IRC logs for #openttd on OFTC at 2020-10-03
⏴ go to previous day
08:41:47 *** berndj-blackout has quit IRC
09:25:43 *** jottyfan has joined #openttd
09:35:47 *** andythenorth has joined #openttd
09:49:33 *** sla_ro|master has joined #openttd
10:18:57 *** andythenorth has joined #openttd
10:20:13 *** Progman has joined #openttd
10:33:59 <DorpsGek_III> [OpenTTD/bananas-frontend-cli] TrueBrain opened pull request #13: Fix: always show a response in the browser during authentication https://git.io/JUQ7W
10:40:42 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain opened pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQ7F
11:00:13 <DorpsGek_III> [OpenTTD/bananas-frontend-cli] LordAro approved pull request #13: Fix: always show a response in the browser during authentication https://git.io/JUQdV
11:02:19 <DorpsGek_III> [OpenTTD/bananas-frontend-web] LordAro commented on pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQdX
11:04:04 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain commented on pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQdF
11:04:14 <DorpsGek_III> [OpenTTD/bananas-frontend-cli] TrueBrain merged pull request #13: Fix: always show a response in the browser during authentication https://git.io/JUQ7W
11:05:26 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain commented on pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQdh
11:07:40 <DorpsGek_III> [OpenTTD/bananas-frontend-web] LordAro commented on pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQFL
11:12:21 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain commented on pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQFg
11:12:56 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain commented on pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQFa
11:12:59 <TrueBrain> too much typing! :P
11:13:07 <TrueBrain> this really broke me head several times today already
11:14:57 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain updated pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQ7F
11:15:04 <TrueBrain> added to the commit message that it is not only the SAST that could get confused :)
11:16:10 * andythenorth has an annoying cold
11:16:50 <LordAro> quick, kick him before he infects us all!
11:23:23 <TrueBrain> LordAro: if you are better suggestions, let me know. But how ever I run this circle .. this is just an annoying flask-implementation-detail .. and I should never have mentioned it is :D
11:23:33 <TrueBrain> if there was a "return" from the start, nobody would think twice about it
11:24:35 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #33: Fix CodeQL warning and a sentry crash report https://git.io/JUQbK
11:30:58 <DorpsGek_III> [OpenTTD/DorpsGek] TrueBrain opened pull request #36: Various of fixes for problems found by CodeQL https://git.io/JUQNl
11:31:38 <TrueBrain> and this concludes all the CodeQL alerts in the repositories I am fixing them in :) I think ... :P
11:31:55 *** SpComb^ has joined #openttd
11:33:26 <LordAro> TrueBrain: it's like "return os.exit(1)" though, you'd definitely say that was weird
11:33:46 <TrueBrain> because you assume every human knows what os.exit() does
11:33:59 <TrueBrain> what flask.abort() does .. I honestly would expect it to return a web-result
11:34:04 <TrueBrain> not raise an exception
11:34:14 <TrueBrain> but this leads to NOT_REACHED() statements
11:34:21 <TrueBrain> which, as you might know, are ugly as fuck :P
11:34:50 <TrueBrain> but for example the redirect() slightly above it, does return a web-result
11:35:16 <TrueBrain> so flask.redirect() does something else than flask.abort()
11:35:32 <TrueBrain> and this is why I mention: any comment would only make the user go: huh?
11:35:40 <TrueBrain> while no comment with a return would go a user like: sure!
11:36:38 <TrueBrain> it really would have been better if their docs said: you should use "raise flask.abort()"
11:36:42 <TrueBrain> THAT would be very clear
11:36:46 <DorpsGek_III> [OpenTTD/bananas-server] LordAro approved pull request #33: Fix CodeQL warning and a sentry crash report https://git.io/JUQAI
11:36:58 <TrueBrain> bananas-server one failed to fix the alert :(
11:37:49 <TrueBrain> tempted to write it off as a false-positive after the PR
11:37:52 <DorpsGek_III> [OpenTTD/DorpsGek] LordAro approved pull request #36: Various of fixes for problems found by CodeQL https://git.io/JUQAm
11:38:27 <TrueBrain> LordAro: I think I know the better solution to abort(): not use it!
