Config
Log for #openttd on 27th April 2020:
Times are UTC Toggle Colours
00:02:13  <DorpsGek_III> [OpenTTD/nml] glx22 opened pull request #108: Codechange: [Actions] Upload standalone executable to GitHub release https://git.io/JfqJv
00:25:34  <mcbanhas> Is it possible to make UI elements semi-transparent at this point?
00:25:56  <mcbanhas> I mean, does the engine support it?
00:40:45  *** mcbanhas has quit IRC
01:14:24  *** adikt has joined #openttd
01:39:34  *** Flygon has joined #openttd
02:01:00  *** Wormnest_ has quit IRC
02:11:29  *** glx has quit IRC
02:15:33  <FLHerne> Is there any case in which the value of a variable read by ActionD can change during one loading pass?
02:15:54  <FLHerne> Other than the grf directly writing to that variable
02:17:49  *** ahmet has joined #openttd
02:17:56  <ahmet> hi
02:18:34  <FLHerne> ahmet: Good morning :-)
02:19:16  <FLHerne> [timezones may vary]
02:19:30  <ahmet> where are you from?
02:19:35  <FLHerne> UK
02:19:53  <FLHerne> It's 3am, I was about to sleep
02:20:09  <FLHerne> And you?
02:20:16  <ahmet> Turkey
02:20:19  <ahmet>  great to see who played this game!
02:21:14  <ahmet> This game is very enjoyable in quarantine days
02:24:04  <FLHerne> Yes, I've been playing more than I had been
02:25:48  <ahmet> The buttons in the game are very small. It hasn't changed for years. My fingers hurt after hours
02:26:05  <ahmet> but i never quit :)))
02:26:27  *** debdog has joined #openttd
02:26:28  <FLHerne> You can make them bigger somehow
02:27:24  <FLHerne> https://wiki.openttd.org/FAQ_troubleshooting#The_User_Interface_buttons_are_too_small
02:27:45  <ahmet> wow
02:27:46  <FLHerne> There's a NewGRF "OpenGFX+ BigGUI" that makes them bigger
02:27:53  <ahmet> thank you!
02:28:21  <ahmet> I don't want to keep you from sleep.. :))
02:28:21  <FLHerne> Enabling it as a 'static' grf as shown on the wiki will mean it's active in multiplayer too
02:28:32  <FLHerne> Even if the server doesn't use it
02:29:13  <FLHerne> Yes, bedtime for me. Goodnight. :-)
02:29:46  *** D-HUND has quit IRC
02:30:16  *** ahmet has quit IRC
03:09:41  *** Wormnest_ has joined #openttd
04:02:01  *** snail_UES_ has quit IRC
04:02:17  *** Wormnest_ has quit IRC
05:18:28  *** berndj has quit IRC
05:18:29  *** syr has quit IRC
05:19:10  *** berndj has joined #openttd
05:20:32  *** syr has joined #openttd
05:21:12  *** Hirundo has quit IRC
05:21:12  *** avdg has quit IRC
05:23:37  *** TinoDidriksen has quit IRC
05:23:56  *** avdg has joined #openttd
05:24:26  *** Hirundo has joined #openttd
05:24:30  *** XeryusTC has quit IRC
05:24:40  *** TinoDidriksen has joined #openttd
05:25:11  *** TinoDidriksen is now known as Guest23449
05:26:56  *** XeryusTC has joined #openttd
05:27:38  *** Guest23449 is now known as TinoDidriksen
06:02:30  *** sla_ro|master has joined #openttd
06:29:00  *** Wolf01 has joined #openttd
06:45:21  *** andythenorth has joined #openttd
07:53:24  <DorpsGek_III> [OpenTTD/nml] andythenorth opened issue #109: Roadtypes and Tramtypes size is 63 not 15 or 64 https://git.io/JfqYj
07:53:52  <DorpsGek_III> [OpenTTD/nml] andythenorth commented on pull request #108: Codechange: [Actions] Upload standalone executable to GitHub release https://git.io/JfqOv
08:22:03  <TrueBrain> morning
08:24:33  <LordAro> can confirm
08:27:39  <TrueBrain> happy you did
08:27:49  <TrueBrain> lets migrate BaNaNaS, I would say .. see how that goes ..
08:28:10  <TrueBrain> lot of moving parts .. lets see what the best order is ..
