IRC logs for #openttd on OFTC at 2020-10-18
⏴ go to previous day
00:00:43 *** tokai|noir has joined #openttd
00:00:43 *** ChanServ sets mode: +v tokai|noir
00:14:36 *** Xunie is now known as Guest3696
00:34:34 <rptr_> is there some overview of all players'/ais' numbers? like num stations, trains, etc, in graph/ overview form?
00:35:06 <Eddi|zuHause> not that i know of
00:35:29 <rptr_> would that be considered not appropriate? i think i read that things should not be added that were not in the original TTD game?
00:35:56 <Eddi|zuHause> that does not sound like a useful rule
00:36:07 <rptr_> i'm sure i read it somewhere!
00:36:54 <Eddi|zuHause> the only real rule i know of is "no war stuff"
00:37:13 <glx> haha there are many things in openttd that were not in original TTD
00:37:20 <rptr_> hm. well i haven't played ttd so i don't know how many of the things in menus etc were not in original TTD
00:37:36 <glx> more visible one being newgrf ;)
00:37:37 <rptr_> maybe that i read was someone's opinion on the forum/bug tracker
00:37:54 <rptr_> i guess i was talking about fundamental stuff like user interface, adding a new menu/window
00:38:03 <rptr_> i guss cargo flow is new and not in ttd
00:38:09 <Eddi|zuHause> we already have a graph menu
00:38:44 <Eddi|zuHause> while you're there, the graphs we have are in serious need of an overhaul :)
00:39:48 <rptr_> maybe you don't find the player overview idea (i am comparing AIs right now) useful. but i had something in mind like graphs for trains/rail pieces/planes or just a table of numbers
00:40:00 <rptr_> sort of like if you click a company and it tells you the number of rail pieces etc, but all in one page
00:40:37 <glx> company windows shows the vehicle numbers IIRC
00:43:13 <rptr_> yes, but not all companies' numbers in one window
00:44:17 <rptr_> Eddi|zuHause, what needs to be fixed with the graphs?
00:48:40 <Eddi|zuHause> rptr_: i don't really remember right now. i think people complained the most about the payment graph
00:49:01 <rptr_> ah yeah. it was a bit confusing to me at first
00:50:29 <Eddi|zuHause> what i was missing in particular was some kind of histogram what actual points of the payment graph my network currently delivers
00:51:26 <rptr_> i usually look at the timetable of a train and check how long it travelled, but not very easy i guess
00:52:24 <rptr_> might be good to have it on the trains themselves as well
00:53:20 <rptr_> hm that sounds really cool, what you said
00:54:42 <rptr_> so like, overlaying the current graph over a histogram? so you could see the volume of each transit-time for e.g. coal
00:54:55 <rptr_> i guess overlaying isn't necessary
00:56:29 <Eddi|zuHause> the graph has these rectangle nodes, they could be made bigger depending on the volume transported
00:58:03 <Eddi|zuHause> another thing that might be useful with the graph is to set the axis to a time, speed, or distance
01:01:16 <rptr_> it would have to take into account the timespan though
01:01:19 <Eddi|zuHause> that might work, but how are you handling multiple selections?
01:01:31 <rptr_> i am not, that's just GIMPed
01:01:41 <rptr_> or you mean hypothetically? not sure
01:02:01 <rptr_> could just draw them in order so they overshadow each other, or accumulate
01:02:05 <rptr_> so you'd get a rainbow box
01:02:09 <rptr_> grey, red, white, all together
01:02:16 <rptr_> (that's not really a rainbow :p)
01:02:44 <Eddi|zuHause> i'm not 100% convinced
01:03:00 <rptr_> might as well make it a new window then
01:04:25 <Eddi|zuHause> that's not a working link
01:04:57 <rptr_> doesn't work even when i allow it in noscript
01:05:09 <rptr_> so that is an "addition" histogram if that makes sense
01:06:17 <Eddi|zuHause> yeah, i knew what you meant.
03:32:43 *** WormnestAndroid has joined #openttd
03:55:48 *** WormnestAndroid has quit IRC
03:56:13 *** WormnestAndroid has joined #openttd
04:19:06 *** Wormnest has joined #openttd
04:20:44 *** WormnestAndroid has quit IRC
04:20:47 *** WormnestAndroid has joined #openttd
05:09:51 *** ChanServ sets mode: +v tokai
05:42:20 *** heffer_ has joined #openttd
07:47:11 <rptr_> is there a configure option to enable debugging symbols?
08:07:20 *** Flygon_ has joined #openttd
08:10:51 *** andythenorth has joined #openttd
09:10:39 *** sla_ro|master has joined #openttd
09:22:53 *** supermop_Home_ has quit IRC
09:30:24 *** longtomjr_ has joined #openttd
09:31:38 *** andythenorth has joined #openttd
09:44:29 *** Progman has joined #openttd
09:45:15 <rptr_> andythenorth, huh? they made a model based on your train?
09:45:25 <andythenorth> IRL no, but convergent evolution
09:45:45 <andythenorth> they did with plastic what I did with pixels
09:46:07 <andythenorth> mine is better TBH :)
09:46:32 <rptr_> i just wish there was a nice friendly server with nice newgrfs ._.
09:47:15 <rptr_> andythenorth, do you draw every perspective from scratch, or render a 3d model?
09:47:46 <rptr_> oh! there's a HK/MTR set. that brings back memories
09:47:52 <andythenorth> mostly actually from copy-paste
09:48:03 <rptr_> not sure if i actually recognise the trains. just the logo
09:48:13 <rptr_> getting an MTR pass took like 2 months
09:48:31 <rptr_> andythenorth, i guess if i did one i'd take an existing and draw over
10:22:18 <TrueBrain> I do not understand why a [[Category:]] is a link, and when a marker of: add this to this category .. mediawiki is really really weird and inconsistent
10:24:06 <andythenorth> well we know _why_ but we don't know _why_
10:24:15 <andythenorth> stuff is made by people
10:28:19 <rptr_> LordAro, re. call stack being printed after catching an exception. i backported 1 LOC from latest squirrel and changed 1 setting in Squirrel.cpp. would that be worth a pull request? i really think if exceptions work properly that would be great
10:30:45 <rptr_> perhaps an AIError.PrintCallStack() would be useful
10:32:25 * rptr_ can run ChooChoo without stack traces \o/
10:42:52 <LordAro> rptr_: i think that would be worth a PR
10:43:25 <LordAro> (certainly the former, and *maybe* the latter)
10:44:05 <rptr_> i am so happy. now i can use exceptions. :p
10:48:58 <andythenorth> so when is daylength?
