Config
Log for #openttd on 4th May 2020:
Times are UTC Toggle Colours
00:02:18  *** Guest24017 has quit IRC
00:19:15  *** gelignite has quit IRC
01:42:11  *** Foveafoxy has joined #openttd
01:42:45  *** Foveafoxy is now known as Guest24020
01:50:18  *** Guest24020 has quit IRC
02:10:51  *** Wormnest has joined #openttd
02:18:41  *** D-HUND has joined #openttd
02:22:01  *** debdog has quit IRC
02:27:41  *** Wormnest has quit IRC
02:28:01  *** Wormnest has joined #openttd
02:34:04  *** Flygon has joined #openttd
02:42:46  *** Thedarkb-Desktop has quit IRC
03:04:43  *** glx has quit IRC
03:30:19  *** Foveafoxy has joined #openttd
03:30:55  *** Foveafoxy is now known as Guest24027
03:36:20  *** Gustavo6046 is now known as Guest24028
03:36:25  *** Gustavo6046 has joined #openttd
03:36:50  *** Guest24028 has quit IRC
03:38:28  *** Guest24027 has quit IRC
03:41:21  *** Wormnest has quit IRC
04:02:06  *** Foveafox_ has joined #openttd
04:05:14  *** Gustavo6046 has quit IRC
05:45:52  *** andythenorth has joined #openttd
05:59:01  *** gnu_jj has quit IRC
05:59:18  *** gnu_jj has joined #openttd
06:06:36  *** sla_ro|master has joined #openttd
06:26:41  *** andythenorth has quit IRC
06:35:38  *** andythenorth has joined #openttd
06:48:06  *** supermop_Home_ has quit IRC
06:58:01  *** Wolf01 has joined #openttd
07:02:56  *** WormnestAndroid has quit IRC
07:03:09  *** WormnestAndroid has joined #openttd
07:09:05  *** WormnestAndroid has quit IRC
07:09:17  *** WormnestAndroid has joined #openttd
07:23:33  *** Smedles has quit IRC
07:24:49  *** Smedles has joined #openttd
07:51:41  *** cHawk has quit IRC
07:55:00  *** Samu has joined #openttd
07:56:55  <Samu> hi
08:19:30  *** iSoSyS has joined #openttd
08:19:36  <DorpsGek_III> [OpenTTD/nml] matthijskooijman commented on pull request #115: Fix #112: use setuptools_scm to determine version https://git.io/JfsHc
08:20:22  *** iSoSyS has quit IRC
08:27:13  <DorpsGek_III> [OpenTTD/nml] matthijskooijman commented on pull request #115: Fix #112: use setuptools_scm to determine version https://git.io/JfsHi
09:04:06  *** heffer has quit IRC
09:06:35  *** cHawk has joined #openttd
09:15:48  <DorpsGek_III> [OpenTTD/nml] LordAro commented on pull request #115: Fix #112: use setuptools_scm to determine version https://git.io/Jfs7r
09:25:10  <DorpsGek_III> [OpenTTD/OpenTTD] James103 commented on issue #7602: There are 22 known save games that crash OpenTTD (latest master) on load. https://git.io/fjBeO
09:25:11  <DorpsGek_III> [OpenTTD/OpenTTD] James103 closed issue #7602: There are 22 known save games that crash OpenTTD (latest master) on load. https://git.io/fjBeO
09:30:32  <DorpsGek_III> [OpenTTD/OpenTTD] James103 opened issue #8108: Loading old TTD save crashes game (station / oil rig related) https://git.io/Jfs5f
09:31:07  <LordAro> i predict lots of new issues incoming
09:37:21  *** D-HUND has quit IRC
09:38:24  *** debdog has joined #openttd
09:43:05  <andythenorth> we should close some :P
09:50:33  <LordAro> possibly
09:50:39  <Samu> found a last minute bug with the canal pathfinder, just a simple wrong variable in the wrong location
09:50:43  <LordAro> i'd prefer it if someone fixed some :p
09:50:46  <Samu> ez fix
09:52:14  <andythenorth> I don't like growing bug counts, they get overwhelming but eh
09:52:19  *** JacobD88 has joined #openttd
09:52:36  <andythenorth> I read a thing about Adobe setting a limit of 10 open bugs before they have to stop and fix them
09:52:41  <andythenorth> not sure they're actually doing that :P
10:00:04  <Samu> there's a hint that 1.10.2 is coming
10:02:38  <Samu> did you fix that desync
10:02:50  <Samu> great job
10:12:35  *** Yexo has joined #openttd
10:12:35  *** ChanServ sets mode: +o Yexo
10:24:30  <Samu> vector iterators incompatible?
10:27:36  <Samu> i tried to load the savegame posted in #8108 and I get vector iterators incompatible crash
10:28:39  <Samu> https://pastebin.com/raw/D6nuevhj
10:33:30  <Samu> crashed here https://github.com/OpenTTD/OpenTTD/blob/master/src/saveload/afterload.cpp#L2196
10:34:51  <LordAro> Samu: that's... weird
10:35:19  <LordAro> and doesn't really make sense
10:35:23  <LordAro> try a full rebuild?
