IRC logs for #openttd on OFTC at 2020-09-08
⏴ go to previous day
00:45:18 *** WormnestAndroid has quit IRC
00:45:35 *** WormnestAndroid has joined #openttd
01:53:19 *** Progman has joined #openttd
02:42:45 *** Lejving_ has joined #openttd
02:47:26 *** ad5twoknebor[m] has quit IRC
02:52:15 *** Heiki[m] has joined #openttd
02:55:28 *** XeryusTC has joined #openttd
02:56:21 *** dihedral has joined #openttd
02:57:58 *** planetmaker has joined #openttd
02:57:58 *** ChanServ sets mode: +o planetmaker
02:59:42 *** ^Spike^ has joined #openttd
03:02:40 *** WormnestAndroid has quit IRC
03:03:14 *** WormnestAndroid has joined #openttd
03:31:58 *** WormnestAndroid has quit IRC
03:32:44 *** WormnestAndroid has joined #openttd
04:14:47 *** snail_UES_ has joined #openttd
06:06:55 *** virtualrandomnumber has joined #openttd
07:48:12 *** andythenorth has joined #openttd
08:04:38 *** sla_ro|master has joined #openttd
08:09:21 *** andythenorth has joined #openttd
08:24:45 *** Hirundo has joined #openttd
08:24:45 *** Hazzard has joined #openttd
08:24:45 *** V453000 has joined #openttd
08:24:45 *** Terkhen has joined #openttd
08:24:45 *** vanessa[m] has joined #openttd
08:24:45 *** rudolfs[m] has joined #openttd
08:24:45 *** robert[m] has joined #openttd
08:24:45 *** pina[m] has joined #openttd
08:24:45 *** patrick[m] has joined #openttd
08:24:45 *** natmac[m] has joined #openttd
08:24:45 *** leward[m] has joined #openttd
08:24:45 *** khavik[m] has joined #openttd
08:24:45 *** josef[m]1 has joined #openttd
08:24:45 *** jeeg[m] has joined #openttd
08:24:45 *** hamstonkid[m] has joined #openttd
08:24:45 *** glothit7ok[m] has joined #openttd
08:24:45 *** fiddeldibu[m] has joined #openttd
08:24:45 *** elliot[m] has joined #openttd
08:24:45 *** christoph[m] has joined #openttd
08:24:45 *** cawal[m] has joined #openttd
08:24:45 *** blikjeham[m] has joined #openttd
08:24:45 *** Aileen[m] has joined #openttd
08:24:45 *** ad5twoknebor[m] has joined #openttd
08:24:45 *** tyteen4a03 has joined #openttd
08:24:45 *** liquid.oftc.net sets mode: +ov Terkhen Terkhen
11:03:40 *** arikover has joined #openttd
12:07:28 *** Progman has joined #openttd
12:11:23 <DorpsGek_III> [OpenTTD/team] LordAro approved pull request #2: Add: initial workflow to automate requested access for translators https://git.io/JUnRj
12:53:39 <Eddi|zuHause> what kind of oops?
12:54:21 <andythenorth> looks like they got routed with a high-roof freight car under a low-roof bridge
16:34:36 *** andythenorth has joined #openttd
16:59:44 *** Wormnest has joined #openttd
17:48:28 *** snail_UES_ has joined #openttd
18:13:23 *** gelignite has joined #openttd
18:39:01 <andythenorth> Timberwolf I have a lot of US-based Horse plans :D
18:39:08 <andythenorth> take my fake roster, voxelise it :P
18:39:48 <Timberwolf> If I build enough trains I'll have enough data to build an AI model for it, right?
18:40:02 <andythenorth> generative trains have been talked about
18:40:10 <andythenorth> there are only so many components and patterns
18:40:43 <andythenorth> I did make a lorry/truck generator for an RV set called BANDIT, but I deleted it in a hissy fit
18:41:10 <andythenorth> it took blocks of pixels and composited them in different ways
18:41:11 <Timberwolf> I did have a go with creating something along those lines years ago (LocoGrammar, which I don't think I ever released) but it ended up being more time than just making models the dumb way.