10:49:16 <rptr_> andythenorth, adjustable daylength?
10:49:42 <rptr_> i don't know what you mean
10:50:36 <andythenorth> new generations of vehicles arrive about too quickly
10:50:46 <andythenorth> ideally they would be about 1.5x slower in my games
10:51:24 <andythenorth> I could just change the newgrfs, but then dates look a bit weird
10:54:06 <rptr_> can't you modify when vehicles arrive? i thought that already existed
11:00:34 <andythenorth> I could just change the dates
11:01:10 <andythenorth> current solution is just to use date cheat and set the year back occasionally
11:05:21 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTWrb
11:13:59 <DorpsGek_III> [OpenTTD/OpenTTD] rptr commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTWoi
11:14:52 <andythenorth> who invented a 'freeze time' mode?
11:17:22 <Eddi|zuHause> that sounds unlikely
11:18:22 <Eddi|zuHause> i would think either Wolf01 or alberth
11:18:25 <TrueBrain> wtf is "freeze time"?
11:19:16 <Eddi|zuHause> TrueBrain: one where the game progresses but the date doesn't advance
11:19:16 <andythenorth> some kind of 'stop the date'
11:19:35 <TrueBrain> what is the obsession with people of stopping time? :P
11:19:36 <Eddi|zuHause> transport fever 2 has that
11:20:15 <Eddi|zuHause> TrueBrain: mostly people who want to play with steam trains instead of maglev
11:20:28 <TrueBrain> so buy a slower CPU :P
11:20:32 <TrueBrain> (I am trolling, btw)
11:20:39 <andythenorth> I tried searching logs but eh
11:20:55 <Eddi|zuHause> TrueBrain: my CPU is probably slower than yours :p
11:21:43 <TrueBrain> ugh @ html library of Python
11:21:52 <TrueBrain> it seems to be fixing "&SF" into "&SF;"
11:21:56 <TrueBrain> but it should be "&SF"
11:22:30 <Eddi|zuHause> sounds like a job for a regex :p
11:22:56 <TrueBrain> it is not wrong perse to do this .. there are a few codes that don't need the ;
11:23:01 <TrueBrain> HTML is annoying like that
11:23:42 <Eddi|zuHause> yeah, i got a bit confused about that when i hand-coded html like 20 years ago
11:31:25 <TrueBrain> I hate libraries that do all kind of things "extra" for you
11:31:30 <TrueBrain> the Python "html" lib is really one of them :P
11:36:30 <DorpsGek_III> [OpenTTD/OpenTTD] rptr commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTW6o
11:36:42 <Eddi|zuHause> you can hopefully configure some of those "extra" things
11:37:14 <TrueBrain> it uses regex to figure out the HTML, and it doesn't give enough information to me to reconstruct the original input
11:37:29 <TrueBrain> so I get "sf", with input "&sf;" and "&sf"
11:37:30 <DorpsGek_III> [OpenTTD/OpenTTD] rptr commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTW6M
11:37:57 <rptr_> wait, is it an actual problem for you that steam trains are phased out?
11:38:19 <rptr_> i prefer them too, but i'm sure i've played on always-steam servers
11:38:20 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTW69
11:39:16 <Eddi|zuHause> rptr_: there is a setting to not phase them out. but then you still have the new trains that are obviously superior
11:39:32 <TrueBrain> and not playing with them is SO DIFFICULT or something :P
11:40:14 <DorpsGek_III> [OpenTTD/OpenTTD] rptr commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTW6p
11:40:26 <Eddi|zuHause> it breaks the immersion :p
11:41:19 <DorpsGek_III> [OpenTTD/OpenTTD] rptr commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTWiJ
11:41:29 <rptr_> sorry for spamming git :P
11:41:43 <rptr_> can't you remove trains with newgrf or something?
11:42:51 <andythenorth> I can control all the trains I want
11:43:21 <Eddi|zuHause> sure, you can make a newgrf that fits one single person's needs
11:43:26 <andythenorth> the specific issue is that upgrades come in waves, it's kind of the only way to do it
11:43:37 <Eddi|zuHause> but the next person would need a slightly different newgrf
11:44:17 <rptr_> so, you want new steam trains to come out, but history should end at a certain point? :p
11:44:26 <Eddi|zuHause> and that doesn't scale well when the number of people goes up
11:45:44 <andythenorth> I find that by the time I've replaced all the trains of one generation
11:45:47 <andythenorth> and tuned the network
11:45:52 <andythenorth> it's time for the next generation
11:46:03 <andythenorth> and unlike building, you can't replace trains in pause mode
11:46:55 <andythenorth> I could just adjust the intro dates, but I'd have to do it for all the vehicle newgrfs I use
11:47:08 <andythenorth> also realism reasons :P
11:47:08 <rptr_> yeah that is annoying. and i never upgrade from electric to mono or mono to maglev. yurk
11:47:17 <rptr_> probably more realistic to not upgrade anyway
11:47:32 <andythenorth> I have to use ffwd a lot when replacing trains
11:47:38 <andythenorth> to get them into depot
11:47:44 <andythenorth> so the game goes by quite fast
11:47:51 <rptr_> what was that game someone mentioned above, some other transport game...
11:48:05 <TrueBrain> that moment you find a random "LEFT-TO-RIGHT MARK" in the input .. wutttttt
11:48:27 <Eddi|zuHause> but that brings us back to the original problem of all daylength patches: you need to dive into the code and decouple simulation time progression from date progression
11:48:44 <rptr_> can you change dates in gamescript?
11:49:52 <Eddi|zuHause> rptr_: transport fever 2
11:52:19 <rptr_> is Cities: Skyline/-s any good?
