IRC logs for #openttd on OFTC at 2020-10-04
⏴ go to previous day
00:01:42 <Afdal> Yeah I'm testing some different newgrfs and different versions of them and it seems it's a pattern in some and not others
00:02:24 <Afdal> oh but it shouldn't be that
00:02:37 <Afdal> I'm looking at the available vehicles list
00:02:44 <Afdal> not current active vehicles in a given game
00:03:06 <Afdal> I do think it's related to a newGRF option change though
00:06:15 <Afdal> furthermore the pattern seems to change depending on my game start date
00:16:23 <Afdal> I dunno, this is really confusing
00:17:05 <Afdal> All I can say is that I've started up multiple games now and played through to find out that I end up with not a single train above 85% from 2000 to 2050
00:17:12 <Afdal> that is an intolerable change
00:19:31 *** snail_UES_ has joined #openttd
00:25:24 <TrueBrain> hmm .. CompileFarm for Windows failed because of a timeout
00:25:34 <TrueBrain> vcpkg failed on pkg-config ..
01:11:19 <glx> but as it's inside a vcpkg packages it's not very a usable workaround
01:21:16 *** tokai|noir has joined #openttd
01:21:16 *** ChanServ sets mode: +v tokai|noir
01:34:19 <glx> #8324 CI run fails for the same reason
01:38:05 <TrueBrain> Lol .. it is a single server?!
01:38:44 <TrueBrain> Vcpkg uses a mirror that is not a CDN? Damn ... even we don't do that for files .. not for the last 10 years at least
03:30:14 <Afdal> I guess v1.9.0 was when the reset_engines command stopped functioning to re-roll reliabilities too
08:04:36 *** andythenorth has joined #openttd
09:42:32 *** Progman has joined #openttd
10:02:28 <TrueBrain> so ... lets move some Python code into its own library ... I guess I make 1 library with all kinds of common bits and pieces or something .. or many libraries for every unique piece .. hmm
10:07:01 <TrueBrain> guess small packages is better :)
10:13:19 *** sla_ro|master has joined #openttd
10:47:44 <TrueBrain> maybe I should just call it "openttd-helpers", as it is a collection of all kind of different things :P
10:57:27 *** HerzogDeXtEr has joined #openttd
11:08:05 *** WormnestAndroid has quit IRC
11:08:18 *** WormnestAndroid has joined #openttd
11:16:22 <andythenorth> I called mine Polar Fox :P
11:16:32 <andythenorth> a 15 year old joke
11:16:42 <andythenorth> that is only in my head
11:52:40 <andythenorth> 'fair use' of song lyrics in the docs for a GPL project? o_O
11:52:48 * andythenorth armchair laywering
12:07:31 *** gelignite has joined #openttd
12:11:20 <andythenorth> internet says 'nope'
12:11:27 <andythenorth> no quoting lyrics
12:21:12 <TrueBrain> [OpenTTD/py-helpers] TrueBrain updated pull request #1: Add: click_helpers, logging_helpers and sentry_helpers https://git.io/JU59V
12:21:32 <TrueBrain> hmm, actions don't kick in
12:21:36 <TrueBrain> okay, multiple PRs it is ..
12:24:42 <TrueBrain> [OpenTTD/py-helpers] TrueBrain opened pull request #2: Add: [DorpsGek] announce issues, pull-requests, and tags to IRC https://git.io/JU597
12:24:47 <TrueBrain> [OpenTTD/py-helpers] TrueBrain opened pull request #3: Add: [Actions] CodeQL, flake8, black, and regression tests https://git.io/JU595
12:25:59 <TrueBrain> it has 100% coverage :D
13:00:43 *** frosch123 has joined #openttd
13:40:17 *** m1cr0man has joined #openttd
14:17:52 *** SpComb^ has joined #openttd
14:20:01 <DorpsGek_III> [OpenTTD/py-helpers] TrueBrain merged pull request #2: Add: [DorpsGek] announce issues, pull-requests, and tags to IRC https://git.io/JU597
14:22:13 <DorpsGek_III> [OpenTTD/py-helpers] glx22 approved pull request #3: Add: [Actions] CodeQL, flake8, black, and regression tests https://git.io/JU5x5
14:22:39 <DorpsGek_III> [OpenTTD/py-helpers] TrueBrain merged pull request #3: Add: [Actions] CodeQL, flake8, black, and regression tests https://git.io/JU595
14:32:47 <DorpsGek_III> [OpenTTD/py-helpers] TrueBrain updated pull request #1: Add: click_helpers, logging_helpers and sentry_helpers https://git.io/JU59V
14:33:52 <glx> lucky regression check was enabled :)
14:35:06 <DorpsGek_III> [OpenTTD/py-helpers] TrueBrain updated pull request #1: Add: click_helpers, logging_helpers and sentry_helpers https://git.io/JU59V
14:35:22 <TrueBrain> it is the reason to have a regression suite :)
14:35:37 <TrueBrain> guess I will be bumping the few 3.7 repositories we have to 3.8!