18:41:29 <andythenorth> Horse composites vehicle chassis, roofs, lights etc
18:41:38 <andythenorth> but they're all quite large components
18:47:49 <Eddi|zuHause> what's the minimum number of input trains for an AI to make reasonable guesses? 1 million? 100 million?
18:49:09 <Timberwolf> I wonder what results you'd get using an adversarial technique with one AI generating voxel objects and the other reversing output from GoRender.
18:49:28 <Timberwolf> Would probably only work for things produced using that workflow.
18:49:58 <Eddi|zuHause> i've no clue what any of those words mean
18:50:54 <andythenorth> Eddi|zuHause pretend you do :)
18:51:19 <Eddi|zuHause> is that the opposite of bullshit bingo?
18:51:21 <Timberwolf> You have two competing AIs. One tries to turn pixels into voxel objects. The other is trying to create voxel objects that, when rendered using my toolchain, cause the first AI to give a result that is wrong.
18:51:42 <Timberwolf> It's one of the ways to overcome the "needing millions of data points" problem.
18:51:53 <Timberwolf> Just have an AI fight until no-one wins.
18:54:43 <Eddi|zuHause> oh, you mean like global thermal nuclear war?
18:54:55 <Timberwolf> local thermonuclear war :)
18:55:12 <andythenorth> yes, train newgrfs occupy the same part of the venn diagram as global thermonuclear war
18:58:06 <Eddi|zuHause> have you seen that movie?
19:00:09 <Timberwolf> Yes. Hacking movies started out reasonable and declined in accuracy ever since.
19:00:26 <Timberwolf> Vector of attack: social engineering!
19:01:30 <LordAro> Timberwolf: Hack the Planet!
19:01:50 <LordAro> only the best hacking movies have a villain who rides around on a skateboard
19:02:22 <Timberwolf> Heh. The '90s ones started to blur into one for me.
19:02:37 <Timberwolf> Was that the one with the yellow 3.5" floppy disk?
19:03:21 <Eddi|zuHause> are there more than 2 good "hacking" movies?
19:07:02 <andythenorth> was WarGames about Von Neumann?
19:07:09 <andythenorth> hacking was incidental
19:10:35 <LordAro> Hackers is the most 90s movie ever
19:11:02 <LordAro> and i refuse to accept that there is any valid criticism of it
19:21:28 <Eddi|zuHause> was tron a hacking movie?
19:22:02 <andythenorth> office space is the definitive hacking movie
19:22:08 <andythenorth> whatever LordAro says
19:34:21 <Timberwolf> Damn, it feels good to be a gangster.
19:43:26 <TrueBrain> "[2020-09-08 19:43:15] dbg: [net] [udp] advertising on master server successful (IPv4)"
19:53:08 *** frosch123 has joined #openttd
20:16:47 <TrueBrain> okay, and the API is working and acting stable ... this is going well .. time to deploy the web
20:19:01 <frosch123> at this point noone is surprised that you are the one deploying THE WEB
20:20:34 <TrueBrain> "Internal Server Error"
20:27:02 <TrueBrain> find the bugs pleaseeeeeee
20:27:43 <frosch123> some servers ban players named "player"
20:27:53 <frosch123> should the masterserver ban servers named "unnamed server"?
