IRC logs for #openttd on OFTC at 2019-01-19
⏴ go to previous day
00:00:10 <nielsm> but can do with a single call that covers all vehicles of each type
00:07:09 <Samu> how do i merge branches without putting one on the other, but on a new one
00:09:10 <Samu> bah, how to undo a commit without commiting a reverse commit
00:09:53 <Samu> rebase without the reverse commit, right? gonna try
00:13:19 <Samu> is there a force fetch or something like that so that I don't get to use recycle bin everytime this happens?
00:13:58 <Samu> i want to revert something that I've done, back to the state I have currently online
00:14:41 <Samu> eh... well forget it, recycle bin it is
00:29:23 <Samu> a branch with the 2 branches merged
00:33:41 <milek7> Samu: >bah, how to undo a commit without commiting a reverse commit
00:34:02 <milek7> git reset --hard commitid
00:36:38 <milek7> (commitid is what you want to reset to, not what you want to revert)
00:39:45 <milek7> (and if you change mind you can just reset back again to newer commit. history is available in git reflog)
01:00:16 *** Thedarkb2-T60 has joined #openttd
01:00:41 *** Thedarkb-X40 has joined #openttd
01:17:41 *** wodencafe has joined #openttd
01:22:16 <Samu> Authentication failed. You may not have permission to access the repository
01:22:32 <Samu> can't push to someone's repository?
01:24:02 <Samu> how do i make a fork of someone's into my own fork?
01:24:15 <Samu> bah i fail at explaining
01:24:47 <peter1138> Add their fork as a remote, then checkout the branch.
01:28:05 <Samu> i created a repository of nielsm on my system, but when i try to push stuff to it, i get authentication failed
01:28:34 <peter1138> yes you need to push to your own repo
01:29:06 <Samu> hmm but it's on my system? this is so confusing
01:29:15 <Samu> how to push it to mine then?
01:34:57 <milek7> (you could do either way, add somebody fork as remote to your repo dir, or add your remote and push to it in somebody repo dir)
01:39:28 <Samu> gonna copy paste nielsm openttd files into mine
01:39:40 <Samu> then i just have to create a branch
01:40:08 <peter1138> adding a remote is very simple.
01:40:53 <Samu> ok, hope i dont ruin my own git bash here
01:41:43 <Samu> do i git bash nielsm's or mine?
01:42:30 <peter1138> use the one you normally work in
01:46:16 <peter1138> replace "origin" with some other identifier
01:46:30 <peter1138> that is your name for the remote
01:46:40 <peter1138> git remote add nielsm ... would make sense
01:47:30 <Samu> but the branches, am i gonna get all his crap on mine? :(
01:47:42 <Samu> im only interested in ship-depot-pathfinding
01:48:02 <peter1138> you only get what you checkout
01:51:20 <Samu> i dont know what's a checkout
01:52:33 <peter1138> git checkout nielsmh/whatever
01:59:17 <Samu> Failed to delete the branch from the remote repository. See the Output window for more details.
01:59:30 <Samu> was trying to delete the branches i'm not interested in
01:59:40 <peter1138> you can't, they're his branches
02:02:51 <Samu> remote: Permission to nielsmh/OpenTTD.git denied to SamuXarick.
02:04:02 <Samu> isn't it on my repository now?
02:04:19 <Samu> what was the point of git remote add then, :|
02:04:46 <glx> the remote is to get the data on your local copy
02:05:31 <milek7> you need to explictly push branch to your remote
02:08:46 <Samu> right click, New Local Branch From... ?
02:09:31 <Samu> A branch with the name "ship-depot-pathfinding" already exists.
02:10:30 <milek7> if origin is your own remote
02:12:39 <Samu> fatal: The current branch ship-depot-pathfinding2 has no upstream branch.
02:12:39 <Samu> To push the current branch and set the remote as upstream, use
02:12:39 <Samu> git push --set-upstream origin ship-depot-pathfinding2
02:22:03 <Samu> i can edit stuff in it now
02:25:39 <peter1138> You don't need to push it to github to edit it locally, of course
02:29:19 *** snail_UES_ has joined #openttd
03:00:25 *** Speedy` has joined #openttd
03:02:54 <Samu> error C2440: 'initializing': cannot convert from 'FindDepotData' to 'const Depot *'
03:03:12 <Samu> can't combine nielsmh code with mine
03:28:11 <Samu> const Depot *depot = (const Depot *)GetDepotIndex(FindClosestShipDepot(v, 0, &failed_depots).tile);
03:41:42 <Samu> const Depot *depot = Depot::GetByTile(FindClosestShipDepot(v, 0, &failed_depots).tile);
03:57:20 <Samu> how do I delete the remote now?
04:13:54 *** snail_UES_ has joined #openttd
05:53:25 *** snail_UES_ has joined #openttd
06:53:10 *** HerzogDeXtEr has joined #openttd
08:50:17 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN dismissed a review for pull request #7072: Change: Add path cache for ships https://git.io/fh03F
09:13:39 *** andythenorth has joined #openttd
09:17:29 *** sla_ro|master has joined #openttd
09:22:54 <andythenorth> so shall I use duck tape?
09:23:00 <andythenorth> or shall I stop doing python wrong?
09:23:09 <peter1138> Time to get ready for bike ride.
09:23:36 * andythenorth testing some random ship patch
09:27:42 <andythenorth> so does it cache dir until the next project turn point?
09:27:46 * andythenorth didn't read the code
09:28:02 <andythenorth> I'm just bulldozing tiles in front of hovercraft to see how it reacts
09:30:19 <peter1138> No, it caches up to 32 tiles.
09:32:25 <peter1138> I just drank a soft drink with 145 kcal in it o_O
09:33:15 <peter1138> Should give me a boost for starting the ride at least, until I crash have way through.
09:38:42 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #7072: Change: Add path cache for ships https://git.io/fhEXF
09:39:30 <andythenorth> so....manually refactor 75 .py files?
09:39:38 <andythenorth> to do things correctly
09:39:51 <andythenorth> or plaster some more @property methods into a single module?
09:40:22 <andythenorth> I think we know the answer :(
09:46:27 <peter1138> andythenorth, oh, and michi_cc already approved it but I decided to fix some comments ;p
09:47:03 <andythenorth> so is deleting OPF next? :P
09:49:45 <DorpsGek> andythenorth: alberth was last seen in #openttd 2 days, 13 hours, 4 minutes, and 1 second ago: <Alberth> o/
09:55:33 *** Progman has joined #openttd
10:07:47 <TrueBrain> that moment that the Azure Pipelines changelog says something is changed, but it isn't .. hmmm
10:10:17 <andythenorth> today might be tax return filing day
10:11:08 <TrueBrain> hmm .. more changes from that changelog are not available .. so .. how do I know when it is .. hmm
10:13:27 <TrueBrain> "These features will roll out over the next two to three weeks."
