IRC logs for #openttd on OFTC at 2020-04-27
⏴ go to previous day
00:47:18 <FLHerne> I don't suppose anyone knows why there's a NOTHASBIT operator?
00:47:36 <FLHerne> As opposed to just... NOT(HASBIT())
01:01:56 <FLHerne> I guess because action7 supports that
01:38:38 <Eddi|zuHause> hysterical raisins?
02:02:13 <DorpsGek_III> [OpenTTD/nml] glx22 opened pull request #108: Codechange: [Actions] Upload standalone executable to GitHub release https://git.io/JfqJv
02:25:34 <mcbanhas> Is it possible to make UI elements semi-transparent at this point?
02:25:56 <mcbanhas> I mean, does the engine support it?
04:15:33 <FLHerne> Is there any case in which the value of a variable read by ActionD can change during one loading pass?
04:15:54 <FLHerne> Other than the grf directly writing to that variable
04:18:34 <FLHerne> ahmet: Good morning :-)
04:19:53 <FLHerne> It's 3am, I was about to sleep
04:20:19 <ahmet> great to see who played this game!
04:21:13 <ahmet> This game is very enjoyable in quarantine days
04:24:04 <FLHerne> Yes, I've been playing more than I had been
04:25:48 <ahmet> The buttons in the game are very small. It hasn't changed for years. My fingers hurt after hours
04:26:28 <FLHerne> You can make them bigger somehow
04:27:46 <FLHerne> There's a NewGRF "OpenGFX+ BigGUI" that makes them bigger
04:28:21 <ahmet> I don't want to keep you from sleep.. :))
04:28:21 <FLHerne> Enabling it as a 'static' grf as shown on the wiki will mean it's active in multiplayer too
04:28:32 <FLHerne> Even if the server doesn't use it
04:29:13 <FLHerne> Yes, bedtime for me. Goodnight. :-)
05:09:41 *** Wormnest_ has joined #openttd
07:24:26 *** Hirundo has joined #openttd
07:24:40 *** TinoDidriksen has joined #openttd
07:25:11 *** TinoDidriksen is now known as Guest23449
07:26:56 *** XeryusTC has joined #openttd
07:27:38 *** Guest23449 is now known as TinoDidriksen
08:02:30 *** sla_ro|master has joined #openttd
08:45:21 *** andythenorth has joined #openttd
09:53:52 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on pull request #108: Codechange: [Actions] Upload standalone executable to GitHub release https://git.io/JfqOv
10:27:49 <TrueBrain> lets migrate BaNaNaS, I would say .. see how that goes ..
10:28:10 <TrueBrain> lot of moving parts .. lets see what the best order is ..
10:30:55 <orudge> I was just coming on here to see how things were going with that and the old VPSes :D
10:30:56 <Wolf01> I think the first thing should be "put everything readonly" :P
10:32:05 <TrueBrain> orudge: I think we just let them expire; for now I don't see a reason to create new ones :)
10:32:12 <TrueBrain> but today will tell ;)
10:32:17 <TrueBrain> otherwise I have 2 days to install new VPSes
10:32:30 <TrueBrain> honestly, only 40 dollar per year for 2, so .. not a real issue ofc, but meh :)
10:34:16 <orudge> Are the contents of the old VPSes (or one of them at least) backed up somehere just in case? :D
10:34:33 <TrueBrain> they are the backups :)
10:34:48 <TrueBrain> it is just an nginx with rsync access
10:34:56 <TrueBrain> and all the binaries we have
10:35:12 <TrueBrain> but I will logon to them today/tomorrow to make sure about that :D
10:35:34 <orudge> I also host an OpenTTD mirror or two of some sort, if I remember correctly
10:35:58 <TrueBrain> these two VPSes do CA and FR
10:36:19 <TrueBrain> but that traffic on all of them should disipate today :P
10:36:37 <TrueBrain> it should already been reduced for a while now btw
10:36:52 <TrueBrain> not sure if you have graphs of these, but like last year or something it should be noticable lower
10:37:59 <orudge> Possibly, I haven't really been paying attention :D
10:38:06 <TrueBrain> okay, time to deploy everything on AWS already .. that requires some certificates to be made .. this is going to take a while :D
10:38:25 <TrueBrain> also not that important; just a bit of stats-nerding :D
10:38:45 <TrueBrain> your DNS admin panel is a bit weird sometimes btw .. it reset all the TTLs :P
10:40:38 <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
10:41:32 <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.
10:41:51 <TrueBrain> over time, maybe :)
10:42:06 <TrueBrain> we already have an AWS zone for openttd.org, but .. not all records are there yet
10:42:13 <TrueBrain> I am happy you are doing it still :)
10:56:41 <andythenorth> happy launch day TrueBrain !
10:56:49 * andythenorth going afk, that work thing I mentioned :P
10:56:56 *** andythenorth has left #openttd
11:11:16 <TrueBrain> possibly I have been blablaing a bit too much, but I would love input :)
11:26:01 <LordAro> TrueBrain: reads fine, imo
11:28:31 <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'
11:35:46 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain opened pull request #32: Cache one endpoint, and tweak text about downloads https://git.io/JfqcG
11:35:56 <TrueBrain> I wonder if frosch123 is going to see this .. :D
11:36:32 <TrueBrain> I kinda need that cache to be in there before we go live :P We have too many NewGRFs :D
11:39:32 <TrueBrain> okay, everything deployed to AWS without issue .. I love CDK :)
11:59:08 <TrueBrain> okay, NSIS installer still works on new infra, that makes me really happy to see :)
12:06:21 <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
12:12:08 <TrueBrain> right, that was an easy fix
12:12:24 <TrueBrain> k, lunch time, after that, I am going to put BaNaNaS v1 on read-only :)
12:13:51 <DorpsGek_III> [OpenTTD/bananas-frontend-web] LordAro approved pull request #32: Cache one endpoint, and tweak text about downloads https://git.io/JfqC7
12:14:53 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain merged pull request #32: Cache one endpoint, and tweak text about downloads https://git.io/JfqcG
12:20:20 <LordAro> TrueBrain: i'd suggest that the more "pythonic" way of doing it would be via a function decorator
12:20:41 <LordAro> then you could just add @cache_api(60 * 5) to the top :)
12:31:22 <TrueBrain> Fully agree. Would be a lot cleaner
12:36:52 <TrueBrain> Possibly Flask has a decorator for it .. owh well, this works for sure for now :p
12:45:41 <TrueBrain> I normally use that .. no clue why I didn't this time .. wanted to have it fixed I guess :p
12:48:34 *** mcbanhas has joined #openttd
12:49:29 <mcbanhas> Is it possible to set transparency for UI elements? Does it the engine support it currently?
12:52:33 <mcbanhas> So how is transparency currently handled in map objects? Are those dithered silhouettes or something similar?
12:52:58 <TrueBrain> alright .. lets put some things in read-only mode, and do this thing ..
12:54:29 <mcbanhas> Hmm, no it doesn't look like dithering.
12:57:16 <TrueBrain> oef ... am I really going to do this? These changes always make me nervous :D
13:14:18 <TrueBrain> RELEASE ALL THE THINGS
13:22:35 <TrueBrain> uploading with 500 mbit/s to AWS :D I love that speed :P
13:33:43 <TrueBrain> haha, it takes too long for bananas-server to read that repo, that is is being killed as "unresponsive"
13:35:56 <TrueBrain> it takes 3 minutes to reload everything :D
13:36:06 <TrueBrain> this is going to be an issue, but not an issue for now :P
13:37:35 <TrueBrain> can you test ingame, if everything works? Browsing, downloading, etc?
13:38:01 <TrueBrain> just some more eyes on this before I change DNS for everyone :)
13:38:35 <TrueBrain> I will be back in like 2 hours or so, and would like to flip the switch by then :)
13:38:56 <TrueBrain> test older clients, test latest, etc etc
14:04:55 *** ChanServ sets mode: +o Yexo
14:45:13 <glx> binaries are now on cdn.openttd.org
14:47:53 <LordAro> glx: the redirect should probably work though
14:48:39 <LordAro> not sure what the issue with content redirecting to bananas is though?
14:49:33 *** Yexo is now known as Guest23464
14:49:39 *** ChanServ sets mode: +o Yexo
14:54:06 <glx> ah yes redirect works, but only for some url
14:59:12 *** Wormnest_ has joined #openttd
15:30:33 <Samu> what is the equivalent of "sizeof(x)" in squirrel? I'm trying to copy a function from c++ to squirrel
15:31:20 <Samu> x is a uint32, if that matters
15:31:33 <LordAro> what do you think, Samu ?