11:40:10 *** SpComb is now known as SpComb_
11:40:28 *** SpComb^ is now known as SpComb
11:41:49 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain updated pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQ7F
11:41:53 <TrueBrain> solved the argument :)
11:42:18 <DorpsGek_III> [OpenTTD/DorpsGek] TrueBrain merged pull request #36: Various of fixes for problems found by CodeQL https://git.io/JUQNl
11:55:53 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain dismissed a review for pull request #33: Fix CodeQL warning and a sentry crash report https://git.io/JUQAI
11:55:53 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain updated pull request #33: Fix CodeQL warning and a sentry crash report https://git.io/JUQbK
11:55:58 <TrueBrain> let's see if this fixes the alert
12:06:33 <TrueBrain> yeah, turns out the problem was even worse .. as not_found() was used in a proper workflow, and abort() was completely unneeded :)
12:07:09 <TrueBrain> so .. did it fix it? or didn't it? wtf? :P
12:10:45 <TrueBrain> what is unclear in documentation, if you want the "open" to be in the __enter__ or in the __init__
12:10:51 <TrueBrain> the example in the Python docs suggest the latter
12:11:00 <TrueBrain> but that still allows leaking of the fd honestly
12:11:15 <TrueBrain> but, if you want to shoot yourself in the foot
12:11:26 <TrueBrain> I am now 99.99% sure the fd is always closed in this codebase, and will consider it now a false positive :D
12:16:59 *** frosch123 has joined #openttd
12:18:38 <DorpsGek_III> [OpenTTD/bananas-frontend-web] LordAro approved pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQjv
12:20:43 <DorpsGek_III> [OpenTTD/bananas-server] LordAro approved pull request #33: Fix CodeQL warning and a sentry crash report https://git.io/JUQj3
12:21:33 <DorpsGek_III> [OpenTTD/bananas-frontend-web] TrueBrain merged pull request #45: Various of fixes for problems found by CodeQL https://git.io/JUQ7F
12:21:42 <LordAro> i also "officially" resolved the "open" as a false positive
12:21:45 <LordAro> as apparently i can do that
12:21:58 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #33: Fix CodeQL warning and a sentry crash report https://git.io/JUQbK
12:22:03 <LordAro> bet it doesn't persist after the merge..
12:23:37 *** andythenorth has joined #openttd
12:29:16 <TrueBrain> Did he delete his comment?
12:29:46 <TrueBrain> Was afraid DorpsGek was acting up :p
12:33:24 <LordAro> oh yes, i didn't quite finish blackening nml
12:34:06 <LordAro> and black CI failed inexplicably
12:36:45 <TrueBrain> well, "inexplicably"..
12:36:49 <TrueBrain> syntax error near unexpected token `('
12:36:55 <TrueBrain> black -l 120 --exclude (generated|nml/actions/action2var_variables.py|nml/actions/action3_callbacks.py) --check nml
12:37:02 <TrueBrain> I can add 1 to 1, and get to 2, with this information ;)
12:37:14 <LordAro> i thought it was during the pip install process
12:37:23 <TrueBrain> not sure if "inexplicably" means "not reading", in your case :P
12:37:38 <TrueBrain> counting to three is difficult, I know :P
12:38:00 <TrueBrain> I like that on the top of a GitHub Actions it tells you now what it is going to execute
12:38:10 <TrueBrain> but yeah, it is not immediately clear which line really failed
12:43:52 <TrueBrain> funny, black does have an GitHub Action made by upstream
12:43:56 <TrueBrain> just .. with no configuration settings
12:44:12 <LordAro> that is their style :p
12:44:49 <TrueBrain> you can make a pyproject.toml with settings in there
12:45:18 <TrueBrain> we might want to do that for all projects tbh, and just use the upstream GitHub action
12:48:28 <TrueBrain> I rather have that composite workflows work with more things besides "run", as that would allow us to put a "black" workflow on a single place
12:49:00 <TrueBrain> actions in actions would really help, basically :)
12:49:39 <TrueBrain> in fact, it would make most of the "actions" repository redundant :D
12:49:54 <LordAro> you want to work on that? :p
12:50:06 <LordAro> given i still haven't gotten my head around how they're supposed to work anyway
12:50:20 <TrueBrain> no, sorry, you misunderstood me: GitHub doesn't support this yet
12:50:31 <TrueBrain> their only "supported" composite workflows are with "run" statements
12:50:35 <TrueBrain> you cannot do actions in actions
12:50:49 <TrueBrain> and I would love to implement it, but I don't work for GitHub :P
12:51:39 *** ChanServ sets mode: +v tokai
12:54:35 <TrueBrain> funny, you can now select that you can only use external GitHub Actions that are made by verified creators
12:54:49 <TrueBrain> but there is no way to become a verified creator atm, unless you have ties with GitHub as organisation
12:54:54 <TrueBrain> that is not fair :P
12:55:40 <TrueBrain> I really do like on forks "schedules" become disabled by default now .. as that was so silly it wasn't :)
12:59:24 <LordAro> imagine knowing more than two languages
13:02:16 <TrueBrain> I am Dutch; that is not hard to imagine :P
13:05:33 <LordAro> I am English; I think I could still do some very basic French if I really worked at it? :p
13:07:58 <TrueBrain> I learnt 6 languages during high school .. 2 of them were dead languages :P
13:08:19 <frosch123> 2 dead languages? what's the other one?
13:08:25 <TrueBrain> Old Greek and Latin
13:08:47 <frosch123> old greek, fancy :)
13:08:51 <TrueBrain> bonus: I can read signs in Greece just fine :P
13:09:02 <frosch123> oh, i can do that too, because maths
13:09:10 <TrueBrain> well, I remember 3 words from old greek .. that is it .. so yeah :P
13:09:34 <TrueBrain> pretty sure it was a waste of brainpower :P
13:09:52 <TrueBrain> Latin is not much better, but I have those flashbacks from time to time :P
13:10:12 <frosch123> i liked latin because i liked roman stuff
13:10:47 <frosch123> and when english moved from learning about talking english to talking about english literature, latin was actually more useful
13:11:01 <TrueBrain> shit like: a ae ae am, sometimes hit my brain for no reason WHAT SO EVER :P
13:11:46 <TrueBrain> wait, that was greek ..
