IRC logs for #openttd on OFTC at 2020-12-18
⏴ go to previous day
00:10:04 *** rptr_ has quit IRC (Quit: Leaving)
00:18:30 <DorpsGek> [OpenTTD/grfcodec] glx22 requested changes for pull request #9: Fix various issues with MinGW build https://git.io/JL8Gc
00:26:03 <LordAro> glx: you're lucky i'm still away
01:47:25 *** y2kboy23 has joined #openttd
01:56:15 *** rptr_ has quit IRC (Quit: Leaving)
02:05:07 *** JustSpectating has joined #openttd
02:08:12 *** JustSpectating has quit IRC ()
03:41:48 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
04:33:27 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
07:02:39 *** sla_ro|master has joined #openttd
08:07:06 *** andythenorth has joined #openttd
08:34:07 <LordAro> that is not a good issue
08:34:14 <LordAro> might be a decent discussion
08:34:32 <LordAro> but even then, very generic
08:35:16 *** WormnestAndroid has quit IRC (Remote host closed the connection)
08:35:49 *** WormnestAndroid has joined #openttd
09:50:27 <Timberwolf> When I'm drawing roads I can't help but notice a subconscious influence from all those 1970s "Roadmaker" sets my parents used to pick up at car boot sales.
10:24:46 *** HerzogDeXtEr has joined #openttd
11:20:28 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick updated pull request #8009: Change #8001: Don't add docking tile cost when ships are still too far from their destination https://git.io/Jv41T
11:21:40 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick updated pull request #7918: Fix 3c047b1: AIGroup.GetProfitLastYear could get values different than those displayed in gui https://git.io/Jvek6
11:23:09 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick updated pull request #7890: Fix #6452: Reset only editable and visible settings from GUI https://git.io/Jexqi
11:24:29 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick updated pull request #7193: Fix #6468: Load correct version of AI as specified during the time of its save. https://git.io/fhHI1
11:32:08 *** Samu has quit IRC (Read error: Connection reset by peer)
12:12:57 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
12:30:03 *** Samu has quit IRC (Read error: Connection reset by peer)
12:40:20 <TrueBrain> LordAro: we can migrate it to a discussion for now
12:43:08 <andythenorth> Timberwolf were they flexible plastic (pvc / nylon?) roadpieces connected by hard plastic t-junctions and joiners?
12:49:38 <Timberwolf> Those were the Matchbox ones, iirc.
12:49:43 <Timberwolf> Roadmaker was cardboard!
12:50:18 <Timberwolf> Proper mid 1960s austerity design (I think it was made for a few years after, most of the boxes I had were '70s colours and typography)
12:50:35 <Timberwolf> It tended to not last long in our house as the dog liked to steal and chew it.
12:56:54 <andythenorth> I must have forgotten the photoshoot
12:57:49 <Timberwolf> The '70s version looks like my "template" road :)
12:59:35 <andythenorth> yeah we had a bit of both I think
12:59:51 * andythenorth steps away from the nostalgia
13:25:15 <TrueBrain> right, I have 3 AWS contacts now as result of the blog-post .. my list is growing :D
13:25:31 <TrueBrain> I rather leave that to your wife, but what happened?
13:26:33 <andythenorth> I will EXPLAIN LATER
13:26:37 <andythenorth> it's very boring
13:31:26 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JLBtG
13:31:51 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JLBt4
13:35:49 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JLBqY
13:38:03 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JIiDr
13:41:12 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8371: Add: [Actions] release workflow for both releases and nightlies https://git.io/JLBmg
13:43:32 *** frosch123 has joined #openttd
13:43:41 <TrueBrain> okay, spend an hour replying to email, and cleaned up my open PR-list ... that is one set of chores done :D
13:43:46 <TrueBrain> a frosch123 here? Omg!
13:44:36 <andythenorth> lol discord handbags
13:44:48 * andythenorth has better things to do, closed that
13:47:59 <frosch123> yeah, it's weekend now, but i have to make sure my food doesn't burn
13:48:39 <TrueBrain> touchy subject, food :)
13:48:57 <TrueBrain> (get it .. it is easily burnt?)
13:51:05 <Samu> question: are these languages new in openttd? persian, ido, macedonian, frisian?
13:54:49 <Samu> wait, I think I understand. I removed a string in one of my PR's from all language files, then 9 months later, these 4 files have the string again. They were missing a translation 9 months ago, and now they do
14:01:07 <Samu> they are unfinished languages
14:01:24 <Samu> aren't they being tracked by the compiler?
14:03:07 <Samu> by the language manager thing you guys got there?