12:01:24 <Eddi|zuHause> if you have lots of RAM...
12:06:07 <Wolf01> <Eddi|zuHause> i would think either Wolf01 or alberth <- one of us probably, my long term memory doesn't work well anymore, I can't load memories and I'm paging constantly, working slowly as my brain is still electrochemical... I might need a soft reset
12:09:40 <Eddi|zuHause> they say covid longterm effects are some kind of brain-fogginess :)
12:10:42 <Wolf01> Nah, this is just stress
12:17:06 <rptr_> the first time i was actually concerned about getting sick w/ covid was after hearing about the brain fog. otherwise "i can probably take it"
12:17:20 <rptr_> i'm still convinced i got some brain damage from that flu at age 14...
12:36:55 *** HerzogDeXtEr has joined #openttd
12:52:43 <DorpsGek_III> [OpenTTD/OpenTTD] rptr commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTWyN
12:57:53 <FLHerne> andythenorth: I can't see anyone paying £30 for that somehow
13:10:26 *** gelignite has joined #openttd
13:11:54 <DorpsGek_III> [OpenTTD/OpenTTD] LordAro commented on pull request #8331: Fix: [NoAI] don't notify caught exceptions https://git.io/JTW9R
13:24:32 <Wolf01> <Eddi|zuHause> i would think either Wolf01 or alberth <- something about it on 4th april, but also I think we referred to the transport fever 2 date freeze feature
13:45:29 *** frosch123 has joined #openttd
13:56:56 * andythenorth considers 'Wales & West Country' FIRS economy
13:57:23 <LordAro> West Country was always my favourite TTD map
13:57:30 <andythenorth> like FIRS Temperate Basic, only not laughably basic
14:06:44 <andythenorth> it's quite hard to build stations to pickup 6 different cargos and deliver 2
14:06:44 <andythenorth> or pick up 6 and deliver 4
14:07:07 <longtomjr_> Being able to use boats makes it easier in a lot of cases though
14:07:38 <longtomjr_> but with default station spread it probably ends up a bit tricky
14:08:47 <rptr_> firs only has 4 economies now?
14:08:54 <rptr_> there used to be some other ones i think
14:09:00 <andythenorth> I removed extreme
14:09:04 <andythenorth> I'm going to add more
14:09:42 <rptr_> more industries = better :D
14:12:00 <rptr_> you could build spaceship parts, Civ 2 style
14:15:02 <Eddi|zuHause> the civ2 spaceship looked weird. civ1 spaceship was nice
14:16:51 <rptr_> didn't know it had one. i wish civ 2 had the resource feature of civ 3, then it'd be perfect
14:18:56 <Wolf01> Never completed one in civ2
14:19:31 <rptr_> i still want to make some mars map. fund farms and dig for water to terraform and win(?). don't know what capabilities game script has
14:22:57 <FLHerne> andythenorth: Yes please, that's my ideal industry
14:23:57 <FLHerne> andythenorth: Have you considered customs posts? :p
14:24:05 <FLHerne> Placed only on land map edges
14:24:32 <andythenorth> Railroad Tycoon had Warehouses
14:24:38 <andythenorth> ports-without-sea
14:25:02 <andythenorth> border posts might solve some problems
14:25:29 <andythenorth> e.g. WAWC would need quicklime, but I'd rather import it
14:30:33 <andythenorth> FLHerne would including Customs Posts on the borders of Wales / Cornwall be political?
14:32:40 <FLHerne> Go to beaches and hotels, and have an extra-slow loading time because of all the surfboards
14:33:08 <longtomjr_> It looks like the pr that changed the display speed in nml introduced a breaking change
14:33:32 <longtomjr_> Error: (TypeError) "new_ottd_display_speed() takes 3 positional arguments but 4 were given".
14:33:32 <FLHerne> I've never seen castles as an OTTD industry, but it would be pretty
14:33:42 <longtomjr_> (sorry not sure if this is a good time)
14:33:52 <FLHerne> longtomjr_: Sorry, that's probably my fault
14:34:06 <FLHerne> longtomjr_: Is the relevant NML public?
14:34:21 <longtomjr_> I saw you added another param there
14:35:04 <longtomjr_> assume the easy workaround is just to make mul 1
14:36:40 <FLHerne> longtomjr_: Oh, you're calling it from Python
14:36:58 <FLHerne> I don't *think* we promise to maintain a stable Python API
14:37:07 <longtomjr_> ah no worries then
14:37:14 <FLHerne> Especially if you're calling odd internals :p
14:37:45 <longtomjr_> working on recompiling the finnish trainset
14:37:51 <longtomjr_> so not my fault :P
14:38:04 <andythenorth> FLHerne I don't think we promise much at all :)
14:38:07 <longtomjr_> anyways, let me try and figure out what is going on.
14:38:13 <longtomjr_> thanks for the help and sorry for bothering
14:38:19 <FLHerne> longtomjr_: And yes, making 'mul' 1 will give you the same behaviour
14:38:24 <andythenorth> mostly we just count ourselves lucky if we keep up with openttd and have some docs :P
14:38:32 <FLHerne> Are you using custom speed units or something?
14:38:39 <FLHerne> Sorry, afk for a bit, RL intrudes
14:38:48 <FLHerne> But I'll read anything you say in a bit :p
14:39:18 <TrueBrain> LordAro: honestly no clue. But do we still want to support i386?
14:39:30 <longtomjr_> Wait, this might still be a bug, the error comes from /lib/python3.8/site-packages/nml/actions/action0properties.py"
14:39:58 <LordAro> not sure what yet though
14:40:08 <LordAro> TrueBrain: no particular reason why not, tbh
14:40:32 <LordAro> i imagine i'd have the same trouble with the 16.04 amd64 container anyway
14:41:22 <longtomjr_> that line is the one throwing the error
14:41:44 <TrueBrain> Fair enough. I can help looking what goes wrong later today.. I have seen it before, just cannot remember what solved it ..
