13:04:25 <Samu> if (IsTileType(n.GetTile(), MP_WATER) && GetWaterDepth(n.GetTile()) == WATER_DEPTH_MAX) c += YAPF_INFINITE_PENALTY; good or bad idea?
13:04:40 <andythenorth> if we start at coast, we can also branch rivers
13:04:44 <Samu> depth 15 means screw pathfinding here
13:04:52 <nielsm> yeah I usually turn off rivers in mapgen, they're not fun just annoying
13:04:57 <andythenorth> I don't actually understand what Eddi|zuHause BFS patch actually does :)
13:04:59 <TrueBrain> decodestring() still needs a byte-string .. I always forget about that :(
13:05:02 <nielsm> Samu doesn't penalty mean "avoid this tile"?
13:05:08 <nielsm> why would you avoid the deepest tiles?
13:05:27 <Samu> to have ships not go too far from the coast
13:05:28 <Eddi|zuHause> it makes a BFS, and counts how many tiles away from the shore you are...
13:05:45 <Samu> they can go far, but not that far
13:05:50 <nielsm> shouldn't that depend on the ship? whether it's oceangoing or meant for shallow water?
13:06:18 <nielsm> e.g. a river paddle steamer wouldn't want to go too far from the coast, but a large oil tanker probably prefers being on deeper ocean
13:06:30 <nielsm> don't put those assumptions into the pathfinder like that
13:06:43 <Eddi|zuHause> well, it technically isn't acutally a BFS, but whatever :)
13:08:02 <Samu> must fast forward like 10 years for depth 15 :(
13:09:00 <nielsm> yeah, I need to make mapgen produce proper depths right away
13:10:25 <nielsm> but ideally it should make "interesting" depths more than just a basic gradient
13:12:01 <andythenorth> I would treat that as a separate concern
13:12:07 <andythenorth> just let TGP do what it does for now
13:12:23 <andythenorth> the landscape gen is a problem above and below water :P
13:12:34 <nielsm> actually doesn't TGP make a full heightmap also for the parts that end up as below water?
13:12:48 <nielsm> i.e. the generator actually takes the heightmap and truncates it at sea level
13:17:26 <peter1138> Yes, that's normally.
13:17:53 <peter1138> Sea level is determined after the heightmap is generated.
13:18:43 <nielsm> so TGP does actually have a water depth that could be applied right away
13:18:56 <nielsm> then it's just making something for original and heightmaps
13:21:43 <peter1138> You can do the same for heightmaps, I guess.
13:21:46 <Samu> they're avoiding the "deep sea"
13:29:47 <peter1138> Most heightmaps are broken by more than 15 height levels anyway.
13:34:23 <nielsm> make a bool trireme flag on ship models, then they can avoid deep sea
13:44:04 * andythenorth plots sandbox mode
13:51:14 <nielsm> Samu: it makes proper depths right at mapgen now
13:51:40 <nielsm> (it can be very slow to do on large maps)
13:52:30 <nielsm> (since it has to visit every tile up to 16 times)
13:58:31 <FLHerne> nielsm: If depth 2 is navigable by anything, 16 seems like overkill
14:03:31 <nielsm> just marshland-like terrain in every direction
14:05:07 <Eddi|zuHause> TGP wasn't much better, it was all uniform hills
14:05:23 <Eddi|zuHause> until variety distribution appeared
14:05:32 <Eddi|zuHause> which also has some flaws
14:05:53 <Eddi|zuHause> like long narrow maps look horrible
14:07:58 <andythenorth> the fundamental of using sprite for mapgen intrigues me though
14:08:01 <andythenorth> maybe just more :P
14:08:23 <nielsm> yeah really you could make newgrf-based map generator sets
14:09:06 <Eddi|zuHause> you need some combinatorics to make it less repetitive, though
14:11:37 <Samu> heh you didn't fix the crash
14:12:16 <nielsm> no I didn't look at that part
14:19:02 <nielsm> yes I just said I didn't fix it yet
14:22:07 <Samu> you could include CheckForDockingTile(tile);
14:22:27 <nielsm> it's a logically separate step
14:23:00 <nielsm> and the compiler will probably do the same thing anyway
14:23:27 <TrueBrain> I am going to switch over the downloads to AWS (from DigitalOcean). This should happen without interruptions .. let me know if that isn't the case :)
14:23:43 * nielsm does not download ottd on most days
14:24:34 <TrueBrain> no, but the people that do will make sure to let you know in very clear language if it is broken
14:29:21 <DorpsGek_III> [OpenTTD/binaries-proxy] TrueBrain opened pull request #10: Fix: no longer log 2XX/3XX with this Docker https://git.io/Jvf65
14:29:28 <TrueBrain> right, that is a bit too much data to send to cloudwatch :D Haha :)
14:30:37 <Samu> something about build water depot turning false
14:30:43 <Samu> it failed to build the ship depot :p
14:32:25 <DorpsGek_III> [OpenTTD/binaries-proxy] TrueBrain merged pull request #10: Fix: no longer log 2XX/3XX with this Docker https://git.io/Jvf65
14:32:59 <crazystacy> nielsm, that mapgen doesn't look *too* ba
14:33:06 <crazystacy> there's some charm in its ugliness
14:40:53 <TrueBrain> okay, binaries-proxy is switched .. logs are clean, so that is good :)
14:43:00 <TrueBrain> time to do DorpsGek .. that will be interesting
14:49:59 <DorpsGek_III> [OpenTTD/DorpsGek] TrueBrain opened pull request #15: Fix: [Actions] Dorpsgek only runs on production, not on staging https://git.io/Jvfiq
14:57:52 <DorpsGek_III> [OpenTTD/OpenTTD] Milek7 opened pull request #7928: Fix: avoid using string pointer after scope end https://git.io/Jvfin
14:59:07 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh approved pull request #7928: Fix: avoid using string pointer after scope end https://git.io/JvfiW
14:59:50 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #7926: Fix: Use after return for stack allocated object https://git.io/Jvfi0
14:59:50 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh closed pull request #7926: Fix: Use after return for stack allocated object https://git.io/JvfE5
15:01:15 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro merged pull request #7922: Improved logic of sharing industry production between stations https://git.io/JvvB1
15:01:20 <nielsm> okay let's see if TGP can do deep water now
15:01:55 <crazystacy> installed visual studio code for the first time. "error code -23913". microsoft, it's good to be back after 10 years
15:05:59 <LordAro> crazystacy: i've gotten good at recognising what 0xC0000005 (segfault) is as a signed integer
15:06:02 <DorpsGek_III> [OpenTTD/DorpsGek] TrueBrain updated pull request #15: Fix: [Actions] Dorpsgek only runs on production, not on staging https://git.io/Jvfiq
15:06:07 <LordAro> (don't recognise that though)
15:07:04 <crazystacy> i was lacking Mono apparently
15:07:33 <DorpsGek_III> [OpenTTD/DorpsGek] TrueBrain updated pull request #15: Fix: [Actions] Dorpsgek only runs on production, not on staging https://git.io/Jvfiq
15:07:45 <crazystacy> i am sure if i googled the error code i'd come to some Microsoft website with a big cryptic headline "error code -12345". with 10 steps in big letters. step 10 being "install mono", which is what i did.