08:30:55  <orudge> I was just coming on here to see how things were going with that and the old VPSes :D
08:30:56  <Wolf01> I think the first thing should be "put everything readonly" :P
08:32:05  <TrueBrain> orudge: I think we just let them expire; for now I don't see a reason to create new ones :)
08:32:12  <TrueBrain> but today will tell ;)
08:32:17  <TrueBrain> otherwise I have 2 days to install new VPSes
08:32:30  <TrueBrain> honestly, only 40 dollar per year for 2, so .. not a real issue ofc, but meh :)
08:33:44  <orudge> Yeah
08:34:17  <orudge> Are the contents of the old VPSes (or one of them at least) backed up somehere just in case? :D
08:34:33  <TrueBrain> they are the backups :)
08:34:48  <TrueBrain> it is just an nginx with rsync access
08:34:56  <TrueBrain> and all the binaries we have
08:34:58  <orudge> OK
08:35:12  <TrueBrain> but I will logon to them today/tomorrow to make sure about that :D
08:35:34  <orudge> I also host an OpenTTD mirror or two of some sort, if I remember correctly
08:35:40  <TrueBrain> you do :)
08:35:45  <TrueBrain> US and UK I think
08:35:58  <TrueBrain> these two VPSes do CA and FR
08:36:19  <TrueBrain> but that traffic on all of them should disipate today :P
08:36:34  <orudge> Yes
08:36:37  <TrueBrain> it should already been reduced for a while now btw
08:36:52  <TrueBrain> not sure if you have graphs of these, but like last year or something it should be noticable lower
08:37:59  <orudge> Possibly, I haven't really been paying attention :D
08:38:06  <TrueBrain> okay, time to deploy everything on AWS already .. that requires some certificates to be made .. this is going to take a while :D
08:38:25  <TrueBrain> also not that important; just a bit of stats-nerding :D
08:38:45  <TrueBrain> your DNS admin panel is a bit weird sometimes btw .. it reset all the TTLs :P
08:40:38  <TrueBrain> k, changed the TTL to 60 for the domains I am about to move .. that should make the transition a bit smoother for most users
08:41:32  <orudge> TrueBrain: yes, probably, while performing a server transfer. It's not really designed for the purpose we're using it for. :) If it's easier to host the DNS on Amazon or soemthing then feel free.
08:41:51  <TrueBrain> over time, maybe :)
08:42:06  <TrueBrain> we already have an AWS zone for openttd.org, but .. not all records are there yet
08:42:07  <TrueBrain> SO MANY RECORDS
08:42:08  <TrueBrain> :P
08:42:13  <TrueBrain> I am happy you are doing it still :)
08:56:41  <andythenorth> happy launch day TrueBrain !
08:56:49  * andythenorth going afk, that work thing I mentioned :P
08:56:53  <andythenorth> GL HF
08:56:56  *** andythenorth has left #openttd
09:01:28  *** Samu has joined #openttd
09:11:04  <DorpsGek_III> [OpenTTD/website] TrueBrain opened pull request #158: Create 2020-04-27-new-bananas.md https://git.io/Jfqn4
09:11:16  <TrueBrain> possibly I have been blablaing a bit too much, but I would love input :)
09:18:35  <DorpsGek_III> [OpenTTD/website] TrueBrain updated pull request #158: Create 2020-04-27-new-bananas.md https://git.io/Jfqn4
09:19:38  <Samu> hello
09:25:50  <DorpsGek_III> [OpenTTD/website] LordAro commented on pull request #158: Create 2020-04-27-new-bananas.md https://git.io/Jfqnx
09:26:01  <LordAro> TrueBrain: reads fine, imo
09:26:51  <TrueBrain> cool, tnx :)
09:28:31  <milek7> 'ABSOLUTELY nothing.' vs 'should solve many of the issues with BaNaNaS, like proper UTF-8 support, being able to change the name of your content, being able to edit existing content'
09:28:34  <milek7> seems contradictory
09:33:19  <TrueBrain> yup :)
09:35:46  <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain opened pull request #32: Cache one endpoint, and tweak text about downloads https://git.io/JfqcG
09:35:56  <TrueBrain> I wonder if frosch123 is going to see this .. :D
09:36:32  <TrueBrain> I kinda need that cache to be in there before we go live :P We have too many NewGRFs :D
09:39:32  <TrueBrain> okay, everything deployed to AWS without issue .. I love CDK :)
09:44:36  <DorpsGek_III> [OpenTTD/website] TrueBrain updated pull request #158: Create 2020-04-27-new-bananas.md https://git.io/Jfqn4
09:59:08  <TrueBrain> okay, NSIS installer still works on new infra, that makes me really happy to see :)
09:59:33  <Artea> hello
10:06:21  <TrueBrain> okay .. sometimes, when moving towards production, you can find these nasty bugs ... in nginx, you can only have 1 "stream{}" block .. that makes writing configurations a bit more difficult :P
10:12:08  <TrueBrain> right, that was an easy fix
10:12:24  <TrueBrain> k, lunch time, after that, I am going to put BaNaNaS v1 on read-only :)
10:12:45  <TrueBrain> LordAro: if you have some time, I would love to get https://git.io/JfqcG merged before we go live :)
10:13:51  <DorpsGek_III> [OpenTTD/bananas-frontend-web] LordAro approved pull request #32: Cache one endpoint, and tweak text about downloads https://git.io/JfqC7
10:14:52  <TrueBrain> ty kind!
