IRC logs for #openttd on OFTC at 2024-04-23
⏴ go to previous day
00:00:37 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
00:44:37 *** Wormnest has joined #openttd
01:40:22 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
02:07:54 *** Wormnest has joined #openttd
02:31:52 *** Wormnest has quit IRC (Quit: Leaving)
02:33:42 *** gnu_jj_ has quit IRC (Ping timeout: 480 seconds)
02:39:34 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:40:40 <DorpsGek> - Update: Translations from eints (by translators)
05:16:34 *** keikoz has quit IRC (Ping timeout: 480 seconds)
05:50:38 *** jenkings has joined #openttd
05:50:38 <jenkings> I have tried to ask on forums, but no reply, What is the main issue with OpenTTD release on Android ?
05:51:02 <jenkings> i have contacted peylya, but the response was, that he is no longer maintainer of android version
05:51:15 <jenkings> i would like to help with the port and release
05:59:40 <jenkings> And one more question. Generally for ottd. is there a chance, there will be ability to place signal on bridges and in tunels ?
06:00:13 <jenkings> i have seen many posts with wishes for this feature, but i have never seen any serious reaction, or some statement for adding something like that
06:01:26 <jenkings> i think it would be one of the best "patches" for this game, because it is really annoying to build many small islands connected with tiny bridges, wich comes across the sea, just to be able to build signals 😦 And it is than really ugly 😦
06:04:20 <kamnet> jenkings: It's available in the JGR Patch Pack, but AFAIK nobody is working on it for a main release at this time.k
06:08:19 <jenkings> kamnet: oh, it is a shame 😦 is there some reason, that it is not intended to work on it ? 🙂
06:09:48 <jenkings> as far as i know, that is the only thing that really annoys me, on this brilliant game. And i play it since the first release by Chris Sawyer in 1994 🙂
06:42:22 <LordAro> there are complicated technical reasons why it's not really possible in the base game h the JGRPP implementation is a hack
06:43:06 <LordAro> (largely, that bridges and tunnels are literally portals)
06:43:34 <LordAro> peter1138: last day and it's pouring :(
06:46:59 <andythenorth> Hmm we know how many trains are on the bridge
06:47:30 <andythenorth> Wondering if we can partition it sonehow, into some virtual track structure
06:48:17 <jenkings> would not be a solution, to split the bridge, to multiple virtual bridges, by placing signals somewhere ?
06:48:45 <jenkings> i mean, that the bridge maybe can stand as two separate bridges, when you place a signal in the middle of bridge?
06:49:05 <jenkings> i have not seen the source of the game, so this is just an idea
06:49:25 <truebrain> andythenorth: yeah, let's do that per tile or something
06:49:30 <truebrain> waaaiiiitttt a minute
06:59:16 <andythenorth> we just need another layer of map 😛
07:00:38 <andythenorth> DOS Extended Memory?
07:12:30 <andythenorth> weird, why can't I find anything on google about the original minesweeper board game?
07:12:45 <andythenorth> was it called something else before Microsoft copied it?
07:14:48 <andythenorth> have I literally hallucinated this? In the UK in the 80s we have a physical version of Minesweeper with 2 plastic boards, each an L shape, with one face for your ships and one for the enemy ships
07:15:12 <jenkings> andythenorth: so is there a chance, that we will get in in some future updates ? 🙂
07:15:28 <andythenorth> peter1138: tells me to never say never
07:15:48 <andythenorth> given that OpenTTD is 20 years of development so far
07:16:36 <andythenorth> yes, that's the one
07:16:50 <andythenorth> ok so we "move" bridges onto an extra layer of map 😛
07:18:10 <jenkings> hmm... but will it work with tunnels too ?
07:18:37 <jenkings> however it would be great progress to have it at least on bridges 🙂
07:19:46 <LordAro> there's the UI to consider too - signals are fiddly enough as it is, imagine them being hidden behind bridge parts as well
07:22:00 <jenkings> is there a difference, it the signals are hidden on bridge, or under bridge ? 🙂
07:24:53 <reldred> extend the map array into the Z axis
07:27:53 <peter1138> Supports signals on bridges, but only path signals.