14:05:23 <TrueBrain> (migrated ^^ to discussion)
14:09:43 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on issue #8389: HDR rendering (12-bit, 16-bit, high brightness) ! https://git.io/JLGT4
14:20:21 <andythenorth> the FML is twitter broke serving for non-JS clients
14:20:43 <andythenorth> the website validation tool we use is non-JS
14:20:51 <andythenorth> so now every page with a link to twitter is reporting as broken
14:20:56 <andythenorth> on a Jenkins job
14:21:06 <andythenorth> and it's the last day I intend to work in 2020
14:21:33 <andythenorth> it's like the most uninteresting test failure ever
14:27:30 <frosch123> poor food. it got eaten. i don't think it can recover from that
14:36:32 <andythenorth> you can always try and rebuild it
14:38:19 <DorpsGek> [OpenTTD/OpenTTD] SamuXarick updated pull request #7822: Fix #7670: Cache the origin tile to prevent recurring calls to the road pathfinder when a vehicle is blocked by another https://git.io/JegbJ
14:43:44 <Samu> I get varying degrees of performance depending on the build, regarding the bug in #7670
14:44:12 <Samu> in 1.10.3, i get 20 ms, in master, i get 47 ms
14:45:35 <Samu> 20 ms makes the game run at 33 fps
14:46:08 <Samu> how do i build without asserts?
14:46:16 <Samu> in visual studio, just wanted to test
15:23:31 *** Progman has joined #openttd
15:27:29 <frosch123> hmm. ottd coding style is just too inconsistent to throw clang-format onto it :)
15:27:44 <LordAro> frosch123: i've tried this in the past
15:27:55 <frosch123> i know. i wanted to try again
15:28:09 <TrueBrain> so change the coding style!
15:28:21 <TrueBrain> I mean ..... personally, having a clang-format that the CI validates is easier
15:28:27 <TrueBrain> than doing it manually every single PR :P
15:28:36 <TrueBrain> (something something bots taking over :P)
15:28:52 <frosch123> it just breaks every pr and fork :)
15:29:13 <LordAro> there's too many manual alignment things that get ruined by it
15:29:14 <TrueBrain> well, forks is another issue, but we can help them
15:29:23 <LordAro> even if you exclude the tables folder :p
15:29:25 <TrueBrain> as it is the same as I do with black for bigger projects
15:29:31 <TrueBrain> I replay forks, blacking every commit
15:29:33 <TrueBrain> and that solves itself
15:34:47 *** m1cr0man has quit IRC (Ping timeout: 480 seconds)
15:36:53 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7962: Improve rendering of large viewport areas https://git.io/JLBiA
15:37:46 <TrueBrain> I was so tempted to just blindly merge it, and say: any problems, you fix it :P :D
15:37:55 <LordAro> we tried that once before
15:38:06 <TrueBrain> I saw the errors of my ways before I did :)
15:38:22 <TrueBrain> with what did you try it btw?
15:38:50 <LordAro> or rather, a previous version
15:38:55 <TrueBrain> so more caution is needed!
15:48:42 <milek7> TrueBrain: there was talk about it few days ago
15:55:06 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler opened pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
16:03:40 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
16:18:43 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBHp
16:32:14 <DorpsGek> [OpenTTD/OpenTTD] Milek7 commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLB5w
16:32:22 *** daspork has quit IRC (synthon.oftc.net singleton.oftc.net)
16:32:35 *** daspork has joined #openttd
16:39:50 <DorpsGek> [OpenTTD/OpenTTD] Milek7 commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBFU
16:50:10 <DorpsGek> [OpenTTD/OpenTTD] Milek7 commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBbD
16:50:32 *** Wormnest has joined #openttd
16:54:21 <Afdal> Perhaps it's worth asking in here too: Would a kind NUTS user happen to have a previous version in the v0.8.x series of releases?
16:56:10 <_dp_> 48410101|F7FDD4541807DFEF0FD669FB73BAE010|nuts_unrealistic_train_set-0.7.8/nuts.grf = 0
16:56:32 <_dp_> should be enough to get it from bananas
16:57:06 <Afdal> Lookin for v0.8.0, 0.8.1, 0.8.2, and 0.8.3 specifically
16:57:24 <_dp_> ah, nvm then, I though you need before 0.8
16:57:34 <Afdal> How do you grab something from BaNaNaS with an md5 sum though?
16:58:01 <_dp_> you can do it with console commands I think
16:58:55 <frosch123> try to find a coop savegame that uses it
16:59:01 <frosch123> then you can get the missing grfs
16:59:13 <Afdal> That's a good idea, already tried it though -_-
16:59:26 <Afdal> they jumped from 0.7.8 straight to 0.8.4
17:00:27 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBAN
17:02:46 <frosch123> what's so special about those versioins?