10:14:53  <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain merged pull request #32: Cache one endpoint, and tweak text about downloads https://git.io/JfqcG
10:20:20  <LordAro> TrueBrain: i'd suggest that the more "pythonic" way of doing it would be via a function decorator
10:20:41  <LordAro> then you could just add @cache_api(60 * 5) to the top :)
10:31:22  <TrueBrain> Fully agree. Would be a lot cleaner
10:36:52  <TrueBrain> Possibly Flask has a decorator for it .. owh well, this works for sure for now :p
10:37:52  <LordAro> https://cachetools.readthedocs.io/en/stable/#cachetools.func.ttl_cache seems quite nice
10:45:41  <TrueBrain> I normally use that .. no clue why I didn't this time .. wanted to have it fixed I guess :p
10:45:58  <LordAro> :)
10:48:35  *** mcbanhas has joined #openttd
10:48:44  <mcbanhas> Howdy
10:49:29  <mcbanhas> Is it possible to set transparency for UI elements? Does it the engine support it currently?
10:50:18  <LordAro> nope
10:52:33  <mcbanhas> So how is transparency currently handled in map objects? Are those dithered silhouettes or something similar?
10:52:58  <TrueBrain> alright .. lets put some things in read-only mode, and do this thing ..
10:54:29  <mcbanhas> Hmm, no it doesn't look like dithering.
10:57:16  <TrueBrain> oef ... am I really going to do this? These changes always make me nervous :D
11:11:50  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain created new tag: 1.0.0 https://git.io/JfqlN
11:13:48  <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain created new tag: 1.0.0 https://git.io/Jfqlp
11:14:02  <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain created new tag: 1.0.0 https://git.io/Jfq8e
11:14:15  <DorpsGek_III> [OpenTTD/binaries-redirect] TrueBrain created new tag: 1.0.0 https://git.io/Jfq8f
11:14:18  <TrueBrain> RELEASE ALL THE THINGS
11:22:35  <TrueBrain> uploading with 500 mbit/s to AWS :D I love that speed :P
11:27:20  <TrueBrain> https://github.com/OpenTTD/BaNaNaS <- in case anyone wants to see our database :P
11:33:43  <TrueBrain> haha, it takes too long for bananas-server to read that repo, that is is being killed as "unresponsive"
11:33:44  <TrueBrain> :D
11:35:56  <TrueBrain> it takes 3 minutes to reload everything :D
11:35:57  <TrueBrain> lolz
11:36:06  <TrueBrain> this is going to be an issue, but not an issue for now :P
11:37:24  <TrueBrain> I need some testers .. add these to your host files: https://gist.github.com/TrueBrain/b4ae5090a9b8aefbf352b6f550386c72
11:37:35  <TrueBrain> can you test ingame, if everything works? Browsing, downloading, etc?
11:37:54  <TrueBrain> don't upload junk, it is production after all; but see if https://bananas.openttd.org works for you, etc
11:38:01  <TrueBrain> just some more eyes on this before I change DNS for everyone :)
11:38:35  <TrueBrain> I will be back in like 2 hours or so, and would like to flip the switch by then :)
11:38:56  <TrueBrain> test older clients, test latest, etc etc
12:04:54  *** Yexo has joined #openttd
12:04:55  *** ChanServ sets mode: +o Yexo
12:15:12  *** glx has joined #openttd
12:15:12  *** ChanServ sets mode: +v glx
12:32:11  <Yexo> TrueBrain: bananas.openttd.org works fine, in-game content works fine, but http://content.openttd.org/ and https://binaries.openttd.org/ don´t work for me
12:32:45  <Yexo> Without changing host-file http://content.openttd.org/ redirects to https://bananas.openttd.org/en, and binaries redirects to an empty page from http://fr.binaries.openttd.org/index.html
12:44:32  <glx> they moved
12:45:13  <glx> binaries are now on cdn.openttd.org
12:47:53  <LordAro> glx: the redirect should probably work though
12:48:39  <LordAro> not sure what the issue with content redirecting to bananas is though?
12:49:33  *** Yexo is now known as Guest23464
12:49:38  *** Yexo has joined #openttd
12:49:39  *** ChanServ sets mode: +o Yexo
12:50:21  *** Yexo has quit IRC
12:54:06  <glx> ah yes redirect works, but only for some url
12:55:36  *** Guest23464 has quit IRC
12:59:12  *** Wormnest_ has joined #openttd
13:03:20  *** Gustavo6046 has quit IRC
13:30:33  <Samu> what is the equivalent of "sizeof(x)" in squirrel? I'm trying to copy a function from c++ to squirrel
13:31:20  <Samu> x is a uint32, if that matters
13:31:33  <LordAro> what do you think, Samu ?
13:36:27  <Samu> 4?
13:37:34  <_dp_> it's already a bat sign if you need sizeof of anything in squirrel...