10:13:33 <TrueBrain> so yeah .. that is a useless changelog :D
10:14:00 <TrueBrain> one of the improvements is that GitHub will deny any Pull Request to work, I think :P
10:14:04 <TrueBrain> so that will be interesting :D
10:20:24 <TrueBrain> meh; wanted to work to publish a nightly ... but I need orudge again to press a button :(
10:20:28 <TrueBrain> annoying Marketplaces
10:28:29 *** ChanServ sets mode: +v tokai
10:31:24 <andythenorth> oof I already wrapped 50% of these files in main() :P
10:31:37 * andythenorth rediscovers previous mistakes
10:31:47 <TrueBrain> :D History tends to repeat itself
10:31:57 <andythenorth> I should have fixed 100% of them last time :P
10:38:53 <TrueBrain> so why didnt you? :P
10:47:28 <TrueBrain> meh; guess that means I should solve the 2 remaining items for the website
10:59:46 <andythenorth> also can we close some issues? :P
10:59:49 <andythenorth> and maybe some PRs
11:10:10 * andythenorth has fewer python modules doing magic on import
11:11:16 <andythenorth> meh, doesn't actually fix my problem :P
11:12:56 <nielsm> nice! got timing to work in my reworked ttd dos music player (that drives the original driver code)
11:16:27 <andythenorth> except there's no risk committee
11:16:44 <andythenorth> we should totally get one
11:31:52 <TrueBrain> andythenorth: something for frosch to judge; I have no clue what the state of the PFs are
11:32:39 <nielsm> imo OPF should be "as close to original behaviour as possible, warts and all"
11:32:46 <nielsm> and the new ones can be extended and improved
11:33:00 <TrueBrain> one more dev to agree, and close them up :)
11:33:05 <TrueBrain> so LordAro, you agree? :P
11:33:57 <LordAro> i'd be tempted to remove OPF entirely, tbh
11:34:14 <andythenorth> I think we should bonfire a few things
11:34:39 <LordAro> andythenorth: like aircraft
11:34:52 <andythenorth> maybe in your patchpack :P
11:34:54 <TrueBrain> so close them up andythenorth
11:35:06 <TrueBrain> clearly this is the new rule now :)
11:35:56 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed pull request #7071: Change: Make OPF check reverse only when exiting depot https://git.io/fhRNP
11:35:57 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #7071: Change: Make OPF check reverse only when exiting depot https://git.io/fhESL
11:36:00 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #7070: Feature: Forbid 90 degrees for OPF https://git.io/fhESt
11:36:50 <LordAro> but once i've looked at it, sure :P
11:37:02 <andythenorth> TrueBrain: do we have a sufficient GDPR declaration for UA parsing?
11:37:05 <andythenorth> should I read it?
11:37:08 <nielsm> andythenorth, copying my comments! INFRINGEMENT!!!
11:37:25 <TrueBrain> GDPR for UA parsing .. that is a first :P
11:37:44 <TrueBrain> are we a data processor now? :D
11:37:49 <andythenorth> we're a controler
11:38:07 <TrueBrain> anyway, fully client-side
11:38:07 <andythenorth> UA parsing is a proper issue, possibly more so than cookies
11:38:19 <andythenorth> anyway, I read privacy policy, I think we're ok
11:38:29 <andythenorth> legitimate purpose, already declared
11:38:34 <TrueBrain> it is fully client-side; so are we the controller? :)
11:38:41 <andythenorth> we're causing it to happen
11:38:57 <andythenorth> blah blah no case law
11:39:08 <TrueBrain> this is really stretching the idea of GDPR :P
11:39:11 <andythenorth> yeah ok, client side risk is tin foil
11:39:15 <TrueBrain> like .. realllllyyyyyy stretching :D
11:39:26 <andythenorth> but we know what binary we served them?
11:39:30 <nielsm> as far as I know, if the data (or a derivative of it) is not being collected it's not an issue
11:39:32 <andythenorth> so we can deduce the result of UA
11:39:44 <andythenorth> so we can infer personal details in combination
11:39:56 <andythenorth> tinfoil even over
11:39:59 <TrueBrain> you smoking too much paranoia today? :P
11:40:07 <andythenorth> no just checking
11:40:17 <TrueBrain> yes, we store how often which binary is downloaded
11:40:32 <andythenorth> no IP addresses in logs?
11:40:41 <TrueBrain> but that is a completely different thing
11:40:56 <TrueBrain> we store logs for debugging purposes for a very limited amount of time
11:41:07 <andythenorth> there's a specific recital in GDPR for that
11:41:29 <andythenorth> 2018 about 30% of my time was spent in GFDPR :P
11:41:37 <andythenorth> the F is not a typo
11:41:39 <TrueBrain> you are not alone, no worries :)
11:42:24 <andythenorth> at the end of that work, zero had changed about the data we process or control, but there was a lot of papework completed :P
11:42:41 <andythenorth> i.e. we weren't doing anything we shouldn't, but lots of trees died
11:43:34 <TrueBrain> I think it is really good companies took a good look internally how they process data
11:43:42 <TrueBrain> we took it too much for granted we could do what-ever-the-fuck-we-wanted with data
11:43:57 <TrueBrain> I have seen many companies that had no way to scrub data
11:44:01 <TrueBrain> or had no retention
11:44:05 <TrueBrain> and all that kind of shit
11:44:19 <TrueBrain> (like seriously .. they had customers that left their company 10 years ago still on record)
11:44:44 <TrueBrain> "but if they come back we still have their data" .. yeah .. exactly
11:45:17 <DorpsGek_II> [OpenTTD/website] TrueBrain opened pull request #20: Add: detect based on User-Agent which download the user wants most likely https://git.io/fhESC
11:49:54 <TrueBrain> guess the rest of my day will be SysOp'ing to fix IPv6 and server-listing ..
11:51:36 <TrueBrain> just that the IPv6 issue is kinda annoying to fix :D
11:52:12 <TrueBrain> IPv4 is served via a LoadBalancer, and downloads will be edge-cached via a CDN
11:52:26 <TrueBrain> and IPv6? It will enter a single machine, that forwards to the IPv4 infrastructure :P
11:53:33 <TrueBrain> I have no other idea how to solve it currently :)
11:53:42 <TrueBrain> our current IPv6 is the same btw, as is our current IPv4
11:53:51 <TrueBrain> so I guess I am improving 90% of the visits :P
11:53:55 <TrueBrain> and I should not worry about the 10% :D
12:03:32 <DorpsGek_II> [OpenTTD/website] andythenorth approved pull request #20: Add: detect based on User-Agent which download the user wants most likely https://git.io/fhES9
12:24:08 *** Arveen2 has joined #openttd
12:27:42 <nielsm> nice, my silly toy also works with other music drivers (e.g. general midi/mpu-401 one) and TTO
13:15:13 <DorpsGek_II> [OpenTTD/website] TrueBrain merged pull request #20: Add: detect based on User-Agent which download the user wants most likely https://git.io/fhESC
13:15:14 * andythenorth wonders about drawing a tech tree with graphviz
13:17:54 <TrueBrain> that moment you are being asked for a code that will be emailed to you, but that ends up in greylisting
13:18:02 *** frosch123 has joined #openttd
13:19:53 <frosch123> i converted the eints repo
13:20:21 <frosch123> not sure what to call the repository on github: eints, webtranslator, WebTranslator?
