IRC logs for #openttd on OFTC at 2018-11-01
⏴ go to previous day
00:03:06 <frosch123> "danish" is a unique identifier, right?
00:03:14 <DorpsGek> planetmaker: Bjarni was last seen in #openttd 7 years, 3 weeks, 5 days, 22 hours, 44 minutes, and 8 seconds ago: <Bjarni> heh
00:22:31 <planetmaker> I guess I fell for the same misjudgement as froshc
00:25:43 <planetmaker> he, a whopping 12 years in the forums without any posting in between
00:26:23 <nielsm> and I'll let you know that for a very long time (until about 1-2 years ago) I absolutely refused to touch anything OTTD at all, for moral reasons (essentially theft/reappropriation of IP)
00:27:02 <nielsm> but decided recently to adopt a ship of theseus model, in that most things have been replaced by now making it a different same thing
00:27:48 <nielsm> I was owner of #tycoon on quakenet for several years too :)
00:29:27 <planetmaker> I wouldn't know that... never been on that network. I guess I dropped-in here somewhen in the ~2006 and got sucked-into :D
00:30:05 <planetmaker> probably I read forum already before that w/o registering
00:30:22 <frosch123> there was a quakenet time before freenode?
00:30:33 <frosch123> or was tycoon on quakenet and openttd on freenode?
00:31:19 <nielsm> #tycoon was originally on quakenet, I think #openttd was too for a little while but not sure
00:31:43 <nielsm> but for the above reasons #tycoon had a strict "do not talk about ottd" policy
00:32:21 <frosch123> oh, indeed, my logs from 2007-05-16 say quakenet
00:33:20 <frosch123> ottd was already on oftc at that tiem
00:35:28 <ST2> lazy way to check if channel was still with +R
01:16:31 *** snail_UES_ has joined #openttd
02:43:19 *** Hobbyboy has joined #openttd
03:32:49 *** Hobbyboy has joined #openttd
08:58:40 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6925: Fix #6574 #6636 #5405 #6493: Aircraft hangar issues https://git.io/fxN2W
09:09:39 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6926: Change: Allow dock to be constructed in more locations https://git.io/fxN2A
09:14:53 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6927: Change: Forbid dock placement on docking area https://git.io/fxNac
09:20:58 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6931: Change: Prevent town growth from blocking ships https://git.io/fxNa2
09:22:39 *** chomwitt has joined #openttd
11:22:59 *** Progman has joined #openttd
11:45:26 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh opened pull request #6956: Fix #6145: Prevent ships without valid orders from moving https://git.io/fxNMB
11:49:38 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6846: Feature: Allow flip of train vehicles in depot independently of NewGRF property https://git.io/fxNMM
11:50:03 <nielsm> can we close that PR if the author doesn't respond within a week or so?
11:51:42 <planetmaker> a week is probably too short. But a month seems reasonable to me
11:51:54 <planetmaker> good morning also
12:12:16 *** matt21347 has joined #openttd
12:59:14 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6923: Fix #6460: [AI] Add start_date parameter for Random AIs on new game https://git.io/fxNQu
13:01:07 <nielsm> okay I think I'm done shooting down samu PRs for today :P
13:03:12 *** snail_UES_ has joined #openttd
13:06:24 <nielsm> (but really, I try reading and understanding them, purpose and method)
13:22:05 *** frosch123 has joined #openttd
14:24:39 <Wolf01> Oh... it's not saturday, I was wondering why there wasn't a FFF this week
14:30:57 <Wolf01> I took 2 days to rest, nothing to do with catholicism :P My coworkers are working
14:33:22 <frosch123> huh? how long have you been in this channel?
14:34:10 <planetmaker> but yes, I don't own Factorio (yet)
14:34:21 <planetmaker> waiting for a good deal somewhere on gog or steam :P
14:34:48 <frosch123> i bought the 10€ version, since 5€ appeared too little
14:34:56 <frosch123> but i guess you are too late for that :p
14:35:23 <Wolf01> Yeah, and I don't think it will be officially discounted
14:35:33 <frosch123> otoh, i got a t-shirt from V, so i made profit
14:36:02 <frosch123> oh, right, kovarex said that he very much opposes sales
14:36:31 <Wolf01> Ha! I want to buy one tee, but that "we plan to add more gear to the shop" is forcing me to wait
14:37:35 <Wolf01> I would like an inserter+belt statuette :P
15:36:19 *** Wormnest has joined #openttd
15:43:44 <peter1138> Urgh, that's that time when that single sale day, Black Friday, starts for the month...