13:37:48  <_dp_> *bad :)
13:38:48  <Samu> I'm copying ROR
13:38:52  <Samu> rotate right
13:39:12  <Samu> it has a sizeof(x) in it
13:41:30  <Samu> seems to be 4
13:41:45  <Samu> but what if it's not always 4? :(
13:42:09  <LordAro> then you've got bigger problems
13:50:30  *** nielsm has joined #openttd
13:50:41  <FLHerne> mcbanhas: That's done using recolour sprites https://newgrf-specs.tt-wiki.net/wiki/RecolorSprites
13:51:49  *** Wormnest_ has quit IRC
13:52:56  <FLHerne> mcbanhas: Basically, it remaps the colour palette of whatever's underneath it
13:53:46  <FLHerne> So it's easy to do one colour of "transparency", but you couldn't make complex multicoloured things transparent in a practical way
13:54:25  <FLHerne> (you'd need a different sprite for each colour in the overlay, AIUI)
13:55:41  <TrueBrain> Yexo_: tnx a bunch for testing, good that it works :D And yes, I stopped routing http for content.openttd.org, because .. nobody should be using that
13:55:51  <TrueBrain> and binaries.openttd.org should give a 404 page .. I guess I can redirect that to the frontpage :D
13:57:56  <FLHerne> You could have semi-transparent backgrounds with opaque text
13:58:03  <DorpsGek_III> [OpenTTD/binaries-redirect] TrueBrain opened pull request #3: Add: redirect / to the cdn too https://git.io/Jfqgq
13:58:10  <FLHerne> But personally I dislike transparency in UI anyway
13:58:41  <TrueBrain> okay, I guess that means it is time to flip the switch ... I am scared :P
14:00:08  *** syr has left #openttd
14:01:58  <TrueBrain> guess we should merge the news post about it first ..
14:02:05  <TrueBrain> WTB: an approval :D
14:06:05  <DorpsGek_III> [OpenTTD/website] TrueBrain commented on pull request #157: Include final URL https://git.io/Jfqgi
14:10:45  *** sla_ro|master has quit IRC
14:12:17  <Samu> my rotate right is always returning -1, I'm doing something wrong
14:21:44  <Samu> btw why am I getting negative numbers?
14:24:40  <glx> -1 is 0xFFFF so all bits are set
14:25:24  *** Wormnest_ has joined #openttd
14:34:14  <Samu> how do i make 0 - 1 = ‭4294967295‬ in squirrel?
14:35:22  <FLHerne> Samu: Squirrel doesn't have unsigned integers
14:38:49  *** Yexo has joined #openttd
14:38:49  *** ChanServ sets mode: +o Yexo
14:47:02  *** Yexo has quit IRC
14:47:20  <Samu> I'm trying to implement the randomizer in squirrel, but i was hoping it to use unsigned integers, I think i managed it somehow
14:47:46  <Samu> this.state[0] = (s + ROR(t ^ 0x1234567F, 7) + 1) & 0xFFFFFFFF;
14:49:02  <LordAro> that's way too low level
14:49:14  <LordAro> just use rand()
14:49:46  <Samu> I need it to be randomizing equally on multiple instances of the same ai
14:50:04  <LordAro> do you really?
14:50:08  <Samu> so, not really random
14:50:43  <Samu> just behave randomlishy
14:51:04  <glx> but if all instance use the same random, it's no longer random
14:53:06  <glx> really just use AIBase.Rand()
14:53:09  <Samu> the AI "randomly" picks 2 locations and tries to connect them
14:53:37  <Samu> i want the 2nd instance of the same AI to also do the same, on the exact locations that the 1st instance tried
14:53:44  <LordAro> why?
14:53:52  <LordAro> that would just mean they're competing with each other
14:53:56  <Samu> going to compare pathfinding speed
14:54:13  <glx> use a scenario and fixed location then
14:54:27  <LordAro> consider the possibility that in a real game you might have 2 instances of your AI running at the same time
14:54:29  <glx> no need to implement a fake random
14:54:33  <LordAro> what you're trying to test here is 2 separate games
14:59:41  <milek7> srand(0); rand();
14:59:49  <milek7> i think squirrel stdlib does have rand?
15:00:04  <LordAro> http://www.squirrel-lang.org/doc/sqstdlib2.html#d0e1519 yup
15:02:43  *** sla_ro|master has joined #openttd
15:02:49  *** rotterdxm has joined #openttd
15:05:26  <DorpsGek_III> [OpenTTD/OpenTTD] giordy commented on issue #7611: Accident/disaster news not always showing https://git.io/fj0cI
15:20:25  <DorpsGek_III> [OpenTTD/OpenTTD] embeddedt commented on issue #7611: Accident/disaster news not always showing https://git.io/fj0cI
15:26:46  <mcbanhas> FLHerne, I see. I do like a little transparency, sorta like how they did in OpenRCT2. It makes the GUI a bit more lightweight on some elements.
15:27:15  <mcbanhas> And it's always nice to let the player customize that.
15:29:16  <mcbanhas> https://github.com/OpenTTD/OpenTTD/issues/8094 You take elements like what's being suggested here, apply some transparency, and they become a lot less obstructive.
15:35:09  <rotterdxm> i am very much in favor of this
15:35:21  <rotterdxm> JGR has something like that already but it´s just yellow box w/ text
15:36:46  <mcbanhas> Yeah I think RCT2/Locomotion has a similar feature.
15:50:11  <DorpsGek_III> [OpenTTD/nml] FLHerne commented on issue #109: Roadtypes and Tramtypes size is 63 not 15 or 64 https://git.io/JfqYj
15:50:21  <FLHerne> peter1138: ^
16:01:50  <Samu> rand() does not exist
16:01:57  <Samu> the squirrel one
16:02:13  <LordAro> math.rand()
16:02:15  <LordAro> read the documentation
16:04:40  <nielsm> the word "math" does not occur in the squirrel 2 documentation
16:04:53  <LordAro> because it's in the stdlib, not the language
16:06:04  <Samu> the index "math" does not exist
16:06:27  <peter1138> Cos the word is maths ;D
16:07:28  <LordAro> src/3rdparty/squirrel/sqstdlib/sqstdmath.cpp:static SQInteger math_rand(HSQUIRRELVM v)
16:07:31  <LordAro> well it's definitely there.