15:37:34 <_dp_> it's already a bat sign if you need sizeof of anything in squirrel...
15:39:12 <Samu> it has a sizeof(x) in it
15:41:45 <Samu> but what if it's not always 4? :(
15:42:09 <LordAro> then you've got bigger problems
15:52:56 <FLHerne> mcbanhas: Basically, it remaps the colour palette of whatever's underneath it
15:53:46 <FLHerne> So it's easy to do one colour of "transparency", but you couldn't make complex multicoloured things transparent in a practical way
15:54:25 <FLHerne> (you'd need a different sprite for each colour in the overlay, AIUI)
15:55:41 <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
15:55:51 <TrueBrain> and binaries.openttd.org should give a 404 page .. I guess I can redirect that to the frontpage :D
15:57:56 <FLHerne> You could have semi-transparent backgrounds with opaque text
15:58:10 <FLHerne> But personally I dislike transparency in UI anyway
15:58:41 <TrueBrain> okay, I guess that means it is time to flip the switch ... I am scared :P
16:01:58 <TrueBrain> guess we should merge the news post about it first ..
16:02:05 <TrueBrain> WTB: an approval :D
16:12:17 <Samu> my rotate right is always returning -1, I'm doing something wrong
16:21:44 <Samu> btw why am I getting negative numbers?
16:24:40 <glx> -1 is 0xFFFF so all bits are set
16:25:24 *** Wormnest_ has joined #openttd
16:34:14 <Samu> how do i make 0 - 1 = 4294967295 in squirrel?
16:35:22 <FLHerne> Samu: Squirrel doesn't have unsigned integers
16:38:49 *** ChanServ sets mode: +o Yexo
16:47:20 <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
16:47:46 <Samu> this.state[0] = (s + ROR(t ^ 0x1234567F, 7) + 1) & 0xFFFFFFFF;
16:49:01 <LordAro> that's way too low level
16:49:46 <Samu> I need it to be randomizing equally on multiple instances of the same ai
16:51:04 <glx> but if all instance use the same random, it's no longer random
16:53:06 <glx> really just use AIBase.Rand()
16:53:09 <Samu> the AI "randomly" picks 2 locations and tries to connect them
16:53:37 <Samu> i want the 2nd instance of the same AI to also do the same, on the exact locations that the 1st instance tried
16:53:52 <LordAro> that would just mean they're competing with each other
16:53:56 <Samu> going to compare pathfinding speed
16:54:13 <glx> use a scenario and fixed location then
16:54:27 <LordAro> consider the possibility that in a real game you might have 2 instances of your AI running at the same time
16:54:29 <glx> no need to implement a fake random
16:54:33 <LordAro> what you're trying to test here is 2 separate games
16:59:49 <milek7> i think squirrel stdlib does have rand?
17:02:43 *** sla_ro|master has joined #openttd
17:02:49 *** rotterdxm has joined #openttd
17:20:25 <DorpsGek_III> [OpenTTD/OpenTTD] embeddedt commented on issue #7611: Accident/disaster news not always showing https://git.io/fj0cI
17:26:46 <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.
17:27:15 <mcbanhas> And it's always nice to let the player customize that.
17:35:09 <rotterdxm> i am very much in favor of this
17:35:21 <rotterdxm> JGR has something like that already but it´s just yellow box w/ text
17:36:45 <mcbanhas> Yeah I think RCT2/Locomotion has a similar feature.
18:02:15 <LordAro> read the documentation
18:04:40 <nielsm> the word "math" does not occur in the squirrel 2 documentation
18:04:53 <LordAro> because it's in the stdlib, not the language
18:06:04 <Samu> the index "math" does not exist
18:06:27 <peter1138> Cos the word is maths ;D
18:07:27 <LordAro> src/3rdparty/squirrel/sqstdlib/sqstdmath.cpp:static SQInteger math_rand(HSQUIRRELVM v)
18:07:31 <LordAro> well it's definitely there.
18:08:24 <peter1138> LordAro, shall I play games or ride bike?
18:08:48 <LordAro> peter1138: have you done your daily exercise yet today?
18:08:58 <peter1138> I... went up the stairs a few times...
18:09:28 <LordAro> ah right, math.{s,}rand are guarded by EXPORT_DEFAULT_SQUIRREL_FUNCTIONS
18:09:29 <peter1138> Hmm, road or trail.
18:09:31 <LordAro> which is not defined.
18:09:43 <peter1138> Road, trail, or chair hehe
18:10:29 <LordAro> probably why AIBase.Rand exists
18:12:00 <LordAro> TrueBrain: redirect - why not just make it a "if anything else, redirect to cdn root ?
18:12:17 <TrueBrain> because I don't like those kind of wildcards; what good does it do?
18:12:38 <TrueBrain> and people not realising they have the wrong URL
18:12:47 <TrueBrain> I rather deal with 404s, which I can monitor, and see if something odd is happening
18:12:58 <TrueBrain> I also don't like certificate wildcards btw :P
18:14:50 <TrueBrain> I like the "commit suggestion" thingy :D
18:19:01 *** sla_ro|master has joined #openttd
18:19:01 <LordAro> TrueBrain: as long as you don't care about the commit history :p
18:19:48 *** andythenorth_ has joined #openttd
18:28:38 *** andythenorth has joined #openttd
18:30:42 <FLHerne> peter1138: Thanks. But if they're shared, why does the map store 6 bits each?
18:31:03 <peter1138> Becuase you can have both on a tile
18:31:16 <FLHerne> Oh, and there can be more than 32 of either one
18:31:50 <FLHerne> Other way - if the map stores them orthogonally, why is there a shared limit? :P
18:32:10 <FLHerne> (besides "no-one wants 128 different categories of road vehicle")
18:32:34 <peter1138> Because they're both road types
18:32:48 <peter1138> Basically it makes lots of things much simpler
18:34:07 <peter1138> There was originally a shed load of code to differentiate them
18:34:18 <FLHerne> Do roadtype and tramtype labels also share a namespace?
18:34:24 <peter1138> But it made loads of things way overcomplicated.
18:34:43 <FLHerne> The NFO docs are wrong too, then :P
18:34:55 <glx> FLHerne: saw the code segment I linked ?
18:35:01 <peter1138> Wouldn't surprise me
18:35:13 <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."
18:36:19 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #9: Fix: use a deterministic way to set a content-id https://git.io/JfqKH
18:36:31 <andythenorth> per the GH issue, the nfo docs are wrong OBIWAN
18:36:39 <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
18:36:57 <FLHerne> That makes it not a trivial NML fix, then :P
18:37:26 <andythenorth> there's always another thing :)
18:37:36 <TrueBrain> LordAro: I would appreciate it if you could check my sanity, and that this is a valid way to make deterministic ids :D
18:39:27 <Eddi|zuHause> FLHerne: nml can probably use "RAIL" (in quotes) but not RAIL (without quotes) as label multiple times
18:40:00 <TrueBrain> LordAro: I meant in relation to my PR. But no, uuid doesnt solve anything. I need it to be deterministic, not random :)
18:40:04 <Eddi|zuHause> as without quotes it is also an Identifier, which cannot be reused
18:40:25 <DorpsGek_III> [OpenTTD/bananas-frontend-web] auge8472 commented on pull request #16: Format the elements in the page header https://git.io/Jfq6J
18:40:25 <DorpsGek_III> [OpenTTD/bananas-frontend-web] auge8472 closed pull request #16: Format the elements in the page header https://git.io/JfIW0
18:41:48 <LordAro> TrueBrain: idk, i'd have to actually spend time working out what what it's doing :p
18:42:06 <TrueBrain> I understand .. I can talk you through it, if that helps
18:42:09 <TrueBrain> but I hope my comments would
18:42:22 <TrueBrain> nothing else code-wise is related btw; this is very much on its own
18:43:19 <Eddi|zuHause> TrueBrain: i don't know the context, but if you want deterministic, you're looking at (cryptographic) hash functions
18:43:28 <TrueBrain> so .. check the PR :P
18:45:02 <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
18:45:12 <FLHerne> Eddi|zuHause: What's the distinction supposed to be?