15:43:56 <peter1138> (The one that was only in the US a couple of years ago)
16:27:24 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6931: Change: Prevent town growth from blocking ships https://git.io/fxAmP
16:59:59 <peter1138> It's like £21 now, so no, I'm not going to buy it.
17:02:19 <peter1138> Might've bought it for 5 or 10 but never knew about it then.
17:05:57 <Wolf01> I think 20€ is the right price for a 1.0 release, at this point 10-15€ is fine, but 26€ is too much
17:12:25 *** Wolf01 is now known as Guest1954
17:19:35 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6927: Change: Forbid dock placement on docking area https://git.io/fxAcT
17:20:56 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #6928: Fix #5713: Use pathfinder to find closest ship depot https://git.io/fxAcl
17:25:36 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6923: Fix #6460: [AI] Add start_date parameter for Random AIs on new game https://git.io/fxAcx
17:29:54 <nielsm> oh yeah, have anyone else noticed "dbg: [misc] String too long for destination buffer" printeds to the console a bunch when starting the game? it appears it's copying the revision name into a gamelog buffer that runs out of space
17:30:04 <nielsm> I'm not sure how much trouble that will cause
17:31:30 <nielsm> but the buffer is 15 bytes and the revision name uses 9 bytes just for date + dash, and then also wants 9 more bytes at least for git hash, and also a branch name between date and revision hash if not on master
17:31:56 <nielsm> so it can definitely cause non-identical versions to show as identical in the gamelog, as far as I understand
17:33:03 <peter1138> Another cause is music packs, there's a fixed length buffer in there for some reason which causes massive spammage if you download a few music packs from bananas.
17:34:33 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6925: Fix #6574 #6636 #5405 #6493: Aircraft hangar issues https://git.io/fxACx
17:34:47 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh approved pull request #6923: Fix #6460: [AI] Add start_date parameter for Random AIs on new game https://git.io/fxACh
17:47:44 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6925: Fix #6574 #6636 #5405 #6493: Aircraft hangar issues https://git.io/fxAlz
17:56:14 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh merged pull request #6923: Fix #6460: [AI] Add start_date parameter for Random AIs on new game https://git.io/fxA8l
17:56:30 *** yellyFish has joined #openttd
17:57:02 <nielsm> there's way too many individual changes in that aircraft hangar PR, I think it'd be easier to handle if split up
17:57:10 <DorpsGek_II> [OpenTTD/DorpsGek-github] pyup-bot opened pull request #19: Scheduled monthly dependency update for November https://git.io/fxA8g
17:57:28 <nielsm> ROBOTS TAKING INITIATIVE
17:58:12 <DorpsGek_II> [OpenTTD/DorpsGek-irc] pyup-bot opened pull request #15: Scheduled monthly dependency update for November https://git.io/fxA86
17:59:15 <DorpsGek_II> [OpenTTD/DorpsGek-runner] pyup-bot opened pull request #10: Scheduled monthly dependency update for November https://git.io/fxA8y
18:06:42 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh merged pull request #6901: Add: Fluidsynth music driver support (jmak) https://git.io/fxA4X
18:12:03 <LordAro> i wonder whether we can differentiate between commebts & reviews
18:12:32 <LordAro> "comments", on the whole, don't need to be echoed here, i'd suggest
18:17:18 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #6566: Very long loading of the maximum "zoom out" level in 4K resolution https://git.io/fxABN
18:17:19 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh closed issue #6566: Very long loading of the maximum "zoom out" level in 4K resolution https://git.io/fxABA
18:30:56 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #6498: Money bug in Detailed performance rating window https://git.io/fxA04
18:33:34 <frosch123> i see no big difference between comments and reviews
18:33:55 <frosch123> unless you mean comments on issues vs. comments on prs
18:35:09 <LordAro> mostly the latter, yeah
18:35:50 <LordAro> you get one review notification, and a comment notification for each response to every bit of the review
18:48:41 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh opened pull request #6957: Fix #6498: Use int64 for all company rating calculations https://git.io/fxAuK