13:20:47 <TrueBrain> I am fine with what-ever btw :)
13:23:33 <milek7> greylisting should die
13:30:53 <LordAro> there's no way around it, when replying to individual review comments
13:30:57 <LordAro> peter1138: how was cycling?
13:31:02 <TrueBrain> slowly getting there ..
13:33:20 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh merged pull request #6990: Fix: Correct display of industry requires/produces in Build Industry window https://git.io/fhTef
13:36:05 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh dismissed a review for pull request #6928: Fix #5713: Use pathfinder to find closest ship depot https://git.io/fxbLs
13:37:06 <nielsm> (I will not approve that PR, someone else will have to do that if anyone wants it)
13:38:11 *** snail_UES_ has joined #openttd
13:40:58 <TrueBrain> frosch123: possibly you constantly want to squash the changes in the openttd branch
13:41:01 <TrueBrain> so it is one or two commits
13:41:05 <TrueBrain> making it clear what the diff is
13:41:26 <frosch123> rebase instead of merge?
13:41:41 <TrueBrain> no, rebase instead of multiple commits
13:41:53 <TrueBrain> and you need to rebase the openttd branch on master btw :) It is missing a README now :P
13:42:15 <TrueBrain> frosch123: what I mean, you have master, and 'openttd', which is master with a few adjustments, right?
13:42:40 <TrueBrain> I suggest to have openttd as close as possible to master; so squash as many as you can
13:42:48 <frosch123> yes, but up to now there are no history changes in "openttd", lots of merge commits
13:43:11 <TrueBrain> yeah; those will be gone with that too :D
13:43:25 <TrueBrain> it is because I do not think the history of the openttd branch is really important
13:43:39 <TrueBrain> but that is just my 2 cents here :)
13:44:17 <frosch123> all your contributions were to that branch :p
13:49:28 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN dismissed a review for pull request #7072: Change: Add path cache for ships https://git.io/fhEyL
13:56:15 <andythenorth> frosch123: no comments from me 👍
13:58:06 <planetmaker> frosch123, I've no objection moving that repo to github
13:59:06 <planetmaker> actually... its description is outdated. It's not only NewGRF translations. It's the main translation service for this whole community
13:59:28 <planetmaker> but that is not a repo issue :)
14:00:44 <peter1138> "error pulling image configuration: received unexpected HTTP status: 502 Bad Gateway."
14:30:08 *** andythenorth is now known as Guest1125
14:30:18 *** andythenorth has joined #openttd
14:44:38 <andythenorth> want me to try windows?
14:45:16 <TrueBrain> nah; I tested win64, that looks fine
14:51:16 <TrueBrain> okay, I can now set content-type on my uploads ... getting there :D
14:59:12 <TrueBrain> next step ... make this available via the website
15:07:40 <TrueBrain> now I remember why we had media.openttd.org .. as multiple subdomains uses the same CSS etc
15:07:56 <TrueBrain> andythenorth: how are we going to solve that part?
15:08:27 <TrueBrain> guess a submodule used by everything that has the same style or something
15:08:55 <andythenorth> absolute css paths?
15:09:06 <andythenorth> or vendor in the css to the repo
15:09:09 <TrueBrain> so back to media.openttd.org :P
15:09:29 <TrueBrain> currently I want to add an index.html to the CDN
15:09:39 <TrueBrain> so if you go there, you can see in pretty ways the files available
15:09:48 <TrueBrain> but ... I have no real way to reference the CSS atm :)
15:09:54 <andythenorth> I would just treat that as subsidiary to the website, and absolute url to the website css url
15:10:18 <TrueBrain> the nice thing about putting the CSS and images on the CDN too, that it is a lot faster
15:10:22 <andythenorth> all the available ways are fine
15:10:47 <andythenorth> I honestly just copy-paste css around for this situation, if left to solve it alone
15:11:01 <TrueBrain> the images are more the issue; have you seen how many we have :P
15:11:02 <andythenorth> there are usually so few cases, it's not worth much engineering for
15:11:07 <TrueBrain> and LordAro is not fixing it any time soon I am afraid :P
15:11:28 <andythenorth> ok so the downside of CDN is that it makes dev fricking hard
15:11:39 <andythenorth> cos the css isn't trivially changed
15:11:49 <TrueBrain> and they need to be versionized
15:11:57 <andythenorth> but eh, let's not over-engineer it, whatever we do :P
15:12:26 <andythenorth> only 44 images, mostly junk
15:12:33 <TrueBrain> most can be replaced with CSS
15:12:48 <LordAro> i could look at PRing it this weekend
15:13:15 <LordAro> TrueBrain: did we not decide on "OpenTTD YYYYMMDD" for nightlies? or is the above link just temporary?
