IRC logs for #openttd on OFTC at 2020-12-08
⏴ go to previous day
00:04:00 <DorpsGek> [OpenTTD/OpenTTD] orudge dismissed a review for pull request #8365: Fix: FreeType is still required on macOS https://git.io/JIBRB
00:19:59 *** TinoDidriksen has joined #openttd
00:20:33 *** TinoDidriksen is now known as Guest7818
00:43:45 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
00:49:29 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIBKS
00:50:41 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #8361: Fix: useless warning with -snull and no BaseSounds available https://git.io/JIBKx
01:09:58 *** Cursarion has quit IRC (Ping timeout: 480 seconds)
01:14:37 *** Cursarion has joined #openttd
01:35:34 *** phil[m] has joined #openttd
02:14:32 *** nartir[m] has joined #openttd
02:19:05 *** robert[m] has joined #openttd
02:32:33 *** Wormnest has quit IRC (Quit: Leaving)
02:53:10 *** cawal[m] has joined #openttd
02:55:52 *** hamstonkid[m] has joined #openttd
03:15:51 *** fjodor[m] has joined #openttd
03:43:20 *** pothyurf[m] has joined #openttd
03:52:53 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
03:55:34 *** gretel[m] has joined #openttd
04:05:47 *** dave[m] has joined #openttd
04:05:47 *** jeremy[m] has joined #openttd
04:05:47 *** magdalena[m] has joined #openttd
04:05:47 *** iarp[m] has joined #openttd
04:05:47 *** natalie[m] has joined #openttd
04:05:47 *** labs[m] has joined #openttd
04:05:47 *** albert[m] has joined #openttd
04:12:04 *** albert[m] has quit IRC (Server closed connection)
04:12:09 *** albert[m] has joined #openttd
04:28:52 *** labs[m] has quit IRC (Server closed connection)
04:28:57 *** labs[m] has joined #openttd
04:41:27 *** Maarten^ has joined #openttd
04:45:54 *** Maarten has quit IRC (Ping timeout: 480 seconds)
04:45:54 *** Maarten^ is now known as Maarten
04:51:16 *** natalie[m] has quit IRC (Server closed connection)
04:51:21 *** natalie[m] has joined #openttd
04:54:44 *** iarp[m] has quit IRC (Server closed connection)
04:54:49 *** iarp[m] has joined #openttd
04:56:04 *** magdalena[m] has quit IRC (Server closed connection)
04:56:05 *** magdalena[m] has joined #openttd
05:05:57 *** tokai|noir has joined #openttd
05:05:57 *** ChanServ sets mode: +v tokai|noir
05:12:39 *** tokai has quit IRC (Ping timeout: 480 seconds)
06:15:09 *** Guest7818 is now known as TinoDidriksen
06:19:00 *** jeremy[m] has quit IRC (Server closed connection)
06:19:04 *** jeremy[m] has joined #openttd
06:41:08 *** dave[m] has quit IRC (Server closed connection)
06:41:10 *** dave[m] has joined #openttd
06:53:56 *** andythenorth has joined #openttd
07:04:01 *** sla_ro|master has joined #openttd
07:13:23 *** keoz has quit IRC (Ping timeout: 480 seconds)
07:17:14 *** andythenorth has quit IRC (Quit: andythenorth)
07:25:41 *** andythenorth has joined #openttd
07:28:06 *** HerzogDeXtEr has joined #openttd
08:47:11 <TrueBrain> milek7_: wait, you don't need autoconfig if you download the tarball? Interesting ...
09:04:02 <TrueBrain> milek7_: it sure does make a lot more console noise :)
09:05:28 <TrueBrain> harfbuzz does that to, so what-ever
09:05:30 <TrueBrain> and it works :D w00p
09:25:03 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain merged pull request #8361: Fix: useless warning with -snull and no BaseSounds available https://git.io/JIWab
09:25:10 <TrueBrain> I CAN MERGE PRS TOO! :P
10:42:38 <TrueBrain> for some reason I have 10k actions executed in my fork of OpenTTD .... did not know :P
10:43:13 <milek7_> usual convention is to ship tarballs with build files pregenerated, so it is not necessary to have autoconf installed
10:44:28 <TrueBrain> milek7_: it really helps, as that requires no changes to the Docker image :)
10:52:56 <TrueBrain> grrr @ GitHub API .. it is giving me a Not Found, and I don't know why :P
10:53:06 *** gelignite has joined #openttd
10:57:52 <TrueBrain> that moment you copy/paste from the API docs, and it still doesn't work, lol
11:28:47 *** Eddi|zuHause2 is now known as Eddi|zuHause
11:33:46 <andythenorth> I still don't understand daylength :)
11:33:57 <andythenorth> groundhog day conversation
11:34:00 <andythenorth> it breaks the vehicles
11:34:05 <andythenorth> it breaks industries
11:34:12 <andythenorth> it breaks town pax generation probably
11:34:18 <andythenorth> why does everybody use it?
11:35:13 <Eddi|zuHause> because people learn to live with the brokenness
11:35:47 <andythenorth> 'because people' probably enough :)
11:36:11 <andythenorth> I'm considering trying JGR again
11:36:20 <andythenorth> to find out what the game people actually play is like
11:36:35 <andythenorth> last time I tried it, I got overwhelmed, it's more bad UI in an already bad UI
11:36:57 <Eddi|zuHause> by "the people" you mean "a significant vocal minority"?
11:38:04 <Eddi|zuHause> the last thing i heard about JGR were the newgrf hacks, and that didn't look very well... or maintainable... or ... i don't know
11:38:48 <andythenorth> apparently I can turn off most things in settings
11:38:59 <andythenorth> so that JGR is pretty similar to OpenTTD
11:39:12 <andythenorth> might learn how to do that
11:39:27 <Eddi|zuHause> "allow all new things be turned off to stay original" is a very TTDPatch approach
11:40:16 <Eddi|zuHause> i see the reason why people would want that. but i also know that this leads into a settings hell
11:40:25 <LordAro> OTTD has too many settings as it is
11:41:15 <TrueBrain> new version: opinionated OpenTTD, you say?
11:41:34 <andythenorth> imagine how broken it would be
11:41:49 <TrueBrain> if you do, please leave this channel; I don't want to hear for the next 20 months how bad everything is :P
11:44:41 <andythenorth> yes social reasons stop me
11:44:44 <andythenorth> I would get kicked
11:44:50 <andythenorth> and I like hanging out here
11:44:53 <andythenorth> talking to myself
11:45:35 <TrueBrain> we both talk a lot to our self in this channel
11:45:38 <TrueBrain> we should fix that :P
11:46:00 <_dp_> settings hell is kinda inevitable with full savegame compatibility
11:47:05 <andythenorth> TrueBrain I think it's a solved problem
11:47:13 <DorpsGek> [OpenTTD/OpenTTD] orudge opened pull request #8366: Change: Don't display OS name when the user is exiting the game https://git.io/JI0Ig
11:47:13 <andythenorth> we're both happy, right?
11:47:19 <andythenorth> TrueBrain tell us, are you happy? :o
11:47:34 <LordAro> _dp_: yes, but old defaults can at least be suitably hidden
11:47:39 <TrueBrain> YES! But that is a bit unrelated to this conversation :P
11:47:55 <TrueBrain> orudge: you are my new hero :D
11:48:37 <TrueBrain> and don't fix other languages sneaky sneaky in your commit :P
11:48:52 <TrueBrain> (the removals are fine, but fixing the languages you know, pffft)
11:49:17 * orudge isn't sure what the procedure is there to be honest :P
11:49:18 <TrueBrain> I wonder how hard it is to fix them for all languages ... this is going to give annoying warnings :D
11:49:26 <TrueBrain> it mostly made me giggle :)
11:49:38 <LordAro> probably best to remove all the outdated strings from the langauges
11:49:38 <orudge> Well, I could go through and attempt to fix some of them, but when it comes to Arabic and Chinese I might struggle, short of just removing the {STRING} text
11:50:02 <TrueBrain> owh, just remove them completely?