10:37:49  <Samu> the < symbol says something about road types
10:37:54  <Samu> let me post screenshot
10:38:35  <LordAro> yeah, that makes no sense at all
10:38:36  <Samu> https://i.imgur.com/OChHbJd.png - bool Roadtypelist const iterator
10:38:38  <LordAro> rebuild
10:38:46  <Samu> ok
10:44:46  <Samu> still using RoadTypeLabelList::iterator
10:44:47  <DorpsGek_III> [OpenTTD/nml] matthijskooijman commented on issue #112: setup.py overwrites nml/__version__.py https://git.io/Jf3Lj
10:46:41  <Samu> interesting, it loads in release build
10:47:00  <LordAro> Samu: it'll work in a release build because asserts are disabled
10:47:07  <LordAro> but your error still makes 0 sense
10:47:09  <LordAro> did you clean & rebuild?
10:47:33  <Samu> yes
10:47:43  <Samu> gonna retry debug build
10:48:45  <Samu> Clean Solution, then Rescan Solution, then build
10:54:33  *** JacobD88 has quit IRC
10:55:18  <DorpsGek_III> [OpenTTD/nml] matthijskooijman commented on pull request #115: Fix #112: use setuptools_scm to determine version https://git.io/JfsbS
10:57:22  <Samu> it iterates the first time
10:57:41  <Samu> then iter++ is not called, the line just down there, because there's remove being == true
10:57:52  <Samu> so it then crashes with iterator incompatible
10:58:17  *** iSoSyS has joined #openttd
11:00:06  <Samu> the part where it says if (remove) { DeleteAnimatedTile(*tile); } else { tile++; }
11:00:20  <Samu> that seems to be the cause
11:00:27  <LordAro> interesting
11:00:35  <LordAro> i wonder if that's technically undefined behaviour
11:00:49  <LordAro> modifying a list while iterating over it often has issues
11:03:10  <LordAro> code is pretty ugly too
11:03:32  <LordAro> Samu: so i've no idea where it's getting roadtypelist iterator from, but that does seem to be a problem
11:04:27  <milek7> "Invalidates iterators and references at or after the point of the erase"
11:16:54  *** gelignite has joined #openttd
11:23:23  <_dp_> yeah, also can probably be replaced entirely with std::remove_if and std::unique
11:30:14  *** Samu_ has joined #openttd
11:35:21  *** Samu has quit IRC
11:42:02  <Yexo> Before https://github.com/OpenTTD/OpenTTD/commit/7dd6027194a5fb12ac5fa074c9c7a055a67ad41a that loop was correct. The index was a simple int and deleting meant moving everything after it forwards 1 element, so not changing the index meant it would point to the next entry
11:45:13  <TrueBrain> SOMEONE MADE A BOO-BOO! :D
11:48:51  <andythenorth> wasn't me
11:48:52  <andythenorth> for once
11:50:39  * Wolf01 hides
11:51:28  <andythenorth> you can go outside again Wolf01
11:51:29  <andythenorth> hide there
11:51:37  <andythenorth> [according to our news anyway]
11:52:19  <Wolf01> Aahahahahah, no, really, we can't, they just opened some more activities, still locked in house
11:54:01  <Wolf01> It's the usual "you could but you shouldn't", also a lot of incoherent babbling from local politics contrasting with the govern declarations
11:54:35  <andythenorth> "You can but you shouldn't" is working fine in the UK
11:54:49  <andythenorth> media etc push for more defined rules, but it's actually just working, at least where I am
11:55:17  <TrueBrain> some in the Netherlands. Use your common sense, and you will be fine
11:55:19  <Wolf01> Yes, but we are in Italy, "you can." People stop listening and do whatever they want
11:55:22  <TrueBrain> except for the beaches, it works :P
11:57:54  <Wolf01> So, it seem that tomorrow I'll be in the office
11:58:57  <andythenorth> bug scoring https://insidegovuk.blog.gov.uk/2019/10/25/how-we-prioritise-bug-fixes-on-gov-uk/
11:59:30  <andythenorth> the approach is 40-50 years old, but eh, they've dug it out and blogged about it
12:08:46  <Yexo> Meh, I just got orders to not leave my house for any reason at all (including groceries) for the next 14 days
12:08:59  <Yexo> Germany is pretty strict
12:16:24  <planetmaker> :-O Hey
12:16:31  <planetmaker> Do you meanwhile live in Germany?