16:08:19  <DorpsGek_III> [OpenTTD/OpenTTD] EgyLynx commented on issue #8083: Add currency Unit - Indian  Rupee https://git.io/Jvjtc
16:08:24  <peter1138> LordAro, shall I play games or ride bike?
16:08:36  <DorpsGek_III> [OpenTTD/nml] PeterN commented on issue #109: Roadtypes and Tramtypes size is 63 not 15 or 64 https://git.io/JfqYj
16:08:48  <LordAro> peter1138: have you done your daily exercise yet today?
16:08:52  <LordAro> if not, then bike :p
16:08:58  <peter1138> I... went up the stairs a few times...
16:09:28  <LordAro> ah right, math.{s,}rand are guarded by EXPORT_DEFAULT_SQUIRREL_FUNCTIONS
16:09:29  <peter1138> Hmm, road or trail.
16:09:31  <LordAro> which is not defined.
16:09:43  <peter1138> Road, trail, or chair hehe
16:10:29  <LordAro> probably why AIBase.Rand exists
16:10:59  <TrueBrain> any chance any of you wants to review either (or both) https://github.com/OpenTTD/website/pull/158 / https://github.com/OpenTTD/binaries-redirect/pull/3 ? :)
16:12:00  <LordAro> TrueBrain: redirect - why not just make it a "if anything else, redirect to cdn root ?
16:12:15  <LordAro> s/ \?/" ?/
16:12:17  <TrueBrain> because I don't like those kind of wildcards; what good does it do?
16:12:28  <LordAro> fewer 404s
16:12:38  <TrueBrain> and people not realising they have the wrong URL
16:12:42  <LordAro> true
16:12:46  <DorpsGek_III> [OpenTTD/binaries-redirect] LordAro approved pull request #3: Add: redirect / to the cdn too https://git.io/JfqoD
16:12:47  <TrueBrain> I rather deal with 404s, which I can monitor, and see if something odd is happening
16:12:58  <TrueBrain> I also don't like certificate wildcards btw :P
16:13:36  <DorpsGek_III> [OpenTTD/binaries-redirect] TrueBrain merged pull request #3: Add: redirect / to the cdn too https://git.io/Jfqgq
16:13:56  <DorpsGek_III> [OpenTTD/website] LordAro commented on pull request #158: Create 2020-04-27-new-bananas.md https://git.io/JfqoQ
16:14:32  <DorpsGek_III> [OpenTTD/website] TrueBrain updated pull request #158: Create 2020-04-27-new-bananas.md https://git.io/Jfqn4
16:14:39  <DorpsGek_III> [OpenTTD/nml] glx22 commented on issue #109: Roadtypes and Tramtypes size is 63 not 15 or 64 https://git.io/JfqYj
16:14:50  <TrueBrain> I like the "commit suggestion" thingy :D
16:18:47  *** sla_ro|master has quit IRC
16:19:01  *** sla_ro|master has joined #openttd
16:19:01  <LordAro> TrueBrain: as long as you don't care about the commit history :p
16:19:09  <TrueBrain> squash ftw!
16:19:21  <DorpsGek_III> [OpenTTD/website] LordAro commented on pull request #158: Create 2020-04-27-new-bananas.md https://git.io/JfqKJ
16:19:35  <DorpsGek_III> [OpenTTD/website] TrueBrain updated pull request #158: Create 2020-04-27-new-bananas.md https://git.io/Jfqn4
16:19:48  *** andythenorth_ has joined #openttd
16:19:57  <DorpsGek_III> [OpenTTD/website] LordAro approved pull request #158: Create 2020-04-27-new-bananas.md https://git.io/JfqKL
16:20:14  <DorpsGek_III> [OpenTTD/binaries-redirect] TrueBrain created new tag: 1.0.1 https://git.io/JfqKt
16:20:20  <TrueBrain> tnx LordAro  :)
16:22:41  <DorpsGek_III> [OpenTTD/website] TrueBrain merged pull request #158: Create 2020-04-27-new-bananas.md https://git.io/Jfqn4
16:28:38  *** andythenorth has joined #openttd
16:28:44  <TrueBrain> wb andythenorth
16:28:46  <TrueBrain> survived your day?
16:29:28  <andythenorth> survivable
16:30:42  <FLHerne> peter1138: Thanks. But if they're shared, why does the map store 6 bits each?
16:31:03  <peter1138> Becuase you can have both on a tile
16:31:16  <FLHerne> Oh, and there can be more than 32 of either one
16:31:50  <FLHerne> Other way - if the map stores them orthogonally, why is there a shared limit? :P
16:32:02  *** andythenorth_ has quit IRC
16:32:10  <FLHerne> (besides "no-one wants 128 different categories of road vehicle")
16:32:34  <peter1138> Because they're both road types
16:32:48  <peter1138> Basically it makes lots of things much simpler
16:33:38  <FLHerne> Hm
16:34:07  <peter1138> There was originally a shed load of code to differentiate them
16:34:18  <FLHerne> Do roadtype and tramtype labels also share a namespace?