15:07:54 <crazystacy> "was this page helpful?"
15:08:03 <crazystacy> we will pipe the reply directly into bill gates' brain
15:09:18 <crazystacy> i think i once called microsoft customer support
15:09:27 <Eddi|zuHause> i don't remember ever coming across a microsoft help page where i would have answered "yes"
15:09:37 <crazystacy> then i started using linux and i never had to ask anyone for help again
15:10:04 <TrueBrain> that is impressive; you invented a brain-interface with that? :)
15:10:17 <crazystacy> it's more like a cattleprod attached to his skull
15:10:54 <TrueBrain> LordAro: LGTM is fucking slow :P It is annoying to wait for its result :P
15:13:35 <DorpsGek_III> [OpenTTD/DorpsGek] TrueBrain merged pull request #15: Fix: [Actions] Dorpsgek only runs on production, not on staging https://git.io/Jvfiq
15:15:39 <Eddi|zuHause> there's no mention of mono on that
15:16:01 <glx> just .NET framework for windows indeed
15:16:17 <glx> (usually means mono on other OS)
15:16:57 <Samu> im doing nasty things to ApplyWaterClassSpeedFrac
15:18:49 <TrueBrain> okay, it appears the new DorpsGek is deployed ... difficult to test, so lets redirect some traffic to it :)
15:18:53 <DorpsGek_III> [OpenTTD/OpenTTD] Milek7 opened pull request #7929: Fix: correct checking of fluid_settings_setnum return code https://git.io/Jvfi9
15:19:38 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh approved pull request #7929: Fix: correct checking of fluid_settings_setnum return code https://git.io/Jvfi7
15:23:28 <TrueBrain> oops, that DNS change was wrong .. we might be losing some GitHub messages here :D
15:25:59 *** DorpsGek_III_ has joined #openttd
15:41:36 <TrueBrain> anyone up for a quick review ^^ ? :)
15:44:01 <glx> looks good to me, but I'm not very good with python :)
15:45:24 <TrueBrain> I am, so that helps :P But I was more wondering if this flow makes sense, so I guess it does :)
15:46:08 <TrueBrain> we are now running on 0.15% CPU usage ... OWH NO! I am not sure if we can handle any more :P
15:46:26 <TrueBrain> and 1.5% memory usage ..
15:46:34 <TrueBrain> I love writing efficient shit :D
15:48:09 <TrueBrain> so if there are no weird reports tonight, I will shut down the k8s cluster .. happy days :)
15:48:29 <TrueBrain> nielsm: make water transparent, so you can see the bottom!
15:48:46 <Samu> dodododo, making ships as slow as 0 km/h, I'm so bad at math
15:50:04 <TrueBrain> LordAro: I just made a tag,but nothing happened :(
15:51:04 <nielsm> as far as water depth goes
15:53:03 <Eddi|zuHause> nielsm: that picture hurts my eyes and i don't see anything
15:55:24 *** DorpsGek_III_ has joined #openttd
15:55:24 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro merged pull request #7929: Fix: correct checking of fluid_settings_setnum return code https://git.io/Jvfi9
15:55:36 <TrueBrain> really stupid bug LordAro :D
15:56:13 <TrueBrain> we both failed, so that makes it a bit better :P
15:56:29 <LordAro> TrueBrain: clearly the CI should've caught it
15:58:24 <Samu> raw_speed *= (256 - (depth >= WATER_DEPTH_DEEP ? this->ocean_speed_frac : this->canal_speed_frac)) / 256;
15:58:25 <Samu> return raw_speed * (256 - ((WATER_DEPTH_MAX - depth) << 1)) / 256;
15:59:24 <TrueBrain> that are two horrible lines of code :P
15:59:27 <TrueBrain> what happened to variables?
15:59:31 <glx> (256 - anything) / 256 is <1 so 0
15:59:32 <TrueBrain> do we punish people now for usingthose?