14:47:16 <DorpsGek_III> [OpenTTD/py-helpers] glx22 approved pull request #1: Add: click_helpers, logging_helpers and sentry_helpers https://git.io/JU5jl
14:47:47 <TrueBrain> "seems okay", lol :D
14:48:21 <glx> well the code is used in all other repos, and it works there
14:48:23 <DorpsGek_III> [OpenTTD/py-helpers] TrueBrain merged pull request #1: Add: click_helpers, logging_helpers and sentry_helpers https://git.io/JU59V
14:48:38 <TrueBrain> now I need to figure out how to publish this :P
14:48:43 <TrueBrain> and then convert all repos ..
14:49:17 <glx> check nml actions, I think there's what you need there
14:49:32 <TrueBrain> yeah, I will copy from there
14:53:46 <glx> and repo.msys2.org is still down
14:54:00 <TrueBrain> I am still really surprised it is a single machine
14:54:25 <glx> they don't have a TrueBrain in the team
14:54:39 <TrueBrain> it is not like msys is small ..
14:55:30 <TrueBrain> just in case you need a giggle
14:56:07 <glx> yeah I saw it, it's linked in the issue
14:58:21 <TrueBrain> and just keep on reading ...
14:58:26 <TrueBrain> mirrors that are out-of-date
15:00:15 <TrueBrain> well, one of those problems with Open Source .. people depend more and more on a small group of people, assuming they have infinite time and money to have everything with 100% availability and 100% integrity .. :D
15:00:27 <TrueBrain> the reason I am rather happy GitHub has their own Container Registry now ..
15:00:40 <TrueBrain> I think that has a higher chance of surviving than Docker Hub has :)
15:08:27 <andythenorth> XKCD has you covered
15:10:32 <glx> msys2 itself is usable, I'm upating my install and it works (except timeout on the down server before trying another one, for each file)
15:10:50 <glx> the main issue is vcpkg hardcoding the server
15:11:24 <TrueBrain> well, no, the main issue is that they were using their "main repo" on a single machine :)
15:11:42 <TrueBrain> even OpenTTD didn't do that .. the main repo was not available for the general public :P
15:12:08 <TrueBrain> but it seems the mirrors are not all up-to-date
15:12:18 <TrueBrain> so people get older versions and stuff
15:12:25 <TrueBrain> something wonky anyway
15:12:37 <TrueBrain> well, I hope they take this opportunity to make their infra better :)
15:12:49 <glx> most mirrors seems to be ok, except sourceforge
15:13:58 <TrueBrain> naming SF in this channel
15:17:56 <andythenorth> isn't it malwareforge?
15:18:00 <andythenorth> or is that just a rumour?
15:18:11 <TrueBrain> well, they did insert stuff in installers
15:18:15 <TrueBrain> that is not a rumour, but a fact
15:18:20 <TrueBrain> think they regretted those actions
15:18:36 <TrueBrain> they are still here ..
15:19:12 <andythenorth> I think it was the banner adds for malware and scamware on every page that I remember
15:19:14 <glx> of course not all projects moved
15:19:23 <andythenorth> big green arrows with 'download now' faking the project download
15:20:19 <TrueBrain> ah, that bullshit .. wouldn't be surprised if they still do that
15:20:30 <TrueBrain> and an english person that says adds instead of ads .. not sure what to think about that :P
15:25:51 <glx> funny how people can't look at other issues before opening theirs, the vcpkg issue list is full of almost duplicates related to msys2 failure
15:26:18 <TrueBrain> yup ... but on the other hand, they could have pinned a single issue with a more clear title :)
15:34:45 *** jottyfan has joined #openttd
16:11:11 <LordAro> "Hi, welcome to Three. You're chatting with Steven from the Complaints team team. I will help you."