19:14:57 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #6322: AI can allocate more memory than the system has, crashing the game https://git.io/fxA20
19:45:00 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6925: Fix #6574 #6636 #5405 #6493: Aircraft hangar issues https://git.io/fxAre
19:46:31 *** andythenorth has joined #openttd
19:51:51 <andythenorth> it's like PR christmas out there
19:57:07 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh opened pull request #6958: Fix #6864: Normalise colour use in Frame rate window https://git.io/fxAoP
20:04:01 <LordAro> nielsm: stop being so productive
20:05:02 <andythenorth> gotta tidy all samu's patches eh :)
20:05:37 <nielsm> I've taken this week off work
20:12:51 <nielsm> now, time to break squirrel completely
20:19:04 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on pull request #6925: Fix #6574 #6636 #5405 #6493: Aircraft hangar issues https://git.io/fxAiL
20:32:11 <michi_cc> andythenorth: You're slacking ;) No love from you for my OSX PR :)
20:32:35 <andythenorth> I can't build reliably currently
20:32:41 <andythenorth> I am stuck on LordAro's branch
20:32:53 <andythenorth> merging multiple forks is odd
20:33:04 <andythenorth> unreliable results :)
20:34:22 <nielsm> alternative hack for building: write a small wrapper shell script for clang that responds more favorably to --version
20:34:42 <nielsm> and tell configure to use that as CC
20:39:40 <andythenorth> could we just fix configure?
20:42:53 <nielsm> configure is PERFECT IN ITS CREATION
20:53:48 <TrueBrain> I am just shocked OpenTTD still uses config.lib, which I wrote 10 years ago :P
20:53:55 <TrueBrain> go replace it with something sane, like CMake, or something :D
20:57:27 <frosch123> do you want to use it to build on windows?
20:57:45 <TrueBrain> that is right, CMake and Windows are friends these days
20:57:48 <TrueBrain> which is pretty nice :)
20:57:51 <andythenorth> if I set LordAro's fork as the upstream to a clone of Michi's fork?
20:57:55 <andythenorth> I might get a build?
20:58:21 <TrueBrain> just cherry-pick them into each other :P
20:58:38 <TrueBrain> hmm, I even forgot that vspkg and CMake are friends
20:58:47 <TrueBrain> meaning you could work towards dropping openttd-useful
20:58:47 <andythenorth> I love providing feedback on branches when I'm testing them in my own personal patch pack :P
20:59:07 <andythenorth> do we all have our own PP these days? o_O
20:59:10 <TrueBrain> now that would be a nice project :)
20:59:28 <andythenorth> TrueBrain: shall I advertise in the forums :P
20:59:46 <andythenorth> 'Project: replace config.lib with buzzwords'
20:59:54 <TrueBrain> CMake is not a buzzword, but sure :P
21:00:09 <andythenorth> is 'bees' a buzzword? o_O
21:00:13 <frosch123> "make openttd build on windows with cmake and vspkg"
21:00:20 <frosch123> "without openttd-useful"
21:00:31 <TrueBrain> that would seriously improve things
21:00:33 <frosch123> i think someone asked for a more precise task
21:02:02 <frosch123> will people take that serious?
21:02:17 <andythenorth> do we want serious people?
21:02:20 <andythenorth> I can more serious it
21:02:21 <frosch123> it's 9 lines which could be shortened into one
21:04:02 <TrueBrain> I love how peopple took my personally trello for shit-to-do-for-OpenTTD as a task-list anyone could pick up :)
21:04:06 <TrueBrain> somewhere some nuance got lost :P
21:04:17 <TrueBrain> can we install Jira to order stuff better? :D
21:05:03 <frosch123> if we install 10 issue trackers, will that order stuff even better? :)
21:05:30 <andythenorth> we can track the issue trackers
21:05:40 <andythenorth> I love how every issue tracker eventually links to a trello
21:05:44 <andythenorth> it's like a modern fact
21:19:12 <TrueBrain> I will see this weekend if I can convert what is on Trello into something more actionable in general
21:19:34 <TrueBrain> or possibly I make an 'infrastructure' project on GitHub or something
21:20:13 <TrueBrain> wish you could move issues between projects on GitHub
21:20:40 <TrueBrain> that is now a beta functionality :D
21:20:49 <andythenorth> or you could API it :P
21:21:47 <TrueBrain> that beta went live ... YESTERDAY :D
21:22:07 <TrueBrain> okay, that is cool and useful
21:23:53 <andythenorth> now we can move all issues to JGR's GH page?