12:24:11  <Yexo> For the past 3 years and the next 3.5 week
12:26:40  <planetmaker> he :) Moving in times of this crisis must be "fun"
12:27:15  <Yexo> All planned before the crisis really took off, or I'd have delayed it until things got better
12:27:56  <planetmaker> yeah. But no point in delaying it, if most is arranged, I guess
12:28:00  <Yexo> The home-owner will not be happy that nobody is allowed to visit for the next 2 weeks. Renting out an apartment without being able to show it to people must be hard
12:28:24  <Yexo> Can't really delay now. Had already cancelled home/work, and have a new job/home lined up
12:28:28  <planetmaker> yeah. But then... not sure that showing the appartment would work even without that order. I'd refuse
12:29:21  <Yexo> I don't mind so much. Not really in a risk group myself, so careful / keeping some distance and that should work
12:29:41  <Yexo> But I can fully understand refusing anyway
12:29:59  <planetmaker> well, me neither exactly. But... you might encounter someone in a risk group unknowingly
12:30:32  <planetmaker> or even knowingly... like if you meet your parents for instance
12:31:04  <Yexo> I just did, but since my dad is a GP he's much more likely to contract it via his job than via me
12:32:06  <Yexo> Also: I just travelled to the Netherlands and back. 8h train + flight back is not good risk-avoidance. Then again, I had to be there to get the keys and prepare new house for moving in (painting, laying flooring)
12:32:49  <planetmaker> yeah. Some things are unavoidable
12:34:16  <Samu_> Yexo, I have an update for this library https://bananas.openttd.org/package/ai-library/51555051 which makes it much much faster.
12:34:23  <Samu_> where do I send it
12:34:39  <Yexo> Does anybody use that?
12:34:55  <planetmaker> Samu_, I saw your request yesterday. Working on pushing the repo to github
12:35:11  <Samu_> not exactly, because it's currenlty super slow
12:35:41  <Yexo> Is your updated version better than FibonacciHeap?
12:35:50  <Samu_> yes
12:36:07  <Yexo> In hindsight I don't know why that became a separate library
12:36:36  <Samu_> it was railwai's work pretty much
12:37:37  <Samu_> he says he used a binary search on insert and combined with squirrel's own array insert
12:38:38  <Yexo> Source used to live at https://dev.openttdcoop.org/ which is currently down, but if I understood correctly planetmaker is already working on that and/or moving the code to github
12:38:48  <Yexo> Once that's done you can open a pull request on github
12:38:54  <Yexo> Happy to discuss further there
12:38:56  <planetmaker> yes. I'm working on that
12:39:07  <planetmaker> I used to push (slowly) all the repos to github
12:39:12  <Yexo> Also not sure who's going to manage that github repo
12:39:28  <planetmaker> https://github.com/orgs/openttdcoop/dashboard
12:39:54  <Yexo> That link doesn't work for me
12:40:01  <planetmaker> I'd just put it there... and have people for it. I think someone with reasonable updates would be given access to bananas to update it
12:40:55  <Yexo> Ah, https://github.com/openttdcoop
12:40:59  <Yexo> Sounds good to me
12:41:29  <planetmaker> https://github.com/openttdcoop <-- does that work? ah, yeah
12:44:20  <andythenorth> without Peter, I fail at lunch :(
12:44:30  <andythenorth> this is a sad outcome
12:45:34  <Samu_> Yexo, the update I got here is based on pq3 of railwai's modifications to pq2: https://www.tt-forums.net/viewtopic.php?p=1231144#p1231144
12:46:29  <Samu_> the "native heap" thing is just a bad implementation, it eats memory
12:47:35  <Samu_> the equivalent solution to "native heap" is PR#8091
12:47:45  <Samu_> which is much more sane with memory
12:49:29  <Samu_> but if the pr isn't merged, then the 2nd best bet would be the pq3 updated
12:50:00  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #20: Add: only buffer at most 5 max-sized packets per client https://git.io/JfshR
12:50:06  <TrueBrain> this PR has sufficient code to comment ratio :D ^^
12:51:24  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #20: Add: only buffer at most 5 max-sized packets per client https://git.io/JfshR
12:54:21  *** glx has joined #openttd
12:54:21  *** ChanServ sets mode: +v glx
12:56:38  <DorpsGek_III> [OpenTTD/OpenTTD] Yexo commented on pull request #8091: Add: [Script] Native priority queue; useful e.g. for pathfinders. https://git.io/JfshM
12:59:55  <Yexo> Samu_: Thanks. Let's wait on pm to move the repo to github first, then you can open a PR against that and we can continue the discussion there. In the meantime you can simply include your updated version within your AI and continue development that way
13:05:34  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain commented on pull request #20: Add: only buffer at most 5 max-sized packets per client https://git.io/JfsjT
13:05:44  <TrueBrain> and some more background, for those people that enjoy that :D
13:08:41  <FLHerne> For the hypothetical native pathfinder, instead of a ton of squirrel callbacks perhaps it could just take a big list of customizable weights?
13:09:20  <Yexo> Once you do that, all AIs will be limited to exactly that list
13:09:37  <Yexo> No way to innovate, since a squirrel pathfinder will be much slower by comparison, hence not a real option
13:10:08  <FLHerne> It wouldn't be any less of an option than it is now
13:10:19  <FLHerne> Except for competitive AI benchmarking
13:10:44  *** cHawk has quit IRC
13:10:56  <Yexo> That alone is a big problem. Not only competitive benchmarking, but any comparison before/after is much harder
13:10:59  <FLHerne> In which case innovative pathfinding is irrelevant, because the answer is to spam airports :P
13:11:43  <FLHerne> Anyway, AIs could do things like get a proposed path from the builtin and then use that as a basis for detail tweaking
13:11:57  <Yexo> And once there is such a native pathfinder, the squirrel version will not be updated (because nobody uses it), hence it might not have the same list of features, so by wanting to test somethign new you first have to backport all other changes
13:12:35  <Yexo> Coming from the other way: is pathfinding speed such a big problem?