20:29:04 <frosch123> ah, clicking on the version filters the list
20:29:19 <TrueBrain> and it is a startswith filter
20:29:47 <frosch123> so versions cannot start with "all" :p
20:29:57 <TrueBrain> you reviewed the PR! :P
20:30:24 <TrueBrain> but so you can filter on all jgrpp servers
20:30:50 <frosch123> i did not read the html, for common reasons
20:31:03 <TrueBrain> it is part of the Python! :P
20:31:15 <TrueBrain> but I will fix this, no worries ;)
20:33:56 <frosch123> like "unknown server id"
20:34:50 <TrueBrain> I agree, sounds like a good-first-issue; will make an issue for that :)
20:36:37 <frosch123> oh... there is a duplicate "UTC"
20:36:52 <frosch123> only noticed that on the 5th read or so :p
20:38:14 <DorpsGek_III> [OpenTTD/master-server-web] TrueBrain opened pull request #5: Several bugs found while bringing this to production https://git.io/JUn7G
20:40:16 <DorpsGek_III> [OpenTTD/master-server] TrueBrain opened pull request #4: Several bugs found while bringing this to production https://git.io/JUn78
20:41:17 <DorpsGek_III> [OpenTTD/master-server] TrueBrain updated pull request #4: Several bugs found while bringing this to production https://git.io/JUn78
20:41:21 <TrueBrain> fixed, and tnx for spotting :D
20:41:39 <TrueBrain> you can also announce your server if you like, but it is a bit of a pita .. on a different port, so you need to recompile OpenTTD
20:42:02 <TrueBrain> nlb.openttd.org / 4978
20:42:23 <TrueBrain> did not test yet if UTF-8 works correctly .. can I do that easily ..
20:43:22 <frosch123> hmm... looks like about everything includes network/config.h
20:45:19 <TrueBrain> ugh, so many caches .. takes so long to see new server data on the web :P
20:47:39 <TrueBrain> soon I will deploy this to production, without switching the infra over to AWS .. that means people can test this with a simple "hosts" modification
20:47:46 <TrueBrain> might be easier to get some testing done
21:05:23 <frosch123> well... i wonder whether my router got a firmware update... and they removed the port forwarding option :/
21:15:33 <andythenorth> I had an external security assessment that put all our staff home routers in scope
21:15:43 <andythenorth> and required them to not have open ports etc
21:16:36 <andythenorth> hmm is Schadenfreude a uniquely English concept, and we've just stolen a German word?
21:17:22 <frosch123> no, there are about 1 million of yt videos of english people discovering that it is a german word
21:19:02 <andythenorth> we should go and laugh at their misfortune
21:19:16 <frosch123> anyway, the manual still lists the option
21:19:20 <frosch123> it just does not exist :p
21:19:38 <andythenorth> congrats, you can now get a certificate
21:20:34 <andythenorth> oh no wait, you have to get a VPN as well for the certificate
21:20:49 <andythenorth> and it has to terminate inside your network perimeter
21:20:53 * andythenorth is not quite over this yet
21:22:03 <andythenorth> the concept of a 'trusted network' could only be dreamed up by people who did not live in student accommodation with security-curious house mates
21:23:06 <andythenorth> or a university accommodation block with 300 people on a 10Mbit ethernet LAN to every room
21:38:32 <Timberwolf> Hah. My cohort, the year before the university had run trials in one of the halls and gone, "we can run this by reserving 20% of our JANET link and doing aggressive QoS rules on ports we know things like Napster use"
21:39:10 <Timberwolf> In between trials and launch, Napster got banned and was replaced by the generation of services that did port-hopping to get around firewall blocks and QoS.
21:39:29 <Timberwolf> You could get Internet in halls if you woke up before 7, or 9 on a weekend.
21:39:54 <andythenorth> that stuff was just starting to happen when I was there, 1997 or so
21:40:33 <andythenorth> mp3.com probably
21:40:55 <Timberwolf> One of the guys in my year group built a site that crawled every open SMB share it could find in its subnet.
21:41:23 <frosch123> how much did he make with blackmailing?
21:41:33 <andythenorth> lol, and I'm still answering procurement questions about 'security at the perimeter'
21:41:51 <andythenorth> and cable inspections
21:42:28 <Timberwolf> None that I recall :). I think his first naive implementation it routed the downloads via the server, which was of course a desktop PC in his... what do you call the individual unit of a hall of residence?