16:34:24  <peter1138> But it made loads of things way overcomplicated.
16:34:25  <peter1138> Yes.
16:34:43  <FLHerne> The NFO docs are wrong too, then :P
16:34:55  <glx> FLHerne: saw the code segment I linked ?
16:35:01  <peter1138> Wouldn't surprise me
16:35:13  <FLHerne> "Note that labels are not shared between features, so the same label can be used for multiple items. For example, the label "RAIL" can be used for a railtype, roadtype, tramtype, and cargotype simultaneously without conflict."
16:36:05  <peter1138> Fix it then :p
16:36:19  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #9: Fix: use a deterministic way to set a content-id https://git.io/JfqKH
16:36:21  *** Flygon has quit IRC
16:36:31  <andythenorth> per the GH issue, the nfo docs are wrong OBIWAN
16:36:39  <TrueBrain> and hopefully the last patch before going live .. this patch allows me to scale the server up, which is going to be needed :D
16:36:57  <FLHerne> That makes it not a trivial NML fix, then :P
16:37:19  <andythenorth> hmm
16:37:26  <andythenorth> there's always another thing :)
16:37:36  <TrueBrain> LordAro: I would appreciate it if you could check my sanity, and that this is a valid way to make deterministic ids :D
16:39:10  <LordAro> https://docs.python.org/3.8/library/uuid.html perhaps?
16:39:27  <Eddi|zuHause> FLHerne: nml can probably use "RAIL" (in quotes) but not RAIL (without quotes) as label multiple times
16:40:00  <TrueBrain> LordAro: I meant in relation to my PR. But no, uuid doesnt solve anything. I need it to be deterministic, not random :)
16:40:04  <Eddi|zuHause> as without quotes it is also an Identifier, which cannot be reused
16:40:25  <DorpsGek_III> [OpenTTD/bananas-frontend-web] auge8472 commented on pull request #16: Format the elements in the page header https://git.io/Jfq6J
16:40:25  <DorpsGek_III> [OpenTTD/bananas-frontend-web] auge8472 closed pull request #16: Format the elements in the page header https://git.io/JfIW0
16:41:48  <LordAro> TrueBrain: idk, i'd have to actually spend time working out what what it's doing :p
16:42:06  <TrueBrain> I understand .. I can talk you through it, if that helps
16:42:09  <TrueBrain> but I hope my comments would
16:42:23  <TrueBrain> nothing else code-wise is related btw; this is very much on its own
16:43:19  <Eddi|zuHause> TrueBrain: i don't know the context, but if you want deterministic, you're looking at (cryptographic) hash functions
16:43:28  <TrueBrain> so .. check the PR :P
16:45:02  <TrueBrain> basically, I have 32bits. So I took the last 24bits from the md5sum, which currently has no collisions. But to be prepared, I added a counter, by sorting any collision based on their upload date
16:45:12  <FLHerne> Eddi|zuHause: What's the distinction supposed to be?
16:45:19  <TrueBrain> so if you are the 4th collision, your number is (4 << 24) + md5sum[-3:]
16:45:22  <FLHerne> I thought that was just a syntax limitation
16:45:28  <TrueBrain> well, 3 << 24, but who is counting
16:45:55  *** gelignite has joined #openttd
16:46:15  <TrueBrain> so if you know the md5sum, you can guess the content-id, which is fine, because we tell you the content-id when we tell you the md5sum :P
16:46:25  <Eddi|zuHause> i don't think md5 counts as a cryptographic hash function anymore :p
16:46:39  <TrueBrain> it very much still is a crypto hash function .. it will never lose that state
16:46:42  <TrueBrain> it is not a safe one, sure
16:46:48  <TrueBrain> but RC4 is also still a crypto function :)
16:47:18  <milek7> you can't just take whole md5 as contentid?
16:47:21  <TrueBrain> I am not looking to be crypto-secure here; I am looking for a way to have deterministic content-ids over multiple runs :)
16:47:34  <TrueBrain> I hate repating myself, but "basically, I have 32bits."
16:47:37  <TrueBrain> I started with that ;)
16:47:53  <Eddi|zuHause> FLHerne: the easiest definition of "cryptographic hash function" is that small differences in input data lead to unpredictable changes in the output data
16:48:19  <FLHerne> Eddi|zuHause: I meant between RAIL and "RAIL" ?