16:11:56 <glx> hmm a team team, that's scary
16:12:10 <LordAro> not as scary as 0.5Mbps download
16:12:30 <glx> I had 3 weeks without internet
16:13:29 <glx> and it was a 5 minutes fix in phone central
16:31:03 <milek7_> sourceforge seems stuck in 00's
16:31:18 <milek7_> with that 'you download will start in 5 seconds...'
16:34:03 *** snail_UES_ has joined #openttd
16:57:31 <LordAro> milek7_: impressively detailed article
17:02:09 <andythenorth> "Then again, Transport Tycoon needs every ounce of good will it can generate — because, taken purely as a piece of zero-sum game design, it’s horribly, hopelessly broken. "
17:02:34 <LordAro> take that OTTD detractors, it's always been broken!
17:18:54 <FLHerne> "This effectively infinite bankroll makes cost-benefit analysis meaningless, causing what ought to be interesting dilemmas — the meat of a good strategy game — to become moot." <- yup
17:19:30 <glx> and that is still unchanged, unless with some newgrf combos
17:20:04 <LordAro> most newgrfs make it *more* broken :p
17:20:40 <FLHerne> You can't really fix it with grfs
17:21:21 <andythenorth> you can't fix it
17:21:30 <FLHerne> The game design means you either make a ton of money or go bankrupt in short order, the trickier grfs just make the second one easier
17:21:37 <andythenorth> the core of playability is pretty much the pipemania puzzle
17:21:43 <andythenorth> fitting stuff into the remaining tiles
17:21:49 <FLHerne> Sustained barely-enough-money isn't a thing
17:22:00 <FLHerne> Might be possible with GS
17:23:38 <andythenorth> it's not even interesting :)
17:23:50 <andythenorth> the constrained resource becomes 'spare tiles with room for stations'
17:24:14 <andythenorth> it's interesting we don't have a second resource
17:24:19 <andythenorth> most games have gems and coins
17:24:21 <glx> room and being accepted by town
17:24:21 <andythenorth> or gold and manna
17:24:40 *** sla_ro|master has joined #openttd
17:24:52 <andythenorth> we just have 'spare tiles'
17:44:07 <Afdal> So... What should I do to try and fix the max. reliability system that's been busted up since v1.9.0? Submit an issue on github? Attempt to write my own fix for it?
17:44:27 <Afdal> I've very serious when I say this approaches game-ruining for me
17:44:48 <Afdal> I can't believe it's been sitting around for so long without other people apparently complaining about it
17:44:57 <nielsm> start with an issue to document it
17:45:12 <Afdal> because it results in regular scenarios where the last several decades of vehicles don't have reliability about the low '80s%
17:46:24 <LordAro> are you certain it was broken in 1.9.0?
17:46:33 <Afdal> I've got a singleplayer save I've been working on right now that I've essentially had to give up on because there's simply no way to. Can't even load the save in an old version of OpenTTD and fix it up with reset_engines.
17:46:34 <LordAro> or are you just going off that random changelog entry?
17:46:48 <Afdal> Yeah I've tested out v1.8.0 and v1.9.0
17:46:57 <Afdal> that's definitely when the change occurred
17:47:25 <LordAro> first step is an issue on GH
17:47:25 <Afdal> I think it might have been an unintented consequence from when the random vehicle introduction year thing got synchronized
17:47:43 <Afdal> I wish I still remembered my old github password ~_~
17:47:58 <LordAro> there are ways of resetting these things...