21:42:35 <Timberwolf> "flat" seems wrong somehow.
21:42:39 <Timberwolf> I guess it's just a room.
21:42:43 <andythenorth> we had flats, 6 rooms
21:42:49 <andythenorth> otherwise corridors
21:43:01 <_dp_> does anyone know how to ssh tunnel openttd server via remote host?
21:43:44 <TrueBrain> okay, automatic delisting after 20 minutes also works \o/ w00p
21:43:45 <andythenorth> with a VPN to an endpoint inside your trusted network perimeter
21:43:48 <TrueBrain> full chain test completed :)
21:44:39 <Timberwolf> We should create something like PCI/DSS for running OpenTTD servers.
21:44:55 <Timberwolf> With levels, so you need a different questionnaire depending on whether you use newgrfs and/or GS.
21:45:24 <andythenorth> and a certification scheme
21:45:37 <andythenorth> oh no wait, the pro move is to certify third party assessors
21:45:52 <andythenorth> charge them for their certification AND take a cut of what they charge
21:46:08 <andythenorth> then ideally get it adopted as a standard by government and privacy regulators
21:47:22 <andythenorth> PCI-DSS oof, that was 10 minutes of my day today, my future contains a questionnaire :(
21:48:09 <andythenorth> we have some card payment processing facility included with our accounting software, so we now have to do a PCI-DSS assessment
21:48:25 <frosch123> i forgot that 1.10.3 does not use cmake :)
21:48:58 <andythenorth> on a totally unrelated point, shall we release something?
21:49:30 <frosch123> release what? noise? gas?
21:49:31 <Timberwolf> It's not too bad, providing you can get a QSA who will let you provide sensible compensating controls for the requirements which are out of date. (I think they're still on 3, right?)
21:50:14 <andythenorth> I have NFI, I'm hoping the self-assessment questionnaire concludes 'move on, nothing to do'
21:50:49 <andythenorth> giving customers a 'pay here' link to Paypal / Strip / Xero / SAGE whatever it is shouldn't bring us into scope
21:51:07 <Timberwolf> I had fewer problems with PCI itself than I did with bosses who would continually repeat, "but do we actually need to do this?" -> yes, it's in several of our supplier contracts.
21:51:32 <andythenorth> Timberwolf I have the inverse of that situation :P
21:51:35 <Timberwolf> That should be simple, it's either A or A-EP depending on how the direction to it works.
21:52:19 <Timberwolf> Both are pretty simple questionnaires IIRC, I think Stripe will even give you a prefilled version to just tick some boxes on.
21:53:20 <Timberwolf> As long as card numbers never enter your network, it's not too bad. At worst a few pages of awkwardly worded "are you managing to avoid being complete idiots?"
21:54:01 * andythenorth skim read the standard
21:54:12 <andythenorth> A-EP seems a lot of work for 0.6% of our revenue
21:55:05 <andythenorth> PCI-A looks more promising
21:55:39 <frosch123> _dp_: is your questionas generic as "how to tunnel a port with ssh" or is there something ottd specific?
21:56:18 <Timberwolf> Yeah, I think the web ones which are all front-end you end up with A mostly? The wording is kind of awkward, and I've only ever done PCI at places which already had card data stored (or at least entering the network)
21:56:44 <Timberwolf> So my experience is mostly, "can we avoid D, can we avoid D, can we... no, we can't."
21:57:07 * andythenorth wonders if PCI-DSS can be avoided
21:57:40 * andythenorth looks which particular mafia is enforcing it
21:57:47 <Timberwolf> You can ignore it until you have a breach, providing no suppliers or clients ask for it :)
21:58:40 <_dp_> frosch123, I guess you can call it that, but it doesn't seem that simple as it's udp port
21:58:41 <andythenorth> ha ha the top result on google for UK is the company who fucked up our Cyber Essentials and gave us illegal requirements
21:58:50 <_dp_> also tcp also needs tunneling iirc
21:59:01 <Timberwolf> Although ISTR if you do have one, you immediately win the strictest possible audit or having Visa/MC refuse to process your payments.