15:13:18 <andythenorth> I copy-paste a UI library for other similar situations, but I rely on me knowing to do that
15:13:26 <TrueBrain> LordAro: we decided on what is on the CDN
15:13:27 <andythenorth> not everyone is me :P
15:14:09 <TrueBrain> LordAro: the final conclusion was that 'master' is just a branch. If you want a nightly of any other branch, that should be possible too
15:14:44 <LordAro> not sure i agreed with that
15:14:52 <TrueBrain> we are NOT going to open this debate again :)
15:16:02 <TrueBrain> (as if OpenTTD ever only moved if things were unanimous, we never moved at all :P)
15:17:17 <TrueBrain> andythenorth: it is a bit of a pickle .. embedding images/CSS is also not really simplistic :D
15:17:27 <andythenorth> this one is never easily solved
15:17:36 <andythenorth> meanwhile, I hate this github dashboard
15:17:44 <andythenorth> the LinkedIn-ification has started
15:17:57 <andythenorth> 'discover repositories you might be interested in'
15:21:56 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on issue #7062: Ship pathfinders ignore the max order distance https://git.io/fhEb6
15:26:50 <TrueBrain> this is ANNOYING .. stupid CSS/images
15:27:01 <TrueBrain> how else am I going to do this .. hmm
15:27:19 <andythenorth> all the available routes are compromised
15:27:28 <andythenorth> I would do the one with least change
15:27:47 <TrueBrain> so, just to summarize what I am trying to do:
15:27:50 <TrueBrain> I now publish the binaries on the CDN
15:27:58 <TrueBrain> the CDN has no dir-listing; only bucket-listing
15:28:13 <TrueBrain> so there is a manifest.yaml per release that tells which files are there
15:28:24 <TrueBrain> currently, 'website' loads those manifest.yaml, and generate the HTML for it
15:28:32 <TrueBrain> that is now possible, as I can list all the releases that exist
15:28:36 <TrueBrain> on the CDN, not so much
15:28:49 <TrueBrain> so I was thinking of generating the html file while uploading to the CDN
15:28:52 <TrueBrain> as it will never change anyway
15:29:12 <TrueBrain> now the 'website' only needs to point to there, and we are done (which is easy, as the URLs are predictable)
15:29:19 <TrueBrain> but ... css/images are annoying there
15:29:27 <TrueBrain> soooooooooo ... what else is there
15:30:17 <andythenorth> - inline all the css and images (convert to inline styles, I am serious)
15:30:26 <TrueBrain> we can put the static files in a submodule, and publish every commit there under 'media.openttd.org' based on the 'git describe'
15:30:58 <TrueBrain> or even publish the static files from 'website' on the CDN every commit
15:31:43 <TrueBrain> I can add .txt files with the needed content to harvest all information
15:31:52 <TrueBrain> but that means I first need to download a file, amend to it, upload again
15:31:55 <TrueBrain> which .. I prefer not to
15:32:24 <TrueBrain> regenerate all the html files on the CDN every time a new 'website' commit happens :D
15:33:02 <andythenorth> regenerating seems fine
15:33:06 <andythenorth> if it's not loads of engineering
15:33:41 <TrueBrain> ~5000 files need regeneration
15:33:54 <TrueBrain> that is not out of the question
15:34:03 <TrueBrain> just a bit of bandwidth, but also not crazy
15:34:32 <andythenorth> sounds like it might scale horribly
15:34:57 <TrueBrain> 1 new file every day
15:35:28 * andythenorth must kill some mice
15:35:33 <andythenorth> need to buy poison and traps
15:35:39 <TrueBrain> that is .. random :D
15:35:45 <andythenorth> mouse in my house
15:36:02 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro approved pull request #7066: Add: [NewGRF] Airport animation trigger for plane landing (#6334, patch by Supercheese) https://git.io/fhENJ
15:36:12 <nielsm> I had mice in my 3rd floor apartment last winter and the two winters before that as well
15:36:18 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7066: Add: [NewGRF] Airport animation trigger for plane landing (#6334, patch by Supercheese) https://git.io/fhliz
15:36:22 <andythenorth> mice are unwanted
15:36:25 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro closed issue #6334: Airports: add new airport tile animation trigger AAT_STATION_AIRPLANE_LAND https://git.io/fhENU
15:37:01 <TrueBrain> do we keep stats who is doing to most merges? :P
15:38:21 <TrueBrain> hmm, custom domains doesn't work yet for DigitalOcean spaces
15:44:39 *** Thedarkb2-T60 has joined #openttd
15:51:06 <TrueBrain> LordAro: did you have plans to use SaSS instead of CSS, or will it still be plain CSS?
15:51:18 <LordAro> plain css, as i recall
15:54:36 <TrueBrain> hmm .. Jekyll doesn't see a difference between 'static' files and generated ones
15:54:43 <TrueBrain> so you cannot move static files to a CDN, for example
15:54:48 <TrueBrain> at least, not what I can find :(
16:00:16 *** andythenorth has joined #openttd
16:09:51 <Samu> oh no, my stuff was closed
16:14:00 * andythenorth wants a scriptable UI :P
16:14:13 <andythenorth> for newgrf dev, I want the defaults:
16:14:30 <andythenorth> - let me toggle expiring vehicles on/off, and reload_newgrfs
16:14:37 <andythenorth> - let me see the base vehicles :P
16:14:56 <andythenorth> OTOH I never use the scriptable UI in *any* of the apps I have that support it :P
16:21:23 <TrueBrain> this is one hard change to test before pushing it to staging :D
16:38:46 <TrueBrain> last one, I think :)
16:38:56 <TrueBrain> tnx for the nice image HeyCitizen :)
16:47:46 <TrueBrain> andythenorth: what do you think about this solution?
16:48:33 <andythenorth> makes sense to me
16:48:53 <andythenorth> go to prod with it, we can find out what we regret
16:49:15 <TrueBrain> I tested it as much as I could; seems to work fine
16:50:54 <TrueBrain> andythenorth: if you can give it a review, just to look for very obvious mistakes, that would be good :)
16:51:38 <TrueBrain> local development is unchanged; it is just the CI/CD that do tricks now :)
16:53:00 *** Wormnest has joined #openttd
16:55:04 <TrueBrain> hmm .. we should run minimizers over static files :D
16:55:09 <TrueBrain> but that is something for another day :P
16:57:15 <TrueBrain> this is not going to work :(
16:57:19 <TrueBrain> no IPv6 support on the CDN
16:57:43 <DorpsGek_II> [OpenTTD/website] TrueBrain updated pull request #22: Minor cleanups and move static files to CDN https://git.io/fhEAr
16:58:15 <DorpsGek_II> [OpenTTD/website] andythenorth approved pull request #22: Minor cleanups and move static files to CDN https://git.io/fhEpC
16:59:54 <TrueBrain> till DO fixes IPv6 on Spaces, I cannot do these kind of tricks
17:00:01 <TrueBrain> guess I have to go a bit more old-fashion ..
17:03:51 <LordAro> Samu: last i checked several of your other PRs needed work/responses
17:16:57 <Samu> I need help here, but I guess nobody understood yet what I'm trying to fix
17:22:21 <Samu> in sum, i need an order expert
17:24:45 <Samu> someone who knows how to determine which location the vehicles goes next and the location the vehicle was before that
17:25:30 <Samu> I didn't look at the case of conditional orders, I'm too scared to even try
17:34:25 <Samu> on that one, i disagree with nielsmh
17:36:20 <TrueBrain> 2 choices: convince him (in that Pull Request) otherwise, or remove that part of the PR
17:36:31 <TrueBrain> he does not have to convince you; you him :)
17:37:40 <TrueBrain> andythenorth: the new version number doesn't fit in the box in the top of the website :D
17:38:17 <andythenorth> url, or it didn't happen :P
17:38:53 <andythenorth> I guess I could pull :P
17:39:08 <andythenorth> I'm drawing pixels and arguing with children currently though
17:41:35 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6926: Change: Allow dock to be constructed in more locations https://git.io/fhEjR
17:42:04 <TrueBrain> Samu: that is NOT convincing anyone of anything
17:42:10 <TrueBrain> doing something "because" is never a good argument
17:42:17 <TrueBrain> explain why you want that change, why you think it is important
17:42:27 <TrueBrain> "because" is the worst argument, and one you can never win :)
17:42:31 <TrueBrain> tell that in the pull request!