15:59:59 <TrueBrain> glx: with the assumption that 'anything' is smaller than 256, ofc
16:00:06 <TrueBrain> (and bigger than 0)
16:00:35 <glx> (I think it's a valid assumption)
16:01:47 <TrueBrain> glx: btw, it is "a * (256 - anything) / 256". That changes it ofc a bit
16:02:02 <TrueBrain> those lines of code are just horrible
16:02:04 <glx> depending on how the compiler reorder things the division can happen first
16:02:21 <TrueBrain> yeah, sorry, the first line makes it 0
16:02:25 <TrueBrain> the second line should be fine
16:02:35 <TrueBrain> I am just going to stop trying to understand what the fuck that is :P
16:08:08 *** DorpsGek_III_ has joined #openttd
16:09:29 <Samu> i really should stop testing newgrfs
16:09:48 <Samu> a ship says 30 km/h in the purchase list
16:09:54 <Samu> then it actually moves at 35 km/h
16:10:38 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh approved pull request #7930: Fix: Silence all notes when stopping song in fluidsynth https://git.io/JvfXL
16:11:34 <nielsm> milek7_: as you can perhaps guess I don't use linux as a day to day system, and it's a big hassle for me to boot up the laptop I use for it :)
16:11:53 <nielsm> so I don't find or have much personal incentive to fix the little things
16:14:30 <Samu> nielsm can u test this? Harbour Point Utility Vessel says it moves at 30 km/h, but then once you have it running, it says 35 km/h in vehicle window
16:15:22 <nielsm> yes there's some weird things with those ship speeds, I don't know what's up with it
16:20:13 <DorpsGek_III_> [OpenTTD/OpenTTD] TrueBrain opened pull request #7931: Add: [DorpsGek] announce new tags on IRC https://git.io/JvfXa
16:28:32 <Samu> return raw_speed * (256 - (min(256, depth >= WATER_DEPTH_DEEP ? this->ocean_speed_frac + (WATER_DEPTH_MAX - depth) : this->canal_speed_frac + (WATER_DEPTH_MAX - depth)))) / 256;
16:29:05 <Samu> the deeper, the faster, the shallower, the slower
16:29:24 <nielsm> you could start by splitting it into more intermediate variables
16:30:13 <nielsm> variables are not dangerous, they don't make things slow
16:30:24 <Samu> it seems im doing this wrong yet, but ok variables
16:30:27 <nielsm> they are there to help you make sense of the code
16:31:04 <nielsm> modern compilers are pretty smart, they can do lots of expression simplification and optimisations across many lines of code
16:31:05 <glx> long lines full of operators are unreadable ;)
16:31:35 <glx> especially if you add ternary in the line
16:31:43 <nielsm> all sorts of things that are difficult to think about and make the code harder to understand if written as it ends up in the machine code
16:32:38 <nielsm> so write code for others to read and understand
16:32:59 <Samu> it's doing the opposite of what i want :o moving faster on shalow, and slower on deep
16:34:22 <andythenorth> I have experimented with pax
16:34:25 <andythenorth> and cdist is broken
16:34:30 <andythenorth> not buggy, just broken
16:36:43 <nielsm> classic pax: passengers go to the nearest station and hop on the first train they possibly can, get off at next stop
16:37:26 <nielsm> cdist pax: passengers go to the nearest station and pull a lottery ticket made by taking all seat numbers in the next 10 departures
16:38:03 <andythenorth> I turned town growth right down as an experiment
16:38:26 <andythenorth> specifically 'slow' with town road building forbidden
16:39:03 <andythenorth> I've connected 8 big cities with 3k-8k popn
16:39:28 <andythenorth> and a bunch of smaller peripheral towns
16:39:48 <andythenorth> with 5 big cities connected, I could clear the stations
16:40:10 <andythenorth> but adding a couple more tipped it over so that there's always 2k-5k pax waiting
16:40:41 <andythenorth> I've added enough trains that most stations now have a queue for entry
16:41:18 <andythenorth> my track layouts aren't co-op efficient but eh
16:42:12 <andythenorth> I don't think there's a solution btw, nerfing town pax production isn't going to be fun
16:43:03 <nielsm> are you doing point to point, or mainline visiting all stations?
16:43:24 <andythenorth> it's a long linear line, with star feeders
16:43:38 <andythenorth> there are trains that do all stations, and some that skip
16:43:50 <glx> if all stations are interconnected cdist is crazy
16:44:01 <andythenorth> the intent of the skipping is that it stops A->G passengers blocking B, C, D, E, F etc
16:45:23 <andythenorth> what I don't know is if that just creates back pressure
16:45:42 <andythenorth> pax might be going A->G->F->E->D->C->B
16:45:48 <andythenorth> I've seen cdist do that with freight
16:45:51 <andythenorth> harder to tell with pax
16:46:17 <Eddi|zuHause> E->A->D->G->H->E?
16:46:24 <glx> yeah it probably adds a link and try to use it for other links
16:46:32 <Samu> cargo over a big number , i think 4096, gets instantly 0'ed
16:46:41 <Samu> i guess all packets are gone
16:47:42 <Eddi|zuHause> cargodist is notably horrible with handling overcrowded links
16:48:21 <Eddi|zuHause> whenever my cargodist overflows passengers, i increase the distance influence
16:50:29 <Samu> yes, it's finally doing what I want
16:53:12 <DorpsGek_III_> [OpenTTD/OpenTTD] SamuXarick commented on pull request #7924: Feature: Water tiles have a depth https://git.io/Jvf18
16:54:12 <Samu> this looks more readable now
16:54:48 <Samu> that uint should be perhaps WaterDepth
16:58:08 <Samu> that byte is maybe wrong
17:01:21 <DorpsGek_III_> [OpenTTD/OpenTTD] TrueBrain merged pull request #7931: Add: [DorpsGek] announce new tags on IRC https://git.io/JvfXa
17:01:44 <andythenorth> the difference between 0% and 20% distance effect on demand
17:03:13 <TrueBrain> around 20%, you think?
17:03:30 <TrueBrain> well, from 0 to 20 is infinite higher, I guess
17:04:31 <andythenorth> I don't know what it's 20% of :)
17:04:39 <andythenorth> I guess the tile distance is in the src somewhere
17:05:44 <peter1138> What's the best engine for playing Doom or Heretic?
17:07:29 <nielsm> depends on what kind of doom or heretic you want to play, I think
17:08:10 <andythenorth> I switched to gzdoom
17:08:12 <nielsm> gzdoom or prboom are those I remember hearing most about
17:08:26 <peter1138> Using gzdoom at the moment. Some things are... not right :p
17:08:45 <nielsm> (but frankly doom is not much of a nostalgia trip for me, I prefer quake 1)
17:10:56 <DorpsGek_III_> [OpenTTD/OpenTTD] kiwitreekor updated pull request #7575: Feature: Add industry production graph https://git.io/fjc3i
17:13:02 *** snail_UES_ has joined #openttd
17:19:06 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro merged pull request #7930: Fix: Silence all notes when stopping song in fluidsynth https://git.io/Jvf1F
17:23:20 <peter1138> Playing original Doom winds up some of my... younger friends.
17:23:31 <peter1138> "You've got an RTX 2080 and you play THAT?!"
17:26:13 *** frosch123 has joined #openttd
17:32:17 <DorpsGek_III_> [OpenTTD/OpenTTD] frosch123 opened pull request #7932: Fix #7925: Reset temporary saveload data at the start of loading. https://git.io/JvfMr
17:36:16 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro approved pull request #7932: Fix #7925: Reset temporary saveload data at the start of loading. https://git.io/JvfM1
17:40:54 *** snail_UES_ has joined #openttd
17:42:45 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro requested changes for pull request #7575: Feature: Add industry production graph https://git.io/JvfMx
17:47:36 *** snail_UES_ has joined #openttd
18:03:03 <Eddi|zuHause> peter1138: what else would you play with an RTX 2080?