17:48:09 <Afdal> and the email associated with it <.<
17:48:22 <Afdal> Guess I'll just have to make a new one
17:51:02 <Afdal> Would it make more sense to just comment on the old pull request
17:51:15 <Afdal> that I'm fairly certain is associated with it
17:51:27 <LordAro> comments on closed pull requests will almost certainly get ignored
17:52:53 <LordAro> remember to include steps for reproducing the behaviour that you're seeing
17:53:42 <LordAro> there are 628 changes between 1.8.0 and 1.9.0 :)
17:54:20 <LordAro> well, 501 non-translation changes
17:54:59 <Afdal> I'm still having trouble understanding exactly how this thing works
17:55:50 <Afdal> because sometimes I'll load up a game at 2200 and there will be tons of 99% max. reliability vehicles
17:56:15 <Afdal> it's almost as if you get a bad max. reliability "seed" and then it propagates across all the vehicles with short year increments between their introduction dates
17:57:59 <LordAro> the sooner we get a list of steps, the sooner we can nail down exactly what's causing it
18:00:02 <Afdal> But whatever the case, if you play with breakdowns on and you get one of these bad seeds, your game is pretty much ruined if you had any plans of using late-game vehicles
18:01:27 <Afdal> with the old system you could at least count on some late vehicles having randomly high reliability. With the new system huge swaths of vehicles can be rendered with reliabilities in the low 80% range
18:01:47 <Afdal> which is pretty bad for dense networking
18:08:25 *** Wormnest has joined #openttd
18:22:44 <andythenorth> it's plausible (without looking at code) that synchronised intro dates means synchronised reliability
18:22:55 * andythenorth would be surprised though
18:23:13 <andythenorth> the lolz of unintended consequences :P
18:23:14 <Timberwolf> From what I can see, if they have the same intro date and type, and come from the same GRF, they'll get the same reliability.
18:23:19 <LordAro> basically, until we get a reproducible report, this is all hearsay
18:23:43 <andythenorth> Horse 2.8.0 time
18:23:43 <LordAro> and i'm not going to spend time speculating further
18:24:35 <Timberwolf> But it's not possible to reproduce it across multiple vehicles unless you can get at least the game generation seed, and the GRF you're using (or at least the grf-defined introduction date for the affected vehicles, and the ID of the grf they come from)
18:54:26 <andythenorth> have a day off andythenorth
18:54:38 <andythenorth> Horse to-do list never gets shorter though :D
19:02:48 <TrueBrain> hmm .. someone tried to upload "rfirs", and failed on BaNaNaS .. lets see why :)
19:03:25 <TrueBrain> hmm, the .tar was never extracted .. that is odd
19:05:26 <TrueBrain> owh, someone tried to upload a tarball in a tarball
19:05:32 <TrueBrain> haha, I did not consider that user-flow ..
19:06:34 <andythenorth> it's the FIRS makefile again
19:06:44 <andythenorth> this isn't me though, this is a new FIRS fork I think
19:06:57 <TrueBrain> it is not you, no :)
19:07:03 <andythenorth> NFI why these makefiles do this :)
19:07:03 <TrueBrain> I see you managed to upload a new version just fine :P
19:07:10 <andythenorth> but not the rationale
19:07:21 <TrueBrain> but okay, the API should not return a 500 in these cases
19:12:56 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain opened pull request #74: Fix: API returned 500 if you uploaded a tarball in a tarball https://git.io/JUdRl
19:13:29 <TrueBrain> we have some silly users :P
19:14:06 <TrueBrain> at least the response time on bugs is quick :P
19:30:53 <DorpsGek_III> [OpenTTD/bananas-api] LordAro approved pull request #74: Fix: API returned 500 if you uploaded a tarball in a tarball https://git.io/JUdEv
19:32:09 * andythenorth awaits applause for shipping newgrf
19:36:51 <frosch123> should we announce bananas uploads in chat?
19:37:06 <TrueBrain> it is not incredible spammy
19:39:18 <frosch123> i gusss andy got more feedback for train whack than for horse?
19:40:54 <andythenorth> I think the quality is more important than the volume :D
19:41:21 <andythenorth> one day, I'll finish FIRS 4
19:42:54 <DorpsGek_III> [OpenTTD/bananas-api] TrueBrain merged pull request #74: Fix: API returned 500 if you uploaded a tarball in a tarball https://git.io/JUdRl
19:43:06 <TrueBrain> andythenorth: rfirs had exactly the same issue you had, yes :)
19:43:21 <TrueBrain> he uploaded it file by file to solve it .. after he tried to tarball the tarball :)
19:52:02 <DorpsGek_III> - Update: Translations from eints (by translators)
19:54:49 <andythenorth> I will fix all my makefiles at some point
19:54:54 <andythenorth> hopefully the forks pull
19:55:08 *** jottyfan has joined #openttd
19:57:01 <TrueBrain> at least it no longer errors out if users are silly :P
20:11:08 <TrueBrain> lovely .. a sentry report about eints, where every variable is [Filtered] :D
20:11:23 <TrueBrain> frosch123: oauth shows an issue again; might taking a look see if you can find it?