17:42:39 <TrueBrain> and not just that, why YOU think it is bugged
17:42:45 <TrueBrain> help others help you
17:42:59 <Samu> you try to place a dock on those places and you are denied, even though there's water there
17:43:13 <Samu> why do I have to explain
17:43:15 <TrueBrain> tell that in the pull request. And adding pictures really helps
17:43:21 <TrueBrain> because you want the change
17:43:24 <TrueBrain> you have to do the work
17:43:29 <TrueBrain> don't expect others to do the work for you
17:44:27 <Samu> i thought it was implied with the title i gave
17:44:57 <TrueBrain> clearly it was not; otherwise it wouldnt have been an issue ;)
17:45:09 <TrueBrain> but here a picture can really help you out; showing what you mean exactly, can help others to understand
17:45:51 <glx> yes pictures showing exactly the denied places
17:46:26 <glx> text is not enough for cases like this
17:48:44 <andythenorth> before and after ideally
17:49:02 <andythenorth> I did read the PR and the code, but I have no idea how I would even test it
17:49:08 <andythenorth> so I didn't fetch it
17:49:17 <glx> because the diff doesn't help to visualise
17:52:14 <TrueBrain> awh, .md files are not shown inline in browsers
17:53:00 <glx> you need a md to html parsing I guess
17:53:42 <TrueBrain> that would be nice :D
17:53:58 <TrueBrain> for now I just rename it to .txt :)
17:55:32 <TrueBrain> owh, the .txt is also not shown inline
17:56:55 <glx> doesn't seem very large too me
17:57:20 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6926: Change: Allow dock to be constructed in more locations https://git.io/fhueq
17:57:31 <TrueBrain> andythenorth: text/plain
17:57:37 <TrueBrain> just no explicit inline
17:57:40 <TrueBrain> because ... that is annoying
17:57:47 <glx> maybe it's the way how browsers handle it
17:58:19 <andythenorth> something DO are setting?
17:58:27 * andythenorth hasn't read any headers or anything
17:58:39 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6926: Change: Allow dock to be constructed in more locations https://git.io/fhues
17:59:14 <TrueBrain> I really have no idea why it makes that difference ..
17:59:51 <TrueBrain> Samu: so much more clear
18:00:47 <Samu> some AIs aren't prepared for the one corner raised cases closing a water connection, they think that by building a dock, the passage still remains
18:00:58 <Samu> it can now close passage
18:02:30 <DorpsGek_II> [OpenTTD/website] TrueBrain opened pull request #23: Add: support binary distribution via our new infrastructure https://git.io/fhuer
18:02:40 <TrueBrain> andythenorth: please do pull this and test it; no rush, but this really needs validation :)
18:02:52 <TrueBrain> if you can run those magic "check-if-links-are-all-valid" scripts over them, would be awesome
18:03:24 <DorpsGek_II> [OpenTTD/website] andythenorth commented on pull request #23: Add: support binary distribution via our new infrastructure https://git.io/fhuei
18:03:26 <andythenorth> might not be today
18:03:39 <TrueBrain> andythenorth: no rush
18:03:59 <TrueBrain> without it, we cannot go live, no; but I wont have much time either, so no worries
18:04:06 <TrueBrain> have to fix 2 other things first anyway
18:04:10 <TrueBrain> which ... will take time ..
18:04:47 <DorpsGek_II> [OpenTTD/website] TrueBrain commented on pull request #23: Add: support binary distribution via our new infrastructure https://git.io/fhue9
18:04:55 <TrueBrain> images to look at :D
18:06:06 <DorpsGek_II> [OpenTTD/website] TrueBrain updated pull request #23: Add: support binary distribution via our new infrastructure https://git.io/fhuer
18:06:23 <TrueBrain> having nightlies is creeping up on us :D
18:07:30 <TrueBrain> hmm ... IPv6 issue is back .. that CDN does not work on IPv6 .. but if I do not give the direct link, CDN is useless
18:07:37 <TrueBrain> and custom domains are not supported yet, so I cannot fake it with a CNAME
18:07:44 <TrueBrain> welllllllllll .... that is a nice blocker
18:08:23 <LordAro> TrueBrain: any idea of what "yet" is?
18:09:04 <LordAro> is there any sort of eta, or is it just a "coming soon" ?
18:09:08 <TrueBrain> I was about to contact them for that :)
18:09:17 <TrueBrain> from what I read, it is in private-beta
18:12:59 <LordAro> alternatively, IPv6 support :p
18:17:54 <TrueBrain> basically, that is what I asked
18:18:19 <TrueBrain> "So I either must route all the traffic via a Droplet (losing edge-caching) or not have IPv6 support (not an option)."
18:18:56 <TrueBrain> the worst is, if they cannot do either, I really have to look into GCP/AWS/Azure
18:19:02 <TrueBrain> and those are so freakinglish unclear
18:19:08 <TrueBrain> I really really don't want to do that :P
18:19:21 <TrueBrain> but okay, ticket sent
18:19:28 <TrueBrain> tonight hopefully I can fix server listing
18:19:38 <TrueBrain> and those are the only 3 remaining things, I think
18:20:22 <TrueBrain> (btw, nightlies are being build every night already; and published. Just not easy to get access to yet)
18:20:53 <TrueBrain> time to feed myself :D
18:21:10 <andythenorth> Samu: it's unlikely to be merged
18:21:16 <andythenorth> it's a boil-the-ocean task
18:21:38 <Samu> it's a change that only makes sense if it's pursued further
18:22:16 <andythenorth> Samu: I'd close that PR and focus on others
18:22:42 <Samu> hmm ok, i'm not in the mood to pursue it further, sorry
18:23:23 <Samu> i think it was Wolf01 that incentivated me to do it
18:23:32 <Samu> the concept of Docking Tile
18:24:12 <Samu> oh, that one is finished, unless you find problems
18:26:23 <Samu> regarding where to place the checks, I have placed where I think it's best
18:26:53 <Samu> but, for me it seems to work
18:48:00 <Samu> it's not any ordinary block, but the block that makes the ship dead-locked, with no means to get out of there
18:54:16 <Samu> imagine that ship is not blue
18:54:33 <Samu> that ship would be dead-locked
18:54:36 <LordAro> Samu: stop explaining things here
18:54:45 <LordAro> things here get lost or just not seen
18:56:19 <nielsm> if you don't feel confident in long-form writing, keep in mind it's a skill that requires practise
18:56:47 <nielsm> explaining technical things concisely and accurately _is_ hard, and you need to do it a lot to get good at it
18:57:06 <nielsm> but it's also a very valuable skill
19:05:18 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6935: Prevent water construction where it could block nearby ships https://git.io/fhufp
19:05:28 <Samu> I suck at explaining, I really do :(
19:06:25 <Samu> I know what's wrong, I see what's wrong, but then I fail at explaining it to others
19:07:15 *** Thedarkb2-T60 has joined #openttd
19:10:19 <Samu> not gonna get approved I guess
19:13:11 <Samu> yesterday's PRs were closed, I really didn't expect it
19:13:48 <Samu> i combined them together, and today I combined them with the find ship depot
19:14:15 <Samu> it further reduced the stall that happens when ordering 5000 ships to depots
19:14:29 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6935: Prevent water construction where it could block nearby ships https://git.io/fhuJY
19:14:32 <Samu> was making progress in that front
19:14:48 <nielsm> the problem is that you're using OPF
19:15:14 <glx> well more like a zombie ;)
19:15:20 <nielsm> OPF exists to mimic the behavior of 1994 TTD, not to be good
19:18:48 <glx> maybe it's possible to allow non moving ships to do 90° turns while moving ones can't
19:19:34 <andythenorth> I commented here multiple times the OPF is dead
19:20:33 <nielsm> it's only going to cause trouble later on
19:22:03 <glx> hmm someone mismeasured something ;)
19:26:33 <Samu> eh, so will yapf find ship depots one day?