18:03:58 <Eddi|zuHause> that card is probably worth more than my whole computer :p
18:05:17 *** Wormnest has joined #openttd
18:20:44 *** snail_UES_ has joined #openttd
18:23:47 <peter1138> Eddi|zuHause, well, costs more, worth is a bit too subjective :)
18:24:37 <Eddi|zuHause> i bought an RX 580 for like 130€
18:25:27 <Eddi|zuHause> and my entire other computer was 800€, 9 years ago
18:26:20 <SpComb> RTX 2080 might be worth it for Transport Fever? :P
18:26:54 <SpComb> going into 1980 with most of the map built out, and FPS is dropping
18:27:22 <Eddi|zuHause> (that's not 100% true, i bought more storage for 300€ inbetween, and ram for 40€)
18:28:06 <Eddi|zuHause> (and another graphics card for 100€, that died half a year ago)
18:28:30 <DorpsGek_III_> [OpenTTD/OpenTTD] James103 opened issue #7933: [Azure Pipelines] #7924: The word "canceled" is misspelled. https://git.io/JvfDb
18:31:47 <nielsm> that's not even a bug in ottd's anything is it
18:33:05 <DorpsGek_III_> [OpenTTD/OpenTTD] PeterN commented on issue #7933: [Azure Pipelines] #7924: The word "canceled" is misspelled. https://git.io/JvfDb
18:34:54 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh commented on issue #7933: [Azure Pipelines] #7924: The word "canceled" is misspelled. https://git.io/JvfDb
18:34:54 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh closed issue #7933: [Azure Pipelines] #7924: The word "canceled" is misspelled. https://git.io/JvfDb
18:36:22 <LordAro> peter1138: tfw you have to look up the key for windscreen wipers in ETS2
18:38:09 <Eddi|zuHause> my dictionary says "canceled(AE)/cancelled(BE)"
18:39:23 <Eddi|zuHause> also, i shouldn't have 2 install processes downloading multiple GB at once
18:39:38 <Eddi|zuHause> ... and watch an online video at the same time
18:41:09 <nielsm> gah it's so hard to get a debugger attached to the regression tests
18:41:35 <LordAro> just run the same command they do?
18:41:41 <Eddi|zuHause> sounds like something that should be added to the makefile?
18:41:58 <LordAro> Eddi|zuHause: attaching gdb is part of the script
18:42:07 <LordAro> attching via VS is harder to do from a makefile though
18:42:14 <nielsm> I don't use gdb on windows with visual studio
18:42:29 <Eddi|zuHause> LordAro: i always find it non-trivial to find out what command is actually being run
18:44:10 <Eddi|zuHause> LordAro: yeah, but looking at a script that is 4 levels deep in some subdirectory runs as "non-trivial" in my book
18:48:17 <LordAro> Eddi|zuHause: not wrong, it could echo the command it's running
18:49:40 <nielsm> I have no idea why this is crashing...
18:50:24 <Eddi|zuHause> so, i just got frostpunk, what are the odds of me failing horribly?
18:50:31 <nielsm> I try building a depot on the same tiles as the script does, and then destroy it from the same tile as the script
18:51:00 <nielsm> but I also can't get the regression AI to load it seems?
18:51:20 <Eddi|zuHause> then you haven't isolated the conditions it crashing from...
18:56:38 <glx> frostpunk is not easy (from streams I've watched)
18:58:13 <Samu> maybe saveload conversion is crashing?
18:59:56 <glx> looks like an infinite loop in the regression
19:02:01 <glx> oh no, I know, it's the crash window
19:02:11 <glx> I remember seeing that before
19:02:44 <nielsm> yeah macos hangs if regression crashes
19:03:34 <glx> but IIRC I disabled crash window in dedicated
19:04:10 <nielsm> it actually uses null_v not dedicated_v for regressions
19:06:11 <DorpsGek_III_> [OpenTTD/OpenTTD] James103 commented on pull request #7932: Fix #7925: Reset temporary saveload data at the start of loading. https://git.io/Jvfyo
19:06:15 <nielsm> and regressions were failing for me locally not because of a crash or because of an actual failure, but because it found two copies of the same game script, which I had one of in my regular (stable) profile folder, and one in my development folder
19:07:27 <nielsm> James103 should really get onto irc so some of those rather over-excited and IMHO somewhat pointless comments and reports could be skipped
19:08:19 <glx> please don't put the [img] tag ;)
19:08:41 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh merged pull request #7932: Fix #7925: Reset temporary saveload data at the start of loading. https://git.io/JvfMr
19:08:42 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh closed issue #7925: A certain save-game corrupts the title screen sprites when loaded. https://git.io/Jvf40
19:09:24 <Samu> and now the awesome code math around it
19:09:50 <glx> but speed on river should be a newgrf property :)
19:10:39 <glx> every data from newgrf must be defined by the newgrf author
19:10:45 <glx> that's the point of newgrf
19:10:55 <Samu> the code that corrects ship speed
19:11:38 <Samu> /* Purchase cost - Max speed */
19:11:38 <Samu> uint raw_speed = e->GetDisplayMaxSpeed();
19:11:38 <Samu> uint ocean_speed = e->u.ship.ApplyWaterClassSpeedFrac(raw_speed, WATER_DEPTH_MAX);
19:11:38 <Samu> uint canal_speed = e->u.ship.ApplyWaterClassSpeedFrac(raw_speed, WATER_DEPTH_MIN);
19:11:44 <Samu> the speeds displayed in gui
19:14:42 <Samu> took me 4 hours to get my maths working the way I wanted
19:15:11 <Samu> the effect is beautiful in my opinion
19:15:34 <Samu> ships move slower/faster, depending on the distance they are to the shore
19:15:44 <Samu> they're no longer pegged to max speed constantly
19:15:53 *** andythenorth has joined #openttd
19:17:13 <glx> oh but it's using engine data so it should be good
19:25:15 <snail_UES_> hey andythenorth, what did you end up deciding about your steamers’ animation?
19:25:37 <snail_UES_> I see, so no rods either right?
19:25:47 <snail_UES_> what made you think so?
