Config
Log for #openttd on 12th May 2021:
Times are UTC Toggle Colours
00:07:01  *** dih has quit IRC
00:07:42  *** dihedral has joined #openttd
00:17:46  *** EER has quit IRC
00:17:56  *** Gustavo6046 has quit IRC
00:23:42  *** Gustavo6046 has joined #openttd
00:37:18  *** EER has joined #openttd
00:43:02  *** FLHerne has quit IRC
00:43:03  *** FLHerne has joined #openttd
00:58:10  *** Hobbyboy has quit IRC
00:58:12  *** Hobbyboy has joined #openttd
00:58:43  *** ericnoan- has quit IRC
00:58:46  *** ericnoan has joined #openttd
01:05:06  *** EER has quit IRC
01:15:04  *** Gustavo6046 has quit IRC
01:15:22  *** Gustavo6046 has joined #openttd
01:26:49  *** Gustavo6046 has quit IRC
01:29:17  *** EER has joined #openttd
01:30:31  *** Gustavo6046 has joined #openttd
01:41:39  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 commented on pull request #9203: Change: Use gender-neutral pronouns https://git.io/JsJNM
01:48:36  *** EER has quit IRC
01:51:18  *** EER has joined #openttd
01:52:38  *** Gustavo6046 has quit IRC
01:54:19  *** Gustavo6046 has joined #openttd
02:17:11  *** EER has quit IRC
02:30:15  *** EER has joined #openttd
02:32:25  *** glx has quit IRC
02:37:58  *** lobstarooo has joined #openttd
02:43:50  *** lobster has quit IRC
02:43:59  *** lobstarooo is now known as lobster
02:49:17  *** D-HUND has joined #openttd
02:50:01  *** EER has quit IRC
02:52:37  *** debdog has quit IRC
02:53:17  *** EER has joined #openttd
03:02:47  *** lobstarooo has joined #openttd
03:10:16  *** lobster has quit IRC
03:10:18  *** lobstarooo is now known as lobster
03:12:50  *** EER has quit IRC
03:14:02  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 opened issue #9249: Airplane takeoff and landing sounds still happen when the "no sound" soundpack is selected https://git.io/JsUqg
03:32:36  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 commented on issue #9249: Airplane takeoff and landing sounds still happen when the "no sound" soundpack is selected https://git.io/JsUqg
03:32:39  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 closed issue #9249: Airplane takeoff and landing sounds still happen when the "no sound" soundpack is selected https://git.io/JsUqg
04:14:23  *** Flygon has joined #openttd
04:43:19  *** snail_UES_ has quit IRC
04:57:18  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 started discussion #9250: [Suggestion] Make console commands and outputs translatable strings https://git.io/JsUuI
05:01:20  *** snail_UES_ has joined #openttd
05:05:36  *** snail_UES_ has quit IRC
05:30:22  *** EER has joined #openttd
05:34:13  *** tonyfinn[m] has quit IRC
05:34:19  *** tonyfinn[m] has joined #openttd
05:37:08  *** andythenorth has joined #openttd
05:37:58  *** supermop_Home_ has quit IRC
05:39:49  *** iarp[m] has quit IRC
05:39:54  *** iarp[m] has joined #openttd
05:40:26  *** EER has quit IRC
06:01:37  *** EER has joined #openttd
06:15:35  *** sla_ro|master has joined #openttd
06:21:35  *** WormnestAndroid has quit IRC
06:21:48  *** WormnestAndroid has joined #openttd
06:29:36  *** EER has quit IRC
06:35:53  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9248: Fix: [Emscripten] Force secure WebSockets over HTTPS https://git.io/JsU9n
06:36:47  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9248: Fix: [Emscripten] Force secure WebSockets over HTTPS https://git.io/JsUHm
06:40:24  <TrueBrain> hmm .. should DEBUG() sentences end with a dot or not
06:40:49  <TrueBrain> seems it is not done often
06:42:13  *** natalie[m] has quit IRC
06:42:17  *** natalie[m] has joined #openttd
06:51:01  *** twom[m] has quit IRC
06:51:04  *** twom[m] has joined #openttd
06:52:14  <Rubidium> well, check the coding style. When it's not there add it and make a PR to fix the others ;)
06:52:37  *** dag[m] has quit IRC
06:52:45  *** dag[m] has joined #openttd
07:11:19  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9226: Fix #9202: Invalid test for unset NewGRF override mapping. https://git.io/J3bMy
07:12:13  *** EER has joined #openttd
07:12:55  <peter1138> LordAro, ah yes, that title.
07:13:32  <peter1138> Because on a news message the caption is fake.
07:13:43  <TrueBrain> can I ask, what is wrong with it?
07:13:58  <peter1138> https://pasteboard.co/K1sAG7U.png < The text is slightly too high.
07:14:09  <peter1138> "Message"
07:14:10  <TrueBrain> ah
07:14:15  <TrueBrain> didn't even spot it the first time :P
07:14:36  <peter1138> I'm assuming it's due to the text alignment stuff I tweaked.
07:15:49  *** menelaos[m] has quit IRC
07:15:54  *** menelaos[m] has joined #openttd
07:16:47  <peter1138> Yup :)
07:17:04  <peter1138> Now, I can either just fix the alignment, or... rewrite so that it isn't a fake caption... :p
07:17:08  <LordAro> :)
07:17:12  <peter1138> Damn it, I did it again.
07:17:44  <peter1138> Also my other half left me some coffee this morning. All of about 2 sips :(
07:21:11  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #9251: Change: reworked the debug levels for network facility https://git.io/JsUxB
07:21:55  <TrueBrain> a very boring PR :P
07:24:44  <DorpsGek> [OpenTTD/OpenTTD] PeterN opened pull request #9252: Fix #9063: Caption of news window incorrectly aligned. https://git.io/JsUhG
07:24:47  <peter1138> Same
07:25:15  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9252: Fix #9063: Caption of news window incorrectly aligned. https://git.io/JsUh4
07:26:03  *** HerzogDeXtEr has joined #openttd
07:32:59  <TrueBrain> finally I can run OpenTTD again with sanitizer on :D w00p :)
07:33:11  <TrueBrain> its annoying that when it runs into any problem, it just quits
07:33:15  <TrueBrain> valgrind at least keeps goingt :P
07:33:19  *** keenriser has joined #openttd
07:37:56  *** EER has quit IRC
07:39:03  <andythenorth> :)
07:39:22  <LordAro> :)
07:39:27  <TrueBrain> :)
07:47:25  *** karoline[m] has quit IRC
07:47:27  *** karoline[m] has joined #openttd
07:50:48  <peter1138> :)
08:08:51  *** keenriser has quit IRC
08:35:45  <TrueBrain> funny .. after connect() succeeds for a STUN request, I send a packet from the client to the GC .. that packet is processed sooner in the GC than the accept() handler of the STUN server :)
08:36:43  <TrueBrain> code-wise that totally makes sense, but it is funny to experience :D
08:37:27  *** guru3_ has joined #openttd
08:38:07  *** guru3 has quit IRC
08:45:19  <peter1138> I was going to do a quick test of 9251 but by manager called earler and then... meh, work o'clock :(
08:45:29  <TrueBrain> :(
08:50:14  <DorpsGek> [OpenTTD/team] Anceph opened issue #213: [tr_TR] Translator access request https://git.io/JsTZs
09:06:13  *** christoph[m]3 has quit IRC
09:06:15  *** christoph[m]3 has joined #openttd
09:26:01  *** Samu has joined #openttd
09:29:27  *** jottyfan has joined #openttd
10:22:59  <TrueBrain> right, had to rework the GC protocol a bit, so it is a lot quicker in detecting errors
10:23:13  <TrueBrain> no longer you have to wait for a timeout if you don't have IPv6, basically :P
10:26:33  <andythenorth> it's definitely not lunch yet
10:26:55  <TrueBrain> almost
10:27:04  <TrueBrain> trying to finish this before I do .. always a tricky thing to set out for yourself :D
10:31:05  *** andythenorth has left #openttd
10:32:13  *** elliot[m] has quit IRC
10:32:14  *** EER has joined #openttd
10:32:18  *** elliot[m] has joined #openttd
10:32:30  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9017: Feature: multiplayer without port forwarding https://git.io/JOk7C
10:32:37  *** jact[m] has quit IRC
10:32:38  *** jact[m] has joined #openttd
10:32:45  <TrueBrain> but so yeah... I rewrote the GC again
10:32:46  <TrueBrain> lol
10:32:59  <TrueBrain> still not easy to make it scalable in a way we can afford ..