21:24:25 <TrueBrain> I wonder what I should prioritize .. and that answer is more complex than I would like
21:24:45 <TrueBrain> having nightlies is nice .. but having a stable infrastructure can very quickly escalate to a must-have :P
21:27:40 <andythenorth> nightlies is not that big a deal right now
21:27:50 <andythenorth> TrueBrain bus factor worries me most :P
21:28:10 <nielsm> having more "regular people" testing everything would be useful though
21:28:16 <andythenorth> it would yes, but
21:28:29 <TrueBrain> andythenorth: I keep hearing that .. just to be perfectly clear: there are more people who have the keys to the castle
21:28:52 <andythenorth> so what are the alternative candidates to nightlies?
21:29:49 <TrueBrain> I dont understand the question :)
21:29:55 <nielsm> I could set up that machine I have with a semi-fried HDMI output to run headless windows and build once in a while
21:30:16 <TrueBrain> nielsm: we have the hardware, that is not the issue :)
21:30:30 <nielsm> WELL SOMETHING NEEDS TO BE DONE!!!!!
21:30:40 <andythenorth> TrueBrain: so what's complex about the prioritisation? o_O
21:30:44 <andythenorth> or should I just play tanks? :P
21:31:04 <TrueBrain> andythenorth: I find it hard to balance .. finish CF, or make sure the infra is stable
21:31:22 <TrueBrain> nielsm: if we like, I can cook up a hack solutions; just it doesn't scale :P
21:31:32 <TrueBrain> no, let me rephrase that
21:31:55 <TrueBrain> nielsm: I try to avoid people creating binaries that are published on openttd.org which run on "home computers"
21:32:03 <TrueBrain> we have some .. terrible experience with that :P
21:32:17 <TrueBrain> openttd.org has the infrastructure to run shit
21:32:23 <TrueBrain> we just need the thing to run :)
21:32:24 <nielsm> yeah, I know it's not great to run production-ish things on residential lines
21:32:37 <TrueBrain> that, and a chain-of-trust, reproducability, etc
21:33:01 <TrueBrain> what I can do, for example, is start up Bamboo again, and let him compile the stuff again
21:33:05 <TrueBrain> only .. he runs compilers from 2015
21:33:11 <TrueBrain> which are not C++11
21:33:18 <TrueBrain> (which people had issues with, for some odd reason :P)
21:36:43 <TrueBrain> nielsm: another reason things are more difficult than they used to be: before GitHub, we could 'trust' the content of the repository to not have a bitcoin miner, for example
21:36:52 <TrueBrain> as such, we just had a Jenkins Agent on a Windows VM, doing its thing
21:37:06 <TrueBrain> with the GitHub move, I ...have .... a bit of issues just blindly running PRs :P
21:37:21 <TrueBrain> that would require a lot more sysadmining, I am pretty sure :D
21:37:31 <TrueBrain> so I would strongly prefer if we could containerize stuff
21:37:43 <TrueBrain> (as we can just kill Dockers that run for more than N minutes :P)
21:37:49 <nielsm> ah right, I was thinking just starting out with master builds
21:37:59 <nielsm> presumably nobody merges random shit without understanding it :)
21:38:13 <TrueBrain> we don't enforce 2FA yet
21:38:18 <TrueBrain> which we should, if we do that
21:38:28 <TrueBrain> plenty of GitHub repositories have been owned the last few months :)
21:39:04 <TrueBrain> maybe we should do that anyway :P
21:39:11 <nielsm> and later, builds of PRs should probably only happen after approval from someone
21:39:23 <andythenorth> I would like to not get owned by OpenTTD
21:39:28 <nielsm> builds that get binaries published, that is
21:39:33 <TrueBrain> last week I got the chance to talk to a few people who run bigger farms than OpenTTD
21:39:35 <andythenorth> it would be preferable to not have to buy a separate mac to run it on
21:39:40 <TrueBrain> asking them how they mitigate the risks
21:39:47 <TrueBrain> is ... interesting ... :P
21:40:15 <TrueBrain> (basically, they have things to prevent abuse, but in the end, it requires their attention to keep it clean)
21:40:49 <TrueBrain> but yeah, having the CI trigger only after some whitelisted person asks for it is possible
21:40:55 <TrueBrain> I was more thinking to use that to publish the binaries
21:41:05 <TrueBrain> as we can fence off that the CI either is killed after N time, or behaves
21:41:16 <TrueBrain> but I am more worries about the binaries we publish for others to download
21:41:25 <TrueBrain> we have to be reasonably sure they are not infected in some matter
21:41:38 <TrueBrain> on the forums, it is nice people post their binaries
21:41:44 <TrueBrain> but .. I would never download any of them :P
21:41:53 <andythenorth> what extra risk is there in GH compared to owning our own repo?