07:28:30 <peter1138> Path signals are always red, so no need for state to show a green signal 🙂
07:29:22 <jenkings> sorry, but i did not got it, how can they be always red ?
07:29:47 <peter1138> Well, "always" as in you barely ever see a green signal.
07:30:34 <peter1138> I'm not being serious here, sorry.
07:31:03 <jenkings> as i am not native speaker, i did not got it at all 🙂 sry 🙂
07:31:31 <kamnet> I've been happy with how JGR handled it, it just seems to work without issues. It's not a feature I regularly use but it's had its moments.
07:33:35 <jenkings> i guess that everything is better, than building micro islands, connected with mini bridges through 100 tiles of sea, just to avoid massive jams, waiting for bridge 🙂
07:34:32 <jenkings> second solution, not less ugly, is to build huge sea dam, just with one or two underpasses for boats 😄
07:34:38 <kamnet> Well we do have fast ships now, so that helps 😄
07:34:42 <emperorjake> There are issues with how JGRPP handles it, such as both bridge heads always having signals on them
07:34:50 <emperorjake> so you can't always control your signal spacing well
07:35:22 <emperorjake> It's far from perfect but a lot better than nothing
07:35:51 <jenkings> still better than sea dam 😄
07:56:45 <_jgr_> emperorjake: This is largely a UI problem. This restriction could be removed, but then how are you going to interact with the bridge/tunnel, as opposed to what is over/under it?
07:57:40 <peter1138> Pixel-perfect positioning 😄
07:58:27 <emperorjake> _jgr_: If there were a way to have a signal on just one end of a bridge, that would solve pretty much all of problems I had with it
07:58:32 <peter1138> (A bit like half-tile track placement)
08:12:04 <_jgr_> emperorjake: When I get around to version N+1 of it, I can look into of
08:21:54 <andythenorth> kamnet: I broke it straight away
08:22:11 <andythenorth> I missed my vocation as a QA tester
08:22:15 <andythenorth> "you're using it wrong"
08:25:28 <peter1138> "broke" or "didn't do what you wanted it to do"?
08:27:18 <andythenorth> I'd have to look in the discord logs, I think most likely I got trains stuck, or crashed them
08:27:52 <andythenorth> I am quite reliable for finding PEBKAC issues
08:28:52 <_jgr_> kamnet: I'm not the original author of it, and would probably do it a bit differently, but it's Good Enough™
08:29:51 <andythenorth> Realistic Braking is also Realistic Signal Aspects?
08:29:56 <andythenorth> or are they unrealistic?
08:35:04 <_jgr_> Multi-aspect signalling isn't supported when not using realistic braking, because realistic braking changes all signalling to PBS and disabled stuff like pre-signals where that wouldn't make any sense
08:37:02 <andythenorth> I don't play much JGRPP, but I tried to break realistic braking, and I couldn't break the braking
08:37:31 <andythenorth> I also didn't get why it's described as 'hard mode' and 'expert only'
08:37:48 <andythenorth> just seemed like a way to not have signals always be red
08:39:01 <_jgr_> Beginners who just place signals at random are potentially not going to have a good time
08:40:07 <_jgr_> In general it needs some extra thought, and it doesn't match what is in all the youtube tutorials, etc
08:44:58 <reldred> I'm pretty smoof brained and I like realistic braking
08:45:13 <reldred> I dunno, seems to just make things work more betterer
09:00:48 <peter1138> Realistic braking for ships.
09:08:42 <alfagamma7> Anti collision for ships in vanilla
09:09:05 <peter1138> They don't collide 🙂
09:17:12 <xarick> as for how many AIs it will break... I don't know, i would need to check
09:19:25 <peter1138> Yeah, but why hungry
09:44:08 <xarick> wondering if the error message returned should be "already built"
09:44:28 <xarick> since AIs do not build the tracks in front of it
09:45:43 <peter1138> Isn't the point of allowing overbuilding to not have an error?
09:46:12 <xarick> yes, but in this case, overbuilding doesn't do anything posterior to it
09:46:50 <peter1138> It never does, right?