10:33:10  <TrueBrain> guess I should load-test a single instance, see what that does
10:33:33  <peter1138> Possible to distribute it?
10:34:16  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9017: Feature: multiplayer without port forwarding https://git.io/JOk7C
10:34:36  <TrueBrain> the main issue is that the GC needs to talk to both the server and client with shared knowledge
10:34:52  <TrueBrain> so very quickly you need a centralized queue of sorts
10:34:54  <TrueBrain> read: redis
10:36:10  <peter1138> Ah yeah
10:36:14  <TrueBrain> so if I run instance A and B, and client 1 connects to A, and server 2 connects to B .. and 1 wants to connect to 2 .. A and B need to cross talk
10:36:18  <TrueBrain> which is always a huge pita
10:36:48  <TrueBrain> normally I wouldn't worry about it at all, but cost is a factor for OpenTTD .. and a crazy one at that :P
10:37:06  <peter1138> Blockchain!
10:37:26  <TrueBrain> either we rent AWS Redis, which is ~20 dollar a month extra
10:37:32  <TrueBrain> or we run redis ourselves (not that hard honestly)
10:37:44  <TrueBrain> I have to think about this a bit more :D
10:38:13  <TrueBrain> I at least now split the logic between a token-controller, and handlers for client/server
10:38:24  <TrueBrain> TURN can run dedicated on its own now too
10:38:34  <TrueBrain> and STUN just needs to a handoff of some information to the GC
10:38:52  <TrueBrain> but otherwise doesn't need the GC too
10:39:05  <TrueBrain> so that already helps, I guess :)
10:40:01  <TrueBrain> and for GC, I think self-hosted redis is the cheapest way to go for us
10:40:10  <TrueBrain> and just run a single instance .. if it has to be restarted, so be it
10:40:20  <TrueBrain> just a temporary loss of server-listing; recovers in seconds
10:40:58  <TrueBrain> but first: lunch!
10:48:29  *** ChaKu has joined #openttd
10:51:46  <ChaKu> hullo good folks! I recently stumbled upon openttd and I have played it for no less than 40 hours in past week. Just wanted to drop by and say thank you for making an incredibly fun game.
10:59:50  <LordAro> :D
11:01:16  * LordAro been reading through the trainwreck of the audacity telemetry PR
11:01:19  <LordAro> found wuzzy
11:01:33  <LordAro> https://github.com/audacity/audacity/pull/835#issuecomment-834405825
11:04:57  *** JacobD88 has joined #openttd
11:08:05  <Xaroth> of course.
11:13:24  <TrueBrain> At least it got changed to opt-in along the way ;)
11:13:48  *** Gustavo6046 has quit IRC
11:13:48  <TrueBrain> Hard to find the moment it got closed via mobile .. lol
11:14:05  <TrueBrain> GitHub shouldn't collapse comments before close event
11:14:58  <LordAro> it's hard on desktop too
11:15:01  <LordAro> i haven't found it yet
11:17:01  *** hamstonkid[m] has quit IRC
11:17:04  *** hamstonkid[m] has joined #openttd
11:17:25  *** Heiki[m] has quit IRC
11:17:29  *** Heiki[m] has joined #openttd
11:17:49  *** EmeraldSnorlax[m] has quit IRC
11:17:50  *** EmeraldSnorlax[m] has joined #openttd
11:18:13  *** rudolfs[m] has quit IRC
11:18:15  *** rudolfs[m] has joined #openttd
11:18:37  *** gretel[m] has quit IRC
11:18:40  *** gretel[m] has joined #openttd
11:19:01  *** einar[m] has quit IRC
11:19:05  *** einar[m] has joined #openttd
11:24:12  <peter1138> LordAro, who?
11:29:23  <LordAro> peter1138: osfx licensing guy
11:38:13  <TrueBrain> https://github.com/audacity/audacity/pull/835#issuecomment-835880865
11:38:14  <TrueBrain> found it
11:54:58  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9251: Change: reworked the debug levels for network facility https://git.io/JsUxB
11:58:44  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9251: Change: reworked the debug levels for network facility https://git.io/JsUxB
12:13:52  *** Speeder_ has quit IRC
12:18:30  *** glx has joined #openttd
12:18:30  *** ChanServ sets mode: +v glx
12:37:01  *** supermop_Home has joined #openttd
12:37:05  <supermop_Home> hi
12:37:15  <DorpsGek> [OpenTTD/OpenTTD] embeddedt updated pull request #9248: Fix: [Emscripten] Force secure WebSockets over HTTPS https://git.io/JsJWz
12:39:46  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9248: Fix: [Emscripten] Force secure WebSockets over HTTPS https://git.io/JskJW
12:51:25  <TrueBrain> https://github.com/TrueBrain/game-coordinator <- at least put the GC code online so it won't get lost :)
12:52:04  <TrueBrain> still far from done, but functional enough for nightlies :)
12:58:19  <LordAro> :)
12:58:41  *** snail_UES_ has joined #openttd
12:58:50  <LordAro> i know i asked already, but i've forgotten the answer - what was py38 specific?
12:59:15  <TrueBrain> openttd-helpers is, can't remember what part :P
12:59:31  <LordAro> really? can't imagine there's anything that new in there
13:00:02  <TrueBrain> what a weird thing to say :P
13:00:31  <LordAro> py38 is very new, with limited amount breaking changes
13:00:35  <LordAro> openttd-helpers is not that new
13:00:43  <TrueBrain> you do know 3.9 is out right? :P
13:00:47  <LordAro> sure
13:00:48  <TrueBrain> but we had this conversation before :)
13:01:10  <LordAro> maybe
13:01:10  <TrueBrain> around the time I created py-helpers, I imagine :)
13:01:13  <LordAro> i have no memory of it
13:01:55  <TrueBrain> it also rarely is important information :D Till now, it was all running via Dockers
13:02:54  <TrueBrain> too bad it all came in as a single commit
13:02:55  <TrueBrain> saddddd
13:05:39  <TrueBrain> either way, I reflected back on how to run it on VPSes, and realised it should be via a container too, if we want to go that way
13:06:58  <TrueBrain> otherwise upgrading would be an ugly mess :D
13:10:14  <TrueBrain> well, it is messy anyway .. we did not build the VPSes with this in mind :(
13:16:33  <LordAro> perhaps
13:16:49  <LordAro> could do things with ssh keys and restricted rsync users
13:16:53  <LordAro> but they're a bit meh
13:17:00  <TrueBrain> its all very meh
13:17:15  <TrueBrain> we just have to find a rich dude that wants to donate lots of money to us :P
13:17:20  <TrueBrain> doesn't have to be a dude btw
13:19:07  <TrueBrain> I did consider making it a setup.py project and pushing the "turn" server part to pip
13:19:11  <TrueBrain> so it can juts be pip installed
13:19:21  <TrueBrain> but .. still .. manual work on a box ..
13:19:31  <LordAro> pip can install from a url directly, no need for pypi
13:19:39  <LordAro> stick that on a cronjob ;)
13:20:08  <TrueBrain> pypi as in: tagged builds that can be reverted easily while having the old builds permanent immutable :)
13:21:31  <TrueBrain> that is what I like about the container images .. no matter what, if the image could run "once upon a day"
13:21:34  <TrueBrain> it can still run :P
13:21:44  <TrueBrain> fuck dependencies that got removed / renamed / ...
13:24:47  <peter1138> Apparently I can't concentrate today :(
13:24:48  <Xaroth> don't we all want to find a rich person that wants to donate lots of money to us?
13:26:25  <rexxars> I'd donate for someone to fix this mouse click lag bug I'm having on macos :p
13:26:49  <TrueBrain> rexxars: sadly that donation has to be the amount of a MacOS laptop to a certain dev (not me) which has the issue, I am afraid :(
13:27:16  <rexxars> I'll donate my laptop when I upgrade to the new M1 ;)
13:27:28  <TrueBrain> MacOS build atm is developed via VMWare .. so yeah .. go figure :P
13:28:14  <peter1138> Did you find the first version it started with?