17:03:34 <Afdal> They may or may not be before a nasty bug was introduced that I'm trying to track down
17:03:36 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBxE
17:04:56 <Afdal> if they were pre-bug introduction, it'd be nice to play with NUTS after some of the major rebalances were done in v0.8.0
17:05:14 <Afdal> Otherwise I'm stuck using v0.7.8
17:05:22 <frosch123> you know that those version numbers mean nothing? :)
17:05:38 <frosch123> 0.7.8 -> 0.7.9 -> 0.8.0
17:05:54 <frosch123> 0.8.0 is no ground-breaking change
17:06:09 <Afdal> it had some pretty massive balancing changes actually, check out the changelog
17:06:56 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBxp
17:06:56 <frosch123> while there were two years between 0.7.9 and 0.8.0
17:07:31 <Afdal> Oh wait, didn't realize there was a v0.7.9 as well. I'd love to get my hands on that too.
17:07:42 <milek7> you could just build it from source, I guess?
17:08:12 <frosch123> looks like V forgot to push after 0.8.1+ :)
17:08:31 <andythenorth> V453000 was in discord yesterday
17:08:45 <andythenorth> NO HEADSPACE FOR OPENTTD V453000
17:08:53 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
17:09:10 <Afdal> V453000 hasn't responded on IRC for months since he went full Factorio artist
17:09:27 <Afdal> does he actually talk on Discord
17:09:30 <frosch123> he hasn't evern written a fff in 4 weeks, that slacker :)
17:09:33 <andythenorth> he did yesterday
17:09:54 <Afdal> Tell him to stop using Discord spyware and come back to his IRC buddies ;_;
17:12:16 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
17:14:20 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
17:15:56 <milek7> it utterly fails at chosing appropriate font color for label backgrounds
17:17:51 <milek7> black text on dark green, urgh
17:19:27 *** m1cr0man has joined #openttd
17:22:42 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
17:23:43 <andythenorth> I've discovered that OpenTTD hangs my mac UI for minutes at a time
17:23:47 <andythenorth> I thought it was Zoom
17:25:12 <milek7> with all these problems, I'm surprised you didn't ditch either openttd or mac yet :)
17:28:22 <andythenorth> both are challenging to drop
17:29:54 *** tokai|noir has joined #openttd
17:29:54 *** ChanServ sets mode: +v tokai|noir
17:32:06 *** daspork_ has joined #openttd
17:33:56 *** daspork has quit IRC (synthon.oftc.net singleton.oftc.net)
17:36:00 *** tokai has quit IRC (Ping timeout: 480 seconds)
17:37:56 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
17:39:13 <milek7> it's only me, or upper one is more readable than bottom?
17:39:28 *** m1cr0man has quit IRC (Ping timeout: 480 seconds)
17:42:33 <frosch123> there are formal tests for this
17:42:47 <frosch123> anyway, we can only configure one color. gh picks the other
17:43:22 <andythenorth> which is more readable will be highly specific to person
17:43:44 <andythenorth> the contrast for the green and the orange is poor with both the white and the black
17:44:00 <andythenorth> the white has better contrast on both, but in a way that some people will find jarring
17:46:41 <milek7> that's backwards to me
17:47:12 *** ChanServ sets mode: +v tokai
17:52:22 *** m1cr0man has joined #openttd
17:53:40 *** tokai|noir has joined #openttd
17:53:40 *** ChanServ sets mode: +v tokai|noir
17:58:57 <Samu> what is prettier to see? 0xFFFF or 65535 or 1 << 16 - 1?
17:59:18 <LordAro> depends on context :p
17:59:21 <dwfreed> which is more meaningful to the context
17:59:47 <Samu> p.second |= (0xFFFF * (a % b) / b) << 16;
17:59:54 <LordAro> the fact that you've even trying to use such a number would suggest that 0xFFFF would be cleaner
18:00:00 <Samu> p.second |= (65535 * (a % b) / b) << 16;
18:00:11 <Samu> p.second |= ((1 << 16 - 1) * (a % b) / b) << 16;
18:00:28 <LordAro> that's just what you're using it for
18:00:35 *** tokai has quit IRC (Ping timeout: 480 seconds)
18:00:37 <LordAro> also i'm pretty sure 1 << 16 - 1 is 1 << 15
18:00:53 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLRUV
18:01:45 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
18:01:49 <LordAro> nevermind gcc, the C specification disagrees
18:02:13 *** m1cr0man has quit IRC (Remote host closed the connection)
18:02:35 <Samu> i didn't know that :( i typed it in the calculator and got it right
18:03:05 *** m1cr0man has joined #openttd
18:05:14 <TrueBrain> tnx for the links LordAro !