16:48:45  <FLHerne> I know about hash functions, fwiw ;-)
16:49:03  <Eddi|zuHause> "RAIL" is a 4-letter string, RAIL is an identifier, which in addition to being used for the "RAIL" label gets mapped to the position in the translation table
16:49:16  *** Progman has joined #openttd
16:50:02  <TrueBrain> okay, looking this over, this should be fine. content-id is stable, and without knowing the md5sum, unknown
16:50:05  <TrueBrain> perfect :)
16:50:30  <TrueBrain> md5 collisions are prevented by the API, so we should be good
16:51:25  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #9: Fix: use a deterministic way to set a content-id https://git.io/JfqKH
16:51:43  <FLHerne> Eddi|zuHause: Where does that apply? In NML's roadtype table code, identifiers are converted to StringLiterals immediately
16:52:12  <FLHerne> There's no difference in how they're handled
16:52:13  <Eddi|zuHause> FLHerne: the RAIL identifier can be used in switches and properties, that take a table index
16:54:08  *** frosch123 has joined #openttd
16:55:33  <milek7> maybe take 32bit from md5, and xor with counter?
16:55:38  <milek7> iterating 24bit is doable
16:55:39  <FLHerne> Which is also true of `railtype("3RDR")` or so, no?
16:56:15  <FLHerne> I still don't see where there's any semantic difference
16:57:03  <FLHerne> Eddi|zuHause: ^
16:57:11  <Eddi|zuHause> yes, it's basically a shortcut
16:57:17  <milek7> but then counter could collide.. so no
16:58:09  <FLHerne> Eddi|zuHause: So why > nml can probably use "RAIL" (in quotes) but not RAIL (without quotes) as label multiple times  ?
16:58:17  <TrueBrain> milek7: honestly, I am more interesting if my current solution works, instead of seeing if there are other solutions ;) Pretty sure there are plenty, but if this works, it works :)
16:58:35  <FLHerne> Internally, they do exactly the same thing, unless I've missed something
16:59:30  <milek7> what happens when more than 2^8 24-bit-part collisions?
17:00:13  <TrueBrain> I buy everyone cake
17:00:34  <TrueBrain> some cosmic event has hit the rays at the exact right spot, and I should have bought a lottery ticket as I would be rich
17:01:09  <Eddi|zuHause> FLHerne: when you define a railtype translation table, and use RAIL, then it creates a RAIL identifier, which evaluates to the position in the RTT, if you then define a tramtype translation table, and try to use the RAIL label there, it tries to create another RAIL identifier, which it cannot, because it already exists
17:01:41  <FLHerne> Eddi|zuHause: Oh, I see
17:01:49  <Eddi|zuHause> and that second RAIL identifier could potentially require you to evaluate to a different position
17:02:06  <FLHerne> That's actually a completely different issue to the one we have :P
17:03:01  <FLHerne> Well, not quite, it makes it less bad
17:03:03  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #9: Fix: use a deterministic way to set a content-id https://git.io/JfqKH
17:03:16  <TrueBrain> but okay milek7 , it now crashes if that happens. Just to let everyone know when I have to buy cake.
17:04:20  <FLHerne> NML *will* let you use "BLAH" for both roadtype and tramtype, despite those actually conflicting
17:04:51  <FLHerne> ...except it won't, because andy forgot to add the `roadtype()` and `tramtype()` builtins, so you can't use them :P
17:05:06  <FLHerne> So it has to be an identifier, and those conflict
17:05:16  <Eddi|zuHause> that's probably a faulty specification because of how it went back and forth between tramtypes being something completely different or being mashed into roadtypes
17:05:19  <FLHerne> andythenorth: Well planned
17:05:49  <FLHerne> Eddi|zuHause: Yes - all of the NFO docs, NML docs and NML implementation are wrong
17:07:59  <FLHerne> Hm, I think it can actually be used wrong atm
17:13:41  <TrueBrain> frosch123: I build a Python script to do the conversion :)
17:14:57  <andythenorth> FLHerne the existence of so many NRT grfs suggest it works despite the poor patch :(
17:14:59  <andythenorth> oof
17:15:12  * andythenorth will be back shortly, beer has run out
17:16:16  <DorpsGek_III> [OpenTTD/bananas-server] frosch123 approved pull request #9: Fix: use a deterministic way to set a content-id https://git.io/Jfqip
17:16:26  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #9: Fix: use a deterministic way to set a content-id https://git.io/JfqKH
17:17:01  <FLHerne> In global_constants: `railtype_table = {'RAIL': 0, 'ELRL': 1, 'MONO': 1, 'MGLV': 2}`
17:17:20  <frosch123> TrueBrain: did you publish that script?
17:17:26  <TrueBrain> see your PM
17:18:32  <frosch123> usually notifications work
17:18:35  <frosch123> not this time?
17:18:50  <TrueBrain> you are asking us about your own IRC client? :P
17:19:29  <frosch123> yeah, notification did not work...
17:19:44  <glx> FLHerne: I think the major issue is the distinction between roadtype an tramtype in nml while in nfo they are the same thing
17:20:46  <TrueBrain> frosch123: my suggestion with that script is that we make PRs out of these changes
17:20:50  <TrueBrain> not that we cannot self-approve
17:20:58  <TrueBrain> but just because it makes the risk on fucking up lower :P
17:21:25  <DorpsGek_III> [OpenTTD/OpenTTD] sheepo99 commented on pull request #7786: Add: Help and manuals window https://git.io/JfqPq
17:21:51  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain created new tag: 1.0.1 https://git.io/JfqP3
17:22:08  <TrueBrain> okay, after this is live, I will scale that server up with 1 .. and after that, time to change some DNS settings :D
17:22:39  <TrueBrain> just over 2 months of hard work leading up to this moment .. PAM PAM PAMMMMMMMMM
17:23:47  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain commented on issue #8: Content-id are not deterministic over runs https://git.io/Jft1F
17:23:47  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain closed issue #8: Content-id are not deterministic over runs https://git.io/Jft1F
17:24:56  *** andythenorth has quit IRC
17:25:09  <Samu> new record!