13:28:32  <peter1138> Possibly something to do with the std::chrono changes, iirc.
13:30:37  <rexxars> yeah, something related to the way it slept for X ms instead of the 1ms it used to do. I looked into it a bit more and from some pretty shallow debugging it seemed like it somehow missed mousedrag events right after a mousedown. I've got some more digging to do
13:32:13  *** nielsm has joined #openttd
13:33:36  <TrueBrain> hmm .. maybe soon we should play a game via TURN, so we know a bit how much bandwidth that actually consumes
13:33:42  <TrueBrain> maybe TURN could just be run on AWS
13:34:10  <TrueBrain> I can estimate what the cost would be, but I don't really have any hard data
13:37:43  *** sla_ro|master has quit IRC
13:58:36  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 opened issue #9253: Sound effects still play through speakers when I have bluetooth headphones on https://git.io/Jsklq
13:59:54  <glx> didn't read the issue, but I think it's expected
14:00:21  <TrueBrain> most games allow you to pick the output to use
14:00:30  <peter1138> That is... enirely down to the sound settings on the computer.
14:00:32  <glx> many don't
14:00:51  <TrueBrain> so there is something there .. just .. lot of work for us, with very little benefit :)
14:00:53  <glx> usually I need to change windows default output device
14:00:56  <TrueBrain> so it is more a feature-request ;)
14:01:03  <TrueBrain> glx: we clearly play very different games :D
14:01:07  <peter1138> Most that do usually have some kinda of voice communication going on.
14:01:11  <peter1138> *Those that do
14:01:19  <TrueBrain> peter1138: when in OpenTTD? :D
14:01:36  <peter1138> Games that allow you to pick the output.
14:01:48  <TrueBrain> no, when voice in OpenTTD? :D
14:01:56  <TrueBrain> hihi :)
14:02:57  <Rubidium> it's already supported in #openttd, isn't that enough?
14:03:10  <LordAro> Rubidium: you've made that joke once already this week
14:03:13  <LordAro> :p
14:03:23  <TrueBrain> yeah, I rate that pun 2 out of 10, sorry
14:03:23  <peter1138> I'm so confused right now.
14:03:46  <TrueBrain> poor peter1138
14:04:11  <peter1138> Well, at least I managed to get an IP onto my server.
14:04:12  <LordAro> TB missed the implicit "voice communication going on (which OTTD does not do)"
14:04:38  <LordAro> assuming i am not also confused by the situation
14:05:00  <peter1138> I had to fire up Internet Explorer to get on to its remote console ;(
14:05:15  <TrueBrain> lolz
14:05:20  <TrueBrain> the moment you realise you need IE
14:05:26  <TrueBrain> brrrr
14:05:26  <LordAro> oh dear
14:06:12  <TrueBrain> ugh ... I now realise we cannot create IPv6 connections on AWS (only accept) .. guess I need that pproxy hack again
14:06:13  <TrueBrain> bah :P
14:06:30  <TrueBrain> (to probe if a server has the game port open)
14:06:53  <LordAro> :(
14:07:00  <LordAro> an odd limitation?
14:07:12  <TrueBrain> not really, the ECS hosts run IPv4 only
14:07:16  <TrueBrain> so the containers do too
14:07:35  <TrueBrain> incoming IPv6 is captured by Load Balancers, and forwarded as IPv4 traffic (with a proxy-protocol header)
14:07:41  <TrueBrain> works fine for 99% of the use-cases
14:07:45  <TrueBrain> we are special :P
14:07:46  <LordAro> ah
14:07:50  <LordAro> bit of a shame
14:07:58  <TrueBrain> you can run ECS with IPv6 these days
14:08:03  <TrueBrain> but it kinda requires me to rebuild everything
14:08:13  <TrueBrain> which, for some odd reason, I don't feel like doing
14:08:23  <LordAro> how intensive is rebuilding things?
14:08:33  <LordAro> i thought it was just a case of clicking (a lot of) buttons
14:08:35  <TrueBrain> as in: rebuild the deployment scripts :P
14:08:47  <LordAro> hmm
14:09:08  <TrueBrain> basically, as they didn't support NLB over IPv6, I build my own
14:09:11  <TrueBrain> (in aws-infra)
14:09:19  <TrueBrain> so that would have to be taken out, and uugggghhhh
14:09:25  <glx> ah found another for loop macro to remove
14:09:50  <TrueBrain> but I might experiment with AWS Fargate
14:09:56  <TrueBrain> that means we no longer have to run ECS
14:10:04  <TrueBrain> which .. saves a lot of other shit
14:10:21  <TrueBrain> then it might be worth just starting from scratch .. and that should be relative quick
14:10:45  <TrueBrain> saves money too!
14:11:31  <LordAro> sounds suspiciously like it might be the better solution long-term ;)
14:11:44  <TrueBrain> yeah .. just work ..
14:11:50  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 commented on pull request #9251: Change: reworked the debug levels for network facility https://git.io/Jsk4w
14:11:53  <LordAro> :)
14:12:13  *** ZirconiumX has quit IRC
14:12:41  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9251: Change: reworked the debug levels for network facility https://git.io/JsUxB
14:12:44  <TrueBrain> incredible, that there are people who's OCD is worse than mine :D
14:13:59  <TrueBrain> "* on this page GB = 1024^3 bytes"
14:14:02  <TrueBrain> I love such remarks
14:14:18  * LordAro mumbles something about it being GiB
14:14:29  <TrueBrain> no matter, at least they made it clear
14:14:59  *** ZirconiumX has joined #openttd
14:15:47  <TrueBrain> AWS pricing is always a bit difficult to read, and can leave you guessing
14:15:56  <TrueBrain> like AWS Fargate, you pay what you use
14:15:58  <TrueBrain> which is nice
14:16:02  <TrueBrain> but .. how do they calculate vCPU
14:16:05  <TrueBrain> is it at least 1?
14:16:12  <TrueBrain> these kind of details are hard to discover :)
14:17:18  <TrueBrain> 0.25 is the lowest possible
14:17:19  <TrueBrain> fine
14:18:06  <TrueBrain> owh, even smaller, 0.125 is the lowest
14:19:09  <TrueBrain> but not on Fargate
14:19:10  <TrueBrain> lol
14:19:16  <TrueBrain> this is always the rabbit hole ......
14:19:30  <TrueBrain> 0.25 vCPU / 512MB RAM is lowest possible setting
14:19:33  <TrueBrain> in that case, not sure it is cheaper :P
14:20:25  <TrueBrain> @calc (0.04656 / 4 + 0.00511 / 2) * 24 * 30
14:20:25  <DorpsGek> TrueBrain: 10.2204
14:20:39  <TrueBrain> that is ... weirdly expensive
14:21:35  <glx> well there's AWS in the name ;)
14:22:39  <Timberwolf> Oh, Fargate is really expensive per unit hour compared to just using EC2 as compute capacity.
14:23:04  <Timberwolf> I think it's designed for use cases where you're running a workload only a couple of times a week and managing the capacity is too much hassle.
14:23:30  <Timberwolf> (and it's not a suitable candidate for lambda-isation)
14:23:37  <TrueBrain> yeah, but I did not expect it to be this bad
14:23:56  <TrueBrain> also don't remember it being this bad :D
14:23:57  <Timberwolf> It is weird how much it costs.
14:24:19  <Timberwolf> Getting into the realm of Azure, "we'll provide it because people keep asking for it but please... don't use it" pricing.
14:24:48  <TrueBrain> I just want to get ride of having to upgrade ECS myself
14:24:54  <TrueBrain> that is such a bullshit job on a cloud
14:25:02  <Timberwolf> Oh yes.
14:25:16  <Timberwolf> There is that side... maybe it's worth it not to have to deal with the ECS agent :p
14:25:32  <TrueBrain> well, I just use their AMIs
14:25:40  <TrueBrain> so I just have to scale up the launch group and down again
14:25:50  <TrueBrain> I mean, it is not that much effort, just .. ugh .. I don't want to have to think about it
14:26:40  <TrueBrain> really, last time I did the math it was not this expensive I thought? Was over a year ago .. but I must have been looking at something else I guess .. ugh, there goes that idea :(
14:27:09  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 commented on pull request #9246: Fix #9242: Tree tick handler did not scale by map size. https://git.io/JskEb
14:27:12  <Timberwolf> They have maximum instance lifetimes for ASGs, does that force the replacement behaviour?