14:46:00 <longtomjr_> or rather that same script, in the finnish trainset repo
14:48:26 <longtomjr_> # Monkey patch for nml to correct display speed rounding errors,
14:48:27 <longtomjr_> # until some nml dev fixes it in upstream
14:48:38 <longtomjr_> I am guessing this is not needed anymore
14:49:23 <rptr_> any industry which generates passengers from passengers? would make sense for a tourist attraction like a castle
14:49:27 <longtomjr_> so yep, error on my end
15:09:37 <FLHerne> rptr_: FIRS has Hotels which are basically the same thing
15:09:42 <FLHerne> But castles would be more scenic
15:11:30 <FLHerne> longtomjr_: Yes, if you monkeypatch internal functions, you get to keep both pieces ;-)
15:11:37 <frosch123> ecs castles are a puzzle game
15:11:52 <FLHerne> longtomjr_: Also, the change that breaks your patch also removes the reason for it
15:12:10 <longtomjr_> yep, figured that ouyt
15:12:19 <andythenorth> industry terraforming :P
15:12:33 <andythenorth> specify height relative to N tile :P
15:12:35 <longtomjr_> and sorry for the false positive
15:24:44 <andythenorth> Wales and West Country economy would be better with regions
15:24:58 <andythenorth> maybe I could use map xy to force industry location
15:25:07 <andythenorth> hmm, but what if some of map is just sea?
15:25:09 <andythenorth> always problems :P
15:25:33 <rptr_> i do find that random locations of industries don't make much sense
15:25:54 <rptr_> makes sense to have farms concentrated mostly in one place, unless it's one big farmland
15:26:22 <rptr_> well i think the last time i played firs there were some clusters of farms. not sure if intentional
15:26:32 <glx> you can craft a scenario manually :)
15:27:33 <glx> usually only one type of industry is allowed per town
15:30:02 <glx> and newgrf industries can use complex checking of surroundings before being built, but that can prevent them totally if the checks are very specific (like in ECS)
15:30:15 *** matt21347 has joined #openttd
15:35:10 <rptr_> is ECS the one with the weird creatures on the industries?
15:50:00 <TrueBrain> LordAro: very weird issue, as executing the commands locally works fine
15:50:16 <TrueBrain> I remember similar issues when using alpine, and not everything being installed
15:50:21 <TrueBrain> but .. that is not this :D
15:51:51 <TrueBrain> LordAro: maybe try 18.04, see if there is any difference?
15:52:24 <TrueBrain> (in general, 16.04 .. maybe we should drop support for that :D 2 new versions are out, and we said to support n-1 :D)
16:17:01 <LordAro> TrueBrain: it's still supported generally, no real reason why we shouldn't support it either
16:17:21 <TrueBrain> LordAro: not that long ago we said we were only going to do n-1, as otherwise it was not doable really
16:17:28 <TrueBrain> we said that for MacOS, but also for Debian
16:17:39 <TrueBrain> not sure why we want to change that again .. but okay, if you want to hurt yourself, I am not stopping you :D
16:17:45 <TrueBrain> but I would start by trying if 18.04 works
16:17:47 <TrueBrain> and work from there
16:18:09 <TrueBrain> (as I think 20.04 has no i386 anymore?)
16:18:48 <TrueBrain> "Ubuntu Linux 16.04 LTS reaches the end of its five-year LTS window on April 30th 2021" <- or just drag this on for a few more months, and you don't have to feel bad anymore
16:20:38 <glx> I think 20.04 still has i386
16:20:59 <TrueBrain> I can never remember .. they were dropping i386 on one of these main streams :P
16:21:30 <TrueBrain> (about freaking time tbh ... i386 is like ppc :D)
16:21:53 <TrueBrain> ah, they wanted to drop, but too many people were: MAMMIE, THIS IS NOT FAIR
16:22:00 <TrueBrain> so they added i386 for 20.04
16:22:38 <TrueBrain> but in that case LordAro , start with 20.04, and work from there :)
16:22:52 <glx> they also have ppc64el and s390x
16:22:54 <TrueBrain> it is a lot easier if you know it is because you trying to support a 4 year old something ;)
16:23:04 <DorpsGek_III> [OpenTTD/eints] Aswn opened issue #29: Unable to use STRING1 or anyother non standard modifiers in translations https://git.io/JTlq9
16:23:20 <TrueBrain> well, I agree with their statement that supporting 0.001% is a bad idea, as it is very much untested at that point
16:24:21 <TrueBrain> LordAro: if it is 16.04, and if your code just works on 20.04, we have a direction to search for what goes wrong :)
16:24:23 <LordAro> glx: no, full i386 was dropped in 19.10
16:24:23 <DorpsGek_III> [OpenTTD/eints] glx22 commented on issue #29: Unable to use STRING1 or anyother non standard modifiers in translations https://git.io/JTlq9
16:25:05 <LordAro> there's no i386 release
16:25:08 <TrueBrain> they were "fully" going to drop i386 :P
16:25:31 <TrueBrain> anyway, please also remember that GitHub Actions Checkout needs a relative new git version (2.18 I believe)
16:25:41 <glx> yeah I guess the installer is 64bit only, but when upgrading existing install you can still use i386
16:25:42 <TrueBrain> otherwise revision detection might fail (as it will download an archive)
16:25:55 <TrueBrain> so in general, I say what I said last time we talked about this: do the checkout on the host, not in a container ;)
16:26:45 <DorpsGek_III> [OpenTTD/eints] TrueBrain commented on issue #29: Unable to use STRING1 or anyother non standard modifiers in translations https://git.io/JTlq9
16:27:45 <LordAro> a lot of repeated "runs-on" though :p
16:32:50 <TrueBrain> but it avoids having to deal with stuff that is not important :)
16:32:57 <TrueBrain> you only want to use the container to compile
16:33:10 <TrueBrain> but again, if you want to hurt yourself, I am not stopping you :D
16:33:22 <TrueBrain> I would advise: get it to work first, optimize it later ;)
16:33:52 <TrueBrain> and talking about hurting yourself. ... comparing HTML with mediawiki is PAINFUL
16:37:42 <rptr_> what other way do parts of openttd check/print based on debug levels apart from _debug_*_level and DEBUG(...)?