09:47:07 <xarick> it connects to the track
09:47:20 <xarick> that's where overbuilding is useful
09:50:32 <peter1138> While I try to understand what Xarick means.
09:50:48 <peter1138> An command that succeeds should also return an error?
09:51:44 <xarick> not sure, i think it should not succeed
09:51:49 <andythenorth> instead of people writing OpenTTD AIs, we should make an AI generator
09:52:00 <andythenorth> and then let people provide the feedback
09:52:33 <xarick> but without checking which AIs attempt overbuilding, it's hard to tell if the new change is fine
09:52:44 <andythenorth> writing these insanely faceted rule-based AIs is weird
09:53:21 <peter1138> Well, I tagged you because you're good at doing lots of AI testing... so... 😄
09:53:34 <peter1138> (Even if it is sometimes irrelevant, hehe)
09:54:11 <peter1138> Why does a watched CI take longer?
09:54:26 <xarick> i still think, for safety, it should not succeed
09:55:13 <xarick> there's nothing to gain from overbuilding as an AI
09:55:20 <andythenorth> peter1138: it's the measurement problem
09:56:11 <andythenorth> if you observe it, the universe has to run that slightly slower
09:56:20 <andythenorth> as you're collapsing the waveform constantly
09:56:42 <andythenorth> it's basically about the same as OpenTTD lagging when the train window is open
09:57:20 <kuhnovic> xarick: I think we should then block depot overbuilding altogether for AIs. Currently they're already able to rotate depots.
09:58:26 <andythenorth> hmm looks like my ships have hit the dock
09:58:32 <xarick> well so far I haven't found any issues with the rotating depots
09:58:51 <xarick> but i didn't check thorough
10:00:04 <peter1138> Same principle though.
10:00:21 <xarick> it's kinda risky if the AI isn't coded properly
10:00:53 <peter1138> kuhnovic: Why can AIs do it but players can't?
10:01:17 <kuhnovic> That's not what I'm saying 😛
10:01:46 <peter1138> I tried with Ship depots, of course. They can't 🙂
10:02:18 <xarick> so... uhm... yeah... I don't know what to say, I'm thinking of overbuilding on a depot connected to another route would be then rotated to connect to the new route... it's ....
10:02:53 <xarick> if the AI isn't coded with other checks such as IsBuildable
10:03:01 <xarick> that might be worrying
10:03:30 <kuhnovic> My thoughts on this: it's already possible for AIs to overbuild (non-ship) depots, so if anything this change makes it more consistent. But whether we want AIs to overbuild in the first place is another matter. I can create a PR to prevent this, but that will be a different PR.
10:05:31 <xarick> maybe update the description for BuildRailDepot and BuildRoadDepot
10:05:59 <xarick> that it will overbuild when rotating
10:07:10 <xarick> but the one that overbuilds into the same direction, this one i really feel it should produce already built
10:07:29 <peter1138> For depots it does.
10:08:00 <peter1138> Unless this PR changes that. Hmm.
10:08:40 <peter1138> I didn't read the title hehe
10:09:01 <peter1138> Or any of it, apparently.
10:09:41 <xarick> i like the overbuilding feature if it changes anything
10:09:53 <peter1138> Isn't the track-bit building client-side?
10:10:05 <peter1138> I mean, by callback.
10:10:20 <xarick> yes, the callback however isn't used for AI
10:11:51 <xarick> there are some very strange post command rules for AIs
10:12:28 <peter1138> So maybe instead of changing the command, the interface can test if the depot already exists, and jump straight to the callback. Instead of issuing the build command again.
10:12:50 <peter1138> No behaviour change for AIs.
10:13:12 <peter1138> It's approved so I'm just bike-shedding here.
10:13:50 <xarick> well, as the screenshot shows, the depot wasn't connected to the track
10:14:48 <xarick> the feature works, but for no benefit
10:16:08 <xarick> if you're asking why it doesn't connect... I dunno, different rules, it's been like that since forever
10:17:31 <peter1138> No, I was absolutely not asking that.