13:11:56 <TrueBrain> a ae ae am a, I guess
13:12:34 <TrueBrain> never understood why they had 5 of those ....
13:12:48 <frosch123> TrueBrain: they had 7
13:13:05 <TrueBrain> well, yes, but you can ignore 2 :P
13:13:18 <frosch123> and even funnier, i learned from smatz that czech has the same cases to this day
13:13:31 <TrueBrain> old Dutch had several too ...
13:13:37 <TrueBrain> happy we are down to 3 :)
13:13:51 <TrueBrain> German has still 4, right?
13:13:54 <frosch123> rescue the genitive!
13:14:20 <frosch123> officially we have 4, but one is talked dead like ottd :p
13:14:52 <TrueBrain> I remember we had to learn that .. for some shitty reason ..
13:15:20 <TrueBrain> owh, Dutch also knows 4 officially, I read ...
13:17:27 <TrueBrain> and of course the shitty things like: tu te tui tibi te .. ugh :P
13:17:38 <TrueBrain> well, honestly, it helps sometimes in understanding languages you don't know ..
13:18:43 <LordAro> like programming languages, i suppose
13:18:44 <TrueBrain> or: sum es est sumus estis sunt
13:19:12 <TrueBrain> or, depending on the .. what is the english word for that .. "time" ?
13:19:18 <TrueBrain> eram eras erat eramus eratis erant
13:19:20 <frosch123> looks like your memory is better than mine :)
13:19:28 <TrueBrain> I am also looking shit up, don't worry :P
13:19:43 <TrueBrain> Indicativus, in Latin :P
13:20:00 <frosch123> indicative, conjunctive
13:20:10 <TrueBrain> I learnt Dutch a lot better by learning Latin ..
13:20:19 <TrueBrain> never knew the mathematical connection between tenses etc
13:20:24 <TrueBrain> (I am dyslectic .. sue me :P)
13:20:46 <TrueBrain> by learning a language from the theory, it really helped understanding my own language
13:22:29 <TrueBrain> Dutch is a very simple language to learn :P
13:22:37 <TrueBrain> also relative pure, in that what you write is what you speak, mostly
13:23:08 <frosch123> i think that applies to most languages, except english
13:23:20 <TrueBrain> some are more pure than others :)
13:23:46 <frosch123> also the cow/beef, swine/pork thingie is funny
13:24:33 <frosch123> the animals use the german-derived names, the meat uses the french-derived names
13:24:46 <TrueBrain> sounds about right :P
13:25:08 <frosch123> which apparently reflects the difference between french rulers and germanish workers
13:25:08 <TrueBrain> I like in Dutch, and I guess other languages have that too, that it is heavily evolving .. we care less and less about certain aspects
13:25:47 <TrueBrain> "van jouw" used to be a big no-no (it is "van jou") .. but less and less people care :P
13:26:12 <TrueBrain> it comes and goes, of course, but it is funny
13:26:27 <TrueBrain> "as long as you can understand each other, who gives a rats ass", basically :P
13:27:13 <frosch123> i only understand dutch in paragraphs. you need enough context to guess the words :)
13:27:24 <TrueBrain> the main part in Dutch where people are still picky on nom/gen/dat/acc shit, is with "zij/hun/hen", dutch for yours
13:27:32 <frosch123> i guess that's also how i translated latin mostly
13:27:34 <TrueBrain> there people go bananas on using it correctly
13:27:53 <TrueBrain> yeah .. I passed Latin on being able to understand context
13:28:02 <TrueBrain> what I often had wrong, was that I picked the inverse
13:28:07 <TrueBrain> as I missed the negate in the sentence
13:28:23 <TrueBrain> so Camilia was NOT angry at the dog
13:29:06 <frosch123> oh, yeah, when you got something wrong in the beginning of a text, and you guessed the rest of the text with that, your translations turns worse and worse
13:29:13 <TrueBrain> I also learnt my appreciation for long sentences, connected with commas; and let's not forget the side-sentences you can create
13:29:25 <TrueBrain> Caesar was my hero, with a single sentence over 6 pages
13:29:50 <frosch123> now you are exagerating
13:29:52 <TrueBrain> s/I also learnt/I also learnt from latin/
13:30:22 <TrueBrain> sadly, my exam was about Caesar .. the first 3 words and the last 4 were pages apart
13:30:25 <TrueBrain> he was a hero in that kind of shit
13:33:14 <TrueBrain> right, what was I doing .. owh, yeah, deploying all the new stuff to production ..
13:36:50 <TrueBrain> hmm, pyup never made PRs .. let me see if I can fix that ..
13:37:38 <TrueBrain> well, that clearly is not working
13:39:23 <TrueBrain> there isn't a button: make PR now :P
13:44:23 *** DorpsGek_III has joined #openttd
13:45:40 <TrueBrain> I guess if pyup doesn't want to do it, I have to, or something
13:47:24 *** andythenorth has joined #openttd
13:51:46 <LordAro> that was a bit of effort
13:53:30 <LordAro> and apparently i missed one
13:55:30 <LordAro> in fact, it's a false positive
14:04:47 <LordAro> FLHerne: how goes that cyclic import fixing?