18:45:19 <TrueBrain> so if you are the 4th collision, your number is (4 << 24) + md5sum[-3:]
18:45:22 <FLHerne> I thought that was just a syntax limitation
18:45:28 <TrueBrain> well, 3 << 24, but who is counting
18:45:55 *** gelignite has joined #openttd
18:46:15 <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
18:46:25 <Eddi|zuHause> i don't think md5 counts as a cryptographic hash function anymore :p
18:46:39 <TrueBrain> it very much still is a crypto hash function .. it will never lose that state
18:46:42 <TrueBrain> it is not a safe one, sure
18:46:48 <TrueBrain> but RC4 is also still a crypto function :)
18:47:18 <milek7> you can't just take whole md5 as contentid?
18:47:21 <TrueBrain> I am not looking to be crypto-secure here; I am looking for a way to have deterministic content-ids over multiple runs :)
18:47:34 <TrueBrain> I hate repating myself, but "basically, I have 32bits."
18:47:37 <TrueBrain> I started with that ;)
18:47:53 <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
18:48:19 <FLHerne> Eddi|zuHause: I meant between RAIL and "RAIL" ?
18:48:45 <FLHerne> I know about hash functions, fwiw ;-)
18:49:03 <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
18:49:16 *** Progman has joined #openttd
18:50:02 <TrueBrain> okay, looking this over, this should be fine. content-id is stable, and without knowing the md5sum, unknown
18:50:30 <TrueBrain> md5 collisions are prevented by the API, so we should be good
18:51:25 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #9: Fix: use a deterministic way to set a content-id https://git.io/JfqKH
18:51:43 <FLHerne> Eddi|zuHause: Where does that apply? In NML's roadtype table code, identifiers are converted to StringLiterals immediately
18:52:12 <FLHerne> There's no difference in how they're handled
18:52:13 <Eddi|zuHause> FLHerne: the RAIL identifier can be used in switches and properties, that take a table index
18:54:08 *** frosch123 has joined #openttd
18:55:33 <milek7> maybe take 32bit from md5, and xor with counter?
18:55:38 <milek7> iterating 24bit is doable
18:55:39 <FLHerne> Which is also true of `railtype("3RDR")` or so, no?
18:56:15 <FLHerne> I still don't see where there's any semantic difference
18:57:11 <Eddi|zuHause> yes, it's basically a shortcut
18:57:17 <milek7> but then counter could collide.. so no
18:58:09 <FLHerne> Eddi|zuHause: So why > nml can probably use "RAIL" (in quotes) but not RAIL (without quotes) as label multiple times ?
18:58:17 <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 :)
18:58:35 <FLHerne> Internally, they do exactly the same thing, unless I've missed something
18:59:30 <milek7> what happens when more than 2^8 24-bit-part collisions?
19:00:13 <TrueBrain> I buy everyone cake
19:00:34 <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
19:01:09 <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
19:01:41 <FLHerne> Eddi|zuHause: Oh, I see
19:01:49 <Eddi|zuHause> and that second RAIL identifier could potentially require you to evaluate to a different position
19:02:06 <FLHerne> That's actually a completely different issue to the one we have :P
19:03:01 <FLHerne> Well, not quite, it makes it less bad
19:03:03 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #9: Fix: use a deterministic way to set a content-id https://git.io/JfqKH
19:03:16 <TrueBrain> but okay milek7 , it now crashes if that happens. Just to let everyone know when I have to buy cake.
19:04:20 <FLHerne> NML *will* let you use "BLAH" for both roadtype and tramtype, despite those actually conflicting
19:04:51 <FLHerne> ...except it won't, because andy forgot to add the `roadtype()` and `tramtype()` builtins, so you can't use them :P
19:05:06 <FLHerne> So it has to be an identifier, and those conflict
19:05:16 <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
19:05:19 <FLHerne> andythenorth: Well planned
19:05:49 <FLHerne> Eddi|zuHause: Yes - all of the NFO docs, NML docs and NML implementation are wrong
19:07:58 <FLHerne> Hm, I think it can actually be used wrong atm
19:13:41 <TrueBrain> frosch123: I build a Python script to do the conversion :)
19:14:57 <andythenorth> FLHerne the existence of so many NRT grfs suggest it works despite the poor patch :(
19:15:12 * andythenorth will be back shortly, beer has run out
19:16:16 <DorpsGek_III> [OpenTTD/bananas-server] frosch123 approved pull request #9: Fix: use a deterministic way to set a content-id https://git.io/Jfqip
19:16:26 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #9: Fix: use a deterministic way to set a content-id https://git.io/JfqKH
19:17:01 <FLHerne> In global_constants: `railtype_table = {'RAIL': 0, 'ELRL': 1, 'MONO': 1, 'MGLV': 2}`
19:17:20 <frosch123> TrueBrain: did you publish that script?
19:18:32 <frosch123> usually notifications work
19:18:50 <TrueBrain> you are asking us about your own IRC client? :P
19:19:29 <frosch123> yeah, notification did not work...
19:19:44 <glx> FLHerne: I think the major issue is the distinction between roadtype an tramtype in nml while in nfo they are the same thing
19:20:46 <TrueBrain> frosch123: my suggestion with that script is that we make PRs out of these changes
19:20:50 <TrueBrain> not that we cannot self-approve
19:20:58 <TrueBrain> but just because it makes the risk on fucking up lower :P
19:22:08 <TrueBrain> okay, after this is live, I will scale that server up with 1 .. and after that, time to change some DNS settings :D
19:22:39 <TrueBrain> just over 2 months of hard work leading up to this moment .. PAM PAM PAMMMMMMMMM
19:23:47 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain commented on issue #8: Content-id are not deterministic over runs https://git.io/Jft1F
19:25:27 <frosch123> TrueBrain: you remove the openttd name. why not keep it?
19:25:38 <TrueBrain> frosch123: is it useful to keep?
19:25:50 <TrueBrain> and because this was easier to script :D
19:26:40 <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
19:26:45 <TrueBrain> but I am not married to either one :)
19:26:56 <frosch123> i guess one can grep r1
19:27:06 <Samu> unless im doing this wrong
19:27:32 <frosch123> anyway, your push is missing a that branch option, that i can't remember, any always c&p from the git error message
19:27:48 <frosch123> but, i guess, that's get fixed on first usage :)
19:28:07 <TrueBrain> no, that was the intention :)
19:28:11 <TrueBrain> git throws an error
19:28:13 <TrueBrain> you copy/paste that
19:28:16 <TrueBrain> and it pushes it :P
19:28:21 <TrueBrain> I hate when scripts push shit for me
19:28:27 <TrueBrain> this was the cheapskate way of showing what to do :P
19:28:39 <TrueBrain> I always depend on that error
19:28:41 <TrueBrain> I cannot remember it
19:28:43 *** andythenorth has joined #openttd
19:28:50 <TrueBrain> and because I copy the spaces in front of it, it also never hits my bash-history
19:29:37 <TrueBrain> okay, scaling up bananas-server .. I tested what I could
19:29:54 <TrueBrain> what could possibly go wrong, right? :D
19:30:05 * Xaroth grabs a bag of popcorn
19:30:21 <TrueBrain> I liked you more when you weren't active on IRC
19:30:46 <Xaroth> And you would've done the same :P
19:30:59 <TrueBrain> in fact, I am doing the same
19:31:02 <TrueBrain> as .. what else is there
19:31:09 <TrueBrain> DNS TTL is 60 seconds, so meh
19:32:36 <Wolf01> Hmm, I can't understand which one of the n-hundred piece of electronics is hissing in my room
19:33:01 <Wolf01> It seem to be... everywhere
19:33:44 <frosch123> maybe it's an insect
19:34:55 <TrueBrain> okay .. there goes the DNS ...
19:35:51 <Samu> ah rats, it was too good to be true after all, I was doing this bad
19:36:07 <Wolf01> It could be the phone charger, I know it does that, but not when I get near :E
19:36:21 <TrueBrain> lol .. and that moment you make a typo in the DNS entry
19:36:33 <TrueBrain> happy with the 60 TTL :D
19:37:32 <TrueBrain> binaries not yet; I will give it some time for people to stop using it, before we start to redirect it :)
19:37:41 <FLHerne> peter1138: Sorry, I'm confused again :-/
19:38:22 <FLHerne> "Do road and tramtype labels also share a namespace" -> <peter1138> Yes.