14:27:36  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 approved pull request #9251: Change: reworked the debug levels for network facility https://git.io/JskEj
14:27:48  <Timberwolf> It's been a while since I did anything involving actual virtual machines I had some sort of control over, I'm not sure that even used to be a feature.
14:28:33  <TrueBrain> well, first I need to run CDK to update the LG with the latest AMI
14:28:37  <TrueBrain> as for some reason that is not automated too
14:29:01  <TrueBrain> and we have a many very small tasks
14:29:17  <TrueBrain> so that is mainly why Fargate becomes expensive
14:29:44  <TrueBrain> the whole cluster combined doesn't even use 0.25vCPU :P
14:29:52  <Timberwolf> Mmm, stuff which tesselates well when you have "an instance"
14:30:35  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 approved pull request #9232: Fix 91b8ce07: dedicated servers could no longer create screenshots https://git.io/JskuS
14:30:51  <TrueBrain> so I guess ECS stays .. owh well
14:31:01  <TrueBrain> still would need to replace our custom NLB with the now-IPv6-enabled NLB
14:31:35  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain closed issue #9231: ``screenshot minimap`` does not generates files in dedicated server https://git.io/J3pD4
14:31:38  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9232: Fix 91b8ce07: dedicated servers could no longer create screenshots https://git.io/J3pDg
14:32:01  <TrueBrain> and rebuild ECS to be IPv6 enabled
14:32:02  <TrueBrain> meh
14:32:07  <TrueBrain> what-ever, not this week
14:32:32  <TrueBrain> I already do have to replace certificates with a hostname less ... which is such a pain in the ass to do via CDK
14:33:37  <TrueBrain> (we used to run an extra domain for CDN business, but we replaced that .. just the certificate still have that name in them, next to the new name .. such a pain to fix ..)
14:34:06  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #9251: Change: reworked the debug levels for network facility https://git.io/JsUxB
14:42:50  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 commented on issue #9253: Sound effects still play through speakers when I have bluetooth headphones on https://git.io/Jsklq
14:43:01  <LordAro> has anyone made an ECS Vectors joke yet?
14:43:35  <TrueBrain> it wasn't normal? (to continue with the "vector" part)
14:44:37  *** ChaKu has quit IRC
14:44:57  *** ChaKu has joined #openttd
14:51:09  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 closed pull request #9212: Change: support passing std::string directly to DEBUG https://git.io/J3HDE
14:54:08  *** JacobD88 has quit IRC
14:54:59  <DorpsGek> [OpenTTD/OpenTTD] glx22 opened pull request #9254: Codechange: Replace TILE_AREA_LOOP with range-based for loops https://git.io/Jskwk
15:01:25  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 updated pull request #9128: Codechange: use std::string exclusively for settings https://git.io/J3TEL
15:19:21  *** lobstarooo has joined #openttd
15:20:25  *** lobstarooo_ has joined #openttd
15:22:09  *** Gustavo6046 has joined #openttd
15:22:27  *** lobstarooo__ has joined #openttd
15:22:46  <DorpsGek> [OpenTTD/OpenTTD] PeterN commented on issue #9253: Sound effects still play through speakers when I have bluetooth headphones on https://git.io/Jsklq
15:24:10  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 closed issue #9253: Sound effects still play through speakers when I have bluetooth headphones on https://git.io/Jsklq
15:25:46  *** lobster has quit IRC
15:25:55  *** lobstarooo___ has joined #openttd
15:25:55  *** lobstarooo___ is now known as lobster
15:26:30  <LordAro> peter1138: yeah, you told him
15:26:32  <LordAro> apparently.
15:27:56  *** lobstarooo has quit IRC
15:29:09  *** Wormnest has joined #openttd
15:29:16  *** lobstarooo_ has quit IRC
15:29:48  <LordAro> https://pasteboard.co/K1zdweb.png this seems like a nice neighbourhood
15:31:31  *** lobstarooo__ has quit IRC
15:32:50  *** y2kboy23 has quit IRC
15:36:37  *** y2kboy23 has joined #openttd
15:47:02  <DorpsGek> [OpenTTD/OpenTTD] unbeatable-101 commented on discussion #9185: [Suggestion] Removing some legacy transparent menus https://git.io/J3ua4
15:47:39  <peter1138> LordAro, oof.
15:49:00  *** Flygon has quit IRC
15:57:59  *** Progman has joined #openttd
16:16:09  *** ChaKu_ has joined #openttd
16:16:41  *** Eddi|zuHause2 has joined #openttd
16:16:41  *** Eddi|zuHause has quit IRC
16:16:41  *** Elouin has quit IRC
16:16:43  *** Elouin has joined #openttd
16:21:52  *** guru3 has joined #openttd
16:22:12  *** guru3_ has quit IRC
16:22:35  *** ChaKu has quit IRC
16:27:54  <DorpsGek> [OpenTTD/OpenTTD] auge8472 commented on discussion #9250: [Suggestion] Make console commands and outputs translatable strings https://git.io/JsUuI
16:43:39  <DorpsGek> [OpenTTD/OpenTTD] PeterN merged pull request #9252: Fix #9063: Caption of news window incorrectly aligned. https://git.io/JsUhG
16:45:43  <DorpsGek> [OpenTTD/OpenTTD] PeterN commented on pull request #9246: Fix #9242: Tree tick handler did not scale by map size. https://git.io/JsIvu
16:59:53  <peter1138> Eh, I don't even have to compile for the static_assert to work. What wizardry is this...
17:00:46  <DorpsGek> [OpenTTD/OpenTTD] PeterN updated pull request #9246: Fix #9242: Tree tick handler did not scale by map size. https://git.io/Jsfjz
17:03:26  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 commented on pull request #9246: Fix #9242: Tree tick handler did not scale by map size. https://git.io/JsIUE
17:03:32  <glx> intellisense magic ?
17:06:12  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 approved pull request #9246: Fix #9242: Tree tick handler did not scale by map size. https://git.io/JsIUx
17:06:24  *** sla_ro|master has joined #openttd
17:07:27  *** frosch123 has joined #openttd
17:13:00  <DorpsGek> [OpenTTD/team] frosch123 commented on issue #213: [tr_TR] Translator access request https://git.io/JsTZs
17:16:40  *** Wolf01 has joined #openttd
17:18:02  <DorpsGek> [OpenTTD/OpenTTD] PeterN merged pull request #9246: Fix #9242: Tree tick handler did not scale by map size. https://git.io/Jsfjz
17:18:05  <DorpsGek> [OpenTTD/OpenTTD] PeterN closed issue #9242: Tropical Lumber Mill doesn't work properly on large maps https://git.io/Jsvnh
17:37:18  <DorpsGek> [OpenTTD/OpenTTD] frosch123 updated pull request #9227: Make Window destruction not rely on undefined behavior https://git.io/J3NvF
17:37:46  <DorpsGek> [OpenTTD/OpenTTD] frosch123 updated pull request #9227: Make Window destruction not rely on undefined behavior https://git.io/J3NvF
17:38:08  <frosch123> the second push is a rebase
17:39:36  <TrueBrain> tnx for the changes, that makes it a lot easier for me to understand now :D \o/
17:39:57  <frosch123> i am used to -m for messages, so hardly use \n :)
17:40:06  <TrueBrain> I noticed :P
17:40:24  <TrueBrain> reason I stopped using -m btw
17:40:36  <TrueBrain> but I might have brought it to another level with complete bedtimestories :D
17:40:54  <frosch123> yes, your message size exceeds the diff size :p
17:41:37  <TrueBrain> we will see in 17 years if that was a good idea or not :D :P
17:42:26  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #9227: Make Window destruction not rely on undefined behavior https://git.io/JsIYz
17:42:27  <frosch123> did you put a reminder in your calendar?