19:26:49 <Samu> it's where the stall is minimal
19:28:03 <Samu> by minimal, it's still a 2.7 seconds stall
19:28:10 <Samu> but considering 5000 ships
19:29:33 <Samu> and only happens once, it's not a check that's done every tile, only done when requested
19:30:59 <Samu> then of course, the ships will be pathfinder later on for the resulting dest-tile
19:31:08 <Samu> but that doesn't involve stalls
19:32:40 <Samu> if petern patch is approved, the path will be cached
19:34:05 <Samu> have to combine petern's cache into my combined stuff for testing
19:35:42 <Samu> but the stall for npf is another different story
19:36:39 <Samu> haven't timed opf stall at the moment, brb
19:42:51 <andythenorth> I don't recognise the case where 5000 ships all need to find a depot simultaneously
19:50:36 <Samu> oh well, it was a nice experiment :|
19:54:21 <Samu> 3.456 ms stall with 90 deg allowed
19:54:38 <Samu> still better than my previous 4.7
20:13:22 <Samu> hmm wasn's the 3x3 roads alignment to happen between towns?
20:13:50 <Samu> some towns dont align with others
20:15:07 <Samu> oops im using random roads, nevermind
20:22:30 <Samu> where do i have the generate friendly rivers patch
20:25:18 <frosch123> do we need different "teams" with merge-permission for openttd/nml/eints/website/compilefarm/dorpsgek/..., or should we continue with a single team? do people want more badges? :p
20:26:38 <LordAro> frosch123: if the need arises, imo
20:34:16 <Samu> peter1138 what happens to cache when switching 90 degrees on and off
20:34:53 <nielsm> why would that matter, nobody does that during normal play
20:34:54 <Samu> or change to other pf and back?
20:35:59 <peter1138> I guess I consider the path is still valid? ;)
20:36:43 <peter1138> I guess switching could clear the path.
20:38:46 <m1cr0man> Does openttd load configs and such from it's install directory or the working directory?
20:39:23 <LordAro> m1cr0man: readme section 4.2
20:39:41 <m1cr0man> ty, I actually forgot how good the readme is
20:39:45 <peter1138> Hmm, need to find my laptop power supply.
20:41:38 <peter1138> ^ LordAro, approved, no commit?
20:44:05 <LordAro> peter1138: i wanted someone else to look at/approve it, i think
20:46:32 <andythenorth> every time I get to 90% I ship Horse
20:46:36 <andythenorth> then I add more stuff :P
20:48:25 *** Thedarkb2-T60 has joined #openttd
20:48:42 <peter1138> Ship early, ship often.
20:48:54 <peter1138> Did I see emails earlier about binaries and nightlies?
20:48:55 <nielsm> but those aren't ships.
20:48:58 <nielsm> train early, train often
20:49:50 <andythenorth> horse early, horse often
20:50:19 <andythenorth> currently it conflicts
20:50:33 <LordAro> or wait, when was it last fixed?
20:50:42 <andythenorth> fixed before we changed road pathfinder
20:51:00 <andythenorth> long-running fork, merge hell
20:51:35 <peter1138> I'll take a look tomorrow.
20:51:53 <LordAro> there doesn't look like an awful lot of conflicts
20:51:56 <peter1138> Currently I'm at the M-i-L's house, on tethered Internet.
20:51:58 <LordAro> nothing scary looking anyway
20:52:09 <andythenorth> no, just above my knowledge to resolve
20:52:17 <andythenorth> I don't want to guess, and introduce subtle bugs
20:52:27 <andythenorth> or even unsubtle ones
20:52:56 <peter1138> Samu, any other configuration settings you can spot which might want the path caches to be invalidated?
20:53:19 <andythenorth> right mexican food time
20:54:17 <Samu> stuff that alters the cost
20:54:39 <nielsm> are there configuration settings that do that?
20:54:52 <peter1138> I don't see a setting to do that.
20:55:24 <peter1138> There's max map height but that can't be changed.
20:55:40 <Samu> canals make some ships move slower, at least i saw some code there
20:55:41 <nielsm> adding checks during landscape manipulation for obstructing any cached ship paths would probably negate any advantage gained by the caching
20:55:46 <LordAro> peter1138: not again :p
20:56:43 <LordAro> More Height Levels took so long to get implemented :p
20:57:03 <peter1138> Eh, I mean the setting can't be changed mid-game, therefore there's nothing to invalidate.
20:57:25 <Samu> your code doesn't compare costs?
20:57:39 <peter1138> Why would it? The pathfinder does all that.
20:57:40 <Samu> at least i would check if the path cost is still the same as before
20:58:17 <peter1138> You'd need to... run the pathfinder again to see if it changed.
20:58:22 <peter1138> And that would be utterly pointless.
20:58:40 <nielsm> path caching might make ships take less optimal routes in some cases where the landscape changes around it
20:58:49 <nielsm> that's the tradeoff for the performance improvement
20:58:53 <Samu> only retrieve the costs of the cached cache and check if it still matches
20:59:29 <peter1138> nielsm, yup, that's why I set a fairly low limit to the cached path length. Actually 32 is quite high.
20:59:41 <peter1138> But I figured we don't need Yet Another Option.
21:00:05 <Samu> only 32 tiles? :( i thought it was caching the whole path
21:00:06 <Eddi|zuHause> but the same thing happens for trains that have long path reservations
21:00:20 <Samu> well i didn't really look at the code in full
21:01:28 <peter1138> 32 tiles means up to 32x less pathfinder calls.
21:01:39 <peter1138> Which is significant.
21:03:14 <Eddi|zuHause> so we allow 6 times longer routes and have 6^2 longer pathfinder calculations :p
21:04:05 <peter1138> Heh, well, I'm wondering now about that earlier patch that refuses to pathfind over max order distance. Would be worth seeing if this cache makes that unnecessary.
21:04:17 <peter1138> Heck maybe max order distance becomes unnecessary.
21:04:40 <Eddi|zuHause> that would be ideal
21:04:52 <peter1138> Now that I think about it... shared path cache...
21:05:02 <peter1138> Maybe this PR is a bit premature :p
21:05:22 <Eddi|zuHause> introduce the concept of routes? :p
21:05:52 <Eddi|zuHause> the route is a complete path, and each ship only needs to follow that path
21:06:39 <glx> shared patch cache seems like a good idea
21:06:46 <peter1138> Cache a complete path, indexed by start and end, and then look that up when finding a path to follow.
21:07:33 <peter1138> Damn. And this patch was approved at two points!