17:25:20  <Samu> 174 ticks!
17:25:27  <frosch123> TrueBrain: you remove the openttd name. why not keep it?
17:25:29  <Samu> but requires 8091
17:25:33  <Samu> pr #8091
17:25:38  <TrueBrain> frosch123: is it useful to keep?
17:25:50  <TrueBrain> and because this was easier to script :D
17:26:40  <TrueBrain> I simply did not consider it useful, tbh .. keeping things for the keep .. meh. It is always in the git history, was my reasoning
17:26:45  <TrueBrain> but I am not married to either one :)
17:26:56  <frosch123> i guess one can grep r1
17:27:06  <Samu> unless im doing this wrong
17:27:10  <Samu> this is a new record
17:27:32  <frosch123> anyway, your push is missing a that branch option, that i can't remember, any always c&p from the git error message
17:27:48  <frosch123> but, i guess, that's get fixed on first usage :)
17:28:07  <TrueBrain> no, that was the intention :)
17:28:11  <TrueBrain> git throws an error
17:28:13  <TrueBrain> you copy/paste that
17:28:16  <TrueBrain> and it pushes it :P
17:28:21  <TrueBrain> I hate when scripts push shit for me
17:28:27  <TrueBrain> this was the cheapskate way of showing what to do :P
17:28:33  <frosch123> haha :)
17:28:39  <TrueBrain> I always depend on that error
17:28:41  <TrueBrain> I cannot remember it
17:28:43  *** andythenorth has joined #openttd
17:28:50  <TrueBrain> and because I copy the spaces in front of it, it also never hits my bash-history
17:29:37  <TrueBrain> okay, scaling up bananas-server .. I tested what I could
17:29:47  <TrueBrain> I have backups
17:29:54  <TrueBrain> what could possibly go wrong, right? :D
17:30:05  * Xaroth grabs a bag of popcorn
17:30:21  <TrueBrain> I liked you more when you weren't active on IRC
17:30:30  <Xaroth> Love you too.
17:30:32  <TrueBrain> <3
17:30:46  <Xaroth> And you would've done the same :P
17:30:52  <TrueBrain> yup :D
17:30:59  <TrueBrain> in fact, I am doing the same
17:31:02  <TrueBrain> as .. what else is there
17:31:09  <TrueBrain> DNS TTL is 60 seconds, so meh
17:32:36  <Wolf01> Hmm, I can't understand which one of the n-hundred piece of electronics is hissing in my room
17:33:01  <Wolf01> It seem to be... everywhere
17:33:44  <frosch123> maybe it's an insect
17:34:55  <TrueBrain> okay .. there goes the DNS ...
17:35:51  <Samu> ah rats, it was too good to be true after all, I was doing this bad
17:36:07  <Wolf01> It could be the phone charger, I know it does that, but not when I get near :E
17:36:21  <TrueBrain> lol .. and that moment you make a typo in the DNS entry
17:36:22  <TrueBrain> oops
17:36:33  <TrueBrain> happy with the 60 TTL :D
17:37:17  <TrueBrain> k .. ingame and https://bananas.openttd.org/ should be switched now
17:37:32  <TrueBrain> binaries not yet; I will give it some time for people to stop using it, before we start to redirect it :)
17:37:41  <FLHerne> peter1138: Sorry, I'm confused again :-/
17:38:22  <FLHerne> "Do road and tramtype labels also share a namespace" -> <peter1138> Yes.
17:38:30  <TrueBrain> oops ... healthcheck killed the server .. hmm .. oops :D
17:38:39  *** mcbanhas_ has joined #openttd
17:39:01  <frosch123> FLHerne: no, they do not. but you cannot define a roadtype and a tramtype with the same label
17:39:24  <DorpsGek_III> [OpenTTD/website] TrueBrain created new tag: 1.3.8 https://git.io/JfqXU
17:39:51  <FLHerne> But there are separate properties to set railtype and tramtype tables, and to test whether a label is defined
17:40:32  <frosch123> FLHerne: from the spec point, they are entirely separate. just peter considered it to complicated, and deleted the implementation that separated them
17:40:41  <frosch123> FLHerne: so, just assume they are separate
17:41:05  <FLHerne> So will Act7 condition 12 "Tramtype label is defined" evaluate as true if the corresponding roadtype label is set?
17:41:11  <frosch123> no
17:41:12  <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #7786: Add: Help and manuals window https://git.io/JfqXm
17:41:33  <FLHerne> frosch123: "but you cannot define a roadtype and a tramtype with the same label" ok, so what happens if you do?
17:41:39  *** iSoSyS has joined #openttd
17:41:4