11:50:06 <TrueBrain> that at least avoids the warnings etc
11:50:18 <LordAro> well it's defaulting back to english as it is
11:50:22 <_dp_> would be nice to send os to servers though...
11:50:36 <TrueBrain> send what to what now?
11:51:25 <LordAro> _dp_: potentially, but it's not something that needs translating :p
11:52:15 <_dp_> TrueBrain, more info about client build to servers for easier debugging. mostly os and patchpack version
11:52:41 <LordAro> 'version' should always match
11:52:53 <LordAro> if it doesn't, that's entirely on you
11:53:25 <TrueBrain> _dp_: sounds that would be solving the wrong problem :)
11:53:45 <TrueBrain> next you want to know RAM installed, diskspace available, ... :P
11:53:49 <_dp_> LordAro, there are at least 2 compatible patchpacks - cmclient and android version
11:54:20 *** Flygon_ has joined #openttd
11:54:51 <_dp_> and even though I can detect cmclient on citymani it would be still useful to have a more universal solution
11:57:56 <LordAro> i suppose we could provide a non-checked client id field that client-side patchpacks could set
11:58:02 <LordAro> but i'm not sure we want to go down that route
11:59:54 *** Flygon has quit IRC (Ping timeout: 480 seconds)
12:03:18 *** WormnestAndroid has quit IRC (Remote host closed the connection)
12:03:31 *** WormnestAndroid has joined #openttd
12:13:53 <_dp_> for one thing I'd like to know how many players use android since it's missing some crucial ui elements :P
12:14:19 <andythenorth> anonymised analytics needed :P
12:25:06 <_dp_> "worst game ever this isnt even close to the classic Sim City"
12:25:12 <_dp_> from android port reviews xD
12:30:14 <TrueBrain> imagine installing OpenTTD expecting Sim City :D
12:52:18 <TrueBrain> a cleaner version will show up soon enough ;)
13:54:18 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8355: Add: support for emscripten (play-OpenTTD-in-the-browser) https://git.io/JIn4l
14:01:10 <TrueBrain> "Secrets are not passed to workflows that are triggered by a pull request from a fork."
14:01:13 <TrueBrain> hmm .. that is an issue
14:01:17 <TrueBrain> I understand that, but it is an issue :)
14:01:50 <TrueBrain> seems I do need to learn DorpsGek to do this, I guess
14:05:55 *** supermop_Home has joined #openttd
14:06:57 *** skrzyp1 has quit IRC (Ping timeout: 480 seconds)
14:07:07 <TrueBrain> I get why these restrictions exist, from a security point of view .. but boy, it makes things more complicated :P
14:08:20 <supermop_Home> andythenorth finally read up on a months+ worth of horse forum
14:09:46 <Eddi|zuHause> that sounds all kinds of wrong :p
14:13:00 <supermop_Home> maybe the horses run the forum?
14:17:08 <TrueBrain> meh .. my idea to automagically create previews is not going to work due to how GitHub build up security .. they could do better, but at least they have security in place :D Basically, from what I can tell, you can only get the deployments to work, if the branch is not in a fork
14:17:24 <TrueBrain> I understand why, but it is a bit of a shame, really ..
14:20:52 <andythenorth> I keep finding more trains to add to Horse
14:21:19 <TrueBrain> ah, no, they have a solution .. just poorly documented ... another project ran into the same issue and they happen to report the solution .. lets try ...
14:23:45 <andythenorth> so wasm OpenTTD, build patchpacks and forks, websaves, MMO
14:23:54 <andythenorth> that just leaves newgrf translator :P
14:24:16 <andythenorth> then everything is finished
14:31:27 *** ChanServ sets mode: +v tokai
14:36:58 *** vejetaryenvampir has quit IRC (Ping timeout: 480 seconds)
14:38:22 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
15:07:20 *** tokai|noir has joined #openttd
15:07:20 *** ChanServ sets mode: +v tokai|noir
15:14:14 *** tokai has quit IRC (Ping timeout: 480 seconds)
15:24:36 *** Flygon_ has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
15:52:45 *** jottyfan has joined #openttd
16:09:27 *** arikover has joined #openttd
16:14:24 <TrueBrain> that moment that while debugging GitHub Actions, GitHub rolls out a new GUI for it :P
16:14:27 <TrueBrain> it does look a lot better
16:22:20 *** arikover` has joined #openttd
16:24:34 *** jottyfan has quit IRC (Quit: jottyfan)
16:27:33 *** arikover has quit IRC (Ping timeout: 480 seconds)
16:32:55 <TrueBrain> okay, previews are now build based on the workflows as in master
16:33:05 <TrueBrain> so no pull request can take over the S3 bucket or something
16:33:11 <TrueBrain> the worst they can do, is upload another .html file, basically
16:33:22 <TrueBrain> but that is why a developer has to add the label to it :D
16:37:21 <TrueBrain> takes ~10 minutes to produce a preview build of a PR
16:41:23 <TrueBrain> it keeps failing after the build, which takes 9 minutes, on AWS CLI :P
16:41:27 <TrueBrain> that is just evil :D
16:42:47 <TrueBrain> "TrueBrain requested a deployment to preview-pr-18 17 seconds ago Pending "
16:42:49 <TrueBrain> that does look good :)
16:45:43 *** blathijs has quit IRC (Ping timeout: 480 seconds)
16:53:23 *** blathijs has joined #openttd
16:55:02 <TrueBrain> hmm .. no clue where "pip3" installs awscli ... not in the normal ~/.local/bin it seems .. annoying :D
16:55:24 <TrueBrain> well, it does locally, but not in a GitHub runner
16:55:28 <TrueBrain> it is hiding somewhere else :D
16:56:41 <LordAro> runners run as root, so it's probably put it in /usr/local/bin
16:56:56 <LordAro> or something with the same permissions as root anyway
16:57:22 *** Wormnest has joined #openttd
17:24:16 <TrueBrain> yeah, but I am using a container to run in
17:24:23 <TrueBrain> and then you are not root, as it turns out :P
17:24:34 <TrueBrain> or at least, nt with this container :P
17:28:47 <TrueBrain> so lets not do these things inside the container, shall we ...
17:28:54 <TrueBrain> oops, dependencies ...
17:30:01 <TrueBrain> so many mails today from GitHub .. so many ... soooo maaannnnyyyyy
17:47:59 <TrueBrain> okay, it is this container that was the issue, changing uid and gid .. ugh .. so now I had to map in the real uid/gid ..
17:49:55 <milek7_> seems in fashion lately
17:51:33 *** otetede has joined #openttd
17:52:47 <TrueBrain> I absolutely hate websites that do not offer dark mode
17:52:55 <TrueBrain> so this makes me smile :)
17:53:05 <TrueBrain> my eyes are no longer burning!