10:59:05 <peter1138> Heh, Linux and Windows don't mind std::make_optional with a private constructor. macOS, though...
11:05:20 <kuhnovic> xarick: The benefit is that the connecting tracks are rebuilt if they weren't built previously
11:06:45 <kuhnovic> Which is exactly what I wanted to fix. I wanted to build a depot next to a piece of track, but a train came whizzing by so the tracks weren't added. I immediately clicked to build the depot again, thinking it would overbuild and add the tracks, but it didn't.
11:12:34 <kuhnovic> If this change turns out to be problematic, which we'll probably only know after some time, then we can do something about it. There's still time until 15.0 is rolled out 😛
12:10:10 <peter1138> Well, I was trying to work out what this was for in my RAII code 🙂
12:58:30 <truebrain> If you keep failing a Captche, does that make you a bot?
13:00:54 <ahyangyi> Can I turn myself into a bot in this way? 😛
13:01:44 <truebrain> I am trying to sign up for the Oracle Cloud, to see what it is about
13:01:52 <truebrain> but it is the most hostile form of signup I have ever witnessed
13:01:58 <truebrain> the Captcha is ... HORRIBLE
13:02:03 <truebrain> but the rest of the signup is also terrible
13:02:08 <_glx_> is it one with photos and object to detect ?
13:02:12 <truebrain> so now I am in a chat with a "live agent"
13:02:23 <truebrain> which just suggested to me to read a page how to ask LLMs a question
13:02:41 <truebrain> now it is actually connecting me to a live agent
13:02:46 <truebrain> with an expected wait time of 3 seconds
13:02:49 <truebrain> which was 5 minutes ago
13:02:56 <truebrain> this is all kinds of terrible, in so many ways 😛
13:03:14 <merni> whatever happened to good old fashioned phone support
13:03:25 <truebrain> owh no, I am connected
13:03:48 <ahyangyi> This is what they mean by "AI increased the efficiency of our customer service system by x times" I guess
13:03:52 <merni> sure it isn't just a marginally better chatbot? :P
13:03:54 <_glx_> merni: it's exactly the same, a bot answers first
13:04:17 <merni> I haven't had voice recognition bot so far, maybe those are not so popular here
13:04:29 <merni> those "press 1 for x, 2 for y" things of course exist
13:05:01 <ahyangyi> By solving the problem "we actually do things", the cost gets cut down 😛
13:05:45 <truebrain> `please follow above steps and try after some time hope it should work`
13:05:51 <truebrain> basically, he suggests me to wait an hour and try again
13:06:07 <ahyangyi> _glx_: Great prediction 😛
13:07:13 <truebrain> I love it when customer services uses those words
13:08:08 <ahyangyi> I remember seeing "independent advisors" offering this kind of suggestions in various Microsoft support forums as well
13:08:22 <truebrain> from what I gather, the signup rate limits you how often you can signup. Which is totally fine ofc. But also when the signup failed. Which is horrible 😛
13:08:28 <ahyangyi> "Try restarting your computer, it should work. If it doesn't, try reinstalling Windows and removing everything else"
13:08:46 <truebrain> Their first suggestion was: try removing cookies, try an incognito window, try another browser
13:08:51 <truebrain> like ... how should ANY of that matter
13:09:04 <truebrain> it reads more as: "please, no, we don't want you as customer, please go awayyyyyyy"
13:09:54 <ahyangyi> Ah, I remember when I received a faulty monitor and sent them a photo showing a BIOS screen with lots of glitches, they suggested me to reinstall Windows 😛
13:10:05 <ahyangyi> surely that would help my BIOS screen
13:10:21 <truebrain> yeah, all nice and all, but this is for a SaaS, where I didn't pay anything yet
13:10:28 <truebrain> you would expect that kind of process to be a bit more ... smooth
13:14:06 <tateisukannanirase> Is there a shortcut to 'snap' or convert the DiagDirection enum to just north or south?
13:15:22 <peter1138> Ah, FS2OTTD/OTTD2FS is also used for Windows calls that use wchar_t, and are unrelated to filesystems. Nice.