19:38:30 <TrueBrain> oops ... healthcheck killed the server .. hmm .. oops :D
19:38:39 *** mcbanhas_ has joined #openttd
19:39:01 <frosch123> FLHerne: no, they do not. but you cannot define a roadtype and a tramtype with the same label
19:39:51 <FLHerne> But there are separate properties to set railtype and tramtype tables, and to test whether a label is defined
19:40:32 <frosch123> FLHerne: from the spec point, they are entirely separate. just peter considered it to complicated, and deleted the implementation that separated them
19:40:41 <frosch123> FLHerne: so, just assume they are separate
19:41:05 <FLHerne> So will Act7 condition 12 "Tramtype label is defined" evaluate as true if the corresponding roadtype label is set?
19:41:33 <FLHerne> frosch123: "but you cannot define a roadtype and a tramtype with the same label" ok, so what happens if you do?
19:41:46 <frosch123> the grf probably gets disabled
19:42:19 <TrueBrain> oops, container is out of memory .. let me fix that quick :)
19:42:20 <glx> tramtype and roadtype are merged in openttd anyway
19:42:49 <frosch123> glx: that is a shitty implementation detail
19:42:59 <FLHerne> glx: But clearly they aren't actually fully merged, because arbitrary bits of the NFO spec treat them separately
19:44:30 <frosch123> FLHerne: again, assume everything treats them separately
19:45:34 <FLHerne> frosch123: Except they aren't, because of the shared set of labels and ID count...
19:45:40 <Samu> 183 ticks, it's still 1 tick less than the previous record
19:45:57 <frosch123> FLHerne: in case you didn't notice. i quit ottd dev because of the shittiness of the nrt implementation
19:46:14 <FLHerne> I think I missed that
19:46:24 <frosch123> FLHerne: so, implement nml according to the intention of the spec. not according to whatever hacks the implementation does
19:46:38 <FLHerne> Anyway, it exists, so nmlc shouldn't let you define broken grfs
19:46:53 <frosch123> no, you are reinforcing bugs, never do that
19:47:09 <Eddi|zuHause> i'm leaning towards frosch123's opinion. the spec wins. if openttd doesn't implement the specs, it needs to be fixed
19:47:56 <Eddi|zuHause> it doesn't help that the half-assed specs were written by the same people that attempted the implementation
19:48:34 <FLHerne> So...just bump the limit to 64 of either
19:48:41 <glx> I can see 2 mapping tables of 64 elements anyway
19:49:32 <FLHerne> And when someone makes a grf with 64 of both, and OTTD rejects it, tell them the spec says that and OTTD is broken
19:49:48 <frosch123> TrueBrain: what magic is that "read more" thing :o
19:50:02 <TrueBrain> we wrote that in for montlhy posts
19:50:27 <FLHerne> frosch123: Should that also apply to railtypes?
19:50:36 <frosch123> lol, i din't notice, i guess i read the news from the vcs
19:50:44 <FLHerne> It looks like currently NML allows a range of 0..63
19:50:48 <glx> railtypes is an old feature
19:50:57 <frosch123> FLHerne: no, 64 railtypes
19:51:14 <FLHerne> So why do roadtypes specifically have one less?
19:51:14 <frosch123> roadtypes have one less to indicate "no road"/"no tram" in the maparray
19:51:31 <frosch123> this case does not exist for railtypes, since there are no rail tiles without rail
19:51:52 <TrueBrain> hmm .. pods are still running the memory limit
19:51:54 <frosch123> FLHerne: grf ids start at 0, ottd concatenates all newgrf ids into a global range
19:51:55 <TrueBrain> that was rather unexpected tbh
19:51:57 <FLHerne> Oh, doesn't matter, OTTD does it
19:52:05 <Eddi|zuHause> well, there's non-track-station-tiles, that actually are a giant hack and do have a track anyway :p
19:53:21 <frosch123> aw, all my old bookmarks with the /en/ are broken :p
19:54:30 <TrueBrain> haha, you never wrote redirects for those?
19:55:08 <frosch123> most endpoints do not match
19:55:31 <frosch123> i think i tried to match /manager/tos, in case it was linked somewhere
19:56:15 <TrueBrain> okay, I see why the pods run out of memory .. they fetch the whole file from S3 .. I was hoping it was streaming it a bit more
19:56:20 <TrueBrain> so when ever someone downloads a huge file, it crashes :P
19:57:56 <TrueBrain> anyway, all DNS entries are changed .. so we will see if we forgot anything :D
20:00:48 <frosch123> ow... i just noticed how silly "only for savegames" is for scenarios/heightmaps...
20:01:00 <TrueBrain> yeah, sorry, my mistake
20:01:05 <TrueBrain> you had "Not Available"
20:01:23 <TrueBrain> so I replaced it with "Ingame only" and "only for savegames"
20:01:30 <TrueBrain> but didn't consider scenarions / heightmaps :D
20:01:32 <frosch123> well, "savegames-only" and "new-games" is as pointless :)
20:01:55 <frosch123> andythenorth: we switched universe
20:01:55 * andythenorth was away-from-brain
20:02:05 <andythenorth> is NRT de-merged then?
20:02:25 <TrueBrain> okay, bananas-server is now up to 512 MB RAM (reservation-wise)
20:02:26 <frosch123> we have a better bananas, but unfortunetely there is a global pandemy instead
20:02:28 <TrueBrain> that seems to be more stable
20:02:39 <frosch123> we are lucky are noone remembers the other universe
20:03:13 <andythenorth> there's a new bananas? :o
20:03:16 <andythenorth> is the UI better?
20:03:33 <frosch123> it got older and wider
20:04:59 <andythenorth> what does reddit think of it?
20:05:08 <TrueBrain> no crashes for 5 minutes, this is getting somewhere :)
20:06:17 <frosch123> can the ships reach the other side?
20:06:51 <frosch123> hmm, or is that wetrail?
20:07:32 <frosch123> nah, only looked like that when zoomed out :p
20:07:58 <andythenorth> needs a ship tunnel
20:09:05 <TrueBrain> ~0.5 mbit/s of traffic I pulled towards AWS
20:09:15 <TrueBrain> how do I check that the ingame content download uses the HTTP?
20:09:22 <frosch123> anyway, i am still unsure about tb's newspost... "make a ticket in the correct repository, and we will look into it." <- is that trolling?
20:10:27 <TrueBrain> you think people are not competent enough? :)
20:10:38 <TrueBrain> (and yes, that post is full of trolls :P)
20:11:09 <TrueBrain> we might want to point to a specific repo for bugs surrounding BaNaNaS, tbh :)
20:11:16 <TrueBrain> and let us move it to the correct repo
20:12:15 <frosch123> maybe -d net=7 shows something
20:13:12 <milek7> it prints error if http failed
20:13:25 <TrueBrain> dbg: [net] [tcp/http] downloading 35522 bytes
20:13:29 <TrueBrain> if that comment is not a lie ..
20:13:46 <frosch123> i looked into the source first :)
20:13:51 <frosch123> looks like it works :)
20:14:03 <TrueBrain> funny, OpenTTD now connects to 3 different IPs to get content :)
20:14:12 <TrueBrain> I am happy AWS just assigns them :P
20:14:32 <TrueBrain> k, redeploying main website, to reduce its memory reservation :D
20:17:36 <frosch123> hmm, we have no dorpsgek.yml in bananas
20:18:19 <TrueBrain> no, I was not sure we would want that
20:18:22 <TrueBrain> feel free to make a PR :)
20:18:33 <TrueBrain> we have no favicon in frontend-web, it seems
20:19:06 <frosch123> i am on ipv4, if that matters
20:19:12 <TrueBrain> GET /favicon.ico HTTP/1.1[0m" 404 -
20:19:29 <TrueBrain> IPv4, IPv6, it all ends up in the same container :)
20:19:35 <frosch123> TrueBrain: clear your cash
20:19:42 <TrueBrain> I am reading logs :)
20:19:43 <frosch123> the url is /static/favicon.ico
20:19:50 <TrueBrain> ah, yeah, that never works
20:20:02 <TrueBrain> when people hit the 404
20:20:06 <TrueBrain> it tries /favicon.ico
20:20:18 <TrueBrain> I am not sure I like my own commit message btw frosch123 :)
20:20:38 <TrueBrain> well, approved anyway :)
20:20:41 <frosch123> sounds like something for the wishlish, when we have nginx in front of it
20:21:30 <TrueBrain> yeah, we should still do something with all that
20:21:34 <frosch123> how long does it take to take effect?
20:21:54 <TrueBrain> your PR? How ever long the GitHub Action takes
20:21:57 <frosch123> does it reload everything, so 2 minutes block?