21:59:03 <_dp_> ssh -v -L 4978:nlb.openttd.org:4978 -R 3979:localhost:3979 -R 4978:localhost:4978 -R 3978:localhost:3978 -L 5978:localhost:5978 -N dp
21:59:24 <_dp_> and two netcat fifos but it's still not working and I've no idea why
22:00:46 <_dp_> ok, I found one more mistake xD
22:01:50 <andythenorth> can't be that every coffee shape and bagel van has PCI-DSS
22:01:54 <andythenorth> for their card machine
22:02:35 <Timberwolf> I think you get away with that on the basis the onus is on the person who supplied you the machine.
22:03:02 <frosch123> _dp_: -R defaults to listing to localhost only, you may need to add a *: in front
22:04:13 <Timberwolf> At least that's how we got round it back when I worked for Priority Pass (someone had to). The individual lounges don't have to certify, but we as the supplier of the card reader did
22:05:40 <andythenorth> every day is a compliance day
22:06:31 <andythenorth> my life for the last 3 years has been 27001 (which I like) Cyber Essentials (which I hate) GDPR (which is lolz) WCAG 2.1 (fine), PECR (total absolute lolz) and now PCI-DSS
22:06:45 * andythenorth only wanted to make Flash games
22:06:47 <Timberwolf> (re Priority Pass, the "you can use your bank card as the access device" was simultaneously the best and worst idea they ever had)
22:06:59 <andythenorth> 'what could go wrong'
22:07:17 <frosch123> andythenorth: you you have a wasm flash emulator?
22:07:25 <Timberwolf> You could have several million customers and no clue who they actually are beyond "they have the right sort of Amex".
22:08:56 <andythenorth> " It makes me sad to think that the next generation of edgy teens wont have the outlet that we old heads had back then."
22:10:21 <frosch123> what? 114 days left? i thought it was already dead
22:10:24 <Timberwolf> Heh. In the Flash era that was what the people who'd grown up on BBC BASIC were saying :)
22:10:59 <andythenorth> I was about 22 when I started using Flash
22:11:06 <andythenorth> so I'd grown up on BBC BASIC :P
22:11:39 <Timberwolf> GW-BASIC, QBASIC and then the DOS port of gpp (DJGPP) with Allegro for me.
22:12:34 <andythenorth> ho, are paypal merchants supposed to be PCI-DSS?
22:12:58 <andythenorth> survey says: YES
22:13:08 <andythenorth> wow I hope someone told all the Etsy people
22:14:23 <Timberwolf> We had an Acorn Electron when I was tiny, although I was too young to do anything much with it.
22:14:36 <Timberwolf> Although I swear I was better at Escape From Moonbase Alpha when I was 5 than I am now.
22:16:30 <frosch123> unlikely. kids are terrible at games, they just don't notice
22:16:36 <andythenorth> oof, coop needs to do PCI-DSS
22:16:40 <andythenorth> good job devzone is down
22:17:50 <andythenorth> do we have a bank?
22:18:39 <frosch123> andythenorth: we only receive money. we do not collect them
22:19:17 <andythenorth> so owen is the merchant?