13:16:46 <_glx_> yes FS2OTTD/OTTD2FS is used whenever we need utf8 char to utf16 char conversion
13:17:36 <_glx_> often for file system, but also for any API use
13:17:43 <peter1138> Not always. In some places we use other calls.
13:19:38 <peter1138> Sometimes we use contro_from_fs/convert_to_fs directly.
13:19:56 <peter1138> Sometimes we use MultiByteToWideChar.
13:19:56 <truebrain> Being consistent isn't fun
13:20:34 <peter1138> I believe Windows 10 allows you to do everything with utf-8, if you set the appropriate incantation.
13:20:58 <peter1138> mac OS is also weird, sometimes is utf-8 and sometimes utf-16, and that may depend on the OS version.
13:21:00 <truebrain> drop support for Win7/win8 you say?
13:21:27 <johnfranklin> I cannot use Alt Code to type symbols on this machine
13:22:31 <Eddi|zuHause> i haven't used alt codes in like 20 years
13:22:34 <peter1138> Well. Turns out the CI finishes sooner if you actually click on Create Pull Request.
13:22:57 <peter1138> win-. is the new meta.
13:23:29 <Eddi|zuHause> whenever i now need "weird" characters i use kcharselect.
13:23:36 <_glx_> the API will still need utf-16, the only thing that may be default with win10 is utf-8 codepage
13:23:57 <peter1138> Oh it's not default 🙂
13:24:45 <_glx_> but legacy stuff can have issues
13:24:47 <peter1138> See the section about -A and -W
13:24:48 <Eddi|zuHause> i'm very certain unicode is one of the inspirations for xkcd "standards"
13:25:13 <peter1138> > Until recently, Windows has emphasized "Unicode" -W variants over -A APIs. However, recent releases have used the ANSI code page and -A APIs as a means to introduce UTF-8 support to apps.
13:26:28 <peter1138> But don't worry, I'm not suggesting we change anything... yet... 🙂
13:28:06 <peter1138> To be fair, most of the MultiByteToWideChar stuff is actually in OTTD2FS/FS2OTTD 🙂
13:28:16 <_glx_> but compiler still defaults to -W variant because -DUNICODE is a default flag in almost all recent VS versions
13:29:27 <truebrain> ```We're unable to complete your sign up. Common errors that prevent sign up include:
13:29:27 <truebrain> a) Entering incomplete or inaccurate information.
13:29:27 <truebrain> b) Intentionally or unintentionally masking your location or identity.
13:29:27 <truebrain> c) Attempting to create multiple accounts.```
13:29:30 <_glx_> oh no we force it in cmake actually
13:29:32 <truebrain> seems I am just not allowed to sign up 😦
13:30:26 <Eddi|zuHause> what kind of signup process asks you to unmask your location?
13:31:26 <_jgr_> I suppose they can afford to have lower prices when they've got no customers 😛
13:32:27 <_glx_> anyway our current utf-8 handling for windows works fine
13:32:53 <peter1138> Oh it works. It's definitely not fine.
13:34:40 <_glx_> ```GDI doesn't currently support setting the ActiveCodePage property per process. Instead, GDI defaults to the active system codepage. To configure your app to render UTF-8 text via GDI, go to Windows Settings > Time & language > Language & region > Administrative language settings > Change system locale, and check Beta: Use Unicode UTF-8 for worldwide language support. Then reboot the PC for the
13:34:40 <_glx_> change to take effect.``` makes me think we should not change anything 😉
13:35:12 <peter1138> It's almost like I said we shouldn't 10 minutes go...
13:35:25 <peter1138> And not that I was ever suggesting we should.
13:35:40 <peter1138> I just found it interesting that the API's stance on it has changed.
13:35:54 <truebrain> _jgr_: that might just be it, honestly. I cannot imagine anyone managing to sign up for this.
13:36:03 <_glx_> they finally moved in the right direction
13:37:05 <truebrain> `Your inquiry has been received. You should expect a response within 48 hours.`. Great.
13:37:15 <_glx_> all this CP silliness coming from DOS
13:37:58 <_glx_> like the forbiden file names
13:46:54 <Eddi|zuHause> hard to get out of
14:10:32 <Rubidium> should that maybe be a Codefix (or introduce Coderevert)?