19:29:27 <andythenorth> general appearance
19:30:03 *** Progman has joined #openttd
19:30:42 <glx> hide the wheels behing a plate, better for air drag :)
19:33:31 <Samu> I wonder how many AIs will fail now thx to the depot requirements
19:34:02 <glx> well it's still a draft ;)
19:35:22 <andythenorth> glx: also means HE will explode on the wheels and eat the shell before it hits main armour
19:35:34 <andythenorth> oh wait, trains aren't tanks? :o
19:35:59 <crazystacy> what's all this talk about trains?
19:38:33 <snail_UES_> andythenorth: so visible, but motionless rods? 9.9
19:39:19 <andythenorth> I'm not happy with steam engine sprites yet, I need to learn how to draw them
19:44:11 <snail_UES_> steamers are all about shading :)
19:45:05 <snail_UES_> it seems you made the 2nd pixel from top as the brightest one...
19:45:19 <glx> or the multiple definitions for ShowMacDialog() in macos.mm should check the video driver
19:45:23 *** WormnestAndroid has joined #openttd
19:45:41 <snail_UES_> I’d suggest to move it down to the 3rd, because of the way we “view” the world in the TTD perspective. It’s quite from the top
19:45:50 <DorpsGek_III_> - Update: Translations from eints (by translators)
19:46:06 <snail_UES_> I used to do the same as you in my older sprites and are changing now.. it looks better (of course, only for the big steamers)
19:48:22 <snail_UES_> also, 45mph for NG steamers in 1860 is just beyond insane :D :D
19:48:44 <snail_UES_> 35 would be more realistic… and even then it’d be “too” fast
19:51:40 <Samu> i'm making buoys relatively inneffective with these speed changes
19:54:03 <andythenorth> the tank wagons are correct, the steam engines are wrong
19:55:01 <Samu> AIs will have to trust pathfinders more now than before, and not relly on buoys... funny how things have turned
19:55:51 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 opened pull request #7934: -Fix: [OSX] Don't show a crash/assertion message box for a GUI-less video driver. https://git.io/JvfSd
19:56:13 * glx likes to write code he can't test
19:56:48 <andythenorth> so now I have to crash openttd on macos? :P
19:59:42 <snail_UES_> andythenorth: yep. I’m doing the same work on my steamers
20:00:52 <andythenorth> I have learnt many tricks to prevent diesels looking like rectangles
20:01:01 <andythenorth> but steam...not so much
20:05:45 <snail_UES_> steamers require exercise
20:05:57 <snail_UES_> more than shaping, it’s about shading
20:15:46 <TrueBrain> right ... just shut down the k8s cluster ... we will see if I forgot anything :D
20:20:52 <TrueBrain> now "all" we have to do is port OpenTTD's CI to GitHub Actions :D
20:21:12 <TrueBrain> with the "cache" step it also means the ugly process of the Windows files can be solved a lot nicer
20:24:13 <DorpsGek_III_> [OpenTTD/website] TrueBrain opened pull request #144: Remove: [AzurePipelines] cleanup as we switched to GitHub Actions https://git.io/Jvf9r
20:25:20 <DorpsGek_III_> [OpenTTD/binaries-proxy] TrueBrain opened pull request #11: Remove: [AzurePipelines] cleanup as we switched to GitHub Actions https://git.io/Jvf9K
20:28:36 <DorpsGek_III_> [OpenTTD/binaries-proxy] TrueBrain updated pull request #11: Remove: [AzurePipelines] cleanup as we switched to GitHub Actions https://git.io/Jvf9K
20:28:46 <TrueBrain> hmm .. I found a bug in the deployment flow .. hmm ... time to figure out what broke :)
20:30:25 <DorpsGek_III_> [OpenTTD/DorpsGek] TrueBrain opened pull request #19: Add: badges to indicate the state of this repository https://git.io/Jvf9Q
20:36:17 <DorpsGek_III_> [OpenTTD/binaries-proxy] LordAro approved pull request #11: Remove: [AzurePipelines] cleanup as we switched to GitHub Actions https://git.io/JvfHf
20:36:24 <DorpsGek_III_> [OpenTTD/DorpsGek] LordAro approved pull request #19: Add: badges to indicate the state of this repository https://git.io/JvfHJ
20:36:31 <DorpsGek_III_> [OpenTTD/website] LordAro approved pull request #144: Remove: [AzurePipelines] cleanup as we switched to GitHub Actions https://git.io/JvfHU
20:37:40 <DorpsGek_III_> [OpenTTD/website] TrueBrain merged pull request #144: Remove: [AzurePipelines] cleanup as we switched to GitHub Actions https://git.io/Jvf9r
20:37:47 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro merged pull request #7914: Town, station and industry directory window sorting performance improvements https://git.io/Jej92
20:38:36 <DorpsGek_III_> [OpenTTD/binaries-proxy] TrueBrain merged pull request #11: Remove: [AzurePipelines] cleanup as we switched to GitHub Actions https://git.io/Jvf9K
20:38:58 <DorpsGek_III_> [OpenTTD/DorpsGek] TrueBrain merged pull request #19: Add: badges to indicate the state of this repository https://git.io/Jvf9Q
20:39:23 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro approved pull request #7791: GS method to control engine availability for a specific company https://git.io/JvfHO
20:41:58 <LordAro> when do we think we should do beta3/RC1 ?
20:47:53 <LordAro> well yes, but taking into account that beta2 was 3 weeks ago
20:48:20 <andythenorth> I have played 150 years of beta 2
20:50:59 <TrueBrain> LordAro: the biggest question is: is 1.10 feature complete?
20:51:03 <TrueBrain> or is there anything else pending?
20:57:48 <milek7_> does any crash for corrupted savegame is reportable bug?
20:57:53 <milek7_> or it is limited to savegames that could be reasonably generated with game?
20:59:08 <LordAro> milek7_: depends how easy it is to fix :p
21:00:58 * LordAro puts andythenorth back in his box
21:02:24 <TrueBrain> I see that the ICU keeps being pushed backwards :D
21:02:30 <TrueBrain> or forwards, what-ever :P
21:03:20 <LordAro> turns out it's quite hard to rewrite the rendering engine
21:03:49 <LordAro> i don't feel it'll be done for 1.10, that's for sure
21:04:28 <TrueBrain> so you can move that to 1.11 already :)
21:04:32 <dekeract[m]> LordAro: text rendering? Is it why industry names are too small in the map gui?