13:12:39  <LordAro> it's worth perhaps reminding people that AI speed isn't really a primary goal
13:12:49  <FLHerne> Yes, IMO
13:12:59  <LordAro> how "good" they are to play against is probably more of a priority
13:13:23  <LordAro> for most usecases, anyway
13:13:31  <Yexo> FLHerne: can you explain why?
13:13:45  <Yexo> Are AIs in general too slow in building when competing against a human?
13:13:49  <FLHerne> It can take them in-game months to design routes, and that takes up computation time they could use to do interesting stuff
13:13:53  <Yexo> Or other obvious related issues?
13:14:11  <FLHerne> And yes, for competition vs humans it's definitely too slow
13:14:51  <FLHerne> Also, you get glitches where it's so slow an industry appeared while the route was being planned
13:14:59  <FLHerne> And then it's in the way
13:15:04  <Yexo> I don't remember if the number of squirrel opcodes/tick was configurable, but does doubling that help?
13:15:22  <FLHerne> (which are probably avoidable by AI authors, but aren't)
13:15:32  <Yexo> Is that for rails only or also for road?
13:16:03  <Yexo> You'll always get changes over your planned route. Planning/building fast just makes the less likely, it doesn't avoid them
13:16:16  <Yexo> And those are relatively hard to work around
13:16:36  <Yexo> So being less likely would be an improvement, but providing example code / showing how to work around those issues is better
13:20:44  <DorpsGek_III> [OpenTTD/bananas-server] Xaroth approved pull request #20: Add: only buffer at most 5 max-sized packets per client https://git.io/Jfsjw
13:21:23  <Samu_> there's also terron priority queue, the comments say it's based on pq2 as well, but then I tested it and the resulting order of the queue is not equal of that of pq2
13:21:34  <Samu_> sometimes it would be faster, sometimes not
13:22:18  <Samu_> it handles ties in a "cheating" manner :p
13:27:29  <Samu_> gonna post on the forum, while at it
13:30:57  <DorpsGek_III> [OpenTTD/nml] matthijskooijman commented on issue #112: setup.py overwrites nml/__version__.py https://git.io/Jf3Lj
13:35:00  <planetmaker> ticks was configurable, yes
13:42:30  <Samu_> Yexo, just posted what I wanted to give u here: https://www.tt-forums.net/viewtopic.php?p=1231631#p1231631 if you wanna test its speed
13:45:06  <Yexo> Samu: Might need to be careful with the license
13:45:31  <Yexo> Original libs are either gpl v2 or gpl v2+ (can't remember). railwAI seems to be gpl v3
13:45:45  <Yexo> Can use those changes and publish the as gpl v2 or gpl v2+
13:46:09  <Yexo> haven't checked the contents of your update, nor do I have time to test now. Maybe later (this week)
13:47:22  <Samu_> oh :(
13:48:13  <Samu_> pq3 is gpl v2
13:48:21  <Samu_> ralwAI is gpl v3?
13:48:28  <Samu_> t.t
13:48:30  <Yexo> https://bananas.openttd.org/package/ai lists it as gpl v3
13:48:36  *** nielsm has joined #openttd
13:48:42  <Yexo> Which means he can't use pq3 if that's v2 (not v2+)
13:49:33  <Samu_> well he used , I also used it
13:49:41  <Samu_> i don't think ppl care much about licenses
13:51:28  <LordAro> until they do
13:53:00  <Yexo> I think you're right a lot of people don't care, and all is fine as long as people are around to fix stuff
13:53:34  <Yexo> There have been several NewGRF's with unclear licensing that have been "lost" over time because the author disappeared and nobody was allowed to fix/update them
13:54:10  <Yexo> In contrast the AI libs for example have an open license for a good reason: we wanted them maintainable even if the original authors were not around
13:54:37  <Yexo> By mixing gpl v2 and gpl v3 code you have essentially an undistributable combination
14:06:24  *** Samu_ has quit IRC
14:06:39  *** Samu has joined #openttd
14:29:57  *** Gustavo6046 has joined #openttd
14:30:55  <Samu> 52776169-RailwAI-25.tar - this came like this from bananas :|
14:31:01  <Samu> what's the big number for
14:31:37  <LordAro> Samu: it's a hash to ensure uniqueness
14:32:02  <LordAro> because packages can theoretically be named identically now
14:32:27  <Samu> :(
14:32:38  <Samu> there goes my perfectly ordered ais
14:33:08  <LordAro> you should stop looking inside the content_download folder
14:33:18  <LordAro> though i'm not sure why it's a prefix rather than a suffix
14:33:20  <LordAro> TrueBrain: ^ ?