17:53:43 <milek7_> mine never do, my displays have brightness setting ;P
17:53:49 *** frosch123 has joined #openttd
17:55:25 <TrueBrain> solving the wrong problem :P
17:55:51 <TrueBrain> okay, "aws" commands now execute, just ... I made another boo-boo ... AND ANOTHER TEST RUN :D
17:55:54 <TrueBrain> 10 minutes on the clock
17:57:14 <andythenorth> I've never used a cookie in my life though, so I don't know how I'd set the prefs
17:57:19 <TrueBrain> sounds like something that is worth your time
17:57:32 <TrueBrain> wait, no, I got that reversed
17:57:35 <TrueBrain> what a total waste of your time :P
17:57:36 <milek7_> @media (prefers-color-scheme: dark) {
17:57:57 <LordAro> milek7_: but analytics
17:58:14 <andythenorth> would be trivial on the query string
17:58:20 <andythenorth> and a js to read it
17:59:25 <TrueBrain> "If you have set privacy.resistFingerprinting to true, prefers-color-scheme preference is overridden to light."
17:59:36 <TrueBrain> haha, yeah .... privacy and customization at war!
18:02:32 *** otetede has quit IRC (Quit: Nettalk6 - www.ntalk.de)
18:03:28 <TrueBrain> well, our prime minister just told us: we are fucked, xmas is cancelled :P (in some nicer words)
18:05:16 <andythenorth> oh we are having a happy xmas in uk
18:05:18 <frosch123> it could have been worse. if they cancel sylvester and new year, we will remain in 2020 forever
18:05:19 <andythenorth> all families can meet up
18:05:24 <andythenorth> everyone can travel
18:05:31 <andythenorth> covid is suspended
18:12:18 <LordAro> covid's gone home for the holidays
18:15:51 <TrueBrain> lot of people around me were saying: yeah, with xmas, it will be gone a bit more, we can have like 8 people over, no problem. Well .. no. Like: no. You can have 3 people over, that is it. Lot of people going to be mad ... everyone knew that was coming, but they now said it out loud ...
18:17:53 <milek7_> here I don't think anybody cares about it anymore
18:18:04 <glx> we'll know in a few days, but it doesn't look promising
18:18:22 <TrueBrain> I cannot believe how many people were thinking: it is just going to be over in a week
18:19:00 <andythenorth> I read the UK scientific forecasting March-time
18:19:04 <andythenorth> reasonably accurate it turns out
18:19:21 <frosch123> black is weird. when i make a line too long, it wraps it. but if i make it shorter again, it does not wrap back
18:20:39 <frosch123> ah, no, it's because it inserts a trailing "," when wrapping, which blocks unwrapping
18:20:45 <LordAro> frosch123: did it add a trailing comma-
18:21:06 <LordAro> can't undo that, as (foo,) != (foo)
18:21:30 <LordAro> TrueBrain: i'll admit, i was one of those who thought it would be over in a couple of months
18:22:49 <TrueBrain> yeah, I was more referring to the people around me that were like last week: no, around xmas we can have people over, it will be gone by then
18:23:03 <TrueBrain> like ... have you seen the numbers lately? :P It did NOT go down at all over the last 6 weeks
18:23:05 <LordAro> oh, yeah, that's pretty dumb
18:23:08 <TrueBrain> how ... is this magically going to change?
18:23:24 <milek7_> you can just ignore the numbers ;p
18:23:29 <andythenorth> I never thought of the 'it will be gone'
18:23:39 <andythenorth> why would it be gone?
18:23:43 <andythenorth> what external factor changed?
18:24:04 * andythenorth isn't very normal, probably :|
18:26:09 <TrueBrain> I assumed in March it was like a month or 2/3 .. I am with LordAro there :)
18:26:15 <TrueBrain> but I don't like to think in doomscenarios :P
18:26:26 * andythenorth automatically thinks doomscenarios
18:26:39 <andythenorth> it's why I'm so happy and optimistic generally
18:26:41 <andythenorth> they rarely happen
18:34:01 *** Progman has joined #openttd
18:35:14 <TrueBrain> frosch123: "With Qt 6 we now require a C++17 compatible compiler enabling the use more modern C++ language constructs when developing Qt and also allows for integration points on the API side." <- if Qt can do it, so can we right? :D
18:35:33 <frosch123> i also got that mail today
18:35:48 <TrueBrain> I didn't get an email :(
18:36:08 <frosch123> so you have no qt licence, or you unsubbed
18:37:09 <TrueBrain> lol @ branch-name .. that might need some fixing
18:38:08 <Xaroth> maybe a bit pedantic, but the pr name in the version string?
18:38:32 <Xaroth> that way when you have a tab open for a specific pr the window title shows the pr number
18:39:10 <TrueBrain> Xaroth: yup, for some reason that is a bit broken
18:39:17 <TrueBrain> guess GitHub doesn't name its branches :D
18:39:18 <frosch123> gcc 7, clang 4, msvc 19.14
18:39:53 <frosch123> last time msvc was an issue, but that is 2 years ago
18:41:17 <LordAro> TrueBrain: would be nice if there was an index page that listed available builds
18:41:37 <LordAro> it'd be better than a 404
18:41:43 <TrueBrain> I understand that you ask it, as I did myself .. but I couldn't think of a good reason :P
18:41:53 <TrueBrain> yeah, I will add an empty page :)
18:42:49 <frosch123> i guess find_version has no chance to detect the PR
18:43:22 <TrueBrain> no, I will name the branch; that should fix it
18:43:49 <frosch123> i know you invested a lot of work into it, but is there a setting to disable the mouse capture? :p
18:44:12 <TrueBrain> it can be disabled, but that also requires changing scroll_mode
18:44:38 <TrueBrain> (otherwise it is really unplayable)
18:44:41 <TrueBrain> not sure how that works together
18:46:21 <TrueBrain> but you both are talking about some minor things .. any opinions about the bigger picture? :P
18:46:59 <Xaroth> Pardon my french, but I fucking love it.
18:47:32 <frosch123> lol, i crashed a train...
18:47:35 <LordAro> as much as i hate the whole "everything is a browser now"
18:47:39 <LordAro> it is pretty damn cool
18:47:51 <frosch123> no idea whether pbs always behaved like that, or whether that patch breaks it :)
18:48:15 <TrueBrain> frosch123: now that is the million dollar question :D
18:48:28 * andythenorth looks in settings
18:48:29 <TrueBrain> right, building a version that should fix branchnames
18:48:36 <TrueBrain> you can change your settings yourself
18:48:48 <TrueBrain> so when ever you come back, IT REMEMBERS IT :D
18:48:55 *** jottyfan has joined #openttd
18:49:06 <frosch123> aww, web browser zoom does not work
18:49:19 <TrueBrain> LordAro: I wanted to make this with Windows binaries first, but .. this is somehow a bit easier :P
18:49:26 <andythenorth> testing MP specific PRs will be 'interesting' :)
18:49:30 <TrueBrain> frosch123: haha, feature requests left and right ;) Good :D
18:49:37 <TrueBrain> owh, networking doesn't work yet btw
18:49:44 <TrueBrain> I still have to copy/paste some work from someone else :P
18:49:48 <frosch123> TrueBrain: no request, i just kind of assumed the browser would magically do that
18:49:55 *** Tirili has quit IRC (Ping timeout: 480 seconds)
18:50:00 <TrueBrain> frosch123: possibly SDL fires the wrong event, or one we don't listen to
18:50:22 <LordAro> TrueBrain: would the previews be deleted after the PR is merged?
18:50:37 <TrueBrain> LordAro: currently they have a 90 day retention
18:50:58 <andythenorth> ok the next thing I need is to infringe copyright with original baseset :P
18:51:03 <TrueBrain> but we can make it so it removes it too, sure
18:51:28 <TrueBrain> I think 90 days is too much btw
18:51:30 <frosch123> with that patch you can skip building a lot of signals
18:51:38 <LordAro> TrueBrain: mm, indeed
18:51:43 <TrueBrain> frosch123: I really really like that patch, honestly :)
18:51:55 <TrueBrain> LordAro: but honestly, that is all a bit tuning :)
18:51:58 <LordAro> TrueBrain: are settings/saves local to a particular preview, or to preview.ottd.org as a whole?