20:12:28 <TrueBrain> "'UserAuthentication' object has no attribute 'callback'"
20:12:41 <frosch123> yes, we already had that
20:12:58 <frosch123> it happens when you get to the oauth callback without a session cookie
20:13:14 <frosch123> so possibly when hitting "back" in the browser after logou
20:13:21 <frosch123> or maybe with some privacy mode
20:13:27 <TrueBrain> well, we might had that, but we haven't fixed it :P
20:13:29 <frosch123> should we silence the problem?
20:13:47 <TrueBrain> in general I prefer servers that don't cause backtraces, yes :D
20:14:06 <TrueBrain> would be nice to just throw an error to the user instead of a 500 :)
20:16:14 <frosch123> hmm, eints became complicated to run
20:18:55 <frosch123> that does not help, i can now longer maintain a simple config.xml
20:19:21 <frosch123> now i have to use an >200 char command line or so
20:22:25 <TrueBrain> disable line 97 till 127 in main.py, and off you go
20:28:42 <DorpsGek_III> [OpenTTD/eints] frosch123 opened pull request #23: Fix: Users without valid session cookie only have an instance of user… https://git.io/JUdac
20:29:38 <DorpsGek_III> [OpenTTD/eints] TrueBrain opened pull request #24: Add: allow "python -m webtranslate.main" to use a static config.xml https://git.io/JUda4
20:29:40 <frosch123> alternatives are: duck typing (check whether instance has member 'callback'), or adding a 'callback' to the base class (which is also used by ldap, redmine, development,...)
20:29:41 <TrueBrain> and for the grumpy old man: ^^ :D
20:31:33 <TrueBrain> pages/login.py, line 82, possibly also add there: "or not req_redirect", or is that silly?
20:32:03 <DorpsGek_III> [OpenTTD/eints] TrueBrain commented on pull request #23: Fix: Users without valid session cookie only have an instance of user… https://git.io/JUda2
20:32:23 <frosch123> TrueBrain: that would be wrong
20:32:37 <frosch123> req_redirect=None just goes to the frontpage
20:32:46 <TrueBrain> ah! That was not clear to me :)
20:32:57 <frosch123> no, it goes to the profile page
20:33:19 <frosch123> anyway, it is fine to login without going to a specific page where you were before
20:34:14 <TrueBrain> it is a bit weird that the userauth can not be GithubUserAuthentication, but I am fine with your solution
20:34:24 <TrueBrain> so slab a comment on there, and off we go :)
20:34:25 <DorpsGek_III> [OpenTTD/eints] frosch123 commented on pull request #24: Add: allow "python -m webtranslate.main" to use a static config.xml https://git.io/JUdaX
20:34:31 <TrueBrain> same user hit it 4 times, the 500 page btw
20:36:56 <DorpsGek_III> [OpenTTD/eints] frosch123 updated pull request #23: Fix: Users without valid session cookie only have an instance of user… https://git.io/JUdac
20:38:01 <DorpsGek_III> [OpenTTD/eints] TrueBrain updated pull request #24: Add: allow "python -m webtranslate.main" to use a static config.xml https://git.io/JUda4
20:38:25 <DorpsGek_III> [OpenTTD/eints] TrueBrain approved pull request #23: Fix: Users without valid session cookie only have an instance of user… https://git.io/JUdaF
20:39:27 <DorpsGek_III> [OpenTTD/eints] TrueBrain commented on pull request #24: Add: allow "python -m webtranslate.main" to use a static config.xml https://git.io/JUdab
20:39:33 <frosch123> sometimes dirt on the screen looks like a ' and makes me think: how did that compile..
20:40:00 <TrueBrain> :D I know that feeling a bit too well ... :P
20:40:56 <DorpsGek_III> [OpenTTD/eints] TrueBrain updated pull request #24: Add: allow "python -m webtranslate.main" to use a static config.xml https://git.io/JUda4
20:41:02 <frosch123> oh... there is __main and main :p
20:41:16 <TrueBrain> eints is the collection of blobs on blobs on blobs :)
20:42:03 <DorpsGek_III> [OpenTTD/eints] frosch123 approved pull request #24: Add: allow "python -m webtranslate.main" to use a static config.xml https://git.io/JUdVL
20:42:44 <frosch123> hmm.. i forgot how long this wiki-migration script runs :)
20:42:57 <TrueBrain> how is it going with the ruby skills?