14:33:53  <TrueBrain> LordAro: because .. who cares? :D
14:34:09  <LordAro> well, Samu
14:34:11  <LordAro> evidentally
14:34:19  <LordAro> evidently?
14:34:22  <LordAro> spelling.
14:34:41  <TrueBrain> ugh, English .. don't talk me about English
14:35:07  <Samu> i knew i downloaded railwai v25, but couldn't find it, turns out it was at the top of the list
14:35:14  <Samu> with an ugly number
14:35:27  <TrueBrain> please file your complain in /dev/null, thank you
14:37:15  <TrueBrain> LordAro: so to answer your question, because from a code-flow perspective that made the most sense. If there is a real issue with it, let me know :) It can also be put in the middle btw, not sure why you went binary on pre vs suf :P :P
14:37:35  <LordAro> :p
14:38:05  <milek7> it would keep alphabetical ordering of content, instead of random shuffle
14:39:46  <TrueBrain> it is very alphabetically ordered; in fact, it is more stable now than it was
14:39:51  <TrueBrain> so that is an argument to have it as a prefix :)
14:54:20  *** andythenorth has quit IRC
15:01:46  *** andythenorth has joined #openttd
15:02:06  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #21: Add: reload the database in a separate process https://git.io/JfGkR
15:02:21  <TrueBrain> you got to love Python :)
15:03:40  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #21: Add: reload the database in a separate process https://git.io/JfGkR
15:03:42  <TrueBrain> why ALWAYS after push, I find a silly mistake I have to fix :P
15:05:58  <TrueBrain> okay, these two fixes should heavily improve the content-server :D \o/
15:11:46  <Samu> all these links in the forums no longer work! https://binaries.openttd.org/bananas/gamescript/City_Founder_GS-3.tar.gz
15:20:08  <Wolf01> Looks like we need a permalink
15:24:15  <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick commented on issue #8108: Loading old TTD save crashes game (station / oil rig related) https://git.io/Jfs5f
15:24:20  <TrueBrain> https://github.com/OpenTTD/bananas-api/issues/29
15:24:29  <TrueBrain> Contributions are welcome, ofc :D
15:27:07  <Samu> the crash about the oilrig, however, is because st->airport.type == AT_OILRIG is false
15:27:29  <Samu> the airport.type is AT_METROPOLITAN
15:27:51  <glx> I'm trying to understand nml source for townnames
15:30:38  <Yexo> Samu: That type didn't exist in TTD. If you want to fix it, add an extra check in Afterload to correct the .type var for oilrigs
15:39:52  *** Wormnest has joined #openttd
16:01:36  *** Smedles has quit IRC
16:13:55  <Samu> Very old savegames sometimes have phantom oil rigs
16:14:02  <Samu> seems to be the case
16:15:29  <LordAro> that is what the comment says, yes
16:24:06  <Samu> well, it no longer crashes due to oilrig being a metropolitan
16:24:20  <Samu> but now it crashes due to iterator crap
16:24:32  <Samu> can't fix one without the other
16:24:54  <Samu> so, i'm not sure i'm making a pr
16:25:55  *** Smedles has joined #openttd
16:27:40  <Samu> vector iterators incompatible, I don't know how to fix this
16:28:40  *** Flygon has quit IRC
16:34:50  <nielsm> if you disable an assertion because it was failing, and code afterwards is now causing errors, it means the assertion was correct
16:35:30  <nielsm> asserts are not there to annoy you, they are there to catch programming errors before something more serious happens
16:36:05  <Samu> that part is fixed, the assert no longer asserts, the oilrig is now an oilrig
16:36:38  <Samu> the assert is enabled
16:36:43  <nielsm> "iterators incompatible" might mean that the iterators point into different containers
16:37:23  <Yexo> Samu: one way to fix the iterators would be to use an int as index instead of using iterators. Then the "delete item at current index or increase index" code would work
16:37:29  *** frosch123 has joined #openttd
16:37:37  <Yexo> There are probably cleaner ways, but those will require somewhat larger changes
16:39:29  <milek7> if (remove) tile = _animated_tiles.erase(tile);
16:39:30  *** Progman has joined #openttd
16:39:56  <milek7> (might also need MarkTileDirtyByTile)
16:41:34  <Samu> the erase is done somewhere else
16:42:01  <Samu> DeleteAnimatedTile(*tile); ibnside this
16:43:12  <Samu> line 31 animated_tile.cpp
16:44:14  <milek7> yes, but you need that iterator returned from erase
16:44:25  <milek7> it's pointing to next element
16:45:39  <Samu> ah, I see, gonna experiment copy paste code
16:45:39  <TrueBrain> frosch123: CloudFlare said no. Their terms are a bit more strict then what I read on the forums :D
16:45:55  <TrueBrain> guess we need to find a better way to host binaries :D
16:46:37  <milek7> doesn't cloudflare have free plan?