16:38:10 <rptr_> can't seem to find a place where _debug_script_level==5 causes output to the in-game console and stdout
16:41:37 *** supermop_Home_ has joined #openttd
16:58:58 <TrueBrain> why do I keep finding bugs in Wiki templates ... most of this stuff could never really work tbfh :P
17:32:28 *** Wormnest has joined #openttd
17:35:25 *** sla_ro|master has joined #openttd
17:36:38 <TrueBrain> owh boy ... <small><<Back to the <a href=""></a></small> ...
17:36:45 <TrueBrain> how to learn anything that the << should be escaped
17:37:09 <TrueBrain> it now sees "back" as a tag, and "to the <a href="">" as the attributes
17:48:57 *** snail_UES_ has joined #openttd
18:24:10 *** snail_UES_ has joined #openttd
18:56:20 <Eddi|zuHause> obviously you need AI, and blockchain... :p
18:59:49 *** andythenorth has joined #openttd
19:10:27 <andythenorth> more cloud needed
19:11:31 <frosch123> it's dark, you can't even see the clouds
19:14:14 <TrueBrain> I am finding so many bugs in mediawiki ... I hope they are all fixes in newer versions ..
19:14:27 <TrueBrain> my new favorite: <dl><dt><dl><dd></dt></dl></dd></dl>
19:15:11 <TrueBrain> cannot believe the crap a browser can still render :P
19:15:44 <andythenorth> browsers put up with all sorts of shit
19:16:07 <andythenorth> they are the multi-wheel run-flat-tyre truck of the industry
19:16:14 <andythenorth> set a few land mines off under them, they keep going
19:16:25 <andythenorth> I mean...they even run javascript
19:17:06 <glx> well most can probably still handle IE6 weirdness
19:17:28 <TrueBrain> right ... more fixes ... more HTML that renders exactly the same ...i getting there .. now lets check all pages starting with an 'a' :D
19:18:55 <frosch123> don't fool us. we know that wiki pages start capitalized
19:19:02 <TrueBrain> I really wonder what is easier .. to just fix every page to be sane, or to make the render accept the insanity :P
19:19:17 <TrueBrain> by now, I just lowercased every URL .. got so sick of those differences :P
19:19:45 <TrueBrain> hmm .. somehow I managed to render </td></li></ul></tr> .. that can never be correct :D
19:20:40 <TrueBrain> so sometimes mediawiki changes "a: b" into "a:b", and sometimes it does not .. I am so happy if I can let this all go, and just go with frosch123's dataset
19:20:50 <TrueBrain> (currently comparing against live-wiki, as .. well, .. ugh
19:23:17 <TrueBrain> how .... why ...... huh?
19:23:59 <TrueBrain> this is just ... wuth?
19:24:25 <frosch123> you mean the [[link]]trails?
19:24:34 <frosch123> they are not supported by wikitextparser, not me
19:24:38 <frosch123> they are just stupid
19:25:44 <TrueBrain> yes, that with "[[negative income with feeder service]]s", the "s" is inside the <a></a>
19:25:51 <TrueBrain> it can go fuck itself
19:26:01 <TrueBrain> (and only for the text it is inside, not for the URL)
19:26:10 <frosch123> yes, that's the stupid mediawiki feature, that noone supports
19:26:26 <frosch123> i only learned about it when figuring out gollum stuff
19:26:29 <TrueBrain> but why .. just why ...
19:26:34 <TrueBrain> "owh, you were stupid, let me fix it for you"
19:27:15 <frosch123> TrueBrain: it's a shortcut: [[foobar]]s == [[foobar | foobars]]
19:27:35 <TrueBrain> I don't understand these things .. it only makes for very unreadable source :(
19:27:35 <frosch123> when the page title is singular, but the grammar requires something else
19:27:44 <frosch123> it works for languages which append stuff
19:27:49 <frosch123> but fails for others
19:27:52 <TrueBrain> it works for English, basically
19:28:36 <frosch123> ask that the guys who added std::chrono:january/05/2005 as valid syntax to c++
19:29:12 <TrueBrain> let me see if I can recover some sanity here, I think it just ran out of the door
19:29:58 <TrueBrain> meh, still no clue how to fix <<Back from working .. how do I know that it is not a tag .. hmmmm
19:30:08 <TrueBrain> guess, check if there is another < in the tag
19:30:11 <TrueBrain> hmm, that would work
19:35:09 <frosch123> how is "<< back" different from "a < b" ?
19:35:21 <TrueBrain> it is <<Back, not << Back
19:35:31 <TrueBrain> it will have issues with a<b too
19:39:01 <TrueBrain> funny, wikitextparser considers ) in URLs as valid
19:40:28 <TrueBrain> +41.682419659735345%</b>
19:40:32 <TrueBrain> owh boy, this will be interesting :D
19:41:49 <TrueBrain> that is also an odd one
19:43:05 <TrueBrain> the source .. really says "0"
19:43:10 <TrueBrain> but .. I guess .. this is ... PHP ...
19:43:19 <TrueBrain> empty("0") == true after all, in that language
19:43:53 <frosch123> thanks, i was happy without knowing that
19:44:15 <TrueBrain> just reading it makes me laugh EVERY single time
19:44:27 <TrueBrain> that "" is empty, and that 0 is empty, okay .. ish
19:44:32 <TrueBrain> that is just hilarious :)
19:44:53 <LordAro> it's the same in python :p
19:45:09 <TrueBrain> if "0" == True: THIS NEVER RUNS
19:45:27 <LordAro> (well, "false", rather than "empty")
19:45:51 <TrueBrain> .... it can go fuck itself then
19:46:11 <TrueBrain> I do not know if I want to fix this or not ...
19:46:18 <TrueBrain> the parameter is "0"
19:46:24 <TrueBrain> it is used in the template as {{2}}
19:46:30 <TrueBrain> and nothing is printed
19:47:01 <frosch123> TrueBrain: some of those wiki parser libs wrote "this does not try to be bug-to-bug compatible with mediawiki"
19:47:08 <frosch123> that's sound like a good approach
19:47:17 <LordAro> sounds like something is using python boolean logic instead of emulating php's
19:47:27 <TrueBrain> yeah, well, let me show you something, just to make you giggle ..