20:22:47 <TrueBrain> so this is weird .. it does a 'git fetch' and a checkout of HEAD
20:22:51 <TrueBrain> and it says: cannot find ref
20:23:44 <TrueBrain> looking at the amount of 404s, a redirect from /en/newgrf to /newgrf would be a good idea
20:24:06 <LordAro> why not /en/* to /* ?
20:25:14 <TrueBrain> takes 3 minutes to fetch an update from GitHub, according to this pod
20:25:21 <TrueBrain> the initial fetch slow, sure
20:25:26 <TrueBrain> but every after .. should not be that slow
20:25:41 *** Gustavo6046 has joined #openttd
20:26:18 <TrueBrain> owh, no, this was a new pod that was created because it took too long
20:26:29 <TrueBrain> the reload mecanism is .. not really okay :P
20:26:54 <frosch123> haha, is there a limit of the number of pods?
20:27:34 <frosch123> when a new pod starts because the previous one took too long to start
20:28:06 <TrueBrain> no, starting is working fine now (just takes 3+ minutes)
20:28:16 <TrueBrain> the reload took such a long time, the pod was killed before it could finish
20:29:01 *** andythenorth has joined #openttd
20:29:43 <FLHerne> frosch123: If we're being really pedantic about the spec, it says there can be 64 roadtypes :P
20:30:07 <FLHerne> So should we *be* that pedantic, or go with 63?
20:30:30 <frosch123> FLHerne: sadly ttdp no longer exists. we always made a majority decision between ottd, ttdp and specs :)
20:31:39 <andythenorth> where is the spec now? o_O
20:32:12 <TrueBrain> WARNING socket.send() raised exception. <- tnx for spamming, socket
20:32:51 <andythenorth> yes there's that one, but it appears to be wrong
20:33:36 <andythenorth> it's nice though that we label the 'spec' the thing that was retrospectively constructed from 'what have we built here'
20:34:02 <andythenorth> maybe we should have RFC bureaucracy and comittees? :)
20:36:06 <frosch123> andythenorth: i changed 6 numbers in the spec, can you find more palces?
20:37:11 <andythenorth> frosch123 yes, I'll change them, thanks
20:37:39 <andythenorth> searching '64' won't work though :|
20:38:23 <TrueBrain> give me a few to clean up some errors, and to see if I can figure out why reloads fail :)
20:38:40 <peter1138> I see I fucked everything up.
20:38:50 *** peter1138 has left #openttd
20:39:52 <DorpsGek_III> [OpenTTD/nml] FLHerne opened pull request #110: Fix: Correct limit of available road/tramtypes. https://git.io/JfqMK
20:39:59 <FLHerne> ^ does that sum it up?
20:40:48 <andythenorth> frosch123 can't find any more
20:41:16 <frosch123> spnda: you can review it, but it should not matter
20:41:20 <andythenorth> 'I fucked everything up' is my default position
20:41:23 <andythenorth> come back peter :P
20:41:34 <andythenorth> it's software, of course it's fucked up :)
20:41:51 <spnda> frosch123: Well, this is what the email says: You can view, comment on, or merge this pull request online at:
20:41:53 <frosch123> spnda: i think you can review all prs on all repos, you just cannot merge
20:42:23 <frosch123> spnda: i guess it's the same text for everyone
20:42:46 <frosch123> spnda: but please, try to press merge :)
20:43:08 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on pull request #110: Fix: Correct limit of available road/tramtypes. https://git.io/JfqM9
20:43:19 <spnda> It needs an approving review first, should I just try it out?
20:43:34 <andythenorth> oh that's the PR not the ticket sorry :P
20:43:46 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on issue #109: Roadtypes and Tramtypes size is 63 not 15 or 64 https://git.io/JfqYj
20:44:24 <spnda> oh yeah, the review doesn't count as an approving review.
20:44:38 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on pull request #110: Fix: Correct limit of available road/tramtypes. https://git.io/JfqMb
20:45:45 <DorpsGek_III> [OpenTTD/nml] FLHerne commented on pull request #110: Fix: Correct limit of available road/tramtypes. https://git.io/JfqMx
20:46:05 <DorpsGek_III> [OpenTTD/nml] frosch123 approved pull request #110: Fix: Correct limit of available road/tramtypes. https://git.io/JfqDe
20:47:44 <andythenorth> it's bizarrely retro
20:49:49 <nielsm> I definitely did make that layout thinking of 90's game menus
20:50:22 <DorpsGek_III> [OpenTTD/nml] andythenorth merged pull request #110: Fix: Correct limit of available road/tramtypes. https://git.io/JfqMK
20:51:01 <DorpsGek_III> [OpenTTD/nml] andythenorth commented on issue #109: Roadtypes and Tramtypes size is 63 not 15 or 64 https://git.io/JfqYj
20:51:15 <FLHerne> andythenorth: I think we should have waited for peter's opinion there
20:51:20 <TrueBrain> okay, the reload fails because git decides to do a gc action, and gitpython doesn't understand that ..
20:51:33 * _dp_ wish someone would actually think of 20's UI design instead
20:51:58 <andythenorth> FLHerne he already said they're shared though? :)
20:53:07 <FLHerne> andythenorth: Yes, but currently we're ignoring that
20:53:50 <FLHerne> i.e. nml will now let you produce grfs that OTTD won't load, until someone fixes one or the other
20:54:09 <andythenorth> What is 20's style UI design anyway? We're waiting for the flat fad to disappear.
20:54:34 <andythenorth> Most UI design is currently low density, not sure OpenTTD works with that
20:54:53 <frosch123> FLHerne: that applies to many things
20:55:07 <TrueBrain> I see that milek7 found a bug :P
20:55:08 <nielsm> I'd like the UI to be somewhat more defined on high res displays, eventually
20:55:08 <Eddi|zuHause> andythenorth: i'm not entirely sure any of the words you use have any meaning behind them
20:55:28 <nielsm> e.g. scale the single-pixel borders into two and four pixel borders
20:55:32 <TrueBrain> people can spam the API with new uploads :P
20:55:35 <andythenorth> Eddi|zuHause I may just be a cat
20:55:44 <TrueBrain> milek7: stop doing that ;)
20:55:44 <DorpsGek_III> [OpenTTD/nml] FLHerne commented on pull request #110: Fix: Correct limit of available road/tramtypes. https://git.io/JfqyJ
20:55:54 <milek7> i'm trying to get that cake :D
20:55:59 <andythenorth> TrueBrain I don't like to release too often, it wrecks my download counts
20:56:10 <andythenorth> I'll never win if I keep shipping, the way to win is to not maintain grfs
20:56:11 <TrueBrain> milek7: pretty sure I have you banned before you get that cake
20:56:29 <Eddi|zuHause> the cake is a lie.
20:56:34 <andythenorth> OMG WHERE ARE THE DOWNLOAD COUNTS?
20:56:37 <andythenorth> who made this UI?
20:56:51 <Eddi|zuHause> do we have up and downvotes yet?
20:57:49 <TrueBrain> milek7: I really do not want to know why you spend time in creating predefined md5sums, but okay :P
20:58:05 <TrueBrain> of course if people want to abuse a system, they can ;) Lucky, we have other policies in place .. like IP bans :P
20:58:46 <milek7> well, if it would crash service, this is rather important issue
20:59:10 <TrueBrain> so .. you want to proof a point? And instead of being constructive, you decided the destructive route?
21:01:02 <milek7> isn't that what staging is for?
21:01:14 <spnda> what is this new user migration for bananas
21:02:16 <spnda> how do I access all my old GRFs huh
21:02:37 <frosch123> spnda: there is a big link on the front page
21:03:00 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain opened pull request #53: Fix: work around an issue with GitPython being confused if the GC kicks in https://git.io/Jfqyr
21:03:02 <andythenorth> the page, not just the link
21:03:07 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #10: Fix some issues found in production https://git.io/Jfqyo
21:03:08 * andythenorth reads the link text over and over again
21:03:19 <TrueBrain> milek7: I rather have you help me find solutions :)
21:03:28 <TrueBrain> but I appreciate your academic mindset ;)
21:03:37 <andythenorth> frosch123 (now may not be the time) any ideas for the terrifying coop account? :D
21:03:42 <andythenorth> I used it to upload all my grfs
21:03:57 <andythenorth> I guess about 20 people have the creds unchanged for 10 years
21:04:02 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 approved pull request #53: Fix: work around an issue with GitPython being confused if the GC kicks in https://git.io/JfqyX
21:04:03 <spnda> well, i saw that i need to post something on the forums that i want my account to be migrated. i'm sure there's a better way to do this though
21:04:18 <frosch123> TrueBrain: i assume terminate on second fail is intentional?