21:05:00 <LordAro> you can blame Chris Sawyer for that one
21:05:16 <TrueBrain> lol, a PR from may? Maybe just close that .. clearly nobody cares enough :)
21:05:52 <nielsm> things I'd like to get in 1.10:
21:05:56 <TrueBrain> okay, fixed a nasty bug that pushed "master" to "production" .. that was an oopsie :) Luckily for "website" "master" was in fact a tag :)
21:06:08 <LordAro> nielsm: pls milestone
21:07:01 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro commented on issue #6922: Replace ICU ParagraphLayout with something else? https://git.io/JvfQ3
21:07:43 <LordAro> nielsm: what are your thoughts on #7870 ?
21:07:59 <LordAro> it's going to create an awful lot of translator churn
21:08:47 <milek7_> LordAro: well, i'm asking because it's rather easy to generate crashing savegames with afl..
21:08:55 <milek7_> but honestly I don't think it is worth to fix all these ;P
21:09:15 <milek7_> otoh some of it could be RCE when connecting to malicious server
21:09:34 <TrueBrain> milek7_: if you run AFL on network traffic, now that is interesting :D
21:10:32 <TrueBrain> we once run a fuzzer for over a week on network level
21:10:36 <nielsm> LordAro: as I wrote in my latest comment, I don't particularly like it, and I guess my main reason is that I think it's unnecessary
21:10:38 <TrueBrain> not sure if that was ever repeated
21:10:58 <nielsm> the style of the original tooltips and GUI in general is part of the flavour of the game
21:12:16 <nielsm> OTTD is a modernised retro game, while the contents is getting updates and improvements, the chrome should generally stay like something from 1994
21:12:29 <LordAro> nielsm: i know what you mean
21:12:39 <LordAro> still, there should be some consistency
21:12:52 <LordAro> basically, it needs breaking down into logic changes that we can review individually? :D
21:13:11 <nielsm> yeah it might be easier to take if it's broken into smaller pieces
21:13:18 <TrueBrain> hmm .. if we make RC, does that mean cmake goes to master? As that would enable LGTM :D
21:13:40 <LordAro> milek7_: the issue is we don't really have any good way of recovering from invalid saves - most of the "correctness" is guarded by asserts, which obviously have no way to recover from, and are disabled in releases
21:14:02 <LordAro> TrueBrain: assuming it's actually mergable, which i'm pretty sure it isn't at this point :p
21:14:13 <glx> TrueBrain: once 1.10 branching is done we can merge cmake yes :)
21:14:21 <TrueBrain> LordAro: I am sure glx disagrees :P
21:14:21 <glx> maybe I need to rebase again
21:14:34 <nielsm> I still haven't tried the cmake branch myself...
21:14:43 <nielsm> have anyone tried how well it integrates with VS build?
21:14:50 <TrueBrain> worked like a charm for me
21:15:06 <TrueBrain> VS2017 needs a bit of work (just a tiny bit)
21:15:18 <nielsm> so VS2015 is getting dropped too?
21:15:27 <nielsm> or, cmake probably has a generator for project files for it
21:15:39 <nielsm> 2015 is still supported, it should have all the C++11 parts we need
21:15:59 <TrueBrain> ah; it was said that it was going to be dropped anyway, when we started with CMake .. so I did not test VS2015 :)
21:16:10 <TrueBrain> honestly, supporting N and N-1 should really be sufficient
21:16:18 <nielsm> yeah there isn't much reason in supporting 2015 still
21:16:19 <TrueBrain> cmake will also drop GCC<old> and shit :P
21:16:29 <Samu> dbg: [sprite] Tried to load character sprite #242 as a normal sprite. Probable cause: NewGRF interference
21:16:45 <Samu> i'm getting tons of '?' over sea tiles
21:16:46 <nielsm> I think 2017 is the last version to have the winxp std libraries
21:16:46 <glx> but cmake -G "Visual Studio 2015" should work
21:17:08 <glx> it's just that since 2017 cmake is integrated in VS
21:17:19 <TrueBrain> lol, cmake even supports VS2008
21:17:26 <glx> and support is even better in 2019
21:17:47 <nielsm> Samu, just disable/remove that block that draws water depths in debug builds
21:17:53 <TrueBrain> milek7_: can you share the patch you used to get afl working? (how are you targeting savegames, basically)
21:18:01 <nielsm> it's not supposed to be there forever anyway, just until a better method comes along
21:18:19 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro commented on pull request #7729: Add #7525: Allow autoreplace with same model vehicle https://git.io/JvfQV
21:18:23 <glx> andythenorth: I could probably add a temp commit to force a crash during regression
21:18:32 <TrueBrain> but yeah, I know LGTM shows a few places where corruption can happen .. so that might be really interesting to see :)
21:19:04 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro commented on pull request #7934: -Fix: [OSX] Don't show a crash/assertion message box for a GUI-less video driver. https://git.io/JvfQr
21:19:30 <TrueBrain> btw, LordAro, "glx22 force-pushed the TrueBrain:cmake branch from 80db1be to 1f5ecea 21 days ago"
21:19:42 <glx> I know LordAro, I know, I copy/paste the win32 message
21:19:47 <TrueBrain> but currently it indeed doesn't merge :P
21:20:02 <TrueBrain> *stupid beta2 release*
21:20:34 <milek7_> i just ugly patched loading intro code to take filename from cmdline
21:20:35 <TrueBrain> I am so happy everything is running in AWS now :D W00p \o/
21:20:53 *** crazystacy has joined #openttd
21:21:11 <TrueBrain> now some old stuff to migrate ...
21:21:32 <TrueBrain> milek7_: oof, so that is going really slow? :D
21:21:34 <LordAro> TrueBrain: be good to get the noai docs moved, they've looked a bit sad for a while
21:21:47 <glx> so should I crash test #7934 (ie add an assert(false) somewhere in the source) ?
21:21:50 <TrueBrain> that should be easy now I guess ..