18:52:05 <TrueBrain> it should be on the whole domain
18:52:17 <frosch123> TrueBrain: no idea, i would worry it hits hard on performance
18:52:39 <LordAro> perhaps the index page should have "clear settings/saves" buttons :)
18:52:44 <TrueBrain> LordAro: I considered it desirable, ironically :)
18:52:59 <LordAro> depends what you're testing, i suppose :)
18:53:05 <TrueBrain> frosch123: the patch reads as an increase in CPU, yes :P
18:53:20 <TrueBrain> LordAro: indeed; I was also thinking it would be nice if a PR can have a savegame-to-test-with
18:53:23 <TrueBrain> which is loaded in there
18:53:29 <TrueBrain> but ... again .. features to add :D
18:53:34 <TrueBrain> lets make the baby-step first :P
18:54:27 <TrueBrain> I see so many things we could add and improve on
18:54:36 <TrueBrain> like allowing you to select other languages
18:54:48 <glx> about c++17, since cmake move openttd force c++17 for msvc and c++11 for other compilers
18:55:30 <LordAro> i thought we'd already moved to c++14, tbh
18:55:41 <LordAro> i'm not quite sure about moving to c++17 just yet though
18:56:03 <LordAro> needs another year or two, imo (for debian stable update, mostly)
18:56:46 <glx> would be nice to jump to c++20, there are nice features
18:56:59 <TrueBrain> there! Now there is a branchname in the version :D
18:57:05 <LordAro> that's definitely too soon :p
18:57:05 <TrueBrain> I just did "pr<number>"
18:57:29 <TrueBrain> I am REALLY happy with this guys :D GitHub Deployments are awesome :D
18:57:42 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8355: Add: support for emscripten (play-OpenTTD-in-the-browser) https://git.io/JIn4l
18:58:02 <TrueBrain> and the preview is done before Mac OS is done, so that is epic win :P
18:58:22 <LordAro> some of us have jobs :p
18:59:04 <TrueBrain> is that your excuse? :P
18:59:19 <TrueBrain> frosch123: without any circles?
18:59:33 <frosch123> LordAro: i am on debian stable, it has gcc8
18:59:51 <TrueBrain> "When I let two trains run in a circle without any circles" <-referring to that :P
19:01:18 <frosch123> so, yeah, it's not c++20 ready
19:01:59 <TrueBrain> okay, next up, fixing network ... not today, but something for this weekend or what-ever :)
19:02:14 <frosch123> but c++20 does not really matter for us. we don't use ranges
19:02:24 <LordAro> there are a few library features that gcc9 only
19:03:08 <LordAro> tell you what, get the std::string "conversion" finished, then we'll upgrade to c++17 :p
19:03:10 <frosch123> parallel stl and polymorphic allocators
19:03:17 <TrueBrain> andythenorth: btw, I fixed your Mac OS problems .. you can now play online .. no more silly issues with news etc :D :D :D
19:03:22 <frosch123> we certainly don't need the former
19:03:38 <frosch123> and the latter... i can't remember whether it was compeltely changed in c++17 or c++20
19:03:42 <TrueBrain> LordAro: how about approving some PRs for that then :P :P
19:04:28 <TrueBrain> ah, sorry, you didn't change the state :P mea culpa :)
19:06:06 <frosch123> TrueBrain: i like how you applied a ttd look to the loading splash screen
19:06:40 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8366: Change: Don't display OS name when the user is exiting the game https://git.io/JIE0p
19:06:54 <TrueBrain> frosch123: :D Tnx :D It took ... more time than I would like to admit :P
19:07:24 <TrueBrain> I actually wanted the intro screen as background .. but that was 2MB image .. which is a bit silly :P
19:07:43 <frosch123> is "github discussions" the new irc?
19:08:02 <TrueBrain> they are rolling out a new version of github.com clearly :P
19:08:05 <TrueBrain> do we want to enable it?
19:09:06 <TrueBrain> it sounds useful, but it strongly depends how they implemented it :)
19:09:25 <andythenorth> TrueBrain it's not even joking
19:09:45 <TrueBrain> it has categories and everything
19:09:47 <andythenorth> there's a default apple way to wrap webkit round a thing to make a fake app
19:09:55 <andythenorth> 'fake' / shell app /s
19:10:46 <Xaroth> I mean, MS Teams, Slack, Discord
19:10:57 <frosch123> TrueBrain LordAro: it does not matter in this case. but eints generally does not care about you removing translations. eints commits the newest valid translation. "no translation" is older in any case, so any removal is reverted. in this case the translation will be invalid (incompatible parameters), so eints won't readd it
19:11:20 <TrueBrain> frosch123: good to know :)
19:11:25 <TrueBrain> we were wondering what the best approach would be
19:11:29 <TrueBrain> don't know if you have a better?
19:11:54 <frosch123> one option is to only change english.txt in the pr
19:12:02 <frosch123> have the build warnings for some hours
19:12:12 <frosch123> and then the eints commit will cleanup the translations
19:12:21 <TrueBrain> it removes the invalid ones?!
19:12:26 <TrueBrain> okay, that sounds like a better way :P
19:12:38 <frosch123> the result after the eints commit will be the same
19:12:47 <frosch123> it's just whether you care about the build warnings inbetween
19:13:12 <frosch123> ofc, you may know how to manually trigger a eints commit :p
19:13:28 <TrueBrain> I really don't care about warning for that bit of time
19:13:35 <TrueBrain> it is just that I am sure not all translations will receive an update before 1.11 :P
19:13:51 <TrueBrain> but that eints fixes that, that is good to know :D
19:14:12 <frosch123> eints will remember the old translation as "invalid" until you wipe the file storage, but it will not be in git
19:15:09 <LordAro> TrueBrain: unlikely to branch for another couple of months, i would think
19:15:19 <LordAro> how are the translator lists doing?
19:15:36 <LordAro> (and updates would be backported anyway)
19:17:21 <TrueBrain> so many words .. so difficult to make sense out of it :P
19:18:37 <supermop_Home> in the category of new and interesting problems to have... I just discovered that malort has been leaking and dripping onto about 30 of my records for 1-2 days
19:19:49 <LordAro> TrueBrain: i was thinking you were referring to the branch point (or first beta), when you were actually talking about 1.11.0
19:20:17 <LordAro> also, i think my grep is wrong, or polish.txt has 500 more strings than english.txt
19:20:43 <supermop_Home> frosch123 different brand but yes
19:21:12 <supermop_Home> apparently its pretty sugary because its really sticky and hard to clean off
19:23:41 <LordAro> oh i see, separate strings for each case
19:25:10 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIEz1
19:27:02 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIEzA
19:28:23 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIEgO
19:28:32 *** sla_ro|master2 has joined #openttd
19:29:45 <LordAro> of course, just as i finish my command, i remember that eints has these numbers already
19:30:22 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIEgw
19:31:20 <LordAro> ( for f in src/lang/*.txt; do echo -n "$f "; grep -o "^STR_[^ .]\+" $f | sort -u | wc -l; done | sort -rk2 | column -t )
19:31:44 *** sla_ro|master has quit IRC (Ping timeout: 480 seconds)
19:33:25 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIE2I
19:33:59 <DorpsGek> [OpenTTD/OpenTTD] michicc commented on pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIE2s
19:36:39 <supermop_Home> what a disaster I've caused for myself here
20:02:42 <TrueBrain> I cannot believe I could make "Preview this PR" :D It is so cool .. GitHub Actions are really awesome :D
20:04:31 <andythenorth> we are living in the future
20:04:35 <frosch123> omg... you can even build feedback loops with that patch
20:04:52 <frosch123> two trains in a circle, both adjusting their speed to each other
20:05:05 <TrueBrain> I see someone is having fun :P
20:05:40 <andythenorth> frosch123 twitch stream or it didn't happen :P
20:05:50 <TrueBrain> build in streaming, you say? :D
20:06:07 <frosch123> TrueBrain: input recording?