14:07:15 <TrueBrain> right, shipped everything I think :)
14:55:55 <Flygon> I got so excited for a moment there too.
14:56:03 <Flygon> I still love you LordAro. <3~
14:56:25 <LordAro> you're quite welcome to try fixing the python cyclic import issue as well, if you like
14:56:45 <Flygon> The last time I coded something, I somehow caused a flamewar about how Nested Loops should function.
14:56:50 <Flygon> I probably shouldn't, sorry. :P
15:15:41 <LordAro> def __lt__(self, other): if not isinstance(other, Change): return False ....
15:15:45 <LordAro> seems oddly defensive, no?
15:34:42 <TrueBrain> pretty common, not?
15:35:28 <LordAro> yeah, i've changed my mind since :p
15:51:12 <DorpsGek_III> [OpenTTD/eints] TrueBrain commented on pull request #22: Fix various issues pointed out by CodeQL https://git.io/JU7Zj
15:52:33 <DorpsGek_III> [OpenTTD/eints] TrueBrain commented on pull request #22: Fix various issues pointed out by CodeQL https://git.io/JU7nI
15:53:01 <DorpsGek_III> [OpenTTD/eints] LordAro commented on pull request #22: Fix various issues pointed out by CodeQL https://git.io/JU7nt
15:53:37 <DorpsGek_III> [OpenTTD/eints] LordAro commented on pull request #22: Fix various issues pointed out by CodeQL https://git.io/JU7n3
15:54:22 <TrueBrain> LordAro: honestly, I am considering making the "click.py" into a small pip package ..
15:54:33 <TrueBrain> it is super-handy, and it is copied in ALL Python projects we have atm :P
15:54:39 <TrueBrain> I might do that tomorrow :)
15:54:49 <TrueBrain> never made a pip package before :D
15:55:41 <DorpsGek_III> [OpenTTD/eints] TrueBrain commented on pull request #22: Fix various issues pointed out by CodeQL https://git.io/JU7nW
15:55:50 <TrueBrain> I think I am going to try it without setup.py
15:55:53 <TrueBrain> seems to be the new thing
15:57:49 <TrueBrain> a return after an abort .. I see here the same weird use of "abort" creeped in
15:58:03 <LordAro> yeah, redirect is the same
15:58:19 <TrueBrain> and they don't even mean "abort" I guess, more "error"
15:58:19 <LordAro> i just left it as is for simplicity
15:58:31 <TrueBrain> yeah, lets not try to fix everything weird in eints :D
15:58:39 <TrueBrain> bottle is one hell of a weird library
15:59:15 <TrueBrain> we now have a mix of aiohttp and flask :)
15:59:29 <TrueBrain> guess flask -> aiohttp would be the first to tackle here :)
15:59:36 <TrueBrain> (in other projects, not in eints)
15:59:43 <TrueBrain> flask is a single-threaded server atm, in production
15:59:52 <TrueBrain> but aiohttp maturity is not there yet, tbh
16:00:20 <LordAro> no? seems to work well enough for all usages i've seen it in
16:00:47 <TrueBrain> there are odd problems not resolved yet last I checked
16:00:51 <TrueBrain> like when a socket closes unexpected
16:01:01 <TrueBrain> but bananas-api runs it, and no production issues so far
16:01:47 <TrueBrain> lol, TIL total_ordering is a thing :)
16:03:11 <LordAro> my thought process was "this is python, there must be a builtin way of specifying the 4 other operators"
16:03:22 <LordAro> it's even mentioned in the CodeQL warning :p
16:03:37 <DorpsGek_III> [OpenTTD/eints] TrueBrain requested changes for pull request #22: Fix various issues pointed out by CodeQL https://git.io/JU7n6
16:03:54 <TrueBrain> found one nitpick :)
16:04:36 <LordAro> hmm, 40 fixes implies that there's 9 unfixed things
16:04:45 <TrueBrain> you can view those if you like
16:04:57 <TrueBrain> it shows on the top now
16:05:00 <TrueBrain> 6 alerts open on that PR
16:06:30 <TrueBrain> it is not that bad, the amount of alerts for eints
16:06:54 <TrueBrain> after you ignored bottle, ofc :)
16:07:58 <LordAro> i have to question why that code block is there if it's not necessary...
16:08:10 <LordAro> or actually, how a KeyError can be raised at all...
16:10:32 <TrueBrain> and this is why you need to add a comment if you do these kind of exception ignoring :P
16:10:40 <TrueBrain> as I am also clueless if that would ever do anything :D
16:10:59 <LordAro> i'm going to go ahead and remove it ¯\_(ツ)_/¯
16:11:09 <TrueBrain> possibly a KeyError is thrown if it already exists or something?
16:11:24 <TrueBrain> honestly, leave it for now
16:11:39 <TrueBrain> frosch123: are we dropping redmine support in eints on GitHub?