21:07:33 <glx> but a change in the middle of the route can be a problem
21:08:27 <peter1138> glx, just invalidate the cache if a ship encounters an obstruction, and maybe add a lifetime to each path.
21:08:57 <peter1138> How common is 100 ships sharing a route?
21:09:45 <glx> cache between 2 orders and not the full list
21:10:58 <peter1138> I'm thinking lookup a cache by start/end tiles, so that non-shared orders still benefit.
21:11:24 <peter1138> Or where orders partially follow the same route.
21:12:07 <glx> ah yes a global cache instead an individual cache
21:12:43 <peter1138> Could have a LRU as well.
21:14:14 <peter1138> Hmm, less optimal for not-found paths.
21:19:00 <peter1138> Sadface. I don't have the proper base graphics installed on this laptop :/
21:22:22 <Gabda> would it help in the ship path finding, if the water domains were indexed?
21:23:04 <Gabda> if the 256 largest individual water surfaces were numbered
21:23:26 <peter1138> What would help is know if start -> end is unreachable.
21:23:27 <Eddi|zuHause> <peter1138> How common is 100 ships sharing a route? <-- i could imagine some people actually do that :p
21:23:34 <Gabda> you wouldn't need to search for most of the path not founds
21:24:09 <Gabda> if you can check at the begining that they are in different water domains
21:24:18 <peter1138> Yeah but "256 largest" isn't that useful.
21:24:38 <peter1138> Find isolated areas and flag them as such, yeah.
21:24:47 <peter1138> Could be tons though :(
21:24:55 *** Thedarkb1-T60 has joined #openttd
21:25:09 <Eddi|zuHause> you don't really know how large an area is when you start indexing it
21:25:15 <Gabda> if one is numbered, the other is nit, there is no route
21:25:20 <peter1138> 256 may or may not be enough.
21:25:32 <peter1138> 65535 may not be enough on a 4096x4096 map.
21:25:38 <Gabda> it none of them is numbered, t
21:26:02 <Gabda> it means that it is one or two, but small ponds
21:26:19 <peter1138> I guess you could mark really small areas the same, and always exclude them.
21:26:32 <Gabda> and in small ponds, it is easier to tell if there is no route
21:26:36 <peter1138> Like an area that's too small to contain a multiple docks.
21:26:58 <LordAro> you'd have to invalidate the cache on almost every terraform action though
21:27:27 <Eddi|zuHause> you can update the indexing on clearing/flooding
21:28:15 <Eddi|zuHause> we discussed this a few days ago, it's a relatively simple flood-fill operation
21:28:22 <Gabda> it might be tricky if you cut a surface into two
21:29:04 <Gabda> but you can find that on the first path not found "error"
21:30:11 <Gabda> but I don't know how often this problem comes up that someone wants to send a ship where it clearly cannot go
21:30:36 <Eddi|zuHause> you just have to loop around the cleared area, and start a new flood-fill action (with a new index) from each tile that still has the old index number
21:31:29 <Eddi|zuHause> peter1138: i mean, it's not gonna be that much code
21:32:16 <Eddi|zuHause> Gabda: i think the more common case is that a route that was once possible gets cut by some unrelated thing
21:32:50 <glx> and sometimes a longer route may exist too
21:40:18 <peter1138> Gabda, so yeah, if you can make a patch for it... ;)
21:41:21 <peter1138> There's loads of space on plain water tiles.
21:41:33 <Gabda> or detecting if a route gets cut?
21:46:47 <Gabda> with a 2 byte indexing it is possible to number each separate water surface above the size of 128 tiles, even on the 4096x4096 map
21:48:28 <peter1138> I think m2 is free, that's 2 bytes.
21:49:48 <Gabda> ships can only go to a different hight if there is a lock build at the transition?
21:54:20 <glx> m2 is not free for ship depots
21:54:56 <glx> but m8 is free for all water tiles
21:55:43 <Gabda> m3 and m4 is free for dock and bouy and depot
21:55:57 <glx> TrueBrain: seems ok for me (and IPv6)
21:56:20 <Gabda> and oilrig (if the ship can go under it)
21:56:49 <peter1138> Well, there's m8 too, yes.
21:57:37 <Gabda> can a ship sail on oilrig tiles?
21:57:56 <DorpsGek_II> [OpenTTD/website] TrueBrain opened pull request #24: Fix: also redirect serverlisting pages to their own subdomain https://git.io/fhuLD
21:58:15 <Gabda> but m3 << 8 + m4 can work
21:58:32 <Samu> what are you discussing?
21:59:49 <Samu> no, indexing water bodies?
21:59:51 <Gabda> but only for a few tiles
22:01:29 <Gabda> well, m8 seems easier, but it is nice to see it so free :)
22:01:29 <peter1138> TrueBrain, looks okay but I don't look at it often enough to know if something is different.
22:01:53 <TrueBrain> if nothing obvious is broken, it should be working, right? :D
22:01:56 <peter1138> Hmm, so global path cache?
22:02:03 <peter1138> TrueBrain, ship it :D
22:02:31 <Samu> what are we caching now?
22:03:07 <Gabda> it seems like everything
22:03:19 <LordAro> TrueBrain: train it :D
22:03:19 <peter1138> Full path between orders, I think.
22:03:21 <Samu> how will the water bodies help?
22:03:43 <TrueBrain> right .. how to setup a ipv6 to ipv4 gateway ...
22:05:07 <TrueBrain> would be nice if it works for TCP and UDP I guess
22:05:12 <Samu> that cost can decide different routes for different ships
22:05:19 <peter1138> Oh, UDP necessary too.
22:05:47 <TrueBrain> yeah, Master Server ...
22:05:48 <Gabda> water bodies could help to find "no routes" faster, if it is possible to implement it well
22:07:59 <LordAro> TrueBrain: remind me how to start/build the website?
22:08:03 <TrueBrain> owh, naptd is not in debian .. hmm
22:08:14 <TrueBrain> LordAro: docker build, for sure
22:08:21 <TrueBrain> looking in there should for sure help you out :D
22:08:28 <TrueBrain> but install Jekyll, and run 'jekyll serve'
22:08:31 <TrueBrain> or use a Jekyll docker
22:09:34 <DorpsGek_II> [OpenTTD/website] LordAro approved pull request #24: Fix: also redirect serverlisting pages to their own subdomain https://git.io/fhut8
22:09:36 <TrueBrain> I will add stuff in the README soon-ish
22:09:41 <DorpsGek_II> [OpenTTD/website] LordAro merged pull request #24: Fix: also redirect serverlisting pages to their own subdomain https://git.io/fhuLD
22:10:08 <TrueBrain> hmmm .. I forgot to change the URL in the menu, I guess
22:12:27 <Samu> caching between shared orders may not work well if the ship models differ
22:13:35 <DorpsGek_II> [OpenTTD/website] LordAro commented on pull request #23: Add: support binary distribution via our new infrastructure https://git.io/fhutP
22:14:45 <DorpsGek_II> [OpenTTD/website] TrueBrain commented on pull request #23: Add: support binary distribution via our new infrastructure https://git.io/fhutD
22:15:04 <DorpsGek_II> [OpenTTD/website] TrueBrain commented on pull request #23: Add: support binary distribution via our new infrastructure https://git.io/fhuty
22:16:36 <TrueBrain> why fixing things if you can make a bug ticket out of it :D
22:17:27 <TrueBrain> okay, an IPv6 is more complex than I assumed
22:17:43 <TrueBrain> guess we have a bit of an unique situation, where we only want to convert a single IPv6 to a single IPv4 :)
22:17:46 <peter1138> Samu, could make that cost part of the cache index.