20:07:12 <frosch123> we should hire that guy for the ottd website
20:07:31 <andythenorth> definition of awesome
20:08:48 <milek7_> Compiled code throwing an exception, 573519688,756348,1866
20:09:43 <milek7_> I'm trying to get network work again
20:10:13 <milek7_> some emscripten flag to generate better messages?
20:10:34 <TrueBrain> you can run it with -s ASSERTIONS=1 and -g4
20:10:45 <TrueBrain> the first has better error handling, and because of that more clear errors
20:10:50 <TrueBrain> the second leaves debug information behind
20:10:58 <TrueBrain> next you have to produce .map file, but I totally forgot how
20:11:02 <TrueBrain> in the end, it is like 10% better
20:11:14 <TrueBrain> for the News bug, it helped me only to find the location in the code
20:11:22 <TrueBrain> debugging in Emscripten is not really .. a thing yet :P
20:12:53 *** Tirili has quit IRC (Ping timeout: 480 seconds)
20:12:58 <andythenorth> eliminating signals you say? o_O
20:13:14 * andythenorth will buy that for $1
20:13:59 <TrueBrain> For some reason I now see you walking with a shotgun in your hometown shooting all signals
20:14:05 <TrueBrain> not sure that is the right xmas spirit
20:14:09 <TrueBrain> but given it is 2020 .. sure
20:14:42 <TrueBrain> it doesn't even give line numbers, it annoyed the hell out of me
20:14:59 <TrueBrain> what did you do? :D
20:15:37 <TrueBrain> you can build a docker with your patch; after that LZMA works :D
20:16:23 <TrueBrain> (and the latest version of my PR, ofc :D)
20:16:53 <TrueBrain> so we can either pick full POSIX sockets support,which is dreadful slow, or emulated POSIX support, which doesn't have all functionality ... righhtttt
20:18:13 <milek7_> but requires special websocket server, to properly handle tcp/udp
20:18:54 <TrueBrain> I was hoping that depending on the target it could use another websocket server
20:19:02 *** jottyfan has quit IRC (Quit: jottyfan)
20:19:04 <TrueBrain> but it seems DNS is resolved on the server-side, if I read the documentation correctly
20:21:19 <TrueBrain> but as always, their documentation is ... well, not complete :D
20:23:02 <TrueBrain> hahahahaha, that comments about // and #
20:25:31 <TrueBrain> hmm .. "Name or service not known" for bananas-1 .. I assume it was an oopsie somewhere :P
20:30:35 <TrueBrain> milek7_: so too bad that Module["websocket"]["url"] is not a function, with "addr" as parameter :)
20:30:51 <TrueBrain> as that would allow to use different websocket servers depending on where you are going to
20:31:09 <TrueBrain> well, "ws" should be an mapping too then
20:31:50 <frosch123> i think i caught up with the rabbit
20:32:01 <TrueBrain> 20 NML patches incoming!
20:32:19 <frosch123> i started out if 1 bug, fixed 4 in total
20:32:46 <TrueBrain> that is not a bad score :)
20:34:31 <frosch123> how does your git-commit-only-some-hunks thingie work?
20:34:54 <TrueBrain> or any modern editor, ofc :D
20:35:51 <frosch123> so many options to choose, but it worked
20:36:04 <TrueBrain> yes for accepting it, no for not, s for splitting it
20:36:06 <TrueBrain> the rest I never use :P
20:36:20 <TrueBrain> s you need rarely, but sometimes, when git combined a few hunks
20:37:04 <TrueBrain> or like me, ctrl+c, but that cancels the current file (not any other file)
20:38:01 <TrueBrain> hmm, no, emscripten makes a new websocket for every connection
20:38:07 <TrueBrain> so having different servers should be trivial
20:39:05 <TrueBrain> I like how by default it just assumed the TCP connection is going to a websocket server, if I am reading this right
20:47:08 <TrueBrain> so what we could do, is for emscripten use a different port, on which in the infra a websocket server is listening, for the content service
20:47:17 <TrueBrain> that is a relative clean and simple solution, I guess
20:47:44 <TrueBrain> milek7_: btw, regarding CSleep, shouldn't the whole CSleep do nothing, I was wondering?
20:48:15 <milek7_> ..but it still bothers me :P
20:48:28 <TrueBrain> in general it is a very bad CSleep location, but meh
20:48:39 <TrueBrain> the whole draw-thread is a bit reversed anyway
20:48:47 <milek7_> click 'add server' -> milek7.pl
20:50:14 <TrueBrain> can't believe how smooth it works
20:50:20 <TrueBrain> it really is incredible, how far emscripten got
20:52:31 <TrueBrain> sorry for the IP of content.openttd.org .. it changes often (like 10 times a year or so) ..
20:55:20 <milek7_> if ($msg->getPayloadLength() == 10 && $payload[0] == "\xff") {
20:55:31 <milek7_> that's for udp handling
20:55:53 <milek7_> I'm not sure if emscripten have this documented somewhere
20:59:08 <TrueBrain> they added it for a different reason, and it happens to work to know it should be an UDP message, I guess? :P
21:01:21 <TrueBrain> you can send back a similar package, to indicate the port on the other side or something
21:02:21 <michi_cc> LordAro: What offends you more in ReadFileToMem, the ::operator new or the reinterpret_cast afterwards?
21:02:35 <LordAro> michi_cc: can i say both? :p
21:02:49 <LordAro> probably the operator new, if i had to pick one
21:04:12 <michi_cc> That one is a small price to pay for not having std::unique_ptr<char, FreeDeleterType> / std::unique_ptr<LanguagePack, FreeDeleterType> in more places.
21:04:23 <milek7_> for some reason, default save location is /user_data
21:04:27 <milek7_> instead of /user_data/save
21:04:32 <TrueBrain> yeah, I noticed that too
21:04:40 <TrueBrain> it happens in normal builds too, if you use -c openttd.cfg
21:04:57 <milek7_> and autosaves go to /
21:05:06 <TrueBrain> if you use -c, the default folder is where the cfg is, despite is also making a "save" folder
21:05:31 <michi_cc> And directly specializing this functions isn't optimal either, LanguagePack is only known inside strings.cpp, which conceptually shouldn't contain file IO code.
21:06:05 <TrueBrain> milek7_: I guess we could avoid "-c openttd.cfg" if we set the folder to /user_data already, or something .. haven't looked into it really
21:06:11 <TrueBrain> it feels like it is buggy in OpenTTD
21:06:22 <TrueBrain> waiting for LordAro to finish his PR, before I look into the folder mess :P
21:06:44 <michi_cc> Anyway, let's (hopefully) kick that warning first.
21:07:10 <DorpsGek> [OpenTTD/OpenTTD] michicc updated pull request #8362: Apply some C++ love to string handling in file and config paths. https://git.io/JIW6A
21:07:32 <TrueBrain> milek7_: I initial tried to keep everything in /, but you cannot mount a file
21:07:38 <TrueBrain> so storing openttd.cfg was really tricky
21:07:39 <DorpsGek> [OpenTTD/OpenTTD] orudge updated pull request #8366: Change: Don't display OS name when the user is exiting the game https://git.io/JI0Ig
21:07:42 <TrueBrain> I tried symlinking it, but boy, what a mess
21:08:12 <TrueBrain> but I have now a better understanding of the FS, so I might try that again, honestly
21:08:54 <TrueBrain> just mount /content_download, /save and /config, where the latter has a symlink from /config/openttd.cfg to /openttd.cfg
21:09:32 <TrueBrain> can you edit hotkeys ingame btw, LordAro / frosch123 ?