17:42:29  <TrueBrain> still no clue what most of the diff does, but it looks fine :)
17:42:59  <TrueBrain> "if constexpr"  was new to me :)
17:43:53  <TrueBrain> so ... I want to test how much bandwidth TURN takes ... who is hosting tonight? :D
17:45:42  *** andythenorth has joined #openttd
17:48:04  <TrueBrain> it is not like I will be any good picking any NewGRFs :D
17:48:31  <peter1138> timberwolf's for the authentic 2x feeling
17:50:27  <TrueBrain> lol: "See readme for more information, including how to us FIRS"
17:50:30  <TrueBrain> well, that is a big fat lie
17:50:47  <frosch123> "if constexpr" forces evaluation at compile time. in this case it is more fancy than necessary, but when it matters it the "else case" does not need to be sematically valid. the compiler can just drop it
17:51:16  <TrueBrain> 60% of that README is licensing, 30% is fluff, en 10% is "thank you"
17:52:29  <TrueBrain> "Many FIRS cargos are not transportable using the default OpenTTD vehicles. Fortunately there are lots of great NewGRF vehicle sets with FIRS support. The OpenTTD content service (Bananas) is a good place to find these."
17:52:32  <TrueBrain> I get it .. but I have no clue
17:53:16  <frosch123> TrueBrain: take my preset from yesterday, and replace ogfx+trains and firs
17:53:23  <frosch123> the rest is always the same :p
17:53:34  <TrueBrain> replace with what?
17:54:11  <frosch123> yeti+nuts, to have something very different compared to yesterday :)
17:54:47  <TrueBrain> didn't make a save of yesterday
17:54:54  <TrueBrain> so "take my preset" is not working :'(
17:56:17  <TrueBrain> I know I am a broken record, and I have been saying it for 5+ years now, but getting NewGRFs in your game is hard :(
17:56:34  <frosch123> https://dpaste.org/ciLT
17:58:11  <TrueBrain> how big is NUTS? Lol
17:58:30  <frosch123> isn't yeti bigger? :p
17:58:42  <TrueBrain> possible I mismatched the sizes
17:58:58  <frosch123> anyway, when you see yeti in-game, you know where the bytes went
17:59:19  <frosch123> nuts has support for all cargos (200?) with a lot of vehicles
18:03:38  <TrueBrain> okay ... lets find out
18:03:45  <TrueBrain> took the weirdest settings I could find
18:04:05  <Rubidium> you disabled foundations?
18:04:11  <TrueBrain> no .. didn't went that far
18:04:18  <TrueBrain> but I am pretty sure some people will be annoyed
18:04:28  <TrueBrain> https://www.openttd.org/downloads/openttd-branches/pr9017/latest.html <- build of 20210512, join me, laugh at my settings :P
18:04:52  <frosch123> no server yet
18:05:40  <TrueBrain> oops, lets make it public
18:05:41  <TrueBrain> :D
18:05:49  <TrueBrain> you can tell me if I did the NewGRFs correct :P
18:05:56  <frosch123> i didn' t even have all grfs :)
18:06:23  <andythenorth> TrueBrain do you have FIRS feedback? :o
18:06:42  <TrueBrain> andythenorth: the description says things about the README that is a lie :P
18:06:53  <frosch123> TrueBrain: well, you added "basecost mod" without any settings :p
18:06:54  <TrueBrain> and in general I just hate how we present NewGRFs to the users .. that is not FIRS related :)
18:06:58  <frosch123> that does nothing :)
18:07:02  <TrueBrain> frosch123: ... I just enabled it :P
18:07:09  <andythenorth> oh dear
18:07:12  <andythenorth> who makes this stuff
18:07:23  <andythenorth> which bits are a lie? https://bananas.openttd.org/package/newgrf/f1250008
18:07:48  <TrueBrain> andythenorth: ingame, when I click the NewGRF
18:07:53  <TrueBrain> it tells me to read the readme for more details how to play
18:07:57  <andythenorth> I did try to fix descriptions a bit recently, but there are a lot of grfs
18:08:19  <andythenorth> oh FIRS 2
18:08:21  <andythenorth> I see
18:09:09  <andythenorth> I would delete the old versions, but seems a bit harsh for the die-hard fans
18:09:43  <TrueBrain> no
18:09:49  <TrueBrain> FIRS 4
18:10:01  <TrueBrain> https://user-images.githubusercontent.com/1663690/118023617-05c49880-b35e-11eb-948e-eade19b9ca12.png
18:10:14  <andythenorth> ooo
18:10:24  <TrueBrain> frosch123: how about now?
18:11:04  <andythenorth> wow how do I even find that window?
18:11:07  <andythenorth> I never use this stuff
18:11:13  <TrueBrain> NewGRF window
18:11:15  <TrueBrain> clicking on FIRS 4
18:11:27  <andythenorth> oh I see
18:11:43  <andythenorth> ok thanks
18:11:44  <TrueBrain> frosch123: better? :)
18:11:49  <TrueBrain> andythenorth: no worries :)
18:12:22  <TrueBrain> wtf is this industry .. lolz
18:13:25  <frosch123> ha, you nerfed trains :)
18:13:37  <TrueBrain> I copy/pasted your settings
18:13:50  <TrueBrain> well .. setting-wise I might have done silly things
18:17:05  <frosch123> you did mostly one silly thing :) i assume it's intended
18:17:18  <TrueBrain> I did at least 2 :P
18:17:25  <frosch123> but if it is not, it's even more funny
18:17:27  <andythenorth> hmm where does description even come from
18:17:32  <andythenorth> maybe it's action 14
18:17:38  <andythenorth> I thought I fixed all this crap recently :D
18:18:07  <andythenorth> ok yeah it's in the grf
18:18:09  <andythenorth> lol
18:18:42  <andythenorth> I should graphviz all the places a grf author has to try and breadcrumb information together so player can actually get + use the grf
18:21:07  <andythenorth> "But at least it's easy to toggle the palette"
18:21:16  <andythenorth> this would be such a good livestream
18:21:26  <andythenorth> "Truebrain tries to load and play grf"
18:21:37  <andythenorth> it would have to be 16+ age rated
18:21:41  <TrueBrain> yup
18:21:43  <TrueBrain> very much so
18:22:08  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 opened issue #9255: Threading issues with the TCPConnecters https://git.io/JsICT
18:26:42  <andythenorth> wait did we fix newgrf scanning speed?  We did right?
18:27:15  <peter1138> Was it ever broken?
18:27:17  <TrueBrain> nah, it only feels better
18:28:16  <andythenorth> I got rid of all the big grfs like Yeti in the hope it would be faster
18:28:27  <andythenorth> but MP games seem to put them bak
18:28:33  <andythenorth> oh my C key is getting annoying
18:31:12  <peter1138> Oof, big newgrfs
18:39:40  <TrueBrain> hmm .. I really now want to implement an authentication mechanism
18:39:45  <TrueBrain> so you can force people to use NORMAL NAMES
18:39:48  <TrueBrain> instead of this .. what-ever this is
18:40:13  <peter1138> I don't think you can get away with making users authenticate with Github though :D
18:40:21  <TrueBrain> sadly, no :P
18:40:49  <TrueBrain> the detail in YETI is amazing
18:40:50  <TrueBrain> and funny
18:41:19  <peter1138> The bounding boxes are lacking though!