16:12:11 <TrueBrain> LordAro: the issue with changing this kind of code, is that you cannot test it .. so removing it might make it non-workable .. :P
16:12:35 <LordAro> well someone would find out when they tried to use it again :p
16:12:47 <LordAro> it's not like an issue that would only reappear a week after they started using it
16:13:44 <TrueBrain> that doesn't sound like a good plan of action :)
16:14:05 <TrueBrain> better add a comment # XXX -- we have no clue why this is here. If you ever test this, can you remove it? :P
16:14:48 <TrueBrain> or .. we can check the history on that file?
16:16:56 <TrueBrain> okay, another way to figure out if this is needed ..
16:16:58 <glx> add a log line to see how often it happens
16:17:07 <TrueBrain> glx: we don't execute that code, so never
16:19:08 <FLHerne> LordAro: I'm not really sure of the right way to do it :-/
16:19:11 *** snail_UES_ has joined #openttd
16:20:00 <TrueBrain> LordAro: yeah, but that would never happen in our case; and nobody else is doing any KeyError check for this kind of code ..
16:20:06 <TrueBrain> so yeah, okay, I think it is safe to drop it :P
16:20:19 <TrueBrain> "if db_schema is not None and db_schema != "":" <- hihi
16:20:26 <frosch123> TrueBrain: bottle and flask are different. iirc flask's abort/redirect throws, while bottle's does not
16:20:41 <FLHerne> I think the right solution is probably to make all/most of the Operators subclasses of BinOp, but that's another big change
16:20:43 <TrueBrain> frosch123: then that code would be broken .. hmm .. lets check
16:21:01 <TrueBrain> raise HTTPError(code, text)
16:21:04 <TrueBrain> frosch123: bottle raises too
16:21:29 <TrueBrain> frosch123: redirect is different: flask returns, bottle raises
16:22:13 <frosch123> oh, the other way around :p
16:22:20 <frosch123> well, confusing in any case
16:22:26 <TrueBrain> :D Yup, confusing it is!
16:24:58 <TrueBrain> LordAro: why the change in indentation for pages/login.py:43?
16:25:15 <TrueBrain> willy function is silly
16:25:22 <LordAro> TrueBrain: yah, return value
16:25:26 <LordAro> was unnecessary in the else
16:25:51 <TrueBrain> that function reads really weird, with some raising exceptions, others returning values :D
16:26:03 <TrueBrain> (not your fault or something to fix)
16:27:16 <DorpsGek_III> [OpenTTD/eints] TrueBrain commented on pull request #22: Fix various issues pointed out by CodeQL https://git.io/JU7Ch
16:29:28 <LordAro> TrueBrain: i'm reasonably certain that the read is unnecessasry
16:29:43 <TrueBrain> I would love to believe you
16:29:48 <TrueBrain> but that would mean you have to test the script :P
16:30:02 <LordAro> just replace urllib with requests :p
16:30:05 <TrueBrain> not a big fan of YOLOing these things :)
16:30:15 <TrueBrain> (as it will be WEEKS before anyone notices it)
16:30:31 <TrueBrain> replace bottle with aiohttp, replace urllib with requests .. you have a big list of things to do :P
16:31:28 <TrueBrain> now at least we know for sure nothing goes BOOM because of this PR ;)
16:31:59 <TrueBrain> did you btw start eints after your changes, just to see if the Python code loads etc?
16:32:31 <LordAro> i feel like you want me to say yes
16:32:40 <TrueBrain> I feel like you didn't, and should ;)
16:32:51 <TrueBrain> "make run" is easy ;)
16:34:10 <LordAro> TrueBrain: it seems to start with an empty data dir and empty rights.dat file :)
16:34:56 <TrueBrain> in my experience, these kind of changes tend to fail in simple: does it start ... from experience of doing the other repositories :P
16:35:08 <TrueBrain> and we are lacking a good test for that
16:35:13 <TrueBrain> (in most repositories)
16:35:25 <TrueBrain> bananas-api is the only one with a regression-test, I believe :)
16:35:47 <DorpsGek_III> [OpenTTD/nml] LordAro commented on pull request #164: Fix: Remove trailing whitespaces in NFO output https://git.io/JU7WP
16:37:42 <DorpsGek_III> [OpenTTD/nml] glx22 commented on pull request #164: Fix: Remove trailing whitespaces in NFO output https://git.io/JU7Wb
16:38:49 <DorpsGek_III> [OpenTTD/nml] LordAro commented on pull request #164: Fix: Remove trailing whitespaces in NFO output https://git.io/JU7lU
16:42:05 <TrueBrain> LordAro: CodeQL says no alerts now :o nice :D
16:42:05 <DorpsGek_III> [OpenTTD/OpenTTD-git-hooks] LordAro approved pull request #14: Fix: Check ObjC files as other source files https://git.io/JU7lZ
16:42:13 <DorpsGek_III> [OpenTTD/OpenTTD-git-hooks] LordAro merged pull request #14: Fix: Check ObjC files as other source files https://git.io/JUQfg
16:42:41 <TrueBrain> oeh, it even shows the CodeQL things inline in the PR
16:43:09 <LordAro> TrueBrain: i'll let you do the openttd-github branch fiddling
16:43:56 <TrueBrain> feel free to rebase openttd-github and force push
16:44:02 <TrueBrain> if possible, with that error fixed :P
16:44:34 <TrueBrain> it is a bit of an ugly way we are doing this atm btw :)
16:44:42 <TrueBrain> nobody really reviews the commits in openttd-github
16:45:12 <TrueBrain> not sure we can make that a bit better
16:45:31 <LordAro> could have PRs for syncing the branches
16:45:40 <LordAro> with merge commits and everything
16:45:50 <TrueBrain> that is the issue, it will be littered with merge commits
16:46:35 <TrueBrain> guess the more fundamental question is, will the eints version from GitHub ever be used for anything else in the (not so near) future :D
16:47:36 <TrueBrain> anyway, I can sync openttd-github branch after dinner LordAro , if you haven't already :)
16:48:56 <TrueBrain> that's what she said
16:51:26 <TrueBrain> the openttd-ldap branch will still throw errors, I guess :P We haven't rebased that in ... months
16:52:04 <LordAro> i should be able to retrieve the old commit hash from my fork if i've broken everything :p
16:52:39 <DorpsGek_III> [OpenTTD/nml] LordAro approved pull request #164: Fix: Remove trailing whitespaces in NFO output https://git.io/JU78Y
16:53:11 <TrueBrain> owh, openttd-ldap is not scanned
16:53:13 <TrueBrain> sweet, so not an issue :)
16:53:27 <LordAro> only becomes an issue if anyone ever updates it :p
16:53:42 <TrueBrain> which is very unlikely to ever happen again :)
16:54:42 <LordAro> glx: re nml#78, you removed the commits that were causing all the slowness, right?