19:47:32 <TrueBrain> LordAro: there is no boolean logic involved, that is the point
19:47:55 <LordAro> ok, "logic to determine whether something is empty"
19:48:13 <TrueBrain> it is printing a parameter ..
19:48:17 <TrueBrain> there should be no logic involved
19:48:19 <TrueBrain> you say: give me a 0
19:48:22 <TrueBrain> you should get a zero :P
19:48:29 <TrueBrain> no clue why mediawiki makes that more complicated
19:48:42 <LordAro> depends what the template is doing
19:48:43 <TrueBrain> so many seds to make comparing possible :D
19:48:49 <TrueBrain> LordAro: printing the variable! That is all!
19:48:55 <TrueBrain> it really just reads: {{2}}
19:49:01 <TrueBrain> just .. printing the damn variable :P
19:49:09 <TrueBrain> ..... glad you caught on :)
19:49:27 <glx> php likes to complicate simple stuff anyway
19:50:16 <LordAro> but ultimately, i'd say a parser does need to replicate php's behaviour
19:50:46 <TrueBrain> I completely and fully disagree :)
19:50:51 <TrueBrain> it needs to output good HTML
19:50:53 <LordAro> actual bugs, sure, leave those. but language 'design decisions' should be followed
19:51:07 <TrueBrain> good luck finding any "design decision" with mediawiki :P
19:51:11 <TrueBrain> it is all ... "what is implemented"
19:51:20 <TrueBrain> even their versions from one version to the other can render different things
19:51:28 <LordAro> that's what the quotes were signifying
19:51:39 <TrueBrain> so it is easy to say from a distance these kind of things, but reality is a whole lot more suckier :)
19:51:55 <DorpsGek_III> - Update: Translations from eints (by translators)
19:52:13 <TrueBrain> there are many things in our wiki that are easier to fix and simply no longer allow, than to implement the PHP behaviour
19:52:28 <TrueBrain> like this "0" .. fixing it in the template is a lot more effective :)
19:52:56 <TrueBrain> the other lovely thing is: when to escape in URLs .. which is completely random, by the looks
19:53:14 <LordAro> but something like empty("0") == true should be followed
19:54:03 <LordAro> you're the one that wants to accurately reproduce the output!
19:54:07 <TrueBrain> something something, high horse, something something .. andythenorth , I have a new NewGRF idea for you :D
19:54:25 <TrueBrain> no, LordAro , I do not; I want to validate that my renderer renders our wiki correctly after conversion
19:54:27 <LordAro> that's not a mediawiki paraer weirdness, that's a fundamental part of php
19:54:32 <TrueBrain> there are several ways to reach that goal
19:54:49 <TrueBrain> you are either being a total dick, or totally misunderstood what I was running into
19:54:56 <TrueBrain> so let me run it again by you, and you can pick :D
19:55:04 <TrueBrain> {{Template|0}} is called
19:55:08 <TrueBrain> Template reads: {{1}}
19:55:12 <TrueBrain> Mediawiki outputs ""
19:55:17 <TrueBrain> my render outputs "0"
19:55:37 <TrueBrain> this is, most likely, due to the fact they use PHP and empty("0") == "" or something silly (I do not know, I have not checked their code)
19:55:48 <TrueBrain> only question is: is "" correct or "0" correct
19:56:04 <LordAro> i thought it was the other way around
19:56:23 <LordAro> -/+ do not make it clear :p
19:56:49 <TrueBrain> - is mediawiki, + is my render, of course :) How else do you compare? Old -> new, right? :D
19:57:00 <TrueBrain> anyway, I assumed you weren't being a total dick, no worries :)
19:57:58 <TrueBrain> it is just this wikitext parser called "mediawiki" ... and our version being ANCIENT ... it is starting to get the better of me :P
19:58:37 <TrueBrain> -<a href="/%d0%9d%d0%b0%d1%81%d1%82%d1%80%d0%be%d0%b9%d0%ba%d0%b8_%d0%98%d0%98/ru" title="настройки ии/ru">
19:58:38 <TrueBrain> +<a href="/%d0%9d%d0%b0%d1%81%d1%82%d1%80%d0%be%d0%b9%d0%ba%d0%b8_%d0%b8%d0%b8/ru" title="настройки ии/ru">
19:58:40 <frosch123> oh, i would bet on current mediawiki still behaving that stupid
19:58:41 <TrueBrain> these diffs crack me up too :)
19:59:16 <TrueBrain> (uppercase vs lowercase)
20:01:35 <TrueBrain> I do understand why wikitext never became a standard, and that markdown exists, btw
20:01:52 <LordAro> TrueBrain: i can't see any difference there at all
20:02:07 <TrueBrain> LordAro: %d0%98 -> %d0%b8, after the _
20:02:19 <TrueBrain> took me a texteditor highlight to find that
20:02:39 <LordAro> ah, that's the last character on the line of my current terminal
20:02:48 <LordAro> which is why i missed it
20:03:34 <TrueBrain> also something I cannot seem to get right, when the "alt" text of an image is the filename and when not
20:03:39 <TrueBrain> I gave up .. who cares anyway ..
20:04:15 <TrueBrain> also "fun", mediawiki renders <pre> blocks in <code> blocks :)
20:05:13 <TrueBrain> -<span style="width:0%; height:2px; background:#7fff00; float:left">
20:05:13 <TrueBrain> +<span style="width:0.0%; height:2px; background:#7fff00; float:left">
20:05:44 <LordAro> frosch123: fyi, with mediawiki 1.31, i get the '0'
20:06:05 <LordAro> hence, a bug that shouldn't be replicated :p
20:06:27 <TrueBrain> I have to say this, but know I am smiling: dickhead
20:06:50 <LordAro> love you too, TrueBrain
20:07:06 <TrueBrain> but good to know they also notice it was a garbage thing to do :P
20:07:20 <LordAro> since i'm now at a computer again, i will do a bit more on the release workflow
20:07:26 <TrueBrain> I like I keep finding bugs this way :) Took for-ever to write the compare script, but it is nice it shows me what I am doing wrong
20:08:25 <TrueBrain> owh, I really have to do something with the <nowiki> stuff ...