21:22:34 <TrueBrain> so .. we used to generate the docs of the latest major/minor
21:22:42 <milek7_> and placed _AFL_INIT() before it to save some time on initialization, instrumentation with afl-clang-fast
21:22:53 <nielsm> btw when are we adding crypto? :D signed savegames (detect tampered and warn about saves from untrusted people), encrypted multiplayer, secure and automatic authentication to companies in multiplayer
21:23:04 <milek7_> and of course using uncompressed savegames
21:23:07 <TrueBrain> nielsm: and HTTPS connections .. yes plz
21:23:12 <milek7_> it runs around 4 runs/s per core
21:23:17 <TrueBrain> milek7_: haha, yes, that is kinda required :P
21:23:24 <TrueBrain> yeah, okay, so really slow; fair enough :)
21:23:25 <LordAro> basic md5 would be a good start for working out whether a save is corrupted or not
21:23:54 <LordAro> it's perfectly good as a checksum
21:24:00 <glx> md5 stored in the savegame ?
21:24:08 <TrueBrain> nielsm: we never added a crypto library, as that kinda grows OpenTTD .. but given its size nowedays, I guess nobody would really notice :)
21:24:13 <LordAro> we're not going to get collisions from savegames
21:24:16 <milek7_> at least CRC32 doesn't pretend it is for security ;P
21:24:16 <nielsm> well corrupted saves usually mean the game writing the save wrote exactly what it intended to and you received that exact file, but the basis for writing the savegame was corrupted already
21:24:18 <LordAro> not ones we care about, anyway
21:24:22 <nielsm> or there was a bug in the version writing it
21:24:28 <LordAro> milek7_: CRC32 works too
21:24:56 <nielsm> with compressed saves, we already have CRC really
21:25:09 <nielsm> unless something decompressed the save and altered it then recompressed
21:25:10 <TrueBrain> adding another hash won't help you
21:25:26 <TrueBrain> if you can decompress a save, you can rewrite the hash
21:25:46 <TrueBrain> but really, https would be very useful to have
21:25:47 <andythenorth> oof big cdist network :)
21:25:53 <TrueBrain> hmm .. cloud-saves ..... hmmmmmm
21:25:59 <TrueBrain> when do we want docs ..
21:25:59 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro approved pull request #7688: Feature: Auto-restart loads the original resources again https://git.io/JvfQp
21:26:01 <nielsm> my idea is actually to generate a private key in a separate file in the user's config/profile folder for ottd, and use that for several authentication purposes
21:26:02 <andythenorth> negative transfer income for vehicles :P
21:26:04 <TrueBrain> should a beta generate 1.10 docs already?
21:26:07 <andythenorth> killing my trains
21:26:23 <LordAro> TrueBrain: i'd be tempted to say no
21:26:44 <TrueBrain> so ... I need an S3 bucket
21:26:55 <TrueBrain> I need to know when a stable release is created
21:26:59 <TrueBrain> and extract the docs there
21:27:37 <andythenorth> that would be a nice reusable thing :P
21:28:09 <TrueBrain> if I remember correctly, we patch up the doxygen files for it to work on the web like this ..
21:28:33 <Samu> difficult to see where depth 300 is :p
21:30:15 <TrueBrain> yeah, a 'sed' is used to add some custom text
21:30:22 <Samu> im not sure how it actually works, seems that a little tiny island
21:30:37 <Samu> can affect a buttload of
21:31:08 <TrueBrain> okay, that addition sounds a bit pointless, honestly
21:32:47 <milek7_> from quick looks on results, most are crashes on assertions, but there are some sneaky corruptions too (runs for few seconds and segfaults) or double-frees
21:32:59 <TrueBrain> so, the only question is: who is going to do the extracting of the docs tarball to an S3 .. hmm .. interesting
21:33:16 <nielsm> Samu well yes, usually islands aren't a narrow pillar of rock in a deep ocean :P
21:33:24 <nielsm> they tend to have a bunch of support around
21:33:27 <TrueBrain> milek7_: the latter two can be interesting. I guess the first won't get much attention currently
21:33:51 <Samu> what have you done to terragenesis?
21:34:15 <Samu> seems that the water isn't at it's full depth
21:34:17 <milek7_> i'm not really sure if it double-free really but messages like "free(): invalid next size (fast)" suggest it ;P
21:36:08 <DorpsGek_III_> [OpenTTD/OpenTTD] frosch123 commented on pull request #7868: Feature: NewGRF callback profiling https://git.io/Jvf7O
21:36:40 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 updated pull request #7934: -Fix: [OSX] Don't show a crash/assertion message box for a GUI-less video driver. https://git.io/JvfSd
21:37:02 <glx> added a crash test, will be faster than waiting for an OSX user ;)
21:37:06 <Samu> everything starts at depth 0?
21:39:43 <frosch123> TrueBrain: it's more useful for noai and nogo
21:41:54 <TrueBrain> yeah, so I think I am going to recycle docs.openttd.org, and put them both on there
21:41:58 <TrueBrain> as that makes it a bit easier
21:43:16 <TrueBrain> on one side I can make a Docker, which collects all the docs of all the releases, and host that
21:43:26 <TrueBrain> on the other side, I can make an S3 bucket and put it on there
21:44:29 <TrueBrain> currently we store docs of every release we ever made .. (well, unless someone forgot to publish a version, ofc). Do we really care about older docs?
21:44:33 * andythenorth is starting to favour S3 :P
21:44:35 <TrueBrain> (to be online, that is)
21:47:19 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 updated pull request #7934: -Fix: [OSX] Don't show a crash/assertion message box for a GUI-less video driver. https://git.io/JvfSd
21:48:03 <frosch123> TrueBrain: i guess docs for master are enough, but it would be nice if old urls would redirect there
21:48:20 <TrueBrain> master? Or latest release?
21:48:52 <TrueBrain> as if I would be to build a NoAI, I would want the docs of the latest release, not the current development version .. as it might have things I cannot use?
21:49:00 <frosch123> the main purpose of hosted docs is to get linked from answers on forums/reddit/irc/whatever
21:49:25 <TrueBrain> so it needs to rebuild every night, basically
21:49:31 <frosch123> if we care we can add @since to all api functions
21:49:38 <frosch123> not sure whether we already do that
21:50:23 <TrueBrain> frosch123: we do not; but that might be a nice addition
21:50:29 <TrueBrain> glx: the changelog would be too far away, honestly
21:50:31 <glx> but adding a tag in the doxygen seems nice yes
21:51:01 <TrueBrain> annoying job to catch up now, but it is just a matter of doing
21:51:39 <TrueBrain> okay, so that makes this pretty simple tbh .. just needs doing, ofc :)
21:51:56 <TrueBrain> and doxygen looks very ugly btw
21:52:19 <LordAro> you're welcome to put some effort into styling it :p
21:52:43 * andythenorth happy to update links in the wiki if urls change
21:52:46 <TrueBrain> I would rather burn it, drown it, launch it into space, and rip it apart
21:52:49 <TrueBrain> but that is just my impuls
21:52:54 <frosch123> TrueBrain: problem with docs only for "master" is that not all doxygen links are permanent
21:53:11 <frosch123> no idea where that hash comes from, and what it means
21:53:14 <frosch123> and how stable it is
21:53:25 <andythenorth> also we'd ideally publish beta / RC docs no?