18:42:25  <TrueBrain> and I wasn't even talking about andythenorth's name
18:42:33  <TrueBrain> as you know andy is there when he joins :P
18:43:29  <milek7> having to choose name to play game is silly :P
18:44:02  <TrueBrain> so not used to anymore that you can pick your own name in a game :)
18:46:22  *** arikover has joined #openttd
18:51:24  *** iSoSyS has joined #openttd
19:03:33  *** iSoSyS has quit IRC
19:05:13  <TrueBrain> YETI industries are fun :D
19:05:15  <TrueBrain> silly, but fun :D
19:06:01  <glx> I guess I will be lost in cargo chains again
19:06:10  <TrueBrain> this one is worse :)
19:06:19  <frosch123> you have to start with pax
19:06:37  <frosch123> i did not know, so i am almost bankrupt
19:07:30  <glx> yesterday I looked at cargo rates, then decided to aim for the most profitable
19:07:44  <milek7> ==26627==ERROR: AddressSanitizer: heap-use-after-free on address 0x616000625ebc at pc 0x564324118f3e bp 0x7fd73266c7d0 sp 0x7fd73266c7c0
19:07:44  <milek7> READ of size 4 at 0x616000625ebc thread T34 (ottd:game)
19:07:44  <milek7>     #0 0x564324118f3d in ReplaceChain(Vehicle**, DoCommandFlag, bool, bool*) (/home/milek7/ottd3/build/openttd+0xc46f3d)
19:08:20  <milek7> https://gist.github.com/Milek7/86cc84da915963a885d97f84fec5ed50
19:08:33  <TrueBrain> our issue tracker is ---> that way
19:08:42  <milek7> this wasn't debug build though ;/
19:08:48  <DorpsGek> [OpenTTD/OpenTTD] DorpsGek pushed 1 commits to master https://git.io/JsIEV
19:08:49  <DorpsGek>   - Update: Translations from eints (by translators)
19:09:30  <milek7> yeah it crashes
19:09:47  <TrueBrain> a spectator is doing ~440 bytes/s
19:09:48  <TrueBrain> lol
19:10:03  <TrueBrain> (after 2600 seconds)
19:10:45  *** gelignite has joined #openttd
19:14:45  <glx> more that 1300 newgrf on my system, and I had to download missing ones yesterday, and again today ;)
19:15:00  <TrueBrain> nice!
19:15:55  <frosch123> looks like i cheated myself out of bankruptcy
19:16:08  <frosch123> only 25 years to repay my loan now
19:16:53  <andythenorth> V453000 should resurrect BRIX https://dev.openttdcoop.org/attachments/download/8631/trees-tropic-02.png
19:17:21  <TrueBrain> that does look pretty :)
19:18:16  <nielsm> yeah BRIX has some real work put into making high-res full colour graphics that looks good and has personality
19:34:55  <DorpsGek> [OpenTTD/OpenTTD] Milek7 opened issue #9256: heap-use-after-free in ReplaceChain https://git.io/JsIaj
19:52:40  *** EER has quit IRC
20:07:16  *** andythenorth has quit IRC
20:12:06  <DorpsGek> [OpenTTD/OpenTTD] vituscze commented on issue #9241: Some tree brushes do not mark tiles as a rainforest https://git.io/JsvZq
20:12:47  *** nielsm has quit IRC
20:18:26  *** andythenorth has joined #openttd
20:24:51  <LordAro> peter1138: cycling through a thunderstorm is great fun. would recommend
20:25:00  <peter1138> I... do not want.
20:25:51  <LordAro> the sort of rain where you're thinking "it can't possibly get any heavier"
20:25:53  <LordAro> then it does
20:26:28  <peter1138> Waterproofs or not?
20:26:34  <LordAro> not
20:26:38  <peter1138> There's a point where having them on is worse...
20:27:03  <LordAro> yeah, i had a gillet with me, decided it'd be better just to wait until it stopped raining to put it on
20:28:58  *** lobstarooo has joined #openttd
20:35:52  <DorpsGek> [OpenTTD/OpenTTD] PeterN merged pull request #9187: Fix #9186: Fix incorrect bounding box height causing station sprite glitch. https://git.io/J3zgJ
20:35:55  <DorpsGek> [OpenTTD/OpenTTD] PeterN closed issue #9186: Station drawing glitch? https://git.io/J3z0v
20:36:41  *** lobster has quit IRC
20:36:49  *** lobstarooo is now known as lobster
20:37:06  <FLHerne> Yeah
20:37:22  <FLHerne> Also, there's a point where you literally can't get any more saturated, so why give a damn?
20:37:51  <FLHerne> I miss cycle commuting :-(
20:39:29  <TrueBrain> would be a neat feature, if the host leaves, another person is elected to be the host
20:39:57  <TrueBrain> anyway ... with a huge tnx to all that joined, I can conclude that on average an OpenTTD game consumes 400 bytes/s (excluding map)
20:40:07  <peter1138> This is where you have to decide if you stop now, or if you should carry on and design the protocol to allow it.
20:41:00  <TrueBrain> @calc 102410241024 / 400
20:41:00  <DorpsGek> TrueBrain: Error: invalid syntax (<unknown>, line 1)
20:41:05  <TrueBrain> grr @ Element
20:41:13  <TrueBrain> @calc 1024 * 1024 * 1024 / 400
20:41:14  <DorpsGek> TrueBrain: 2684354.56
20:41:24  <TrueBrain> @calc 1024 * 1024 * 1024 / 400 / 3600 / 24
20:41:24  <DorpsGek> TrueBrain: 31.06891851851852
20:41:33  <TrueBrain> that is 1 GB for a full month of play
20:41:44  <TrueBrain> 1 GB via AWS costs 9 dollarcent
20:42:19  <TrueBrain> so yeah .. we might be fine hosting TURN on AWS .. but it strongly depends on how much % is map-exchange
20:42:22  <TrueBrain> and how many people need TURN :D
20:42:33  <TrueBrain> anyway, YETI + NUTS was great fun, tnx frosch123 , nice idea :)
20:42:54  <glx> too much choice in trains
20:43:15  <frosch123> TrueBrain: i totally forgot how yeti worked, or it worked different last time i used it
20:43:16  *** ChaKu_ has quit IRC
20:43:26  <frosch123> glx: that's tbs fault
20:43:35  <TrueBrain> yeah, sorry, never expire was on
20:43:35  <frosch123> it's easy if vehicles expire
20:43:37  <TrueBrain> that was my bad
20:43:40  <frosch123> one engine per class
20:43:43  <frosch123> 5 classes or so
20:43:50  <_dp_> yeti is kind of pointless for a short game though :p
20:44:04  <TrueBrain> speak for yourself
20:44:24  <frosch123> TrueBrain: gs like siliconvalley and nocargoal work better in coop imo
20:44:42  <TrueBrain> will check that out next time :)
20:44:48  <TrueBrain> right, GC now uses STUN again if possible :P
20:44:59  <_dp_> it takes quite a lot of time to grow towns and set up full feedback loop
20:46:06  <TrueBrain> really surprised by the low bandwidth
20:46:26  <glx> mostly sync packets and commands
20:46:32  <TrueBrain> I knew that already from testing the other day with 1.10, but .. it is really true :P
20:46:43  <TrueBrain> yeah, and we can reduce the amount if we want to
20:46:49  <TrueBrain> but .. 400 bytes/s .. lol
20:47:22  <TrueBrain> I mean, it really is just 1GB if you play 24/7 for a full month .. insane if you think about it :)
20:47:55  <andythenorth> pikka sets up nice nocargoal games
20:47:58  <frosch123> well, if people join 20 servers until they find one they like, you easily get into a 1% of users use 99% of the bandwidth situation again :)
20:47:59  <glx> and in most case TURN won't be used
20:48:03  <andythenorth> I played a few the last few weekends
20:48:12  <andythenorth> 7 years, about 2 hours
20:48:17  <TrueBrain> frosch123: absolutely
20:48:38  <glx> we really should force join as spectator ;)
20:48:51  <TrueBrain> I still think we should remove the "lobby" window
20:49:00  <TrueBrain> and just join the game as spectator, with Network GUI window open
20:49:08  <frosch123> +1
20:49:23  <TrueBrain> that lobby window has no function, like .. none
20:49:30  <frosch123> milek7: what is the easiest method to tell cmake to add asan?
20:49:44  <TrueBrain> cmake .. -DCMAKE_CXX_FLAGS=-fsanitize=address
20:50:04  <TrueBrain> (possibly from an empty build-folder)
20:50:18  <glx> or a new OPTION so we can supprt more compilers ?
20:50:22  <TrueBrain> there is surprisingly little slowdown with debug builds when using that btw
20:50:29  <glx> I think MSVC has one too
20:50:32  <TrueBrain> (on GCC)
20:50:46  <TrueBrain> just when it finds a problem, it aborts, there and then
20:50:49  <TrueBrain> which is rather annoying
20:50:56  <frosch123> msvc has it since march
20:51:21  <glx> https://devblogs.microsoft.com/cppblog/addresssanitizer-asan-for-windows-with-msvc/
20:51:25  <frosch123> TrueBrain: aren't there env variables to control that?