18:05:37 <Samu> but it's a programming calculator :p
18:06:12 <Samu> p.second |= (((1 << 16) - 1) * (a % b) / b) << 16;
18:06:53 <LordAro> you haven't said what you're actually doing yet
18:07:00 <LordAro> it almost looks like you're setting a specific bit
18:08:17 <DorpsGek> - Update: Translations from eints (by translators)
18:08:57 <Samu> /* Extend the resulting remainder to a precision of 16 bits and store it in the same variable in the following manner:
18:09:19 <Samu> * bits 16..31: remainder, with an extended precision of 16 bits
18:09:59 <Samu> im working on an alternative to MoveGoodsToStation
18:11:33 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7962: Improve rendering of large viewport areas https://git.io/JLRkG
18:12:52 <TrueBrain> a typical help me help you reply of mine ^^ :)
18:14:27 <milek7> it somewhat improves unzoom performance at high resolutions
18:14:29 *** m1cr0m4n has joined #openttd
18:15:15 <dwfreed> Samu: your equation doesn't exactly make sense, though
18:15:37 <TrueBrain> for these kind of PRs, we need more context :) How does it perform on low res, for example, etc etc ... they are just a bitch to judge, basically .. which is a bit sad, as they are the most fun to make :D
18:16:28 <dwfreed> Samu: unless that's supposed to be fractional remainder, not integer remainder ?
18:17:18 *** m1cr0man has quit IRC (Ping timeout: 480 seconds)
18:17:49 <TrueBrain> I remember back in the days we did in the end not apply a patch that was a HUGE increase in performance ... for anyone with 2+ cores ... anyone with 1 core, it was a 15% drop in performance :P
18:18:04 <dwfreed> Samu: that is not a % b
18:18:39 <milek7> sounds like reasonable tradeoff ;P
18:19:34 <TrueBrain> milek7: if we do it again, yes, it would
18:19:44 <TrueBrain> back then, a small percentage of users had more than 1 core :)
18:19:59 <dwfreed> right, it is an integer numerator in the fraction x/65535 that is closest to (a % b) / b
18:20:10 <Samu> people also got confused the other time I tried this, dwfreed
18:20:28 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLRIK
18:20:57 <dwfreed> why not just put a % b there; presumably b is going to be available if the floating point value of a / b is meaningful
18:22:51 <andythenorth> when is livestream?
18:27:44 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLRLN
18:28:35 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLRtv
18:29:28 <Samu> another example: 10 / 8 = 1. Then 65535 * (10 % 8) / 8 = 16383
18:33:50 <Samu> i don't put a % b, because I won't know the values of a and b later on in the function. So I extend them all the common x/65535 fraction
18:35:41 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLRtp
18:37:24 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7028: Feature: Option to group vehicle list by shared orders https://git.io/JLRqI
18:37:46 <DorpsGek> [OpenTTD/OpenTTD] stormcone commented on pull request #8308: Feature: Add cargo filter support to vehicle list. https://git.io/JLRqm
18:40:11 <TrueBrain> as a pre-livestream party, https://github.com/OpenTTD/OpenTTD/pull/8354 , the more I think about it, the less I like it. It feels like a lot of magic ... like ... it is missing a game mechanic for this to make sense (so ignoring the implementation totally). Like, yellow signals, I can understand. But magically adjusting the speed to a train in front of you .. feels wrong. Is that just me? Opinions? :D
18:40:38 <andythenorth> it's mystery meat
18:41:02 <andythenorth> it probably looks cool, but the mechanic is weird
18:41:07 <TrueBrain> most of the train handling is very determined ... PBS signals, you can see reserved tracks .. red signals, trains stop, going up hill, you go slow, etc etc
18:41:22 <TrueBrain> deterministic, not determined, lol
18:45:07 <TrueBrain> frosch123: nice extension on the default suggestions ... :D Really nice :)
18:45:56 <TrueBrain> funny .. a Discussion mentioning an Issue does not generate a lnk
18:46:09 <TrueBrain> I am disappointing in you GitHub
18:52:13 <frosch123> took me some seconds to understand "default suggestions" :p
18:52:24 <frosch123> i ignore frequent suggestions :)
18:54:00 <TrueBrain> english good yes yes
18:54:09 <TrueBrain> owh, this will be hilarious on livestream
18:56:49 <milek7> 'real world' is not very convincing argument in ottd :D
18:56:50 *** Laedek has quit IRC (Ping timeout: 480 seconds)
18:57:02 <TrueBrain> if that is your argument, denied! :D
18:57:22 <TrueBrain> (as that was not my argument :P)
18:57:59 <TrueBrain> I should have kept it for the livestream, stupid stupid stupid
19:00:20 <frosch123> the "split answers into multiple parts" is very important :)
19:00:42 <frosch123> if the discussions were linear, they would be trash
19:00:45 <TrueBrain> having individual replies helps so much ...