22:21:26 <peter1138> Maybe best to stick with the caching we have for now. It's still a massive improvement.
22:22:10 <peter1138> Anyone want to approve it? ;)
22:22:31 <peter1138> Before I decide to change anything else.
22:25:07 <Samu> i don't know how to approve
22:26:54 *** Gustavo6046 has joined #openttd
22:28:34 *** gelignite has joined #openttd
22:32:08 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick opened pull request #7073: Feature: Generate lock ready rivers upon world generation https://git.io/fhuqz
22:34:33 <Samu> lock ready or lock friendly
22:35:25 <Samu> it's not ready for forbiden 90 degrees
22:40:29 <DorpsGek_II> [OpenTTD/website] TrueBrain opened pull request #27: Update: README.md is now more verbose and even contains a FAQ https://git.io/fhuqd
22:40:31 <TrueBrain> LordAro will be so proud :D
22:40:37 <LordAro> TrueBrain: `Pagination Error: Detected invalid url "/news" for "pages/news_archive.html" xpected "/" or "/index.html"` any ideas?
22:40:56 <TrueBrain> none; never seen that
22:41:09 * peter1138 attempts to remove path_dest
22:42:07 <TrueBrain> LordAro: which version of Jekyll?
22:42:16 *** Speedyn has joined #openttd
22:42:40 <peter1138> Can someone go for a pee for me?
22:42:51 <TrueBrain> hmm, I have the same LordAro ..
22:43:12 <TrueBrain> LordAro: you did install the Gemfile.lock deps?
22:43:21 <TrueBrain> especially the jekyll-feed and jekyll-paginate-v2 ?
22:44:37 <LordAro> TrueBrain: pretty sure i have
22:46:31 <LordAro> oh hang on, i've updated, rather than installed
22:46:55 <LordAro> mayhaps jekyll-paginate-v2 2.1 has changed something
22:47:10 <DorpsGek_II> [OpenTTD/website] TrueBrain updated pull request #27: Update: README.md is now more verbose and even contains a FAQ https://git.io/fhuqd
22:47:19 <TrueBrain> I updated the README to show what you should do
22:47:31 <TrueBrain> should used the pinned versions
22:47:55 <TrueBrain> not completely sure, because ruby
22:48:30 <LordAro> yeah, got it running now
22:49:09 <DorpsGek_II> [OpenTTD/website] TrueBrain updated pull request #27: Update: README.md is now more verbose and even contains a FAQ https://git.io/fhuqd
22:49:27 <TrueBrain> README is not perfect yet, but it is a start :)
22:49:51 <TrueBrain> this 'website' became pretty .. huge :D
22:50:47 <LordAro> not that that's a particularly useful bug report
22:51:17 <TrueBrain> will check it out when I bump the versions .. cant be arshed atm :D
22:55:12 <Samu> generating 13k towns when there's more water than land is tooo slow
22:58:04 <Samu> why doesn't the world editor allow me to generate random rivers?
22:59:18 <peter1138> Because you didn't add that feature yet? :D
23:03:52 <TrueBrain> "2a03:b0c0:2:d0::dc6:7000 - 2a03:b0c0:2:d0::dc6:700f" <- LOL @ IPv6 space that DO assigns to a droplet .. that is ... cute?
23:05:16 <TrueBrain> I think they are scared they run out
23:06:26 <glx> I have a lot more for just a home connection
23:07:31 <peter1138> I think I have a /48 for home, although it's not all routed there.
23:10:39 <milek7> scaleway is even more lol, /128...
23:11:16 <Samu> i wish i could drag and drop branches onto each other to combine patches
23:12:44 <peter1138> You could make a tool that lets you do that :p
23:15:10 <TrueBrain> but but buttttttt OMG
23:19:23 <TrueBrain> okay, this IPv6 stuff is more of an issue than I expected; if I had spend a few more brain-cycles, I would have realized ithat :D
23:20:26 <TrueBrain> IPv4 to IPv6 is easier, as an IPv4 fits in an IPv6 :P
23:25:38 <milek7> ah, but probably doesn't do udp
23:26:03 <TrueBrain> yeah, but if you look in the code, there is one minor issue :D
23:26:09 <TrueBrain> you can fork-bomb that very very quickly
23:27:14 <TrueBrain> 6tunnel has a global connection limit, instead of a per IP or something
23:27:40 <TrueBrain> guess nginx will have to do .. (not the HTTP part, but nginx can also route TCP and UDP)
23:34:39 <milek7> hard to have meaningful per-IP limit on ipv6 ;D
23:35:07 <TrueBrain> nginx does this pretty well
23:35:15 <TrueBrain> right .. now I need an IPv6 enabled host .. hmm
23:39:23 *** Progman has joined #openttd
23:45:51 <glx> http is redirected to https
23:46:07 <TrueBrain> I guess that means it works :P
23:46:34 <glx> firefox says ipv6 but the openttd icon doesn't ;)
23:46:44 <TrueBrain> no, the icon will no longer change :)
23:46:50 <TrueBrain> I forgot we added that gimmick :)
23:47:23 <TrueBrain> okay ... so I now have a DigitalOcean LoadBalancer which is configured automatically via kubernetes, which is awesome
23:47:35 <TrueBrain> and ... an IPv6 LoadBalancer I had to setup myself and manually sync with kubernetes :P
23:47:38 <TrueBrain> this will NEVER go wrong!
23:47:44 <peter1138> No IPv6 on my tethered ghetto connection :/
23:48:27 <TrueBrain> but for now this allows us to have IPv6 as before, so I can continue the migration .. means only the CDN issues are left .. hmm
23:48:48 <TrueBrain> I guess I can add a redirector there, which splits IPv4 from IPv6
23:49:00 <TrueBrain> sub-optimal, but at least means we can go live with the new stuff
23:50:54 <milek7> it seriously needs loadbalancer? one machine for simple site is not enough?
23:52:13 <TrueBrain> I think you heavily underestimate how busy OpenTTD is, and especially can be :D
23:52:27 <TrueBrain> and many parts of the OpenTTD services are far from 'simple'
23:52:38 <TrueBrain> the 'website' project is just the first and simplest :)
23:53:34 <TrueBrain> but also, LB is common practice, and as we are running k8s, also simplistic to add :)
continue to next day ⏵