20:43:10 <DorpsGek_III> [OpenTTD/eints] TrueBrain merged pull request #24: Add: allow "python -m webtranslate.main" to use a static config.xml https://git.io/JUda4
20:43:31 <frosch123> oh, i rather not tell what hacks i put into the dockerfile
20:43:55 <frosch123> but i tuned my migration script to output more gollumish content and add lfs to it
20:44:03 <frosch123> just i forgot that it takes 2 hours to run
20:46:38 <DorpsGek_III> [OpenTTD/eints] TrueBrain merged pull request #23: Fix: Users without valid session cookie only have an instance of user… https://git.io/JUdac
20:46:56 <frosch123> oh dear, sometimes i wish i did not know german... then i would not have to deal with what some translators write
20:51:28 <frosch123> italians took the pole position: 4 translators
20:53:30 <TrueBrain> frosch123: your PR broke GitHub login :(
20:53:36 *** sla_ro|master has joined #openttd
20:54:51 <frosch123> hmm, right, i only checked the negative case
20:55:05 <TrueBrain> isinstance instead of is?
20:55:24 <frosch123> long ago since i checked a class in python :)
20:55:39 <TrueBrain> I assumed userauth was the class .. but if it is the instance:
20:55:46 <TrueBrain> isinstance(userauth, GitHubblabla)
20:56:06 <TrueBrain> I should have realised it during review :P
20:56:25 <TrueBrain> but, this is what we have staging for :D
20:56:27 <frosch123> i do not longer have my test github app :)
20:56:45 <frosch123> but i guess i can fake it
20:56:57 <TrueBrain> funny, neither do I :P
21:01:12 <frosch123> hmm, it's not that easy to fake
21:01:46 <TrueBrain> make a new app, like 5 clicks? :D
21:04:29 <frosch123> that's way more clicks, i also would have to fake localhost and more again
21:04:54 <TrueBrain> I never faked localhost to test it ..
21:05:00 <TrueBrain> I just redirected it to http back :P
21:05:32 <frosch123> i needed a replace("http","https") somewhere
21:05:39 <TrueBrain> (on GitHub, redirecturl can be http ONLY for localhost :))
21:06:20 <frosch123> oh, gh accepts localhost? either that is new, or i just assumed that it would not work...
21:06:35 <frosch123> but i mean the requests_oauth, which denies http
21:06:36 <TrueBrain> it is "new", 2017 I believe :P
21:07:07 <TrueBrain> owh, yes, you need to set an ENV variable for requests_oauth :)
21:07:28 <frosch123> haha, i just faked the caller url in the source :p
21:07:47 <TrueBrain> OAUTHLIB_INSECURE_TRANSPORT=1 :)
21:07:49 <frosch123> you want https? i pretend it was :)
21:08:00 <TrueBrain> I found that because of stackoverflow
21:08:05 <TrueBrain> not because ... it was clearly documented :P
21:15:35 <frosch123> also when pressing "back"
21:17:13 <TrueBrain> to production it is!
21:17:25 <TrueBrain> 2 sentry errors on 1 day .. fixed within an evening \o/
21:17:37 <andythenorth> lovin it lovin it lovin it
21:18:22 <frosch123> on devzone i had custom crafted code to file issues in redmine on exceptions :)
21:22:17 <TrueBrain> sentry allows you to do that manually :)
21:22:27 <TrueBrain> and their deduplication is awesome
21:23:16 <TrueBrain> I really cannot understand projects not using Sentry :D
21:23:36 <TrueBrain> like for this bug, I marked it as: resolved in next release
21:23:52 <TrueBrain> it will not send me any alert for that bug until it has seen a new version
21:23:54 <TrueBrain> which is rather nice :)
21:24:20 <TrueBrain> (and this is why the slightly complicated version-bla is in the Python repositories :D)
23:01:58 *** WormnestAndroid has quit IRC
23:01:59 *** WormnestAndroid has joined #openttd
continue to next day ⏵