20:51:59  <TrueBrain> possibly
20:52:14  <TrueBrain> haven't really looked into it any deeper
20:53:09  <frosch123> ok, triggered 9256 somehow
20:54:03  <frosch123> hmm, writes no coredump by default
20:54:43  <TrueBrain> no clue which sanitizers are enabled with just -fsanitize btw
20:56:09  <TrueBrain> owh, that is not allowed
20:56:10  <TrueBrain> lol
20:56:14  <TrueBrain> there are so many more sanitizers :P
20:56:38  <LordAro> maybe should run regression with them enabled ;)
20:57:21  <TrueBrain> do we dare using -fsanitize=undefined
20:57:40  *** Samu has quit IRC
20:59:19  <frosch123> 9256 exists since 2019
20:59:31  <TrueBrain> that is not too old
20:59:34  <TrueBrain> we had older bugs :P
21:01:31  *** gelignite has quit IRC
21:06:36  <TrueBrain> okay, -fundefined is not that bad
21:07:20  *** ChaKu_ has joined #openttd
21:08:25  <TrueBrain> lot of this: src/window.cpp:1429:5: runtime error: member access within address 0x55acd2078cd0 which does not point to an object of type 'Window'
21:08:48  <TrueBrain> guess why?
21:08:49  <TrueBrain> note: object is of type 'ZeroedMemoryAllocator'
21:08:50  <TrueBrain> :D
21:09:01  <LordAro> sounds like you need frosch's PR :p
21:09:09  <TrueBrain> didn't he merge yet?
21:09:32  <TrueBrain> src/core/bitmath_func.hpp:105:21: runtime error: shift exponent 255 is too large for 64-bit type 'long long unsigned int' <-that one sounds silly
21:09:40  *** Tirili has joined #openttd
21:12:35  <LordAro> sounds a bit '-1'
21:12:37  <peter1138> Incremental LTCG not compatible with Address Sanitizer, build without incremental LTCG.
21:12:40  <peter1138> Crap :p
21:16:18  <TrueBrain> LordAro: indeed, frosch123's PR solves that :)
21:16:26  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 opened pull request #9257: Codechange: document/fix virtual function resolving in con/destructors https://git.io/JsIHI
21:19:49  <DorpsGek> [OpenTTD/OpenTTD] frosch123 opened pull request #9258: Fix #9256, 12e43c697d2: invalid read after free. https://git.io/JsIHM
21:19:55  <glx> need more details about the HasBit() call
21:20:29  <frosch123> glx: do you remember the intent of #7716?
21:22:05  <glx> it was to inform AI about replacement, so it can link old with new
21:22:37  <frosch123> so is it okay to not create the event when replacing wagons, i.e. when the front remains the same?
21:22:46  <DorpsGek> [OpenTTD/OpenTTD] frosch123 merged pull request #9227: Make Window destruction not rely on undefined behavior https://git.io/J3NvF
21:24:42  <peter1138> Hmm, of course, the chat history is not a window.
21:30:52  <frosch123> we don't support transparent windows
21:31:43  <frosch123> the viewport drawing checks which areas are covered by windows, and tries to not draw them
21:31:45  <glx> frosch123: yeah, I think it was the intended meaning, just badly placed
21:31:59  <frosch123> though no idea, whether that is acutally a speed improvement or whether it makes it worse
21:34:44  *** Wolf01 has quit IRC
21:35:25  <DorpsGek> [OpenTTD/OpenTTD] glx22 approved pull request #9258: Fix #9256, 12e43c697d2: invalid read after free. https://git.io/JsI5w
21:36:15  <peter1138> Which makes it odd then that windows in front of viewport areas that have things changed get updated.
21:40:07  <DorpsGek> [OpenTTD/OpenTTD] frosch123 merged pull request #9258: Fix #9256, 12e43c697d2: invalid read after free. https://git.io/JsIHM
21:40:10  <DorpsGek> [OpenTTD/OpenTTD] frosch123 closed issue #9256: heap-use-after-free in ReplaceChain https://git.io/JsIaj
21:43:13  *** arikover has quit IRC
21:45:29  * andythenorth listening to news about vaccines in Europe 
21:45:38  <andythenorth> not so easy to get?
21:45:59  <glx> I'm too young for it
21:46:38  <andythenorth> UK is only just now doing 35-40 year olds
21:46:55  <glx> they are just started 50+ here
21:47:03  <andythenorth> and AZ isn't being given to < 40 so much here
21:47:09  <andythenorth> to encourage uptake
21:47:14  <frosch123> here doctors can give it to anyone, starting next week. (if they have any vaccines)
21:47:18  <glx> 55 min for AZ
21:47:30  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #9259: Fix #9257: [Network] TCPConnecter crashes when hostname not found https://git.io/JsIFU
21:47:36  <andythenorth> yeah news said 55 min
21:47:42  <andythenorth> France
21:47:46  <frosch123> the government centers which have the stuff stick to the order, and are still at 60+people
21:48:26  <TrueBrain> why do I keep linking the wrong tickets?
21:48:30  <frosch123> but you also get vaccinated if you have contact to a pregnant women
21:48:40  <frosch123> which spawns some new friendships :p
21:48:58  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsIFU
21:49:17  <TrueBrain> eeuuh .. frosch123  .. that doesn't sound like a good reason to "befriend" females :P
21:49:45  * andythenorth had AZ, still alive so far
21:49:46  <frosch123> you do not need to be the father. pregant people can name 2 contact persons
21:49:58  <TrueBrain> that makes more sense :D
21:50:04  <TrueBrain> it sounded rather shady :P
21:50:09  <andythenorth> oof
21:50:26  <frosch123> my mother also had AZ
21:50:35  <glx> my mother had moderna
21:50:48  <andythenorth> AZ risk is 5-10 days after injection in most cases, so see if I'm still here on Monday :P
21:51:00  <frosch123> TrueBrain: isn't it still shady?
21:51:34  <TrueBrain> nah, that is fine :P
21:51:52  <andythenorth> oof dishwasher is full, I will have to wash stuff
21:51:58  <andythenorth> or get a bigger dishwasher
21:52:21  <frosch123> is it broken and you use it like a cupboard?
21:53:20  <TrueBrain> writing a PR, looking at it, going .. no, that is just silly code
21:53:24  <andythenorth> no but it's a funny image frosch123
21:53:24  <TrueBrain> grr, I hate my self-review
21:53:32  <andythenorth> worst cupboard ever
21:54:42  <TrueBrain> we really should take a look at our includes ..
21:54:55  <TrueBrain> so silly that changing some files make everything recompile
21:55:14  * peter1138 ponders switching to bleeding-edge software on his routers.
21:55:57  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsIFU
21:56:13  <TrueBrain> now that looks a lot cleaner
21:56:42  <frosch123> TrueBrain: wait for c++20 support :)
21:57:33  *** jottyfan has joined #openttd
21:57:49  <peter1138> When do we start using enum class instead of enum?
21:58:14  <TrueBrain> frosch123: what horrors does that bring me?
21:58:18  <TrueBrain> peter1138: good point!
21:58:44  <peter1138> And also stop using enums for things that are clearly constants.
21:58:57  *** jottyfan has quit IRC
21:59:20  <frosch123> TrueBrain: modules. no more header files
21:59:34  <frosch123> or "better header files"
21:59:36  <TrueBrain> peter1138: pretty sure this should be an enum, if you are talking about my PR :)
21:59:52  <TrueBrain> I very much would liek the switch to complaint when a new entry is added :D
21:59:55  <peter1138> That, yes :)
22:00:13  <peter1138> Plenty of places everywhere that aren't :)
22:00:19  <TrueBrain> fair :D
22:00:47  <TrueBrain> also added missing doxygen ... oops :D
22:00:58  <TrueBrain> frosch123: bit like Rust?
22:00:58  <peter1138> I'm assuming you've already had the Connector vs Connecter discussion.
22:01:06  <TrueBrain> I keep typing it wrong, so annoying
22:01:16  <TrueBrain> well, "wrong"
22:01:45  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsIFU
22:01:48  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 commented on pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsIb1
22:01:53  <TrueBrain> there, enum in the class!
22:02:02  <peter1138> Brave.
22:02:25  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsIFU
22:02:30  <peter1138> Oh! In the class, not enum class.