19:01:22 <milek7> "train slows down because the train before him is slower" is rather easy to explain
19:01:24 <milek7> I have more issue with that it won't work when junctions are ahead
19:01:25 <frosch123> TrueBrain: it's the wrong strategy
19:01:54 <TrueBrain> milek7: from a gaming perspective, that is hard to explain .. as in: someone has to go like: wtf?, ask about it on the forum, and get an answer :)
19:01:59 <TrueBrain> that is a bad gaming mechanic :)
19:02:07 <TrueBrain> frosch123: in what sense?
19:02:09 <frosch123> the sprite sorter only works because it does almost nothing. the main ingredient to correct sprite order is that the landscape drawing traverses tiles from back to front
19:02:38 <frosch123> so any approach that tries to "sort" sprites usually destroys it
19:02:41 <andythenorth> oof sprite sorter
19:02:45 <andythenorth> one day I should figure out why 50% of FIRS flickers
19:02:56 <TrueBrain> I once understood the sorter .... that day is well behind me
19:03:01 <frosch123> the sprite sorting is a very unstable thing that only works by luck
19:03:05 <andythenorth> currently I solve the flickering by moving the viewport :P
19:03:19 <frosch123> i would rather attempt to put the sprite sorting and drawing into a different thread
19:03:23 <frosch123> and ignore how long it takes
19:03:56 <TrueBrain> so the argument is: the time of sprite sorting should not influence rendering time, how ever long/short it is?
19:05:04 <frosch123> you can speed up things by different means, than the spritesorter core loop
19:05:25 <frosch123> changing the sprite sorter itself is a lose-only strategy :p
19:05:52 <TrueBrain> what I learnt from years ago: don't touch these functions unless we really have to :P In other words: I am scared shitless :P
19:05:53 <milek7> otoh that original sorter is slooow and that works acceptably only because there is some legacy chunking that is otherwise not needed is.. icky
19:05:57 <frosch123> for example: the spritesorter has bad complexity in number of input sprites
19:06:19 <frosch123> but you can split the rendering area into smaller parts, and draw them independently
19:07:50 <frosch123> nothing stops you to just dump all sprites to draw into a list (already the case), sort them in a thread, and draw them on the next tick
19:08:32 <frosch123> there are many parts in the rendering code that could be parallelized, i.e. does not depend on gamestate after an initial "snapshot"
19:08:58 <TrueBrain> okay .. so on one hand, it is likely this PR does what it says, and it increases speed for at least a few cases. On the other hand, it needs a bit more checking on other hardware/constraints to validate it does for all. But from a concept point of view, we should focus more on reworking the whole engine, as it were.
19:09:55 <TrueBrain> is that a fair description?
19:10:37 <frosch123> i have no doubts that the PR speeds things up, also on other hardware
19:10:53 <frosch123> but whether it breaks sorting is nothing you can "review" by looking at the code
19:11:02 <frosch123> you have to play some weeks with it. and noone here plays :p
19:11:17 <TrueBrain> it runs on cmclient, what I understand; is that considered a good enough testbed?
19:11:37 <frosch123> there are some typical scenarios to check, like foundations, fences, diagonal tracks, bridges, ...
19:11:42 <andythenorth> do we have baselines for performance?
19:11:50 <andythenorth> and stable test cases?
19:11:52 <frosch123> rail catenary, signals, tram catenary
19:12:08 <TrueBrain> I think we gathered enough info to conclude it does increase performance, so that is fine
19:12:32 <TrueBrain> I am just trying to balance how much time and effort our already small review team needs to spend on this vs the reward it delivers
19:12:38 <TrueBrain> which is especially for this one a though nut to crack
19:13:00 <TrueBrain> as the blowback, looking at the past, is HUGE
19:13:12 <TrueBrain> (people do not take kind to errors in either the blitter or the sorter)
19:13:14 <FLHerne> Why not just ask _dp_?
19:13:20 <TrueBrain> we did, in the issue :)
19:13:29 <FLHerne> Yeah, but IRC works better
19:13:33 <TrueBrain> which is a very well defined answer :)
19:13:34 <frosch123> TrueBrain: that PR removes the splitting of the drawing btw. :) that is kind of the opposite of what i suggested for parallelizing :p
19:13:44 <TrueBrain> I do not know what to ask him more FLHerne , but feel free to ask him anything :)
19:14:03 <TrueBrain> frosch123: I do not understand the patch 1 bit, so that is a bit of an issue for me :)
19:14:17 <michi_cc> Another sprite sorter test: large aircraft that move on tiles with foundations.
19:14:36 <TrueBrain> michi_cc: lolz, reliving past memories now? :D
19:15:00 <michi_cc> I think we "fixed" that more than once by now.