20:08:34 <TrueBrain> I just hope wikitextparse does something with it :P
20:08:37 <TrueBrain> I honestly do not know :)
20:09:21 <TrueBrain> plenty of tests mentioning it
20:12:41 <TrueBrain> -<table border="1" class="wikitable sortable">
20:12:41 <TrueBrain> +<table border="1" padding="4" class="wikitable sortable">
20:12:50 <TrueBrain> how ... just .. why ..
20:13:54 <TrueBrain> well, you are not alone, that is good
20:14:08 <TrueBrain> still curious if it does work with newer Ubuntus
20:14:13 <TrueBrain> as the amd64 did work in your build
20:14:29 <LordAro> apparently not - i386/ubuntu:18.04 had some problem
20:14:33 <TrueBrain> I wonder if it is an issue where a syscall cannot be executed
20:14:43 <TrueBrain> or something in those lines
20:15:05 <TrueBrain> worth a shot possibly?
20:15:25 <LordAro> i386/ubuntu:20.04 doesn't exist
20:15:27 <TrueBrain> so why is mediawiki ignoring "padding", I wonder ..
20:15:33 <LordAro> which is why i never added it to CompileFarm :p
20:15:37 <TrueBrain> it doesn't? Awh ... I assumed they would have made it :P
20:15:53 <TrueBrain> dist-upgrade an 18.04 :P
20:16:02 <TrueBrain> what is a bit odd that you can execute the "apt update" etc
20:16:38 <LordAro> well we had to - cached apt packages got too out of date for the mirrors
20:16:40 <TrueBrain> but really .. a checkout on the host is most likely your solution, as I think the cmake will also just work
20:17:03 <TrueBrain> yes, you have to; what I mention, is that it is odd that you can execute apt update, but the actions/checkout doesn't work
20:17:10 <TrueBrain> so the docker is working, and responding okay
20:17:24 <TrueBrain> it is somehow the way actions/checkout interacts with it, that is the issue, it seems
20:17:42 <LordAro> it's really difficult to do mixed host/container commands though, as far as i can tell?
20:18:01 <LordAro> there's some comments in linked issues that suggest that it's due to node being missing in the container
20:18:17 <TrueBrain> install it and find out :P
20:18:27 <LordAro> that's my next step :)
20:18:36 <LordAro> step after that is to work out why FindVersion.cmake is not working
20:19:10 <TrueBrain> because git is too old
20:19:15 <TrueBrain> we have been over that, haven't we? :D
20:19:43 <LordAro> it's returning a result that suggests that git doesn't exist at all (.git dir doesn't)
20:19:59 <TrueBrain> because you used an old git, and it fetched an archive
20:20:05 <TrueBrain> check the run results, the checkout step
20:20:07 <TrueBrain> it mentions it in there
20:20:14 <LordAro> yes, that makes sense
20:20:24 <TrueBrain> not the first time we have this conversation about that :P :D
20:20:40 <TrueBrain> this is why I suggest the checkout on host :) Can't you set per step if it needs to run in the container or something?
20:20:41 <LordAro> it's a very confusing fallback behaviour!
20:20:52 <TrueBrain> and I would rather have it just fails
20:21:02 <LordAro> i don't think so, as far as i can tell from the docs
20:21:21 <TrueBrain> If you do not set a container, all steps will run directly on the host specified by runs-on unless a step refers to an action configured to run in a container.
20:21:26 <TrueBrain> so it is possible, but how .. hmm
20:21:29 <TrueBrain> and yes, docs are not clear :P
20:22:47 <TrueBrain> owh, an action can be a Dockerfile
20:22:50 <TrueBrain> right, forgot about that
20:23:18 <TrueBrain> install ca-certifications :)
20:23:34 <LordAro> curious that it's only required for that one though
20:23:48 <TrueBrain> git clone https: ..
20:23:52 <TrueBrain> apt normally uses http still
20:23:58 <TrueBrain> you need to install .. https-transport ?
20:25:40 <LordAro> yeah, i thought it might be that
20:25:45 <TrueBrain> it seems you can do: make a single job to do a checkout, and use that for the others
20:25:52 <TrueBrain> might even save a bit of time :D
20:26:29 <TrueBrain> I think it requires you to stash the source btw
20:26:36 <TrueBrain> I know it does share some data automatically
20:26:49 <TrueBrain> bit of trying, lot of `run: ls` steps, etc
20:28:22 <LordAro> dunno how that will work with a matrix either
20:28:31 <LordAro> don't exactly want multiple dockers trying to compile the same sources :p
20:28:51 <TrueBrain> make the build folder out-of-tree and you will be fine
20:29:30 <TrueBrain> I love cmake for this :)
20:30:53 <TrueBrain> hmm .. can I make a sed that removes all spaces from the style attribute alone? :D
20:44:14 <TrueBrain> made .. something .. for that, lol
20:44:20 <TrueBrain> it is ugly :D Who cares .. just to test stuff :)
20:47:19 <LordAro> TrueBrain: i386 containers fail when downloading artifacts in the same way as they do when trying to checkout repo :p
20:47:57 <TrueBrain> other solution is a tiny bit more meh
20:48:04 <TrueBrain> but you can make a Dockerfile in os/linux or what-ever
20:48:20 <TrueBrain> which does the installing, and an entrypoint which does the cmake
20:48:25 <TrueBrain> what is in CompileFarm basically
20:48:37 <TrueBrain> and you can do "uses: ./os/linux/bla"
20:48:42 <TrueBrain> which will build and run the Dockerfile
20:52:42 <TrueBrain> mod.c3.a8le_d.27acc.c3.a9l.c3.a9ration_des_trains
20:54:21 *** y2kboy23 has joined #openttd
20:59:07 <TrueBrain> time to overload our wiki again ... the API is freaking slow :P
21:01:33 *** Feuersalamander has quit IRC
21:01:37 <TrueBrain> sometimes I get OSError: [Errno 101] Network is unreachable
21:01:43 <TrueBrain> not sure .. what to think of that
21:03:54 *** arikover has joined #openttd
21:05:14 <frosch123> are you using a bot account?