21:41:57 <TrueBrain> (well, only to check for malware, from time to time :P)
21:42:08 <TrueBrain> andythenorth: fully whitelisted
21:42:13 <andythenorth> compromised user creds is same risk both places?
21:42:14 <TrueBrain> SVN didn't have pull requests
21:42:28 <TrueBrain> yes; but it was non-trivial to login
21:42:36 <TrueBrain> (svn+ssh over an URL not published)
21:42:51 <andythenorth> so you'd have to craft a specific attack?
21:42:54 <TrueBrain> GitHub is widely known ... so a much easier vector :)
21:43:03 <andythenorth> GH is commodity attack
21:43:16 <TrueBrain> people who do these things (and can harvest credentials) don't do targeted attacks :P
21:43:26 <andythenorth> shall we just move to client-server architecture / MMORG :P
21:43:33 <andythenorth> and distribute a signed client in app stores
21:43:48 <andythenorth> that escalated quickly
21:44:14 <TrueBrain> well, left or right, publishing binaries does require some level of confidence
21:44:21 <TrueBrain> signing the binaries is not the worst idea
21:44:32 <TrueBrain> sadly, not many good common ways to validate the signature
21:44:45 <TrueBrain> (on Windows I am sure 99% of the people never ever validated a signature :P)
21:47:37 <TrueBrain> ah, here the CMake bla comes showing up again : currently for a Windows container, I have to rename files to match the expected name by the MSVC project file :D
21:47:45 <TrueBrain> vspkg calls it libpng16.lib
21:47:51 <TrueBrain> OpenTTD expected libpng.lib :D
21:48:00 <TrueBrain> (and there are a few more)
21:48:25 <nielsm> alternatively set up the msbuild project to accept overrides of the library names
21:48:53 <TrueBrain> I know nothing about the MSVC project :D
21:49:32 <TrueBrain> so let me show you a few things .. I have to type the URLs, this will be fun:
21:49:58 <nielsm> several years ago I did some pretty snazzy project setup for aegisub, with custom property pages and all :)
21:49:58 <TrueBrain> that is the "common" part for Windows Container
21:50:23 <TrueBrain> and there is the rest of it
21:50:34 <TrueBrain> line 28 .. 73 are "fixes"
21:50:47 <TrueBrain> just to conform to our build-system
21:51:00 <TrueBrain> but I can summarize that different:
21:51:08 <TrueBrain> "make MSVC work with packages from vspkg"
21:51:16 <TrueBrain> (dropping openttd-useful)
21:51:37 <TrueBrain> maybe even more actionable than changing to CMake :P
21:51:50 <TrueBrain> (as you noticed in these Dockerfiles, no openttd-useful is installed)
21:51:50 <nielsm> is vspkg part of the win10 sdk?