19:15:09 <frosch123> TrueBrain: i do not either. but there is not much to understand about sprite sorting. the problem itself is fishy and undefined. so any code is as good as any other. you can just experiment with it
19:15:22 <frosch123> TrueBrain: you can also just merge it, and later revert it again
19:15:24 <TrueBrain> bug-compatible-code is never a good indication :D
19:15:38 <Xaroth> Now look what you made me do, LordAro :P
19:15:40 <TrueBrain> well, given that already happened once, what I have been told, I tried to balance this a bit more :)
19:15:57 <frosch123> TrueBrain: but in any case, when dp says it's tested on citymania, i do not trust that one bit. they have a very opinionated playstyle
19:15:59 <TrueBrain> but I love all this input :D
19:16:07 <frosch123> that excludes many features that others may use
19:16:14 <TrueBrain> but they too will spot sorter errors :P
19:17:07 <_dp_> you talk so fast I can't catch up :p but for one thing I didn't "change" sprite sorter, I just improved the performance
19:17:15 <_dp_> it does exact same thing, just faster
19:17:31 <_dp_> using structures instead of loops basically
19:17:36 <TrueBrain> bug-compatible :D (which I appreciate btw :P)
19:18:21 <_dp_> btw, there are still glitches with original sprite sorter even without newgrfs
19:18:48 <frosch123> _dp_: there will always be glitches
19:19:04 <frosch123> spritesorting is an invalid problem. it is not solvable
19:20:33 <TrueBrain> getting it to work correctly on Unity took me a few guides :D
19:21:44 <FLHerne> Xaroth: Needs deprecation warnings really, I have an idea for that...
19:22:39 <Xaroth> FLHerne: It does, but by the time I've figured out how/where to best implement them, so I left it with comments.
19:22:50 <Xaroth> Feel free to push to my PR
19:24:06 <TrueBrain> frosch123: I am still on the fence about it .. it sounds like: "it works, so why not", but it makes future work more difficult as you mentioned, but it might be that nobody is ever going to pick that up :D
19:25:01 <frosch123> nah, i don't think it makes future stuff more difficult
19:25:45 <TrueBrain> owh, you meant it is only the wrong direction for a future solution, but that is not a real issue in that sense?
19:25:46 <frosch123> anyway, if dp says the result is the same, it makes it easier to argue: merge it, and revert when that claim is disproven :)
19:26:13 <TrueBrain> doesn't happen often, that we trust PR authors sufficiently to YOLO this :)
19:26:21 <TrueBrain> _dp_: take it as a compliment :)
19:26:23 <frosch123> TrueBrain: the first commit in that PR makes a change, that would be reintroduced in the future, but probably in a different place
19:26:54 <frosch123> TrueBrain: it does not affect savegames, and it does not affect apis. so YOLO is fine
19:27:04 <frosch123> it's revertable without lasting damage
19:27:58 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #7962: Improve rendering of large viewport areas https://git.io/JLRG5
19:28:25 <frosch123> in contrary: (1) merge old pbs, move map bits to make room for more signals (2) revert old pbs, move map bits back (3) merge new pbs, move map bits the same way again as in (1)
19:28:25 <TrueBrain> _dp_ has many different names on different places :P
19:28:36 <frosch123> i never understood why the map bits were changed in (2) :p
19:28:50 <TrueBrain> stop talking about PBS. Just .. never again. Please. :P
19:28:56 <TrueBrain> what a shitshow that was ...
19:29:13 <frosch123> TrueBrain: never talk about nrt, and we have a deal :)
19:29:17 <andythenorth> terrible amalgam name
19:29:21 <TrueBrain> :D That is fair frosch123 :)
19:29:22 <andythenorth> oof frosch123 :P
19:29:27 <andythenorth> do I have to leave in shame again?
19:29:36 <TrueBrain> did you misread? Don't talk about it
19:29:54 <milek7> Samu: that's very obtuse
19:29:54 <milek7> why do you need to shift remainder to MSB bits, not other way around? could you do it without packing two things into one int? that would be more readable
19:29:54 <milek7> btw, that task sounds like election system thing. cargo packet is available seat in parliament, stations are different parties, and station ratings are votes, right? :P
19:30:32 <TrueBrain> okay, tackled 2 difficult PRs tonight ... now lets see if we can just merge some instead :D
19:30:40 <andythenorth> hmm how do industries work
19:34:34 <_dp_> TrueBrain, the downsides of a 2-letter nickname :p
19:34:39 <_dp_> I'll rebase it bit later
19:36:18 <TrueBrain> (and this is why we need a livestream! :P :D So much more fun to do this live :P Tomorrow I will plan more ... :D)
19:37:00 <TrueBrain> GitHub is following us btw ... they now also removed all cookies except for functional ones :P
19:37:04 <andythenorth> hmm, scaling FIRS production over time to suit my train set
19:39:03 *** ChanServ sets mode: +v tokai
19:41:26 <TrueBrain> who claimed "OpenTTD" on twitch?!