14:12:06 <peter1138> I mean it's not a direct revert sure..
14:37:19 <tateisukannanirase> reldred: yesterday in Discord channel #openttd we were talking about improving YAPF performance for RVs. Would appreciate if you could test my modifications here with your tram maps. Cheers! https://github.com/OpenTTD/OpenTTD/pull/12558
15:10:51 *** HerzogDeXtEr has joined #openttd
15:17:27 *** Wormnest has joined #openttd
16:05:00 <peter1138> It was VS 2005? Figures, lol 😄
16:09:37 <alfagamma7> Nml supports stations now, right?
17:16:16 <peter1138> The PR didn't link.
17:17:04 <kuhnovic> I believe it was traced back to an std::map using pointers as keys, which led to indeterminate behaviour
17:17:24 <peter1138> Was in the PR title, but nowhere else, cos I didn't get a chance to edit the commit title on merge.
17:18:10 <peter1138> Sorry, this developer was tired and didn't notice 😦
17:20:00 <peter1138> Okay, how can I download the NotPerfectAi mentioned in #8230?
17:21:05 <peter1138> Hmm, it's in Github.
17:27:21 <efessel> I thought I saw the root cause was found but wasn't sure
17:48:51 <_glx_> efessel: It was a fun one to track, mainly the unexplained randomness of replaying commands from the same starting save
17:52:51 <efessel> So it was the non-deterministic map indexes using pointers instead of id's?
17:55:19 <efessel> I swear game bugs can be the most frustrating (and most rewarding) bugs to find/fix.
17:55:38 <_glx_> To be replaced véhicules where sorted by address in memory
17:56:37 <peter1138> Got sidetracked by a pool change that was not the problem...
17:58:31 <_glx_> Based on the effect it was a possible cause
18:02:24 <Rubidium> peter1138: not really. I don't think it's usefull/worth it to keep updating a PR that seemingly nobody looked at in depth, except for comments in the sense of "it would be better to do a proper refactor of the affected code"
18:19:15 *** peter1139 has quit IRC (Remote host closed the connection)
18:19:34 *** peter1139 has joined #openttd
18:31:47 <peter1138> My 10+ year old patches have something to tell you...
18:49:41 <peter1138> Hm,m do all bananas tar files end up as "blah.tar/blah/content"?
19:00:21 <truebrain> peter1138: Name of tar file and folder in it?
19:12:30 <_glx_> it used to not reference tar name IIRC
19:22:15 <peter1138> So if we're trying to find a file called GM-TTO.CAT and it is, at that point, just GM-TTO.CAT, there's no point in searching the tar list, right?
19:24:19 <truebrain> Files downloaded via BaNaNaS will always have a single root folder in the tar too, so even a/b.cat will never match those tars. But a custom tar might violate that concept, ofc
19:26:02 <truebrain> But as glx mentioned, the prefixing of the tarname is relative new
19:26:56 <_glx_> yeah and now you can't have original TTD files in a tar 😉
19:27:45 <truebrain> It did fix a very old bug
19:28:23 <_glx_> I was probably one of the rare person to use a tar for original files
19:49:50 <Eddi|zuHause> "workaround for being correct" what? :p
19:51:12 <_glx_> some newgrf relying on our incorrect implementation of the specs
19:51:27 <Eddi|zuHause> yes. i've seen that discussion
19:51:38 <Eddi|zuHause> i meant the wording :p
19:52:49 <Eddi|zuHause> essentially you're forcing "undefined behaviour" to the assumed value
19:53:18 <peter1138> I'm not, because it's closed.
19:55:12 <Eddi|zuHause> well, you were. but now you aren't
19:57:20 <peter1138> Feel free to add comments to the PR now.
19:58:28 <peter1138> Instead of 8 minutes after it was closed after being open for 9 days.
20:00:39 <peter1138> Right that's it, the window for comments is now closed 😉
20:00:44 <Eddi|zuHause> i'm not sure i have any "constructive" comments
20:10:53 <peter1138> Hmm, now to find a tar file with symlinks...