21:52:00 <TrueBrain> vspkg is a Microsoft project on GitHub
21:52:21 <TrueBrain> fucking hell, typing is hard
21:52:57 <TrueBrain> cannot believe I kept typing that wrong :P
21:53:11 <TrueBrain> andythenorth: fix that in your post plz :D
21:53:44 <TrueBrain> I think I can drop the whole ICU hack, and just not install it
21:55:04 <TrueBrain> okay, slowly crawling to more actionable things to get the CF alive:
21:55:38 <TrueBrain> 1) remove openttd-useful (meaning we have up-to-date libraries ... which sounds like a good idea I guess :P)
21:56:01 <TrueBrain> 2) check if the VM on our server is Windows 10 Pro
21:56:24 <TrueBrain> 3) validate that my 'windows' branch of OpenTTD-CF still works; fix what is broken (see 1)
21:56:24 <orudge> I did intend recently to try to look at CMake for OpenTTD. As usual, time is the problem :(
21:56:41 <TrueBrain> orudge: yup; I made a quick mockup months ago. It is pretty straight-forward
21:57:21 <orudge> CMake + vcpkg sounds like it would make things much easier for MSVC users
21:57:37 <TrueBrain> it drastically improves starting with OpenTTD on Windows
21:57:57 <TrueBrain> (instead of having to go through a wiki tutorial :P)
21:58:37 <TrueBrain> either way, to continue my brain-dump
21:58:51 <TrueBrain> 4) find a controller to run release-* every night on the 'master'
21:59:01 <TrueBrain> 5) publish those results ... 'somehow', for people to download
21:59:25 <TrueBrain> 4) is most likely Jenkins; not the best, but will hav eto do
22:00:14 <TrueBrain> 5) is more tricky than I would like .. currently binaries are published in an internal VM, to be moved to another internal VM, to be mirrored
22:00:43 <TrueBrain> 2) is an easy check ... let me login
22:01:30 <TrueBrain> haha, not enough memory available
22:04:32 <TrueBrain> yup, it is Windows 10 Pro; good, means we can run Hyper-V on it
22:04:49 *** Wolf01 is now known as Guest1967
22:05:25 <Wolf01> Internet disservice provider
22:05:30 <Wolf01> Oh, a wild TrueBrain: since you seem good with docker, I would like to know what it could do for me, I want to publish a SAAS cloud infrastructure, every time I need to publish a new version I could do it once or for each container?
22:06:48 <Wolf01> Does it work on windows server?
22:07:02 <TrueBrain> Docker on Windows in production I havent seen yet
22:08:01 <TrueBrain> either way, a container is just what is running an image (in docker terms)
22:08:27 <TrueBrain> if you make a new image, you of course have to recreate the containers
22:08:36 <TrueBrain> there are people who update the containers themself
22:08:42 <TrueBrain> but .. that is a bit weird
22:08:46 <TrueBrain> (why are you using docker in that case)
22:09:10 <TrueBrain> left or right, the answer is not simple. But one piece of advise: don't use docker because it is hot
22:09:23 <TrueBrain> use it because you have a need for it
22:09:34 <Wolf01> It was one of the ideas :P
22:09:59 <TrueBrain> don't solve a non-existing problem by throwing a solution against it
22:10:18 <TrueBrain> first wonder what you are trying to solve; after that look for solutions that supply that :)
22:10:45 <TrueBrain> (this currently really happens a lot .. people using Docker because "it is hot" .. some for cloud ..)
22:11:03 <andythenorth> Wolf01: where are your instances deployed?
22:11:09 <andythenorth> VMs on a cloud provider?
22:11:36 <Wolf01> We have a server on a datacenter
22:11:49 <andythenorth> and do you deploy multiple instances of the same app?
22:11:52 <TrueBrain> I hope it is in it, but that is just me :D
22:12:40 <TrueBrain> nielsm: I hope you understand a tiny bit more that for me at least it is non-trivial to have a working nightly compile farm :D
22:12:41 <Wolf01> Currently it is just one app with multiple databases, but the users are growing and the machine doesn't seem to keep up
22:13:11 <andythenorth> we run single-tenant apps, in single tenant VMs
22:13:21 <TrueBrain> nielsm: but if you know MSVC projects a bit, if you could look into removing openttd-useful, that would really really help me
22:13:35 <andythenorth> a long time ago we ran multi-tenant apps on shared hosting with other customers
22:13:46 <andythenorth> over time we've moved through multiple solutions
22:13:50 <Wolf01> That was my idea too, but our problem is to minimise the weekly update time and work
22:14:24 <andythenorth> single-tenant app, single-tenant VM on private cloud offers the best infosec for us
22:14:33 <andythenorth> but it also has some scaling issues
22:14:38 <nielsm> TrueBrain: let's start by removing the project files for VC older than 2015 :)
22:14:52 <TrueBrain> sounds like a plan :D
22:15:22 <andythenorth> Wolf01 our app instances are deployed from git tags, we have a utility that polls for new tags and runs the update process if one is found
22:15:30 <andythenorth> this scales well to N
22:15:57 <andythenorth> it could also be done with Ansible or a similar orchestration tool
22:16:11 <andythenorth> we use orchestration to manage the VMs
22:16:28 <andythenorth> we look at docker sometimes because 'everyone else uses it' but we miss what we would gain
22:16:41 <andythenorth> it just seems to add complexity and attack surface in our case
22:17:35 <TrueBrain> using only docker for production at this day and age is silly
22:17:43 <TrueBrain> you need at least a wrapper around it to help you out :)
22:17:49 <andythenorth> Wolf01 in our model we pay relatively high ££ for VMs, for workload that could just be run on one huge box
22:17:52 <TrueBrain> (kubernetes is flavor-of-the-month)
22:18:10 <andythenorth> but generally outages are limited to fewer customers at once, unless the hosting provider loses a whole rack
22:18:32 <andythenorth> TrueBrain: I kind of prefer steam age technology for some reason :P
22:19:05 <TrueBrain> I am pleasently surprised you use an orchestration tool for VMs :)
22:19:09 <TrueBrain> that too, I see not often enough
22:19:13 <TrueBrain> with hilarious results :D
22:19:38 <andythenorth> cattle not pets, no?