21:04:28 <TrueBrain> frosch123: yes; it should only happen once, basically
21:04:32 <spnda> "we will change your content over to your GitHub account". What does that mean anyway?
21:04:46 <spnda> Is it just an internal thing or do you create new repositories or something
21:04:58 <TrueBrain> andythenorth: we need to add "Teams" :)
21:05:07 <frosch123> spnda: we are removing openttd's account system
21:05:29 <TrueBrain> andythenorth: funny enough, that is not a bad idea!
21:05:38 <TrueBrain> andythenorth: not difficult to implement, so yeah ...
21:05:56 <TrueBrain> spnda: after you understand what is going on, and you have suggestions to improve that text so others understand it quicker, please do let us know!
21:06:32 <frosch123> andythenorth: do you want to keep the coop account, or make it all yours?
21:07:05 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain dismissed a review for pull request #53: Fix: work around an issue with GitPython being confused if the GC kicks in https://git.io/JfqyX
21:07:05 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain updated pull request #53: Fix: work around an issue with GitPython being confused if the GC kicks in https://git.io/Jfqyr
21:07:09 <spnda> In your example forums post you said "Link to thread: <copy here the URL of your thread>", which thread do you mean. Just any thread that I created for my own GRF?
21:07:37 <andythenorth> also did anyone volunteer to write a UI framework yet? :P
21:07:40 <TrueBrain> spnda: it is refering to
21:07:41 <TrueBrain> Is the URL of any of your content linking to tt-forums.net?
21:07:41 <TrueBrain> Are you the topic starter of this thread?
21:08:02 <DorpsGek_III> [OpenTTD/bananas-api] frosch123 approved pull request #53: Fix: work around an issue with GitPython being confused if the GC kicks in https://git.io/Jfqyh
21:08:08 <TrueBrain> again, any rewording, very much welcome :)
21:09:52 <TrueBrain> frosch123: still not really sure why reload failed .. but lets start with these fixes, and see where that brings us
21:10:31 <TrueBrain> and we are new to this too, so explore with us what works :)
21:10:44 <spnda> Why not have a form on the new bananas page that automates this?
21:11:51 <andythenorth> spnda that starts from 'why is everything not finished yet?' premise :)
21:11:59 <andythenorth> why is not world peace?
21:12:21 <andythenorth> nah, it's a valid suggestion
21:12:24 <TrueBrain> spnda: we considered it; given less than 200 people will use it, it was considered this being the least amount of work overall for everyone; still on the fence about it :)
21:12:42 <TrueBrain> we can make a Google Form, I guess :P
21:12:43 <LordAro> and setting up a system to send emails is quite tricky
21:13:00 <TrueBrain> not the worst idea tbh .. just some external tool to ask you a few questions
21:13:05 <spnda> yeah google forms could make it a bit easier aswell
21:13:08 <TrueBrain> did not consider it :)
21:13:15 <TrueBrain> feel up to setting that up? :D :)
21:13:33 <TrueBrain> (personally a bit busy fixing these darn production bugs :P)
21:14:03 <TrueBrain> and hoping people review my PRs :P
21:14:10 <TrueBrain> I hate needing approval of others :P
21:15:51 <Eddi|zuHause> imagine seeking approval from others, and never getting it...
21:16:14 <Eddi|zuHause> i think they wrote books about that :p
21:17:12 <TrueBrain> dickhead :P (sorry, you deserved that :D It was a good joke :D)
21:17:45 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain merged pull request #53: Fix: work around an issue with GitPython being confused if the GC kicks in https://git.io/Jfqyr
21:17:49 <TrueBrain> 1 PR down, 1 to go :)
21:18:32 <frosch123> oh, that one got lost in the spam
21:18:46 <TrueBrain> I can imagine that happening
21:18:51 <TrueBrain> my mailbox has 400+ emails of the last week
21:19:17 <spnda> TrueBrain: Would you say an application that asks for the OpenTTD account e-mail and a GitHub profile url should suffice?
21:19:21 <frosch123> mailbox is compeltely hopeless
21:19:50 <TrueBrain> spnda: no; those questions we asked are with a reason :)
21:20:05 <TrueBrain> we need some way to proof you are you on GitHub too
21:20:11 <TrueBrain> and knowing someones email .. is not one of those ways
21:20:25 <TrueBrain> that is why we suggest a PM from tt-forums
21:20:47 <TrueBrain> as when you created a thread about some content, and that thread is in the content, and you are the topic starter .. it is VERY likely you are who you say you are
21:20:50 <DorpsGek_III> [OpenTTD/bananas-server] frosch123 approved pull request #10: Fix some issues found in production https://git.io/JfqSw
21:20:58 <TrueBrain> the same, sending an email from that address to us, shows you own that email address
21:21:01 <frosch123> all of them are copies from the api, right?
21:21:03 <TrueBrain> just knowing it, not enough
21:21:17 <TrueBrain> I am very good at copy/paste :)
21:21:40 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #10: Fix some issues found in production https://git.io/Jfqyo
21:21:44 <TrueBrain> I fucked that up :P
21:21:50 <TrueBrain> now we have a weird commit :P
21:22:07 <TrueBrain> owh well, we will have to live with it
21:22:15 <TrueBrain> pretty sure andythenorth won't approve of it when ever he makes the changelog :P
21:23:27 <frosch123> ah, that's what hell looks like... you have to write all the changelogs that you postponed in your life
21:23:43 <TrueBrain> that .. would be horrible
21:24:08 <spnda> just noticed that this sentence makes no sense "I would like to migrate the author of my BaNaNaS content to GitHub."
21:24:16 <Eddi|zuHause> i think that's the intention of hell... to be horrible :p
21:24:29 <TrueBrain> spnda: to me it does :D But I am pretty sure from an outside perspective it is weird :P
21:24:33 <TrueBrain> every BaNaNaS content has an author
21:24:37 <TrueBrain> all authors are now linked to OpenTTD accounts
21:24:45 <TrueBrain> and the question is: change this into a GitHub account
21:24:50 <TrueBrain> again, better wording, pleaaasseee :D
21:24:52 <spnda> Maybe rephrase it like "I would like to migrate my BaNaNaS content to GitHub."
21:25:01 <TrueBrain> the content is not being migrated
21:25:06 <TrueBrain> (well, we already did, sort of :P)
21:25:15 <TrueBrain> it is only the "who has access to make changes"
21:25:29 <spnda> Ok, then "I would like to migrate my BaNaNaS account to GitHub"
21:26:32 *** Lejving has joined #openttd
21:27:08 <TrueBrain> seems sentry doesn't eally like my version numbering .. it thinks 1.0.0 is older than 0.0.0-53-<hash>
21:28:32 <spnda> Oh nice you've got a repository for it
21:28:55 <TrueBrain> everything is now publicly available, source-wise, and auto-deployable :)
21:29:12 <TrueBrain> (we might have spend a bit too much time on a system used by ~200 people more than once :P)
21:29:15 <spnda> Also sent you a account migration pm on tt-forums
21:29:17 <TrueBrain> I am still depressed by that fact frosch123 :D
21:29:37 <frosch123> TrueBrain: luckily you did not notice before it was done :)
21:29:40 <milek7> how to use that server locally?
21:29:45 <TrueBrain> clearly those emails no longer arrive in my mailbox .. I got 5 PMs :P
21:29:46 <milek7> 2020-04-27 21:22:03 ERROR Server returned invalid status code 400. Authentication failed. Error: {'message': "audience is not one of the following: dict_keys(['developer'])"}
21:29:47 <frosch123> TrueBrain: anyway, think of the download users :)
21:30:43 <Eddi|zuHause> how is that an invalid status code?
21:31:29 <LordAro> Eddi|zuHause: invalid status, code 400
21:32:00 <TrueBrain> I see even andythenorth send me a PM on the forums ... seems office365 doesn't accept mails from tt-forums .. something for orudge possibly? :)
21:32:56 <TrueBrain> milek7: not sure what you mean, your information is a bit fragmented
21:33:09 <TrueBrain> it seems you have the API running, as a local dev, which is good
21:33:21 <TrueBrain> after that .. I guess you connect with your own tool?
21:33:30 <TrueBrain> which handshakes via GitHub normally
21:33:34 <TrueBrain> but in a dev-API, that is not enabled
21:33:46 <TrueBrain> not sure if there is code for the bananas-frontend-cli to work around that
21:33:52 <milek7> so unmodified frontend-cli won't work?