22:19:40 <andythenorth> I love these mafia schemes like PCI
22:19:41 <frosch123> yes, owen lives on a south sea island with all the ottd money
22:19:52 <andythenorth> it's not actually a law, but someone might be VERY MEAN TO US
22:20:32 <andythenorth> well maybe the new MicroProse can do a PCI-DSS for owen
22:20:39 <andythenorth> as a sponsorship
22:22:59 <TrueBrain> andythenorth: please stop making noise where no noise is needed ;) That you are neck deep in PCI shit .. don't drag OpenTTD into that void :P ktnx :)
22:24:26 <andythenorth> I was waiting for the sensible TrueBrain view
22:24:44 <andythenorth> I had to type quite a lot to get it, usually the service is quicker
22:25:11 <andythenorth> Truebrain-as-a-Service
22:32:22 <andythenorth> meh all these PCI controls are just 27002 so far
22:32:25 * andythenorth having a fun evening
22:32:53 <andythenorth> maybe I should do pixels
22:33:57 <frosch123> maybe try pcie x16. it's a lot faster
22:35:09 * andythenorth finds new things to do
22:35:18 <andythenorth> can I glue styrene-styrene with cyano?
22:35:29 <andythenorth> I don't have any plastic weld type stuff
22:36:51 <TrueBrain> okay, no more errors on sentry .. guess that is a good thing :)
22:37:24 <TrueBrain> now I just need to configure GitHub to auto-deploy etc .. but that is something for another day
22:37:40 <TrueBrain> owh, and ofc ask pretty please for someone to review my bugs-in-production PRs :D
22:40:35 <Timberwolf> I should do some pixels but I keep stalling at the idea of setting up varaction2s for station boundary fences.
22:40:50 <andythenorth> Timberwolf you are now approaching my level of newgrf complexity :D
22:40:56 <andythenorth> you got there fast, congrats
22:41:06 <Timberwolf> NML helped a lot :)
22:41:20 <andythenorth> first you make progress quick, then suddenly all projects and ideas seem related :D
22:41:49 <Timberwolf> Fences ought not to be that difficult, a few station sets have it (MB's newstats particularly), I'm just not very good with NFO.
22:42:34 <Timberwolf> I mostly cope by turning everything into a Go program, which I can happily fumble about in.
22:43:01 <andythenorth> I never really understood station nfo
22:44:15 <Timberwolf> It's not as well-trodden as the various vehicle options. Like with NML, I was able to build on a lot of stuff other people had done and documented, like Eddi's stuff with oversize vehicles and all the IRC chats about how sprite stacks worked.
22:45:00 <frosch123> you use irc logs as docs? :)
22:45:04 <Timberwolf> But for NFO there's the AuzStations tutorial, some of the forum threads around the CZTR team doing things, and then newgrf specs and hard work.
22:45:13 <Timberwolf> There's a website which aggregates them somewhere.
22:45:31 <Timberwolf> When you search some of the more obscure NML variables they often end up being the 2nd or 3rd result.
22:46:39 <Timberwolf> Sometimes they're very useful, especially if it's a discussion with someone whose thought process is similar to mine, and they ask the same questions in the same order.
22:48:25 <frosch123> it even compiles nml-style lang files into nfo
22:50:58 <FLHerne> frosch123: Who doesn't?
22:51:21 <FLHerne> There's a *reason* I have several gigabytes of the thngs
22:51:55 <frosch123> it's just text, how do you get gb?
22:52:39 <FLHerne> I did try to write a tool to answer queries from them semi-automatically, but it didn't work :p
22:52:52 <FLHerne> Years of logs of many channels
22:52:54 <frosch123> 1037801 15139532 105137794 oftc_#openttd.log <- ok, i understand :)
22:53:09 <FLHerne> They'd probably compress pretty well
22:53:50 <FLHerne> Actually, given current I/O vs CPU speed, it might even be faster to grep them compressed...
22:54:34 <frosch123> store them on an pcie ssd?
22:55:32 <FLHerne> I do, but it's a waste of space
23:15:41 *** tokai|noir has joined #openttd
23:15:41 *** ChanServ sets mode: +v tokai|noir
23:48:14 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz commented on pull request #8285: Rewrite `OverflowSafeInt` to avoid undefined behaviour https://git.io/JUnhU
continue to next day ⏵