22:19:49 <andythenorth> we have half a plan to just image them and redeploy images
22:20:12 <andythenorth> but on OpenStack it seems to occasionally mean things that should be quick are very very slow
22:21:37 <TrueBrain> still have to try out things like OpenStack
22:21:44 <TrueBrain> I only know about it, never used it
22:22:00 <TrueBrain> same for openvswitch, also from what I understand very useful in cloud-like setups
22:22:38 <andythenorth> we use it via the rackspace web UI :P
22:23:26 <andythenorth> one day I find an SRE
22:23:30 <andythenorth> then all my problems are solved
22:23:46 <TrueBrain> keep dreaming my friend :D
22:24:02 <andythenorth> do I just have 2 problems?
22:24:14 <andythenorth> and I have to manage an SRE :P
22:24:21 <andythenorth> problems doubled
22:24:56 <andythenorth> I had an idea that cloud would free us from servers
22:30:05 <yellyFish> you are telling me that data is not stored in the sky?
22:30:31 <TrueBrain> hmm .. I also should really setup a staging area for DorpsGek_II .. I cannot test the PyUp change :D
22:30:34 <planetmaker> I'm telling you there are data in the sky :P
22:30:48 <TrueBrain> and crashing down on us!
22:30:52 <TrueBrain> OWH NO, WE ARE BOMBARDED !
22:31:12 <yellyFish> actually, china's satellite...
22:31:50 <TrueBrain> right, that is my cue to leave :P
22:37:57 <nielsm> the bash 'generate' script adds fluidsynth to the windows project files
22:38:20 <nielsm> neither of them should defined FLUIDSYNTH
22:40:28 <nielsm> oh, the bash script does it backwards
22:42:57 *** Wolf01 is now known as Guest1971
22:45:41 <nielsm> and apparently the pre-commit check script does not handle deleted files
22:52:01 *** Wolf01 is now known as Guest1972
22:52:03 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh opened pull request #6959: Fix: Fluidsynth should not be added to VC project files https://git.io/fxANi
22:53:12 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh opened pull request #6960: Remove: Visual Studio project files for earlier than 2015 https://git.io/fxANX
22:54:49 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 approved pull request #6959: Fix: Fluidsynth should not be added to VC project files https://git.io/fxANQ
22:56:47 <nielsm> sometimes I wonder if it would be worth adding a check to the CI so it doesn't attempt to build for a platform when nothing affecting that platform has changed
22:57:25 <glx> will be fixed with the windows compile farm ;)
22:57:50 <glx> at least the build won't be useless
22:58:31 <nielsm> and maybe can run some more of the builds in parallel? :/
22:59:07 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 approved pull request #6960: Remove: Visual Studio project files for earlier than 2015 https://git.io/fxAAn
23:16:09 *** Wolf01 is now known as Guest1975
23:19:18 *** andythenorth has left #openttd
23:20:47 <DorpsGek_II> [OpenTTD/OpenTTD] michicc merged pull request #6959: Fix: Fluidsynth should not be added to VC project files https://git.io/fxANi
23:21:21 <nielsm> as expected, CI fails on the other
23:21:31 <nielsm> it doesn't handle deleted files
23:21:50 <nielsm> rejects /dev/null for having no terminating newline
23:36:23 <LordAro> TrueBrain: orudge: my branch is progressed slightly further
continue to next day ⏵