21:08:16 <frosch123> who knows, it's php
21:08:28 <TrueBrain> I now use a session; that is already a huge improvement
21:08:31 <TrueBrain> 1 socket to do everything with :)
21:09:31 <TrueBrain> so much pages are just absolute crap
21:09:40 <TrueBrain> both from a content, but also from a structure
21:12:03 <TrueBrain> still some tables that fail parsing
21:12:12 <TrueBrain> some places where I parse a list, and mediawiki does not
21:13:08 <TrueBrain> and I absolutely LOVE how things like {{{updated}}} is also translated into all the languages ...... lol
21:13:26 <TrueBrain> can we agree to keep all variables and functions just in english? :)
21:14:56 <frosch123> we can only make the english pages look good. translators have to fix the rest by hand
21:15:51 <TrueBrain> owh, so the error was not from the line above
21:16:49 <TrueBrain> just to check for node12 or node12_alpine ..sounds sane :P
21:17:55 <TrueBrain> frosch123: I am also scanning other languages, and even categories now I see, to check for my render .. but yeah, I was hoping to avoid fixing other languages :D
21:18:05 <LordAro> not sure why that works for the x64 containers though... am investigating..
21:18:08 <TrueBrain> I am basically just hitting every URL in your full.json :P
21:18:54 <TrueBrain> btw, LordAro , in case you missed it: there is a TODO in the part that should produce .deb files in CMake
21:19:08 <TrueBrain> seems nobody really implemented it / tested it
21:19:13 <TrueBrain> just so I have said it explicit :D
21:19:49 <LordAro> TrueBrain: i was intending on testing the debs before making a PR :p
21:19:54 <LordAro> but i am a long way from that yet :p
21:20:02 <LordAro> interestingly, none of the containers have a `node12`
21:20:07 <LordAro> so why is it working for some of them..
21:20:19 <TrueBrain> not sure it needs it .. :P
21:20:34 <TrueBrain> you can fake the /etc files and see what that does? :D
21:20:55 <LordAro> it's nothing to do with the /etc files
21:21:00 <LordAro> other than the alpine switch
21:21:24 <LordAro> it must actually be failing in the calling function
21:21:36 <TrueBrain> frosch123: do you also have a json of pages (on their original URL) you are keeping?
21:21:43 <TrueBrain> or are we keeping them all currently?
21:21:55 <TrueBrain> "xargs: unmatched double quote; by default quotes are special to xargs unless you use the -0 option" <- pfft, what a way to go :(
21:22:58 <frosch123> TrueBrain: fullmap contains everything that is not in redirects.json, trash.json or in one of the namespaces i ignore (user pages, talk pages, mediawiki/internal pages)
21:23:25 <TrueBrain> ah .. I am using "jq" to read the JSON, so matching stuff is not easy :P
21:23:28 <frosch123> trash.json will grow slightly, but not much
21:23:37 <TrueBrain> I don't have trash.json atm, but that will be fine :)
21:27:17 <TrueBrain> meh, to get ride of the capital issue, I need to do a lookup after all .. bah
21:28:05 <frosch123> doesn't my conversion fix all capitalisation issues?
21:28:37 <TrueBrain> I dunno; I am testing against the live site atm, as because of the changes, it was very hard to validate HTML
21:28:55 <TrueBrain> (I am checking the live HTML with the rendered one, basically)
21:29:05 <TrueBrain> meant I had to rewrite the URLs again :P
21:32:39 <glx> LordAro: wouldn't it be better to build source package with cmake and use that in containers ?
21:41:30 <LordAro> of course a x86 container can't execute a 64bit binary
21:41:47 <TrueBrain> I believe I mentioned earlier that it looks like invalid syscalls :P
21:41:58 <TrueBrain> bit silly of a GitHub Runner to do this, I guess
21:42:09 <TrueBrain> means you cannot execute node-actions on a container
21:42:13 <TrueBrain> but you can do "run"
21:42:19 <TrueBrain> and most likely Dockerfiles
21:42:44 <TrueBrain> but as your git is too old anyway, this is a problem left or right :)
21:43:30 <LordAro> it'd be easier to make it work if it didn't hardcode a magical mysterious path
21:43:51 <TrueBrain> _w is workspace on host
21:44:14 <LordAro> well that removes the need to use upload/download artifacts
21:44:36 <LordAro> kinda hacky and all "using undocumented features", but eh
21:44:49 <TrueBrain> still not sure how you are going to get over the git-version issue :)
21:45:07 <LordAro> well i can do the checkout on host
21:45:12 <LordAro> git version isn't an issue for cmake
21:45:50 <glx> oh and as I think the workspace is shared for all containers you can't use the same "build" dir
21:46:22 <LordAro> mkdir build${{matrix.container}}
21:46:39 <TrueBrain> at least put a - in there
21:51:56 <andythenorth> oof naming of things
21:52:07 <andythenorth> completely not obvious
21:56:09 <LordAro> CMake Error at /usr/share/cmake-3.5/Modules/TestBigEndian.cmake:51
21:56:16 <LordAro> i'll be honest, i wasn't expecting that
21:57:59 <TrueBrain> hmm, even if I look up what the correct name is for wiki pages, even than I have a diff because of casing difference
22:00:53 <TrueBrain> ah, redirects .. ugh
22:02:58 <TrueBrain> another thing I haven't figured out .. sometimes it makes [1] from link-text
22:03:01 <TrueBrain> no clue when/why/how
22:05:09 <TrueBrain> okay, first 20 pages validate fully now .. \o/
22:25:51 <TrueBrain> okay .. ran 1100 pages .. 680 pass, 450 have a diff, and 25 fail to render .. over 50% done, I guess :P
22:27:08 <TrueBrain> +<a href="/user:0" title="user:0">
22:27:12 <TrueBrain> lol .. guess why that happened :P
23:21:07 <Eddi|zuHause> TrueBrain: that sounds like "Error Code: 0. No Error"
continue to next day ⏵