16:55:08 <glx> need to check, was a long time ago
16:55:26 <LordAro> Yexo approved it, so presumably so
16:55:50 <DorpsGek> LordAro: Yexo was last seen in #openttd 16 weeks, 4 days, 18 hours, 16 minutes, and 3 seconds ago: <Yexo> Oh well, not worth spending more time on, still interesting
16:56:10 <TrueBrain> that does some up #openttd tbh :P
16:56:18 <LordAro> i was thinking the same :p
16:56:34 <LordAro> glx: might as well rebase it while you're at it anyway
16:56:59 <LordAro> certainly can't see any loops anymore
16:57:15 <glx> yeah I'm rebasing #155 now, will do #78 after
16:57:45 <TrueBrain> lets push it to production, why not
16:57:59 <FLHerne> LordAro: I need to finish RL stuff now (right now my housemate and I can't get to the fridge because there's a half-built canoe blocking it)
16:58:10 <FLHerne> But I'll fix my PR later
16:58:17 <TrueBrain> sounds like a TERRIBLE place to build a canoe
16:58:42 <FLHerne> Well, it was outside, but it's raining and forecast to do so all week
16:58:58 <FLHerne> And I can't put it where I was going to inside, because there's stuff in the way
16:59:16 <TrueBrain> sounds more important than NML for sure :)
16:59:35 <FLHerne> So it's like one of those 15-piece puzzles, and the empty square is currently not the kitchen :p
17:00:56 <TrueBrain> LordAro: 2 alerts left in the openttd-github branch
17:01:13 <TrueBrain> seems CodeQL doesn't know about delayed imports
17:01:57 <TrueBrain> second one I don't really understand
17:02:12 <TrueBrain> owh, no, they are the same
17:02:15 <LordAro> isn't it just the same issue?>
17:02:23 <TrueBrain> going to ignore them both honestly
17:02:26 <TrueBrain> it is not the pretiest
17:03:20 <LordAro> fixing the warnings by disabling the warnings
17:03:38 <LordAro> andythenorth: should be able to
17:03:45 <LordAro> though adding flake first might be easier
17:05:08 <andythenorth> I'm just curious how far I've diverged from anything considered 'good'
17:05:25 <andythenorth> I use pyflakes and auto-pep8 sometimes, but not systematically
17:05:39 <andythenorth> and I PEP8 troubles me also
17:06:20 <LordAro> i think there's some settings you have to hit on GH itself - TrueBrain knows about them
17:06:37 <LordAro> andythenorth: oh, and change the path that flake8 is using
17:07:17 <andythenorth> I was playing warcraft 1 on fastest speed, and now my brain is all out of time
17:07:25 <andythenorth> everything seems weirdly speeded up
17:08:26 <TrueBrain> No other settings needed. Just that file in the right place :)
17:08:48 <LordAro> TrueBrain: so, merging my Actions CI PR
17:08:59 <LordAro> i think you said there's something i need to disable?
17:10:53 *** HerzogDeXtEr has joined #openttd
17:13:20 <TrueBrain> To merge it you have to change which checks are mandatory
17:13:33 <TrueBrain> In the branch protection
17:14:51 *** Wormnest has joined #openttd
17:15:03 <LordAro> do the new ones need enabling? (after it's been merged)
17:15:35 <glx> new ones need to be required yes
17:16:03 <TrueBrain> And no need to do it after, just before is fine too
17:17:26 <glx> and now all PRs need a rebase :)
17:26:24 <LordAro> why has vcpkg failed?