16:46:49  <TrueBrain> it is mainly a few files btw .. like 80% of the bill is generated by 5 BaNaNaS entries :)
16:47:50  <Samu> is DeleteAnimatedTile called from some other places?
16:48:04  <Samu> wondering if I can just remove it
16:48:19  <milek7> it is
16:48:29  <Samu> aha, it loaded, thx milek7
16:49:13  <frosch123> TrueBrain: so back to vpn?
16:49:24  <TrueBrain> or backblaze
16:49:27  <TrueBrain> 10 times cheaper than AWS
16:50:07  <TrueBrain> or offloading those few files to a dedicated VPS indeed
16:50:29  <milek7> btw backblaze have free bandwidth when serving through cloudflare
16:54:48  <Samu> interesting find
16:55:02  <_dp_> wasn't company id issue fixed? why is ResolveCompanyID still returning ScriptCompany::CompanyID ?
16:55:03  <Samu> TTD doesn't build houses on half-land, half-water tiles
16:55:33  <TrueBrain> current AWS estimate is ~460 dollar .. backblaze would be 40 dollar (and I took only the BaNaNaS CDN in these numbers), per month. I guess 40 dollar a month is reasonable, and what I prefer, against setting up our own VPS and "maintaining" it frosch123 :) Good thing is, today backblaze released that they now have an S3 compatible API, so it should be easy to migrate :P
16:57:03  <_dp_> and all those double type conversions... what a shitshow :(
17:00:57  <frosch123> TrueBrain: what would we need for the wiki? i remembered that there are git-based wikis. but no idea how mature they are
17:01:13  <TrueBrain> good question; haven't really looked into it, tbh
17:02:19  <TrueBrain> if we keep using mediawiki, there are 3 parts: authentication, content, and file storage
17:02:26  <TrueBrain> authentication we already solved
17:02:30  <TrueBrain> content .. is some kind of database
17:02:42  <TrueBrain> file storage .. on AWS can be EFS or S3, I think
17:07:30  <frosch123> https://github.com/gollum/gollum <- that appears popular
17:07:57  <andythenorth> frosch123 that's interesting
17:08:05  <frosch123> the editing is somewhat ambiguous. the description talks about local editing, but the screenshots show web editing
17:08:27  <andythenorth> I hadn't considered alternatives to media wiki
17:08:37  * andythenorth assumed there was only one option
17:08:43  <frosch123> it claims to be able to display mediawiki markup
17:08:51  <frosch123> but for sure there are issues :)
17:10:56  <Samu> Crash log (after demolishing bridge at 0x299A): crash_floating_train.zip i can't reproduce this crash
17:11:10  <Samu> james103 messed up or is it me?
17:13:24  <frosch123> most other git-based wiki are markdown-only
17:14:39  <TrueBrain> nothing wrong with markdown :)
17:16:16  <TrueBrain> honestly, I am surprised these kind of wikis exist :)
17:16:25  <TrueBrain> they would solve 2) and 3) I guess, with GitLFS
17:16:32  <frosch123> have you ever seen a table written in markdown?
17:17:11  <frosch123> if you want a pure markdown-based wiki, you can use github
17:17:11  <TrueBrain> yes ... that part is laughable :D
17:17:13  <frosch123> 's own wiki
17:17:17  <TrueBrain> wasn't it as bad in mediawiki? :P
17:17:41  <frosch123> mediawiki can do rowspan and cellspan, and often has plugins for js-sortmagic
17:18:18  <TrueBrain> so more attention to tables; nice :)
17:18:45  <TrueBrain> bit unclear how editing in gollum works, as in: does it commit?
17:19:11  <frosch123> looks like the second place is jingo (Jingo is not Gollum)
17:19:30  <frosch123> but it definitely needs conversion of the content to new markup
17:19:45  <frosch123> but it mentions github-auth as standard
17:19:50  <DorpsGek_III> [OpenTTD/nml] glx22 opened pull request #117: Fix #116: towname used bits could be the same for sub-parts https://git.io/JfGYB
17:20:05  <TrueBrain> its funny, gollum-demo is 10 years old :P No commit since .. that is not useful :D
17:20:24  <frosch123> there should be something on dockerhub
17:21:00  <TrueBrain> "Automatically push to a remote (optionally)" <- Jingo is explicit about it :P
17:21:56  <frosch123> hmm, jingo is only gh-flavoured markdown, so no tables?