19:43:30 <Afdal> When are you guise gonna move off MicrosoftHub anyway
19:44:11 <andythenorth> when the default is something else
19:44:17 <andythenorth> or it stops working
19:44:27 <FLHerne> Afdal: Given that the wiki and translations have only just finished migrating *onto* it, not soon
19:44:31 <Afdal> Embrace, Extend, Extinguish
19:44:41 <Afdal> The wiki was moved to github?
19:44:59 <Afdal> Is this the reason a ton of wiki articles are currently broken
19:45:02 <FLHerne> The new wiki is a very cunning frontend to a git repository
19:45:11 <andythenorth> 'git is the database'
19:45:16 <andythenorth> such a good pattern
19:45:21 <Afdal> I'm kind of mad that an article I put some work on for a while is busted at the moment
19:45:38 <FLHerne> The brokenness is probably a result of the move, but not directly related to the git part
19:46:12 <Afdal> does it exist somewhere else or was it lost completely in the migration
19:46:18 <TrueBrain> that reminds me, lets close that PR ..
19:52:41 <Timberwolf> Do drive-through road stops always have a town zone of 0?
19:53:52 <andythenorth> oh I broke auto-separation somehow
19:53:59 <andythenorth> all my trains are waiting in stations :P
19:54:49 <andythenorth> bug report is no use, unreleased grfs, reload_newgrfs extensive
19:55:35 <andythenorth> fixed by stopping all vehicles, then starting all vehicles
19:57:14 <FLHerne> What should NML coders be using instead?
19:57:38 <frosch123> Timberwolf: draw a ticket number for yet another nrt bug. but we agreed to not talk about them
19:59:08 <frosch123> FLHerne: which ones, i checked the first 3, and they have new names
20:00:01 <Timberwolf> Cool, I'll make my town zone 0 "unpaved track" cope with being drawn over a road tile and maybe take a look at the code problem at some future point.
20:02:26 <FLHerne> frosch123: BYTE_S, CURRENCY_QWORD, UNPRINT, BYTE_HEX, WORD_HEX, QWORD_HEX
20:06:40 *** tokai|noir has joined #openttd
20:06:40 *** ChanServ sets mode: +v tokai|noir
20:12:04 <frosch123> i think UNPRINT was trashed from ottd
20:12:16 <frosch123> so there is no replacement, and also no use case
20:12:45 *** tokai has quit IRC (Ping timeout: 480 seconds)
20:13:34 <frosch123> no, it's still there
20:13:44 <frosch123> must have confused that with something else
20:17:11 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
20:18:50 *** ChanServ sets mode: +v tokai
20:18:52 <frosch123> one reason is that translations only work for word-sized stuff
20:19:09 <frosch123> so that removes BYTE_S, BYTE_HEX, and UNPRINT is makes even less sense
20:19:23 <frosch123> UNPRINT is probably a ttdp-internal code :)
20:20:58 <frosch123> the second reason is probably that nml should be able to fill the text stack itself at some point (was never done). so qword parameters do not really work
20:21:36 <frosch123> and the optimisation to store small numbers in words instead of dwords was considered unnecessary
20:21:57 <frosch123> so in summary: there is no replacement, noone is supposed to use or even need them anyway
20:27:59 *** gelignite has joined #openttd
20:30:24 <andythenorth> The BEST Feature!
20:37:29 <Timberwolf> Uh-oh, GUI sprites.
20:41:45 <frosch123> you are drawing with a dos painting program in dosbox? :o
20:41:59 *** tokai|noir has joined #openttd
20:41:59 *** ChanServ sets mode: +v tokai|noir
20:42:38 <frosch123> i guess you want the full sf experience
20:44:40 *** tokai has quit IRC (Ping timeout: 480 seconds)
20:51:32 <TrueBrain> okay, my first non-live stream worked fine :) Tnx Xaroth , much appreciated :D
20:52:03 <TrueBrain> we have a twitch account, I can show my pretty face, and I know how to show other people's faces :P
20:52:15 <TrueBrain> the only thing that will be a bit of an issue is audio management, but we will work something out
20:52:22 <TrueBrain> I can show OpenTTD, GitHub, etc
20:52:34 <TrueBrain> I just have to be rather careful not to press this darn: GO LIVE, button :D
20:53:25 <TrueBrain> so tomorrow I write down what I think the format should be .. and we can test a bit :D
20:53:31 <milek7> I almost forgot it still exists ;P
20:53:56 <TrueBrain> welcome to planet earth and 2020 :)
20:54:41 <TrueBrain> I will work out some nice backgrounds .. yeah, this will be a total disaster, but it will be a fun one nevertheless :)
20:55:13 <milek7> I feel their marketshare was totally eaten by youtube when they introduced livestreaming
20:55:16 <TrueBrain> but I am surprised how easy this all is, IF you know what buttons to press :P
20:55:18 <milek7> but might be my bubble :P
20:55:38 <TrueBrain> stream will be at 1080p / 60fps
20:55:47 <TrueBrain> my CPU did not even blink when I generated a 4kx4k map ...