21:09:36 <DorpsGek> [OpenTTD/OpenTTD] orudge commented on pull request #8366: Change: Don't display OS name when the user is exiting the game https://git.io/JIE1h
21:09:43 <TrueBrain> and what is windows.cfg exactly?
21:09:50 <DorpsGek> [OpenTTD/OpenTTD] LordAro updated pull request #8357: Codechange: Drop libxdg-basedir dependency in favour of finding the directories ourselves https://git.io/JIWz8
21:10:32 <TrueBrain> milek7_: that initial "port" packet you receive, is to indicate on which port the "server" was binding, so the remote could try to use that too
21:11:12 <TrueBrain> the server can send back a "port" packet, to indicate what port the remote is using
21:12:21 <frosch123> TrueBrain: no, hotkeys are only editable via texteditor hotkey.cfg
21:12:31 <TrueBrain> okay, so I don't have to store that, nice :D
21:12:32 <frosch123> TrueBrain: windows.cfg stores default window sizes and sticky state
21:12:42 <TrueBrain> ah; so I should store that, k k :)
21:12:43 <frosch123> you can change those in-game by ctrl-clicking the window buttons
21:13:05 <TrueBrain> funny; that is new! (for me) :D
21:13:55 <frosch123> it may be newer than half the age of ottd :)
21:14:15 <frosch123> not sure, it's close
21:15:08 <TrueBrain> from after my time :P
21:15:17 <TrueBrain> the whole window rewrite is btw ;)
21:15:36 <frosch123> i think there were 3?
21:17:30 <TrueBrain> milek7_: the fact folders don't seem to work, seems to be caused by xdg
21:17:41 <TrueBrain> it seems to be messing some stuff up
21:19:51 *** Speeder has joined #openttd
21:20:43 <TrueBrain> guess we violated that statement :D
21:21:32 <TrueBrain> and there is some really weird code going on ..
21:22:01 <TrueBrain> the personal dir becomes the config dir
21:22:11 <TrueBrain> but we already knew that was a total mess :P
21:22:53 <milek7_> maybe just setup real directory structure
21:22:57 <TrueBrain> so I think we should put the files in openttd.data in some folder, instead of /
21:22:58 <milek7_> .openttd in homedir etc.
21:23:20 <TrueBrain> easier to mount, just ~/.openttd
21:23:37 <TrueBrain> and the openttd.data files we can put in /usr/share/openttd or something
21:23:58 <TrueBrain> but really, LordAro 's patch should land, as that makes this easier to read wtf is going on
21:24:03 <TrueBrain> xdg doesn't really help in readability :P
21:26:52 <TrueBrain> yeah, basically, the setup already does "the right thing", from what I can tell
21:27:11 <TrueBrain> the global dir is only /usr/local/share/games/openttd
21:29:13 <TrueBrain> ah, XDG is not configured, that explains :)
21:33:27 <TrueBrain> milek7_: what I do not understand, why the autosave dir is wrong; it really is set to /usr_data/save/autosave .. hmmmmmm
21:34:26 *** Wormnest has quit IRC (Quit: Leaving)
21:34:47 *** Wormnest has joined #openttd
21:37:57 <andythenorth> now we're sucking diesel :)
21:38:26 <frosch123> lol, comment too long
21:38:39 <frosch123> black does not wrap comments
21:46:30 <frosch123> are we the only ones using select() ? or do people like porting ottd to stuff?
21:47:32 <TrueBrain> I did not even read that part :P
21:47:34 <frosch123> maybe everyone else uses event queues
21:47:45 <TrueBrain> okay, this is a lovely wtf ... there is a variable _personal_dir
21:47:52 <TrueBrain> which is NOT the search path PERSONAL_DIR
21:47:57 <TrueBrain> or at least .. it can be, but it doesn't have to
21:48:05 <TrueBrain> when creating folders to prepare the system for everything
21:48:09 <TrueBrain> _personal_dir is used
21:48:17 <TrueBrain> writing etc later on, uses PERSONAL_PATH from the Search Path
21:48:24 <TrueBrain> crashlogs? Always use _personal_dir
21:48:34 <TrueBrain> s/PERSONAL_PATH/PERSONAL_DIR/
21:48:37 <TrueBrain> like what the actual fuck
21:49:05 <TrueBrain> and this is why when using -c things go weird
21:49:09 <TrueBrain> as -c changes _personal_dir
21:49:11 <TrueBrain> but not PERSONAL_DIR
21:49:21 <LordAro> super fuckin confusing
21:49:27 <TrueBrain> you say "yup" as you would expect that :P
21:49:42 <TrueBrain> very tempted to remove _personal_dir in favour of the SearchPath value
21:49:49 <LordAro> i've looked into it before as part of the -c bug
21:49:52 <TrueBrain> just to have a single personal dir :P
21:50:13 <milek7_> I went through network quirks
21:50:22 <milek7_> all are still necessary
21:54:05 <TrueBrain> looking at the git (read: subversion) history, it seems an oversight, to leave _personal_dir
21:54:10 <TrueBrain> not sure .. hard to tell sometimes :D
21:56:33 *** milek7_ is now known as milek7
21:56:40 <TrueBrain> okay, I kinda understand what is happening ... basically, if you would have an openttd.cfg in a weird location, you normally wouldn't want your personal-dir to point there
21:57:24 <TrueBrain> so if I collapse _personal_dir into _searchpaths[SP_PERSONAL_DIR]
21:57:35 <TrueBrain> it would change for people who have openttd.cfg in some read-only location
21:59:26 <TrueBrain> it is so clearly wrong, that it is difficult to see what is right :D haha
22:03:00 *** Tirili has quit IRC (Remote host closed the connection)
22:03:20 <milek7> path handling in legacy projects is always horrid
22:03:36 <milek7> in eu07 there was code that was using both / and \, and it actually mattered because it was later doing strpos path splitting and expected to skip one type of slashes :D
22:07:40 <andythenorth> hmm random triggers when I don't want them
22:08:30 <milek7> the issue with emulated sockets is that if I supply complete url in config
22:08:48 <milek7> then I don't see a way to tell destination address
22:08:54 <milek7> so no proxying possible
22:09:48 <andythenorth> hmm what do dual headed vehicles do with random bits?
22:09:55 <TrueBrain> indeed, that was what I complained about earlier milek7 :D
22:10:16 <TrueBrain> they should implement proxy-protocol tbh
22:10:51 <andythenorth> I have single vehicles triggering in stations, but not depots (desired)
22:11:13 <andythenorth> and I have dual-headed vehicles triggering in stations and depots (depots not desired)
22:11:26 <andythenorth> the trigger...is in the same procedure
22:14:11 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain opened pull request #8367: Fix: create folders in personal-dir on startup when -c is given https://git.io/JIE7t
22:14:20 <TrueBrain> okay .. this fixes -a- problem, but not -the- problem
22:14:44 <TrueBrain> mostly, the xdg flow is totally bananas now .. it already was, but now it is more clear it is :D
22:15:18 <TrueBrain> the only real change this PR does, is change the location of content_download
22:15:22 <TrueBrain> as that was set to the folder of -c
22:16:00 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8367: Fix: create folders in personal-dir on startup when -c is given https://git.io/JIE7t
22:16:58 <TrueBrain> now the real question is, with -c, should config-dir become your personal-dir or not?