17:22:08  <frosch123> next up is dokuwiki
17:22:57  <andythenorth> mixed html and md? :P
17:23:30  <andythenorth> I find all markdown / rst / bbcode type languages harder than html :)
17:23:35  <andythenorth> I am not the intended audience
17:24:09  <glx> markdown is nice for simple formatting
17:25:09  <TrueBrain> okay, gollum only does commits, no pushing, it seems
17:25:12  <TrueBrain> but that is easily fixed :P
17:25:40  <andythenorth> is this for nml, or openttd, or general wiking? o_O
17:26:02  <frosch123> openttd first
17:26:24  <Samu> im not sure which commit is the blame
17:26:29  <frosch123> we want to get rid of openttd-auth and servers :)
17:26:34  *** tokai has joined #openttd
17:26:34  *** ChanServ sets mode: +v tokai
17:28:04  <Samu> im currently looking for which commit causes the iterator incompatible thing, is it this? https://github.com/OpenTTD/OpenTTD/commit/ab711e6942757d775c08c31a6c32d488feba1dba
17:28:27  <TrueBrain> ah, you can add hooks in gollum
17:28:33  <TrueBrain> it is a nice little wiki, tbh
17:28:42  <TrueBrain> seems we can do the same trick as with bananas-server, so it can scale out nicely
17:28:57  <TrueBrain> reminds me, I have 2 PRs that needs reviewing :P
17:29:22  <frosch123> i can also create 2 PRs :)
17:29:27  <Samu> oh, sorry, this commit is huge, let me get a smarter link
17:29:29  <frosch123> shall we trade?
17:29:43  <TrueBrain> frosch123: deal :)
17:30:24  <TrueBrain> we can always just make a "wiki" in Jekyll :D
17:30:30  <TrueBrain> we already have everything for that in place :P
17:30:31  <frosch123> anyway, when i slept on nml town names, i concluded it's really hard to fix. yet glx made a PR within one day :)
17:30:37  <Samu> https://github.com/OpenTTD/OpenTTD/commit/ab711e6942757d775c08c31a6c32d488feba1dba#diff-5d7621e6bc0894916f1492a729fc7ec9
17:31:38  <glx> frosch123: it's not optimal as sometimes overlap are ok but that's quite impossible to check in current implementation :)
17:33:21  *** tokai|noir has quit IRC
17:33:42  <TrueBrain> ironicly, gollum uses GitHub Wiki :D
17:34:18  <glx> but yeah townname implementation is a hell, most of the debug_print() are pure non working garbage, but I won't fix that ;)
17:38:43  <Samu> Need help with commit message now
17:38:46  <Samu> Fix #8108, ab711e6: Convert TTD oil rigs encoding, and fix incompatible iterator crash
17:38:56  <Samu> heh, terribad, I know
17:39:56  <DorpsGek_III> [OpenTTD/bananas-server] frosch123 approved pull request #20: Add: only buffer at most 5 max-sized packets per client https://git.io/JfGOI
17:41:05  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain closed issue #15: Raise exception if connection is gone https://git.io/JfqFH
17:41:05  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #20: Add: only buffer at most 5 max-sized packets per client https://git.io/JfshR
17:41:09  <TrueBrain> I hope you liked my documentation frosch123  :)
17:41:20  <frosch123> very unusual :)
17:41:33  <TrueBrain> it was needed :P
17:42:23  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #21: Add: reload the database in a separate process https://git.io/JfGkR
17:42:25  <TrueBrain> (just a rebase ^^)
17:42:47  <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick opened pull request #8109: Fix #8108, ab711e6: Convert TTD oil rigs encoding, and fix incompatible iterator crash https://git.io/JfGOm
17:43:40  <frosch123> the other PR is funny, only a few lines changed, but whole methods are moved into different classes :)
17:44:05  <TrueBrain> yeah .. minimizing what cpickle had to do :)
17:44:12  <TrueBrain> it was .. a challenge
17:45:26  <frosch123> oh dear, sometimes black formatting is a challenge for me :)
17:45:49  <DorpsGek_III> [OpenTTD/OpenTTD] DorpsGek pushed 1 commits to master https://git.io/JfGO3
17:45:49  <DorpsGek_III>   - Update: Translations from eints (by translators)
17:45:57  <TrueBrain> Honestly, I don't see it anymore :P I just run it, and say: what-ever
17:46:04  <TrueBrain> but it was funny what it did with the return tuple :)
17:46:04  <frosch123> wrapping the left side of an asignment over multiple rows broke my parser
17:46:39  <frosch123> haha, at least you tripped over the same place :)
17:46:49  <TrueBrain> it stood out, for sure :)
17:48:15  <frosch123> the api never talks to the server, right?
17:48:34  <frosch123> so there are no race conditions between server/api?
17:49:11  <frosch123> can the api spam the server with reloads?
17:49:13  <TrueBrain> all communication goes via GitHub
17:49:33  <TrueBrain> and the reload is guarded, only 1 reload can be active at the same time
17:49:44  <frosch123> usually the api pushes multiple commits at once, does the server then get 1 reload, or many?