17:27:24 <LordAro> ...because repo.msys2.org appear to be down
17:29:16 <LordAro> doesn't look like there's a "rerun CI" button anywhere either
17:29:47 <glx> there's a rerun failed checks
17:30:13 <LordAro> ah yes, it's there now
17:32:19 <LordAro> and shows up all the warnings from the OSX build :p
17:33:41 <glx> having them displayed in all PRs will maybe let someone decide to fix them :)
17:34:06 <LordAro> a couple of them are relatively doable
17:34:12 <LordAro> have no idea about this Gestalt stuff though
17:40:46 <DorpsGek_III> [OpenTTD/nml] glx22 dismissed a review for pull request #155: Add: Support for parameters in switches https://git.io/JU7ls
17:42:58 <TrueBrain> So add the MSVC already
17:43:09 <TrueBrain> Of the same dude, predictable repo
17:43:35 <LordAro> TrueBrain: look closer
17:43:59 <TrueBrain> So why isn't it working? :p
17:49:45 *** innocenat_ has joined #openttd
18:01:15 <DorpsGek_III> [OpenTTD/nml] LordAro commented on pull request #28: Fix: Allow access to towns as parent object https://git.io/JU7uA
18:02:37 <glx> oups forgot to check regression before pushing
18:14:55 <glx> ok optimising doesn't work well with the recent NML additions
18:17:30 <glx> nfo seems correct, but looks so unoptimised
18:21:17 <glx> well it inlined the content of a procedure
18:22:10 <glx> and kept the procedure code too
18:23:06 *** gelignite has joined #openttd
19:51:40 *** snail_UES_ has joined #openttd
19:51:52 <DorpsGek_III> - Update: Translations from eints (by translators)
19:57:10 <andythenorth> what's the best format for uploading to Bananas?
19:57:19 <andythenorth> individual files? tar? zip?
19:57:34 <andythenorth> currently Horse makefile produces a tar which bananas can't accept
19:58:25 <andythenorth> it contains license.txt
19:58:34 <andythenorth> which is needed when I distribute the tar manually
19:58:43 <andythenorth> I might need a bananas specific target
19:58:46 <frosch123> ah yes, my work around was: upload tar, delete license via the gui
20:00:55 <andythenorth> least work is just to manually upload required files I think
20:08:48 <andythenorth> I can drop license.txt easily enough
20:08:58 <andythenorth> but if I accidentally distribute that it's a GPL variation
20:09:09 <andythenorth> and Ill probably forget
20:34:32 <andythenorth> I've made a tar which bananas will accept
20:34:36 <andythenorth> but it won't parse the files in it
20:40:24 <andythenorth> frosch123 what did I break now? :) grf.farm/misc/bananas-dist-iron-horse-2.8.0-beta-1-15-g88a888fc.tar
20:41:22 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain opened pull request #34: Fix: don't report SocketClosed when sending https://git.io/JU7DZ
20:41:50 <TrueBrain> I love sentry for reporting useful errors :D
20:43:24 <TrueBrain> andythenorth: folder in a folder in a folder in a folder
20:43:30 <TrueBrain> I hear you like nesting :)
20:44:02 <TrueBrain> but this is obvious after upload, not?
20:44:24 <andythenorth> oh that's interesting
20:44:30 * andythenorth wonders why it does that
20:44:52 <TrueBrain> I mean .. that was easy to spot :P
20:44:53 *** Gustavo6046 has joined #openttd
20:45:04 <TrueBrain> tar tvf bananas-dist-iron-horse-2.8.0-beta-1-15-g88a888fc.tar
20:45:04 <TrueBrain> iron-horse-2.8.0-beta-1-15-g88a888fc/iron-horse-2.8.0-beta-1-15-g88a888fc/iron-horse.grf
20:45:27 <TrueBrain> the upload does what it should: preserve folders .. well, it removes the first, because that is a very common mistake
20:45:31 <andythenorth> I wonder why the tar was made to do that
20:45:36 <TrueBrain> it does not remove multiple, because .. it has to stop somewhere :)
20:45:53 <andythenorth> there was probably a reason
20:46:19 <andythenorth> I guess I have to learn how this works
20:46:49 <andythenorth> what makes tars?
20:47:41 <TrueBrain> it is your code; I don't know who made you that tar :)
20:48:16 <TrueBrain> the tar is unlikely to be accepted by OpenTTD too btw :P
20:48:36 <andythenorth> I guess people have just been unpacking it manually
20:48:37 <TrueBrain> well, the grf might, but the .txt files not :D
20:48:45 <andythenorth> nobody reported it broken when I distributed in forums
20:49:32 <andythenorth> L10-11 look relevant
20:51:36 <TrueBrain> enable verbose, and you can see what it is doing?
20:53:17 <TrueBrain> I do not understand most of that code, so I cannot help you with that :)
21:04:27 * andythenorth read the docstrings
21:48:50 *** HerzogDeXtEr has joined #openttd
21:59:29 <TrueBrain> frosch123 / LordAro : would you mind approving https://git.io/JU7DZ ? Would like to have this fixed on production ASAP :D
22:01:32 <DorpsGek_III> [OpenTTD/bananas-server] LordAro approved pull request #34: Fix: don't report SocketClosed when sending https://git.io/JU77Z
22:01:51 <DorpsGek_III> [OpenTTD/bananas-server] TrueBrain merged pull request #34: Fix: don't report SocketClosed when sending https://git.io/JU7DZ
23:08:24 <Afdal> Can I just say I really hate the new Maximum Reliability algorithm
23:08:41 <Afdal> The rule now seems to be that the newer your vehicle the more it sucks
23:10:03 <Afdal> For example, why would I ever upgrade to a Lev2, Lev3, or a Lev4 when they apparently can't beat the reliability of a Lev1 as a rule?