21:33:52 <TrueBrain> I only tested it against the staging API
21:34:04 <TrueBrain> not sure it has code for the development flow; honestly, I do not know
21:34:16 <frosch123> TrueBrain: ok, i have access to my content again :)
21:34:35 <frosch123> meanwhile we have 2 testcases for next round
21:34:42 <Samu> new record! 178 ticks, this time for real
21:34:59 <TrueBrain> milek7: you can enable GitHub authentication for the API; you just need to create a GitHub OAuth App for that
21:35:23 <TrueBrain> or write the code for the CLI to work with development flow. We welcome any PR, ofc :)
21:36:38 <frosch123> hmm, i think it would have been smarted if we had asked for the bananas content site, instead for the forum topic
21:36:59 <TrueBrain> now we have to CTRL+F indeed :)
21:38:41 <TrueBrain> hmm .. I don't get it
21:38:45 <TrueBrain> the api works fine, also for reloading now
21:38:48 <TrueBrain> the server .. not so much
21:38:53 <TrueBrain> it is also a lot slower to start up
21:38:57 <TrueBrain> with the same repository checked out
21:40:08 <TrueBrain> owh, one is over SSH, the other over HTTPS
21:40:18 <TrueBrain> the SSH is so much slower, it seems
21:41:04 <TrueBrain> could it be IPv6 or something .. hmm
21:41:54 <TrueBrain> no, github.com has no AAAA record
21:44:00 <andythenorth> what eh, did I do something wrong in a former life?
21:44:06 <andythenorth> is that why I have to write the changelogs?
21:44:13 <andythenorth> TBF LordAro does the hardest one
21:46:08 <TrueBrain> ssh vs http are equal in speed .. so .. what is going on .. I think GitPython is just retarded ..
21:46:43 <TrueBrain> guess ... I am just going to use "os.system"
21:47:10 <DorpsGek_III> [OpenTTD/bananas-frontend-web] frosch123 opened pull request #33: Change: Slightly change migration message template, to make it easier… https://git.io/Jfq9i
21:48:03 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain approved pull request #33: Change: Slightly change migration message template, to make it easier… https://git.io/Jfq9P
21:48:07 <frosch123> TrueBrain: funny, libsubversion and pyperforce are also meh
21:48:29 <TrueBrain> I cannot find what the difference is between API and Server for this speed difference
21:49:21 <DorpsGek_III> [OpenTTD/bananas-frontend-web] frosch123 merged pull request #33: Change: Slightly change migration message template, to make it easier… https://git.io/Jfq9i
21:51:32 <TrueBrain> hmmm .. 30k log messages in a few seconds
21:55:21 <_dp_> first 1.9.3 then jgrpp... 1.10.1 last :(
21:56:03 <frosch123> i can put 1.10.1 on top
21:56:16 <frosch123> but it will still sort alphabetic, not numeric :)
21:56:42 <Eddi|zuHause> i thought natural sorting was a solved issue?
21:57:09 <Eddi|zuHause> as in, there should be a library method to do that?
21:57:15 <frosch123> yes, but smart-assery isn't
21:57:49 <Eddi|zuHause> i'm sorry, but that is my raison-d'être
22:01:09 <orudge> [20:32:23] <TrueBrain> I see even andythenorth send me a PM on the forums ... seems office365 doesn't accept mails from tt-forums .. something for orudge possibly? :) <-- hmm, that might explain why I've not had any e-mails from TT-Forums lately, only just noticed :D
22:01:23 <andythenorth> what are emails? :P
22:02:46 <Eddi|zuHause> i think i've totally lost track of what weekday it is
22:03:17 <TrueBrain> orudge: my lack of emails dates back to march 2019 :P
22:03:39 <orudge> TrueBrain: I only moved my personal e-mails to O365 about a month ago :)
22:04:46 <orudge> "The IP address you submitted 94.130.0.50 was successfully delisted. This may take up to 30 minutes to take effect."
22:14:01 <TrueBrain> okay, found the bug that is causing a lot of logs to be generated
22:16:37 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #11: Fix: stop sending data if the other end closed the connection https://git.io/JfqHH
22:16:51 <TrueBrain> okay, now to find why GitHub is so slow for one, and not for the other ..
22:20:02 <DorpsGek_III> [OpenTTD/nml] glx22 updated pull request #108: Codechange: [Actions] Upload standalone executable to GitHub release https://git.io/JfqJv
22:21:48 <DorpsGek_III> [OpenTTD/bananas-server] frosch123 approved pull request #11: Fix: stop sending data if the other end closed the connection https://git.io/JfqQf
22:21:59 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #11: Fix: stop sending data if the other end closed the connection https://git.io/JfqHH
22:22:29 <TrueBrain> except for the HTTPS vs SSH, I don't see why they differ so much in timing ..
22:23:37 <TrueBrain> switch them both to SSH, and see if that solves the issue, I guess?
22:24:07 <TrueBrain> it is very consistent in its timing at least
22:24:32 <TrueBrain> 4 seconds vs 110 seconds .. lol
22:27:47 *** andythenorth has joined #openttd
22:30:54 <Eddi|zuHause> of all the weird things steam dose, the thing i find worst is the "community" page...
22:30:54 <frosch123> ah, windows users :)
22:31:17 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #12: Add: also allow git to be checked out over SSH https://git.io/JfqQz
22:31:21 <Eddi|zuHause> if i go to the start page, it shows me random screenshots of games i don't have, don't play, and am not interested in
22:31:25 <TrueBrain> frosch123: hopefully the last one of the night; this is mostly a copy from the API
22:31:39 <frosch123> people started putting their newgrfs onto github. someone pushed nml and 7z binaries with it
22:31:54 <Eddi|zuHause> and also fan-translations into german of random games that i also don't have and i'm not interested in
22:32:33 <DorpsGek_III> [OpenTTD/bananas-server] frosch123 approved pull request #12: Add: also allow git to be checked out over SSH https://git.io/JfqQr
22:32:43 <frosch123> TrueBrain: i didn't dare make a tag for the frontend :)
22:33:13 <frosch123> to not disrupt stuff
22:33:36 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #12: Add: also allow git to be checked out over SSH https://git.io/JfqQz
22:35:41 <TrueBrain> release all the things!
22:35:56 <TrueBrain> okay, the logs of the last 30 minutes are empty; that is a good sign at least
22:36:03 <TrueBrain> and nobody yelled that anything is burning
22:36:12 <TrueBrain> I am very curious about our AWS bill tomorrow
22:36:14 <TrueBrain> but that will be fine
22:37:05 <frosch123> did you reply to the forum pm after merge?
22:37:08 <orudge> TrueBrain: AWS is sponsored at the moment, right? At least, I haven't paid any bills lately...
22:37:20 <TrueBrain> we have free credits, yes :)
22:37:29 <TrueBrain> frosch123: I did not, sorry
22:37:33 <frosch123> wasn't it a single-time lump sum?
22:37:43 <frosch123> TrueBrain: i will, when we have all 3
22:37:59 <TrueBrain> after this deployment, I hope merges work :)
22:40:47 <TrueBrain> at least over 5000 downloads already over the new network
22:41:06 <TrueBrain> because we use CloudFront, stats are a bit delayed .. takes a while for all regions to report in
22:41:22 <frosch123> that sounds like a lot
22:41:27 <TrueBrain> the load balancer has almost nothing to do, which is funny
22:41:30 <frosch123> but it is prime time
22:41:37 <TrueBrain> yeah .. and I see a lot of files downloaded exactly the same amount
22:41:41 <TrueBrain> I still find that fishy
22:41:52 <TrueBrain> but I now have better logs, so I can do some statistics on it :)
22:43:27 <frosch123> where does servers.openttd.org run these days?
22:43:46 <frosch123> i thought it was the last thing on the old box
22:44:23 <orudge> the wiki is still on the old box too, I assume
22:44:40 <TrueBrain> wiki is not on the old old old box
22:44:42 <TrueBrain> if that makes sense :P
22:44:56 <frosch123> then i don't get it. i wanted to update the "latest version" to "1.10.1", but it did not work
22:45:03 <frosch123> i killed it, but it still runs the old thing
22:45:10 <frosch123> it doesn't even die :p
22:45:14 <orudge> How many old boxes do we have? :D
22:45:22 <TrueBrain> do you really want a count?