22:17:03 <TrueBrain> the code suggests that was the intention
22:17:06 <TrueBrain> it only utterly failed to do so
22:17:22 <TrueBrain> LordAro: I would love your opinion on the matter :D
22:17:50 <LordAro> i'm not sure i'm qualified to have an opinion on it :p
22:17:57 <frosch123> "FioCreateDirectory(config_dir);" is directly followed with "FioCreateDirectory(_config_dir);" :p
22:18:37 <LordAro> i am fairly certain using -c should override any user config/data directories
22:18:44 <LordAro> and possibly the global ones too
22:18:53 <TrueBrain> frosch123: yes :) Idiotic, isn't it? :P
22:18:58 <TrueBrain> with xdg, they can be 2 different folders
22:19:32 <TrueBrain> _config_dir = _searchpaths[SP_PERSONAL_DIR_XDG]; <- pretttttyyyy sure that was not the intention of that line
22:19:34 <frosch123> i think the intention was to keep using .openttd when it exists, even with XDG is available
22:19:53 <LordAro> it should no longer create that dir though
22:19:58 <andythenorth> random triggers can wait :P
22:20:07 *** andythenorth has quit IRC (Quit: andythenorth)
22:20:48 <TrueBrain> I honestly do not know what xdg is btw
22:21:00 <TrueBrain> I just see code that most likely never worked as intended :D
22:21:14 <frosch123> traditional we use ~/.openttd
22:21:24 <LordAro> the spec that defaults to ~/.local/share and/or ~/.config
22:21:38 <frosch123> but since people hate their home dir getting spammed, they came up with ~/.local/openttd
22:22:02 *** Progman has quit IRC (Remote host closed the connection)
22:22:11 <TrueBrain> so that is where the lovely: if PERSONAL_DIR[0] == ".", skip first char, comes from
22:22:15 <TrueBrain> so .openttd becomes .local/openttd
22:22:32 <TrueBrain> so that really never worked as intended
22:22:39 <TrueBrain> the only thing that got stored in .local/openttd, are crashlogs :D
22:22:46 <frosch123> where does it check whether .openttd exists?
22:22:59 <frosch123> TrueBrain: no, it worked
22:23:03 *** sla_ro|master2 has quit IRC ()
22:23:10 <frosch123> i don't know about -c, but .openttd and .local work well
22:23:12 <TrueBrain> I mean ... how?! :P
22:23:42 <TrueBrain> owh, wait, yes, it is in the searchpath too
22:24:00 <TrueBrain> and is in priority above .openttd
22:24:07 <TrueBrain> so new files are created in .local/openttd
22:24:11 <TrueBrain> old files can be read from .openttd
22:24:27 <frosch123> ottd should check both, but only use one, and only create .local if .openttd does not exist
22:24:52 <frosch123> these two paths work very different to the other search paths :)
22:24:54 <TrueBrain> (if you don't use -c)
22:25:21 <frosch123> pretty sure that would have been noticed
22:25:47 <TrueBrain> euh, no, you are right: it creates .local/openttd always, and config-dir after
22:25:53 <frosch123> i think you jsut removed that check
22:25:56 <TrueBrain> config-dir is, if you do not have .openttd/openttd.cfg, .local/openttd
22:26:25 <TrueBrain> the same still happens
22:26:39 <TrueBrain> the if() before the FioCreateDirectory was not super useful
22:26:42 <frosch123> if (config_dir != _personal_dir) FioCreateDirectory(_personal_dir); <- isn't that "if" the important part?
22:26:47 <frosch123> that prevents creating both
22:26:54 <TrueBrain> FioCreateDirectory(config_dir); is before it
22:27:00 <TrueBrain> so either they are the same
22:27:02 <TrueBrain> or they are different
22:27:04 <LordAro> (.local/share/openttd, not .local/openttd)
22:27:11 <TrueBrain> either way, doing CreateDirectory twice doesn't change anything
22:27:35 <TrueBrain> LordAro: sure, what-ever floats your boat :D
22:27:54 <TrueBrain> frosch123: that ticked me off about that if(), it made it feel really important, but really, it doesn't do anything :D
22:28:29 <TrueBrain> (please correct me if I am reading this wrong btw :P This is fucked up code :D)
22:28:36 <frosch123> hmm, ok, but why do you add a createdir for SP_PERSONAL_DIR ?
22:28:50 <TrueBrain> and that is where I go in the wrong, yes
22:28:54 <TrueBrain> but that has to do with -c
22:29:00 <TrueBrain> and I now get why this code is so bbbbbbbbrrrrrrrrr
22:29:10 <LordAro> TrueBrain: just making sure you don't get confused :p
22:29:32 <LordAro> sounds like the code needs a larger comment
22:30:06 <TrueBrain> no, there are 3 flows waved into each other here :P
22:30:10 <LordAro> and possibly docs/directory_structure.md should be expanded
22:30:22 <TrueBrain> okay, so the question: what should -c do, is important
22:30:41 <TrueBrain> should -c change the personal-dir
22:31:39 <LordAro> i think it should *remove* the personal-dir, rather than change it
22:31:54 <TrueBrain> where would it store files in that case?
22:32:11 <TrueBrain> so ... -c changes personal-dir in config-dir in that case, not?
22:32:19 <TrueBrain> there currently is no config-dir on the searchpath
22:32:31 <TrueBrain> which is most likely what is causing this problem :D
22:36:45 <TrueBrain> okay, so, if openttd.cfg exists in the search-path, that location is used as configuration
22:36:54 <TrueBrain> otherwise, if xdg, .local/share/openttd is used
22:37:14 <TrueBrain> if not, it looks for a place to store it in personal-dir, binary-dir, working-dir, shared-dir, and in the end installation-dir
22:37:45 <TrueBrain> in the same folder as openttd.cfg, the other 3 configuration (highscore is not a configuration!, hotkeys and windows) is stored
22:38:19 <TrueBrain> if xdg, the paths like "save" and "save/autosave" are created in .local/share/openttd, so they are used with preference
22:38:32 <TrueBrain> otherwise, what-ever openttd.cfg was found is prepared to be used as such
22:38:44 <TrueBrain> right ... that kinda makes sense .. written in the most annoying way possible, but I get it
22:38:57 <TrueBrain> so if you have an openttd.cfg in your working-dir, it will prepare that directory to receive all other files too
22:39:03 <TrueBrain> assuming the other folders were not prepared already
22:39:07 <TrueBrain> head spinning? GOOD!