23:10:23 *** Wormnest_ has joined #openttd
23:11:08 <Afdal> At least with the more random older system you could reasonably expect at least -some- top tier vehicles to have really good reliability
23:12:20 <Afdal> In fact even a basic Max reliability of 90% seems incredibly rare now
23:12:29 <Afdal> That is extremely not cool :(
23:13:14 <Afdal> You're giving me yet another reason to learn C++ and finally give OpenTTD a sensible breakdown system
23:13:22 <Afdal> I don't have time for that -_-
23:16:30 <Afdal> I actually kind of can't believe more people haven't complained about this change. This is close to game-ruining for me.
23:16:53 <Afdal> 85% reliability is NOT an acceptable end-game reliability
23:20:27 <FLHerne> Afdal: What new algorithm is this?
23:20:46 <Afdal> I'm trying to figure out when this change was made
23:20:52 <FLHerne> I think most serious players have breakdowns disabled anyway
23:21:14 <Afdal> I think it's been sometime around 1.0.8 or 1.0.9 maybe
23:21:45 <Afdal> is that a page with links to all the changelogs
23:22:18 <FLHerne> Well, the breakdown mechanics suck, so most people who've been playing OpenTTD long enough to get fed up of it turn it off :p
23:22:30 <FLHerne> (judging empirically by comments over the years)
23:22:42 <FLHerne> I'm not saying that's a good thing...
23:23:17 <Afdal> Seems like 1.9.0 was when this major change was made
23:24:53 <Afdal> Change: Synchronise introduction date and reliability randomness across vehicles with the same base introduction date (#7147)
23:25:14 <FLHerne> I don't see how that change would cause the behaviour you describe
23:25:17 <Afdal> what this does now is ensure that every vehicle in the same class has worse reliability than the previous one
23:25:40 <Afdal> Basically meaning it's always a tradeoff between more speed/power/cargo and your vehicles breaking down more
23:26:17 <Afdal> At least I think this is what's causing this behavior
23:26:41 <FLHerne> Or at least, it's not intended to do that and I can't see how it possibly could
23:27:01 <Afdal> It's definitely real though, I've observed it in many games now and the max. reliability randomizer definitely used to work defferently
23:27:11 <FLHerne> It has the effect that all vehicles introduced *on the same date* have the same random adjustment to reliability
23:27:42 <Afdal> Well that seems like a fine enough change on its own
23:28:13 <Afdal> Nothing else says "reliability" on this changelog list though I:>
23:28:13 <FLHerne> (which is basically a side-effect of making them all have the same random adjustment to the intro date)
23:30:42 <Afdal> My preferred change to the breakdown algorithm is to simply allow players to set a %reliability at which breakdowns are impossible
23:31:06 <Afdal> So as long as you keep a vehicle serviced above that %, your train networks can run deterministically :)
23:32:36 <FLHerne> But assuming your servicing is deterministic, that's just 'breakdowns disabled' with added inconvenience :p
23:32:58 <FLHerne> (yes, I know, the entire game is a pointless inconvenience...)
23:33:16 <Afdal> At least it preserves breakdowns as an actual mechanic
23:33:35 <Afdal> anyway that's a different topic
23:33:47 <Afdal> I wanna stay on this new max. reliability algorithm
23:33:56 <Afdal> Could it be related to the nerfing of reset_engines?
23:34:07 <Afdal> (it no longer alters reliabilities randomly)
23:36:57 <Afdal> I'm looking at a huge vehicle list by introduction date and it seems it is very strongly correlated to maximum reliability
23:37:18 <Afdal> So I think it might be related to that change we discussed, even if it doesn't quite describe it
23:37:49 <FLHerne> I still don't see how
23:38:00 <Afdal> nothing introduced after 2018 for example in this list can get max. reliability above 83%
23:38:02 <FLHerne> Reliability has always been correlated to intro date
23:38:17 <Afdal> No it used to be more randomized
23:38:23 <FLHerne> It starts out terrible, increases over the first few years of model life, then slowly tails off
23:38:38 <FLHerne> So the most recent vehicles are always unreliable
23:38:45 <Afdal> I mean the maximum reliability -after- it's crept upward to reach its final value
23:39:06 <Afdal> usually takes a couple years to reach its final value
23:39:23 <Afdal> I'm comparing all maximum reliabilities at the year 2100
23:39:35 <Afdal> so no need to worry about that effect
23:40:44 <Afdal> this train designed in 1926 at 90% max reliability is the most reliable in this entire list of ~50 engines
23:55:18 <Afdal> I'm noticing on some newGRFs
23:55:40 <Afdal> the time~max.reliability correlation holds except past a certain range of years
23:55:54 <Afdal> it might be within a 255 year window
23:56:44 <Afdal> like for example something introduced in 1910 might have the best max. reliability among most other vehicles but then something introduced in 1850 has junk reliability
continue to next day ⏵