22:45:37 <TrueBrain> frosch123: what did you kill?
22:45:50 <TrueBrain> I see openttd_live still running since this morning
22:45:51 <frosch123> orudge: it's a Matryoshka
22:46:13 <frosch123> TrueBrain: svc -d django_live
22:46:28 <TrueBrain> I never learnt how svc worked
22:46:51 <TrueBrain> I mostly do: ps aux | grep openttd_live
22:46:57 <TrueBrain> and I kill .. what-ever process feels most sensible
22:47:05 <TrueBrain> the "python ./manage" ones always come to mind
22:48:44 <TrueBrain> oops, I made a boo-boo ... configuration is not being accepted ..
22:50:00 <frosch123> spnda: sorry, i changed one sentence an hour ago
22:51:28 <TrueBrain> that moment the system tells you you screwed up, but you cannot find why
22:52:48 <Samu> arf, the idea of caching the path so far for every node, is becoming a bad idea the more I think of it
22:53:03 <frosch123> haha, now i killed too much
22:53:04 <Samu> it requires too much memory
22:53:35 <spnda> TrueBrain: Any documentation on the bananas api yeT?
22:53:47 <TrueBrain> see the news post :)
22:53:49 <frosch123> spnda: read the news post, it's all there
22:54:00 <TrueBrain> any questions, let me know!
22:54:11 <Samu> granted, I'm testing absurd distances, who in their right mind do a 500~ tiles away ship route
22:55:33 <TrueBrain> I do not understand what I did wrong ..
22:55:45 <TrueBrain> on the cluster, it dies a horrible death
22:56:27 <glx> when you can't reproduce locally
22:56:27 <frosch123> _dp_: 1.10.1 is now on top, but it will be forgotten again next release :)
22:57:26 <TrueBrain> especially as a near-identical copy is running fine
22:57:37 <TrueBrain> and I hate just trying again, assuming it magically fixed itself or what-ever
22:58:24 <glx> and it just fails without telling anything I assume
22:58:28 <TrueBrain> nevertheless, that is what I am going to do ..
22:58:39 <TrueBrain> it tells me it cannot do "git fetch", but I have no clue why not
23:00:02 <TrueBrain> nope, it really fails
23:00:04 <TrueBrain> ugh .. what am I missing
23:01:07 <frosch123> TrueBrain: api or server?
23:04:39 <TrueBrain> it seems something to do with my CDK (AWS deployment
23:04:42 <TrueBrain> but I cannot see what ..
23:08:34 <TrueBrain> I am guessing it does
23:08:51 <TrueBrain> but the exact same code works for the API
23:08:53 <glx> as that's the most recent change
23:10:10 <TrueBrain> it is also the change I am pushing, where I change the GitHub URL from HTTP to SSH
23:10:26 <TrueBrain> when I use the same setup, the same variables etc, it does work here fine
23:11:26 <TrueBrain> this took waaayyyy too long
23:12:59 <TrueBrain> that moment you fixed an issue before
23:13:14 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #13: Fix: add github.com to the known_host https://git.io/Jfq7N
23:14:19 <DorpsGek_III> [OpenTTD/bananas-server] glx22 approved pull request #13: Fix: add github.com to the known_host https://git.io/Jfq7h
23:14:46 <TrueBrain> hmm ../bin/sh: 1: ssh-keyscan: not found
23:15:18 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain dismissed a review for pull request #13: Fix: add github.com to the known_host https://git.io/Jfq7h
23:15:18 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #13: Fix: add github.com to the known_host https://git.io/Jfq7N
23:15:24 <TrueBrain> I think I am done with the day :P Starting to slip up more and more :P
23:15:31 <TrueBrain> yeah ... and I failed at copy/pasting
23:15:43 <TrueBrain> that was the second reason it didn't work .. no ssh client :D
23:16:32 <TrueBrain> okay, this was just really really silly :P
23:16:34 <TrueBrain> well, it builds now :)
23:16:36 <DorpsGek_III> [OpenTTD/bananas-server] glx22 approved pull request #13: Fix: add github.com to the known_host https://git.io/Jfq5I
23:17:18 <glx> yeah ssh client helps a lot to use ssh :)
23:17:21 <spnda> Sorry about that commit, accidentally put the changes into frosch's commit
23:17:23 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #13: Fix: add github.com to the known_host https://git.io/Jfq7N
23:17:52 <TrueBrain> okay .... a new attempt ...
23:19:02 <TrueBrain> lately I have been so focused on writing proper Dutch, that clearly my English is slipping even more .. who am I kidding, I never wrote good grammar :P
23:19:42 <spnda> Me neither, asked my mum for some help as she's an english teacher
23:21:06 <TrueBrain> frosch123: that indeed works a lot better, the bananas link
23:21:53 <TrueBrain> orudge: mail works again, sweet :)
23:23:30 <frosch123> and yes, a lot easier
23:24:19 <TrueBrain> I am pretty happy with the setup .. every new deployment of the server, it first starts a new one
23:24:23 <TrueBrain> it updates the NLB with the new one
23:24:24 <orudge> TrueBrain: I've actually moved mail sending for tt-forums to mailgun, though I have asked that the server IP be unblocked anyway
23:24:30 <TrueBrain> if deprecates the old one .. it updates the NLB with it
23:24:38 <TrueBrain> it waits 30 seconds for people to be done
23:25:00 <TrueBrain> orudge: nevertheless: \o/ :D
23:25:23 <TrueBrain> okay .. lets roll out: use-ssh, again ..
23:26:38 <TrueBrain> anyway, mostly what I was rambling about: despite the server being upgraded, nobody should have any notice of this happening .. at least, that is what should happen. No clue if that is true ofc :D
23:28:02 <frosch123> still no angry mob on the forums... why do you force us to register to github?
23:29:05 <andythenorth> there is time yet :)
23:29:19 <orudge> You could have used TT-Forums authentication ;) (At least, TT-Forums now acts as a SAML provider for TT-Wiki)
23:29:27 <TrueBrain> nope, even over SSH it is as slow ... wtf ...
23:29:50 <TrueBrain> how can one pod be this much slower than another .. that makes absolutely no sense to me
23:33:35 <TrueBrain> this is something to look into further another day, I guess
23:33:44 <TrueBrain> I ... simply don't know why one is this much slower than the other
23:34:02 <frosch123> spnda: can you squash those 2 commits somehow. the diff github displays makes no sense at all
23:34:20 <frosch123> github has a lot of issues if diffs have non-ascending time stamps
23:37:07 *** ChanServ sets mode: +o Yexo
23:38:02 <TrueBrain> frosch123: I am going to hold off on merging those access-tickets, as currently when people modify their content, I am not sure everything is reloaded properly
23:38:14 <TrueBrain> I cannot find why the server is this much slower
23:43:40 <TrueBrain> in my local docker it works fine, which suggests it is a network issue .. but how would that be possible ..
23:44:30 <frosch123> the server consumes the complete bandwidth :)
23:45:20 <TrueBrain> AH! It is not the GitHub connection!
23:45:21 <frosch123> how often does the server access the disk? maybe it's i/o limited?
23:45:28 <TrueBrain> I forgot there is one HUGE difference between API and Server
23:45:34 <TrueBrain> the Server does an S3 file listing
23:45:53 <TrueBrain> that is what is taking so long
23:46:36 <TrueBrain> takes a while to list, it seems
23:47:01 <frosch123> does it only do that on restart? or also when notified about new content?
23:47:07 <DorpsGek> Yexo: zuu was last seen in #openttd 3 weeks, 2 days, 0 hours, 29 minutes, and 19 seconds ago: <Zuu> Yes some of the devzone projects have been moved there but not all yet.
23:47:31 <TrueBrain> frosch123: both; it needs to know the new md5sums too
23:47:37 <TrueBrain> yes, this I can reproduce
23:48:20 <TrueBrain> it walks all the folders, one by one
23:49:17 <glx> could spawn workers recursively I guess
23:49:57 <TrueBrain> no, I don't need to make this many API calls
23:50:02 <TrueBrain> I thought I already fixed that .. I did not :P
23:50:45 <TrueBrain> when you list on S3, you get all childs, if you want to or not :P
23:50:47 <glx> "todo" post-it should have fall on the floor behind desktop ;)
23:53:41 <TrueBrain> ah, no, it does not recurse .. that was why I did it like this .. but that is just silly
23:53:44 <TrueBrain> as doing .. euh ...
23:53:57 <TrueBrain> is going to be slow
continue to next day ⏵