22:02:36  <peter1138> Not quite so drastic.
22:02:40  <TrueBrain> there is also anum class?
22:02:41  <TrueBrain> enum
22:02:44  <TrueBrain> owh dear ...
22:03:19  <TrueBrain> sorry, I thought you meant enum in the class, as that was a good idea (to me at least)
22:03:27  <TrueBrain> but wtf is an enum class, and why does it exist ..
22:04:13  <peter1138> Type safety
22:04:20  <frosch123> and no namespace pollution
22:04:41  <TrueBrain> the latter is pretty nice
22:04:51  <TrueBrain> but in the class also avoids that, kinda, I guess? :D
22:05:26  <frosch123> well, defining enums inside classes was the old workaround, before enum classes were added :p
22:05:34  <peter1138> And you can do things like `enum class X : uint8`
22:05:50  <frosch123> that also works without the "class"
22:05:52  <TrueBrain> enums-ala-Python, basically
22:06:05  <peter1138> Ah ok.
22:06:19  <TrueBrain> means we don't have to prefix the enum values with the enum type, you say? :)
22:06:28  <frosch123> oh, but when you define the underlying type, you can forward declare enums now
22:06:39  <frosch123> TrueBrain: exactly
22:06:51  <TrueBrain> that would be a welcome change, honestly
22:06:55  <TrueBrain> that feels so dumb
22:06:58  <frosch123> and inside switches you can still do "using TCPConnectorStatus" to locally skip the prefix
22:07:22  <TrueBrain> well, shall we just start doing that? :P
22:07:37  <peter1138> Dunno will it stop compilation on XP?
22:07:45  <TrueBrain> I hope so
22:07:55  <peter1138> heh
22:07:57  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 approved pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsINs
22:09:51  *** lobstarooo has joined #openttd
22:10:20  <TrueBrain> using enum is C++20
22:10:25  <TrueBrain> YOU MADE ME HAPPY ABOUT NOTHING
22:10:37  <frosch123> what? no
22:10:39  <Rubidium> when can we use C23?
22:11:05  <TrueBrain> https://en.cppreference.com/w/cpp/language/enum <- using enum nested-name-specifier(optional) name ; 		(since C++20)
22:11:10  <TrueBrain> also failed to compile :P
22:12:11  <DorpsGek> [OpenTTD/OpenTTD] rubidium42 approved pull request #9254: Codechange: Replace TILE_AREA_LOOP with range-based for loops https://git.io/JsINX
22:12:17  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain dismissed a review for pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsINs
22:12:20  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsIFU
22:12:20  <TrueBrain> so something like this ^^ ?
22:13:20  <TrueBrain> does look like a lot less yelling, which is nice :)
22:13:24  <frosch123> hmm, i thought i used it before...
22:13:30  <frosch123> shows how few enums i use at work
22:13:40  <TrueBrain> hehe, sorry :)
22:13:45  <TrueBrain> can we switch to C++20? :P
22:13:58  <DorpsGek> [OpenTTD/OpenTTD] glx22 merged pull request #9254: Codechange: Replace TILE_AREA_LOOP with range-based for loops https://git.io/Jskwk
22:14:04  <TrueBrain> from C++11 to C++20 in a year yes yes? :P
22:14:20  <TrueBrain> it really does look better, enum classes
22:14:23  <TrueBrain> and it is even in a class
22:14:50  <TrueBrain> owh, can I also just call it Status now?
22:14:54  <TrueBrain> as the TCPConnecter prefix is implied?
22:15:00  <TrueBrain> or is that pushing it?
22:15:24  <frosch123> yep, from the outside it would be TCPConnector::TCPConnectorStatus::INIT :)
22:15:49  <TrueBrain> case TCPConnecter::Status::RESOLVING:
22:15:50  <TrueBrain> or
22:15:53  <TrueBrain> case Status::RESOLVING:
22:15:53  <TrueBrain> ?
22:15:54  <TrueBrain> :)
22:16:04  <frosch123> the latter
22:16:16  *** lobster has quit IRC
22:16:22  *** lobstarooo is now known as lobster
22:16:23  <TrueBrain> look at that, us being all modern :D
22:16:35  <glx> it's used from inside so no need to repeat class name
22:16:46  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #9259: Fix #9255: [Network] TCPConnecter crashes when hostname not found https://git.io/JsIFU
22:16:52  <TrueBrain> why I like this most, as it means my comment space-alignment works again :P
22:17:05  <frosch123> Rubidium: i do not follow the C standard. only when c++ copies stuff from it. but afaik c also mostly copies from c++ :p
22:17:27  <peter1138> TrueBrain, best reason for it
22:17:50  <TrueBrain> my motivations are sometimes a bit weird :P
22:18:18  <TrueBrain> right, all done, sorry Rubidium :) It became a bikeshedding PR
22:18:25  <TrueBrain> I wanted to go to bed early today .. lolz .. that failed :P
22:20:50  <peter1138> Anyway, looks like you move the thread creation out of the constructor and into the status checik
22:21:20  <peter1138> And my thought is ... thread creation in a constructor... what...
22:21:47  <TrueBrain> it has bitten me so many times by now
22:21:57  <andythenorth> yes! bedtime!
22:21:58  <TrueBrain> I am so happy I have a legit reason to move it out
22:22:02  <andythenorth> I was going to sleep at 10pm
22:22:04  <andythenorth> oops
22:22:17  *** andythenorth has quit IRC
22:22:23  <TrueBrain> it is so annoying that you create an object, and poof, it starts to do all these things for you
22:22:27  <TrueBrain> like: NOOOOO, NOT YET!!
22:22:36  <TrueBrain> so yeah ..
22:25:27  <glx> destructor are as bad :)
22:26:00  <peter1138> This one seems to be polarizing.. "yes of course you can" and "no never!"
22:26:48  <TrueBrain> it mainly was an issue when I tried to overload the behaviour
22:26:52  <TrueBrain> as in .. FUCK YOU NO
22:26:55  <TrueBrain> was the answer :P
22:27:23  <frosch123> is there a similar problem when exiting?
22:27:24  <peter1138> Yes, calling virtual members from a thread before the constructor has finished is... eh... no.
22:27:51  <TrueBrain> well, from a thread wasn't the issue there .. the fallback when threads were not there were a bit of an issue :D
22:28:04  <TrueBrain> with threads the issue was that it could call the onfailure callback without having a lock on the game-state
22:28:08  <TrueBrain> with equal BOOOMMMM effects
22:28:27  <peter1138> Pffft
22:28:48  <TrueBrain> threads are hard yo
22:28:54  <peter1138> We should just have 10GHz CPUs by now.
22:29:01  <milek7> connecter?
22:29:30  <TrueBrain> frosch123: what do you mean, when exiting?
22:29:57  <milek7> hm, it apparently exists in both variants
22:30:58  <frosch123> TrueBrain: previously OnFailure was called from the thread, right?
22:31:09  <TrueBrain> yes
22:31:29  <frosch123> that would have been a problem when exiting. as the thread would could call the callback, while the callback is already destroyed
22:31:39  <frosch123> but now it seems like everything is in the main thread
22:31:44  <TrueBrain> exactly :)
22:31:55  <TrueBrain> it is now really clear there cannot be a threading issue :P
22:32:26  <TrueBrain> force-everything-in-the-main-thread-with-signals-and-be-done-with-it :D
22:33:59  <TrueBrain> I like enum class in class, it is pretty
22:34:05  <TrueBrain> going to do that in 9017 tomorrow too :D
22:35:11  *** Tirili has quit IRC
22:37:06  <frosch123> it is already tomorrow
22:37:53  <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #9257: Codechange: document/fix virtual function resolving in con/destructors https://git.io/JsIp5
22:37:59  <TrueBrain> "later today"
22:38:17  <TrueBrain> well, at least many people looked at 9259 now :P
22:38:29  <TrueBrain> night all
22:45:07  *** Progman has quit IRC
22:48:16  *** frosch123 has quit IRC
23:12:11  *** sla_ro|master has quit IRC
23:22:49  *** snail_UES_ has quit IRC
23:42:07  *** snail_UES_ has joined #openttd

Powered by YARRSTE version: svn-trunk