20:55:50 <TrueBrain> so that is promising :P
20:56:15 <michi_cc> No idea how the "market" for game streaming is nowadays, but Twitch has a lot more than just games now
20:57:13 *** ChanServ sets mode: +v tokai
20:57:24 <_dp_> nah, twitch is still alright
20:57:30 <_dp_> mb not for long but for now it is
20:57:37 <TrueBrain> milek7: "Twitch's 75.1% share and YouTube's 22.1% share." <- 2019 stats
20:59:32 <TrueBrain> btw, we might need a chat moderator .. I might poke Discord for that
20:59:43 <TrueBrain> (as in, one of the mods from Discord to moderate Twitch)
21:01:00 <Timberwolf> frosch123: Indeed, Deluxe Paint 2 is the thing I had as a kid and spent more time in that any other paint package, as a result it's still the only one I have decent muscle memory for.
21:01:30 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
21:01:38 <andythenorth> I am doing audio only :P
21:01:42 <andythenorth> so sick of video this year
21:01:47 <andythenorth> 3-5 hours every day
21:02:16 <TrueBrain> andythenorth: but this is fun :)
21:02:22 <TrueBrain> you don't have to act all professional
21:02:25 <TrueBrain> in fact .. I prefer you don't :P
21:03:01 <TrueBrain> webcam is only for parts of the stream where it benefits, which won't be that long, honestly
21:03:10 <andythenorth> the idea that I act professional
21:03:37 <andythenorth> it's the watching my own face that kills me :)
21:03:43 <andythenorth> I turn it off in the apps that support that
21:03:46 <TrueBrain> you don't have to, I promise :)
21:04:24 <andythenorth> I have seen me, I know what I look like already
21:05:51 <TrueBrain> I also put myself in a tiny corner
21:07:30 <Xaroth> There's a benefit to being on camera
21:07:42 <Xaroth> means people can yell at you when you're talking without using your push-to-talk key
21:07:51 <TrueBrain> we did NOT have that TWICE today :P
21:07:55 <TrueBrain> it did NOT happen :P
21:08:37 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
21:22:38 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLBDc
21:24:22 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8395: Change: Towns don't build dead-end road bridges https://git.io/JLRg1
21:25:30 *** tokai has quit IRC (Ping timeout: 480 seconds)
21:26:10 *** ChanServ sets mode: +v tokai
21:31:50 <TrueBrain> on Discord we now have a "dev" channel for the livestream, which is invite only (except for .. devs! :P)
21:31:55 <TrueBrain> row row row row this boat
21:33:22 *** tokai|noir has joined #openttd
21:33:22 *** ChanServ sets mode: +v tokai|noir
21:35:05 *** tokai has quit IRC (Ping timeout: 480 seconds)
21:36:53 <DorpsGek> [OpenTTD/OpenTTD] andythenorth commented on pull request #8342: Feature #8095: Allow automatically separating vehicles in shared orders https://git.io/JLRaa
21:37:34 <andythenorth> so Daylength discussion?
21:37:37 <andythenorth> might as well eh
21:40:48 <Samu> but it has a dummy now :(
21:40:51 <milek7> but that 0xFFFF * (a % b) / b weirdess likely requires more explaination
21:41:08 <milek7> at least because it's not 'remainder' anymore
21:44:53 <Samu> I'm unsure how to explain it
21:46:34 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
21:46:49 *** ChanServ sets mode: +v tokai
21:47:49 <Samu> i could use the whole 32 bits now
21:50:01 <milek7> there have to be some bits left for computation
21:53:35 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
21:55:30 *** tokai has quit IRC (Ping timeout: 480 seconds)
21:57:13 <milek7> TrueBrain: so about crypto RFC
21:57:13 <milek7> do we want servers to keep pubkeys private?
21:57:13 <milek7> or do we want to share pubkeys to other clients as I suggested? (with that optional privacy mode etc.)
21:58:54 *** ChanServ sets mode: +v tokai
22:00:37 *** tokai|noir has joined #openttd
22:00:37 *** ChanServ sets mode: +v tokai|noir
22:07:05 *** tokai has quit IRC (Ping timeout: 480 seconds)
22:07:18 *** Samu has quit IRC (Quit: Leaving)
22:08:36 *** ChanServ sets mode: +v tokai
22:15:00 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
22:18:51 <andythenorth> oof comment from Alberth in 2017
22:18:52 <andythenorth> "As for 2020, don't know what will happen, but will be interesting to find out :)"
22:54:01 *** andythenorth has quit IRC (Quit: andythenorth)
23:18:09 *** sla_ro|master has quit IRC ()
23:25:36 *** Progman has quit IRC (Remote host closed the connection)
23:37:57 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
continue to next day ⏵