22:39:32 <TrueBrain> so _personal_dir is not so much the SP_PERSONAL_DIR, but more the "I would prefer it if files go here" directory
22:39:44 <TrueBrain> so crashlogs etc are always stored next to the openttd.cfg that started the game
22:40:11 <TrueBrain> and so is content_download, which is a very odd duck
22:41:45 <TrueBrain> owh, that is because it is on his own searchpath
22:42:13 <TrueBrain> so if you have a .openttd/content_download with xdg, it is NOT used
22:42:19 <TrueBrain> only .local/share/openttd/content_download is used
22:43:54 *** arikover` has quit IRC (Remote host closed the connection)
22:44:07 *** arikover` has joined #openttd
22:44:48 <LordAro> .openttd should only ever be read from, imo
22:45:06 <TrueBrain> for content_download, it is not even read with xdg :)
22:45:09 <LordAro> there's no particular reason to exclude it from content_download reading though
22:45:12 <TrueBrain> but as it is a searchpath, that is also hard to avoid :P
22:45:28 <TrueBrain> or we need to add an SP_AUTODOWNLOAD_DIR_XDG :D
22:45:50 *** keoz has quit IRC (Ping timeout: 480 seconds)
22:46:24 <TrueBrain> well, the same trick was used with SP_PERSONAL_DIR and SP_PERSONAL_DIR_XDG :)
22:46:32 <TrueBrain> to make sure .openttd was still read
22:46:55 *** frosch123 has quit IRC (Ping timeout: 480 seconds)
22:47:20 <TrueBrain> ugh, changing PERSONAL_DIR with -c is nasty as fuck
22:47:27 <TrueBrain> as that means it also doesn't read .openttd
22:47:33 <TrueBrain> but with xdg, it does read .local/share/openttd
22:47:40 <TrueBrain> that is just all kinds of wrong
22:48:33 <TrueBrain> going through the motions here, but at least I understand this now ... somewhat :P
22:48:35 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
22:48:49 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8367: Fix: create folders in personal-dir on startup when -c is given https://git.io/JIE7t
22:49:19 <TrueBrain> changed that PR into "Codechange: to create or not to create a folder, that is the question"
22:49:46 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8367: Codechange: to create or not to create a folder, that is the question https://git.io/JIE7t
22:50:32 <TrueBrain> so, I guess, the only way out is indeed to add a search-path
22:50:37 <TrueBrain> or to change the working-dir
22:51:49 <TrueBrain> .openttd vs .local is btw only a problem if you switch from non-xdg to xdg
22:51:52 <TrueBrain> guess it is not really that relevant
22:53:33 <TrueBrain> all these kind of odd little things ... if you run "./openttd" with an "openttd.cfg" next to it, it will still rather user ".openttd/openttd.cfg"
22:53:46 <TrueBrain> but it would love to store the autosave in ./save/autosave, if that folder exists
22:53:54 <TrueBrain> I ... just .... don't .... wuth?
22:55:00 <TrueBrain> hmm, no, I am wrong, it does not .. ugh, this code is horrible :( I get why so many people have tried and failed :D
22:57:50 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:59:10 <glx> or to rewrite entirely how path are handled
22:59:23 <glx> (but I won't do that myself)
22:59:50 <TrueBrain> I guess the issue a bit is, that it is not really wrong what it is currently doing
22:59:55 <TrueBrain> just written in such a way it is impossible to figure out
23:00:30 <glx> yeah this part is really are to follow
23:01:24 <LordAro> i... think my release workflow might be working
23:01:33 <TrueBrain> and the way content_download is solved is .. even adding to that pile :P
23:01:39 <TrueBrain> LordAro: \o/ \o/ \p/
23:01:50 <TrueBrain> means I have to get to work to figure out how to upload the files, I am afraid :(
23:02:15 <LordAro> no more fun emscripten stuff for you!
23:06:57 <glx> hmm ideally first step would be to create source bundle, including .ottdrev
23:07:11 <glx> then other steps could use it
23:07:27 <TrueBrain> that is what I used to do, yes
23:07:32 <TrueBrain> making sure they all do the same version etc :D
23:07:51 <glx> and you remove the need for git in the containers
23:08:03 <LordAro> i've still got the separate linux-checkout job, so that's easy enough to adapt
23:09:12 <glx> and checksum can be a final common step
23:09:34 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JIE7t
23:09:35 <TrueBrain> okay, I think this should be it ^^
23:10:10 <TrueBrain> LordAro: didn't check, but remember we need a single job at the end that combines all the results together
23:10:21 <TrueBrain> (as we don't upload till all jobs are done; otherwise we might upload a partial :P)
23:11:04 <LordAro> i was gonna say, copying the artifacts around all over the place would be unnecessarily expensive otherwise
23:11:28 <LordAro> but look! a list of artifacts!
23:12:22 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JIE7t
23:12:37 <TrueBrain> okay ... with my PR, config_dir is now ALWAYS one of the search-paths
23:12:44 <TrueBrain> that should fix most of the weirdness tbfh
23:12:57 <TrueBrain> LordAro: that is really nice :D
23:13:04 <TrueBrain> and pretty quick too
23:13:14 <TrueBrain> the CI takes the same amount of time, lol
23:13:30 <TrueBrain> tomorrow I will check if the output is what I expect ;)
23:14:40 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JIE7t
23:17:52 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8367: Fix: change the working-dir searchpath when using '-c' https://git.io/JIE7t
23:17:56 <TrueBrain> fixing wording ... lalala :D
23:18:25 <TrueBrain> so when using -c, the content_download was stored next to the configuration
23:18:32 <TrueBrain> it was the only folder that DID work :P
23:21:26 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain approved pull request #8366: Change: Don't display OS name when the user is exiting the game https://git.io/JIExp
23:21:30 <TrueBrain> as nobody objected, lets go! :P
23:22:48 <TrueBrain> milek7: I upgraded to the latest SDK ... the game is noticeable slower for me
23:23:51 <TrueBrain> this is a debug build
23:24:08 *** ChanServ sets mode: +v tokai
23:24:18 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8355: Add: support for emscripten (play-OpenTTD-in-the-browser) https://git.io/JIn4l
23:24:23 <TrueBrain> in that case, milek7 , ^^, with fixed search-path, so autosaves go in /user_data
23:24:35 <TrueBrain> but I will still fix it to use ~/.openttd :P
23:30:49 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain updated pull request #8355: Add: support for emscripten (play-OpenTTD-in-the-browser) https://git.io/JIn4l
23:31:06 <TrueBrain> tnx for the info btw on xdg LordAro and frosch123 (for when you read your backlog :P)
23:31:08 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
23:36:39 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8357: Codechange: Drop libxdg-basedir dependency in favour of finding the directories ourselves https://git.io/JIEhK
23:40:02 <TrueBrain> wow, that is a boring read :P
23:40:05 <TrueBrain> just how it looks alone
23:40:32 <FLHerne> At least it's pretty short
23:40:49 <TrueBrain> other cleanups we could do, I guess, is remove the ifdefs around PERSONAL_DIR, as I believe it is now always set :D
23:40:50 <FLHerne> And not actually 10 mutually contradictory specs
23:42:06 <TrueBrain> yeah, PERSONAL_DIR is set for all OSes
23:42:18 <TrueBrain> unless you explicitly set it to "(not set)", as in, that exact string
23:43:41 <milek7> eh, I thought about including websocket in ottd servers
23:43:46 <milek7> but it would be a pain due to https
23:45:02 <DorpsGek> [OpenTTD/OpenTTD] TrueBrain commented on pull request #8357: Codechange: Drop libxdg-basedir dependency in favour of finding the directories ourselves https://git.io/JIEjZ
23:45:23 <TrueBrain> I had the exact same thought earlier this week :D
23:45:33 <TrueBrain> but .. there are a lot of buts
23:45:44 <TrueBrain> also finding a good implementation
23:45:50 <TrueBrain> telling people to open yet-another-port
23:45:59 <TrueBrain> but yeah, it might be the correct solution, honestly
23:46:04 <TrueBrain> I even considered adding a field in the master-server
23:46:19 <TrueBrain> so you can see in the listing if a server has websocket support
23:51:40 <milek7> it's possible to connect to http websocket but only from http site
23:51:44 <milek7> hosting client on http because of that.. that is bit silly
23:54:20 <TrueBrain> Yup ... but we have to bite this ssl bullet sooner or later anyway
23:54:53 *** nielsm has quit IRC (Ping timeout: 480 seconds)
23:59:22 *** gelignite has quit IRC (Quit: Stay safe! Stay at home! Stop the chain reaction!)
continue to next day ⏵