17:50:00  <TrueBrain> GitHub Actions listen to a push
17:50:01  <TrueBrain> not commits
17:50:21  <TrueBrain> as can been seen here: https://github.com/OpenTTD/BaNaNaS/actions :)
17:50:38  <TrueBrain> or easier: https://github.com/OpenTTD/BaNaNaS/commits/master
17:50:43  <TrueBrain> not all commits have a green flag :D
17:51:24  <frosch123> ah, i always wondered what the flags meant, they have no tooltip
17:51:40  <TrueBrain> click on it :)
17:52:00  <DorpsGek_III> [OpenTTD/bananas-server] frosch123 approved pull request #21: Add: reload the database in a separate process https://git.io/JfGO8
17:52:47  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #21: Add: reload the database in a separate process https://git.io/JfGkR
17:52:52  <TrueBrain> ty kind
17:53:04  <TrueBrain> after testing on staging and deploying on production, I will merge yours :D
17:53:35  <DorpsGek_III> [OpenTTD/OpenTTD] Yexo commented on pull request #8109: Fix #8108, ab711e6: Convert TTD oil rigs encoding, and fix incompatible iterator crash https://git.io/JfGOR
17:57:04  <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain created new tag: 1.0.8 https://git.io/JfGOz
17:57:43  <TrueBrain> 2 minutes of silence in the Netherlands in a few, to remember freedom .. which is a weird thing with this Corona stuff ..
17:58:04  <TrueBrain> normally you notice the silence ..
18:02:30  <DorpsGek_III> [OpenTTD/nml] frosch123 commented on pull request #117: Fix #116: towname used bits could be the same for sub-parts https://git.io/JfGOi
18:03:43  <TrueBrain> owh boy ... it runs on local, it runs on staging .. but not on production .. oh-oh
18:05:22  <TrueBrain> how is this possible .. lol
18:05:42  <TrueBrain> well, lets roll back for now :)
18:09:04  *** WormnestAndroid has quit IRC
18:09:45  *** WormnestAndroid has joined #openttd
18:10:10  <DorpsGek_III> [OpenTTD/OpenTTD] Yexo commented on pull request #8109: Fix #8108, ab711e6: Convert TTD oil rigs encoding, and fix incompatible iterator crash https://git.io/JfG3v
18:10:40  <TrueBrain> ah .. I understand .. hmm ..
18:21:13  <DorpsGek_III> [OpenTTD/OpenTTD] ldpl commented on pull request #8109: Fix #8108, ab711e6: Convert TTD oil rigs encoding, and fix incompatible iterator crash https://git.io/JfG3w
18:22:27  <TrueBrain> there, fixed :)
18:22:30  <TrueBrain> it works frosch123  :D
18:22:44  <frosch123> no commit needed, just magic :)
18:22:44  <TrueBrain> the API does still stall, it needs a similar solution
18:22:49  <TrueBrain> but that priority is .. just a lot lower
18:22:56  <TrueBrain> yeah, it was a configuration error on AWS :P
18:23:04  <TrueBrain> if staging passes and production does not, it rarely is a coding issue :P
18:23:40  <frosch123> oh, i have for sure seen "if (production)" or "if (staging)"
18:24:12  <TrueBrain> yeah .. I have the same .. I removed the "if production" line :P Fixed it :D
18:24:27  <TrueBrain> okay .. this should also fix the OOM events .. we will see tomorrow :D
18:28:42  <frosch123> the forums start to throttle me sending pms :p
18:28:49  <TrueBrain> hahaha, really?
18:28:50  <TrueBrain> lol
18:38:56  <TrueBrain> hmm .. it seems you cannot put a custom domain on backblaze .. well, yes, via cloudflare, but that is a no :D
18:41:26  *** Wormnest has quit IRC
18:43:11  <andythenorth> 'cloud they said'
18:43:41  <TrueBrain> still cannot believe GB prices are this high :P
18:43:52  <TrueBrain> I remember when 1 mbit/s was 15 euro a month
18:44:04  <DorpsGek_III> [OpenTTD/OpenTTD] SamuXarick commented on pull request #8109: Fix #8108, ab711e6: Convert TTD oil rigs encoding, and fix incompatible iterator crash https://git.io/JfGsE
18:44:04  <TrueBrain> @calc 1 * 3600 * 24 * 30 / 10 / 1024
18:44:04  <DorpsGek> TrueBrain: 253.125
18:44:18  <TrueBrain> well, with 95%, so .. ~300GB per month
18:44:45  <TrueBrain> so 20 euro per GB .. lol .. those were the days :P
18:45:01  <TrueBrain> guess 0.09 euro per GB is nothing, compared to that :D
18:47:13  <andythenorth> GB = coal :|
18:47:16  <andythenorth> but coal is cheap now!
18:48:50  <frosch123> what crisis does it need for negative GB prices?
18:48:54  <TrueBrain> yet .. it is very expensive, 0.09 euro per GB :P
18:49:24  <andythenorth> €0.10 per zbase download
18:49:38  <andythenorth> 'or run our coin miner'
18:49:41  <TrueBrain> a bit less, but yes
18:49:52  <TrueBrain> so that puts it a bit in perspective :)
18:50:00  <TrueBrain> I like that andythenorth  .. 10 eurocent per download for zBase :)
18:50:07  <TrueBrain> (and abase .. and there are a few more :P)
18:50:13  <andythenorth> buy a monthly pass
18:50:28  <andythenorth> like all the 'free to play' games with €0.99 premium pass
18:50:39  * andythenorth isn't serious
18:50:43  <TrueBrain> we could start by asking for donations, instead of them accidentily happening :P
18:50:55</