20:11:36 <peter1138> Back in the SVN days when we had less meta-data 😄
20:12:15 <truebrain> Why did we need symlinks? 😛
20:12:27 <truebrain> Tomorrow I can check if there is such upload in BaNaNaS
20:12:42 <peter1138> Someone must have had one for it to be added.
20:13:09 <andythenorth> can we rename "Dock"?
20:13:14 <andythenorth> I want it for an industry name 😛
20:13:32 <truebrain> peter1138: Big assumption 😛
20:15:41 <Eddi|zuHause> it's not even referencing a flyspray ticket
20:16:58 <truebrain> I am wondering the use cases where symlinks are useful, and I can't think up any 😦
20:17:07 <truebrain> For NewGRFs it seems pointless
20:17:16 <truebrain> For AIs .. maybe, but not really
20:17:52 *** nielsm has quit IRC (Ping timeout: 480 seconds)
20:17:55 <truebrain> Maybe so info.nut can be in a subfolder, but how about just not doing so 😛
20:18:06 <truebrain> Owh well, will see tomorrow if we have any 🙂
20:18:45 <peter1138> > OpenTTD does not find data files when lowercase (per Wine installation)
20:19:43 <Eddi|zuHause> is that even openttd's fault?
20:20:13 <truebrain> Using Wine to run OpenTTD, how cute is that 🙂
20:20:41 <peter1138> Trying to find something relevant on the froom 🙂
20:21:32 <_jgr_> truebrain: How else am I going to check that Windows builds actually work 😛
20:22:00 <truebrain> Hahaha, nice reversal 😄
20:22:09 <peter1138> Well, this was certainly a rabbit hole.
20:22:30 <peter1138> Not yet. RAII was hiding in there somewhere.
20:26:53 <peter1138> My _tar_linklist is 0 for all subdirs.
20:27:21 <peter1138> If we had "download all" button I could check this easily 😉
20:31:55 <peter1138> - Fix: Buffer overruns in handling of symbolic links inside tars (r26514)
20:32:40 <peter1138> So someone must have found a bug with it.
20:32:57 <peter1138> Or Rb just happened to find it when looking.
20:34:04 <peter1138> Only 10 years ago. I think I had the RGB CC patch at that point.
20:36:56 <Eddi|zuHause> i'd think finding buffer overruns is totally unrelated to the feature they're in.
20:37:41 <Eddi|zuHause> like, using some code analysis tool, or something
20:47:35 *** matusguy has joined #openttd
20:50:15 <peter1138> That's probably `[Bug] CargoDist breaks CargoDist`
20:51:33 <truebrain> frosch123: has been busy 😛
20:51:45 <truebrain> Hopefully I have time soon to integrate it 🙂
20:52:11 <frosch123> it's easy as long a rust is new an shiny 🙂
20:52:33 <truebrain> Keep trying those edits, you will get there
20:53:18 <frosch123> poor IRC people won't understand it
20:53:26 <truebrain> The error messages look nice 🙂
20:54:00 <frosch123> oh, right, i report error locations as byte offsets
20:54:14 <frosch123> if js prefers unicode code points, i guess i can also return those
20:54:43 <peter1138> frosch123: I'm not on IRC and I don't understand...
20:55:03 <andythenorth> did we have lunch?
20:55:11 <truebrain> Code points sound more useful, but let me integrate first, and we can check after 🙂
20:55:16 <truebrain> This weekend, hopefully
20:57:10 <truebrain> Btw, Javascript is UTF-16, so byte offsets are weird anyway.
20:57:22 <truebrain> (Wasm bindgen takes care of the conversion)
20:59:07 <frosch123> oh, did not know that. i would have assumed utf-8 because internet. but i guess js is from 90s
21:00:55 <truebrain> Normally you will never notice in Javascript 🙂
21:30:46 <_glx_> truebrain: monthly quota is more than what we would need for a full year 🙂
21:38:17 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:41:47 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:32:32 *** matusguy has quit IRC (Remote host closed the connection)
22:39:45 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
22:41:50 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
continue to next day ⏵