21:53:30 <TrueBrain> owh, that is just bullshit, holy crap ..
21:54:18 <glx> oh nice with my fix OSX regression output even shows useful info hidden from other platforms (basically it displays the crashlog)
21:54:43 * andythenorth needs daylength 1.5 or so
21:54:50 <andythenorth> maybe I just scale in the newgrf
21:56:07 <TrueBrain> frosch123: so publishing releases-only solves that issue, I guess
21:56:49 <frosch123> let me try something, it may just be a doxygen setting
21:58:04 <frosch123> noai tournaments are a hobby among teenagers
21:58:14 <andythenorth> snail_UES_: daylength via fake intro dates? o_O
21:58:47 <snail_UES_> andythenorth: introducing a 1950 engine in year 2000?
21:59:17 <TrueBrain> it was fun to build that tournament stuff :) Some AIs did weirdly well :)
21:59:29 <TrueBrain> frosch123: okay, let me know if you can fix this hash stuff :) Would be nice :)
21:59:51 <snail_UES_> the problem with daylength is the rebalancing of incomes and costs… it involves the whole game, not just a newgrf
22:00:08 <snail_UES_> though I do hate seeing steamers getting extinct so soon :p
22:00:25 <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro approved pull request #7934: -Fix: [OSX] Don't show a crash/assertion message box for a GUI-less video driver. https://git.io/Jvf7y
22:00:48 <snail_UES_> and the other issue is, I use the “total town replacement” newGRF that evolves the looks of houses etc. as time goes by
22:01:15 <snail_UES_> so it wouldn’t work… you’d still have 1890s steamers running across a town with modern skyscrapers
22:01:50 <snail_UES_> I sometimes use the “year cheat” in my games, reverting back to the previous year on jan 1st :p
22:02:13 <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 merged pull request #7934: -Fix: [OSX] Don't show a crash/assertion message box for a GUI-less video driver. https://git.io/JvfSd
22:04:37 <frosch123> TrueBrain: apparently it is a md5sum of the function signature, so is stable as long as the function does not change
22:05:07 <TrueBrain> so master-only it is
22:05:11 <TrueBrain> we can iterate from there
22:05:30 <andythenorth> hmm, house dates could be adjusted
22:06:08 <andythenorth> costs, income etc wouldn't affect my grfs significantly
22:07:17 <snail_UES_> house date could only be adjusted by editing the TTRS newGRF… right? even bridge introduction dates etc
22:08:47 <frosch123> 1.9 docs have the same hashes as my local docs
22:08:54 <frosch123> but the 1.8 docs have different ones
22:09:12 <frosch123> the 1.8 docs also have a very different layout, so maybe it is specific to the doxygen version or something
22:11:11 <frosch123> at least they are not random
22:11:16 <andythenorth> what we need is a patch that lies to the newgrf date var
22:11:34 <TrueBrain> so ... S3 websites doesn't support https for custom domains ...
22:11:42 <andythenorth> oof that's very lame
22:12:17 <TrueBrain> so it needs CloudFront in front of it
22:14:09 * andythenorth was just reading that
22:14:14 <andythenorth> wtf is cloudfront even
22:14:19 <andythenorth> I love the cloud, I hate the cloud
22:17:07 <LordAro> Computers 👏 were 👏 a 👏 mistake
22:17:19 <TrueBrain> never do that again plz
22:17:34 <TrueBrain> hmm ... so basically creating a Docker and running it with ECS is easier :P
22:17:53 <TrueBrain> (as I have that infrastructure ready)
22:22:36 <TrueBrain> meh; one way or the other, I need to solve this S3 issue anyway, as I want to move the binaries to there too ...
22:22:45 <TrueBrain> right, something to think about
22:48:57 <Samu> crap, i had a but in those speed_fracs
22:49:09 <Samu> speed_frac = min(255, speed_frac);
22:49:22 <Samu> i had 256, but 255 is what it should be
22:50:47 <Samu> in fact, i don't think i even need to min anymore
23:24:37 <Samu> confirmed, don't need the min(255, speed_frac); at all
23:39:58 *** hythlodaeus has joined #openttd
23:44:37 <Samu> woah, the station patch is in!
23:46:57 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh commented on pull request #7729: Add #7525: Allow autoreplace with same model vehicle https://git.io/JvfFD
23:47:23 <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh merged pull request #7688: Feature: Auto-restart loads the original resources again https://git.io/fj7vr
23:48:07 <andythenorth> nielsm: if sea has depth....can I lower land below sea level? o_O
23:49:13 <nielsm> andythenorth: not really no, but otoh the OTTD map model does allow water at any tile height
23:49:31 <nielsm> sea, river and canal tiles are really not very different at all
23:49:39 <Samu> btw nielsm, your terragenesis is producing weird depth levels
23:50:10 <nielsm> I think in the original ttdpatch implementation, rivers and canals were only distinguished from sea from being at above level zero
23:50:36 <nielsm> samu depends on height settings
23:51:26 <Samu> hmm i think i had alpinist, not sure
23:51:41 <nielsm> try alpinist with medium or high sea level
23:56:22 <DorpsGek_III_> [OpenTTD/OpenTTD] sheepo99 commented on pull request #7729: Add #7525: Allow autoreplace with same model vehicle https://git.io/JvfbJ
continue to next day ⏵
![https://i.imgur.com/bQg4Uvm.png[/img]](https://i.imgur.com/bQg4Uvm.png%5B/img%5D){kind=link}
![https://i.imgur.com/htW9X8e.png[/img]](https://i.imgur.com/htW9X8e.png%5B/img%5D){kind=link}
![https://i.imgur.com/Qt3WSWO.png[/img]](https://i.imgur.com/Qt3WSWO.png%5B/img%5D){kind=link}
