IRC logs for #openttd on OFTC at 2023-06-23
⏴ go to previous day
00:25:49 <Eddi|zuHause> sometimes i want to just facepalm at the forums...
00:30:13 <iqTop[m]> Eddi|zuHause: Why? Because I haven't been allowed there for over two years so whatever just happened, it wasn't me.
01:55:10 <emperorjake> The multiplayer server is down atm
02:18:24 *** playback2396[m] has quit IRC ()
02:41:25 *** debdog has quit IRC (Ping timeout: 480 seconds)
02:44:57 *** D-HUND is now known as debdog
04:39:18 <talltyler> `how can you put so much nonsense into so little words?` Well said, Eddi 😛
05:01:38 <Rubidium_> peter1138: no, not really. Either go the way with std::vector<StringParameterBackup> (#11052) or if it isn't really that likely that you change something in the dropdown by just formatting to std::string (#11058)
05:08:25 *** olionkey[d] has joined #openttd
05:08:25 <olionkey[d]> This isn't right. Right?
05:09:41 <Rubidium_> peter1138: a relatively thoroug check of the code doesn't give me the idea that there are multiple places where you conceivably want to change the string in a dropdown. I was thinking about the company list, but that seems to have its own DropDownListItem. So, maybe the story book page list? I'm assuming you change the language with clicking on the dropdown, so there isn't any dropdown open
05:12:38 <ahyangyi> andythenorth: If it works, it's Iron Horse; if it doesn't, it's Irony Horse
05:12:53 <Rubidium_> k-man: maybe the track isn't electrified (for a short bit)? or the vehicle is stopped?
05:16:10 <k-man> good theory, but didn't fix it
05:18:15 <k-man> oh i think the trains were just in stopped mode for some resaon
05:40:47 <k-man> is there some way to save and quit without prompting for anything?
05:46:52 <jfs> You can depend on the autosave and just kill the game from the OS but that's an entirely different hassle
05:57:00 <iqTop[m]> <talltyler> "`how can you put so much..." <- Okay, I'm an idiot for not logging out of my banned account all this time, but not that dumb. Yeah, there's a setting for plane crashes, not sure why they didn't just ask if there was before asking for a "new feature".
06:13:23 <Rubidium_> iqTop[m]: the point is that the "nonsense" that is talked about was given in a post that quotes the post that says there is a setting to disable crashes. So, the person knows (has quoted) the fact that there is a setting to disable crashes, and then continues on with that it is impossible to play (multiplayer) without plane crashes.
06:14:13 <iqTop[m]> Rubidium_: That... is way worse, yeah.
06:16:13 <Rubidium_> suggesting there to be a setting to disable plane crashes (or something along the line) is not necessarily non-sense, maybe just a "sigh, why didn't you look through the existing settings" sentiment. But continueing to proclaim something cannot be done after you have been made aware of the setting, that's more like "cannot comprehend it is possible, and will continue in the train of thought that it is
06:25:30 <peter1138> Hmm, prerendering (to std::string) and then skipping the string system feels a bit wrong, but is probably ultimately simpler and more efficient.
06:26:52 <peter1138> I like that we already have a dropdown item that takes a std::string, but it's called, and documented, as a C string...
06:39:36 <Rubidium_> the question is: how wrong is it? Will the place in the UI survive changing languages/fonts? Is there anything that could change the strings? If so, does that trigger the dropdown to layout again in case the text got longer, or is that a bug/oversight?
06:40:40 <peter1138> As far as I know, anything you might do that could change the content of dropdowns would also close the dropdown anyway.
06:41:13 <Rubidium_> now I think of it... tooltips could probably also just prerender the string as I doubt you want the tooltip to change mid-showing
06:44:59 <Rubidium_> news and text effects can survive changing the language, so they should not be (always) prerendered. The strings on the viewport will be prerendered, as the only reason for them to change between creating and drawing is UI and game logic running in parallel which would probably end badly
07:09:52 *** keikoz has quit IRC (Ping timeout: 480 seconds)
07:10:19 <truebrain> so all instances of the game-coordinator decided to stop working? Now that is interesting .. 😛
07:12:13 <truebrain> they are both alive and kicking .. lol
07:13:08 <truebrain> owh boy, AWS decided to make a change in the lambda handlers ..
07:13:54 <truebrain> `import json ..... json.dumps` AAHHH CRASH BURN JSON DOESNT EXIST
07:15:00 <iqTop[m]> So, um... Can someone give me an opinion on these glowy but not-so-well-done "futuristic roadtypes" I made?
07:19:30 <truebrain> odd, the NLB notices there are no upstream instance to handle traffic, but ECS says they are alive and kicking ...
07:20:17 <peter1138> Hmm, dropdown lists with company names can't prerender as the name could change.
07:20:19 <truebrain> one of the two is lying 😛
07:20:28 <truebrain> the NLB is not the one that is lying ... time to cycle the instances
07:20:33 <peter1138> (But the precalculated width will be wrong)
07:23:14 <Rubidium_> precalculated width of what would be wrong? Why?
07:24:37 <peter1138> When you open one of the main toolbar menus with a company list, the list width is calculated to fit the longest company name. So if a company is added/removed/renamed while the list is still open, the width is going to be wrong.
07:25:23 <Rubidium_> ah, I wasn't thinking about that toolbar "now" as that has its own fully custom code
07:26:25 <truebrain> hmm .. no, the NLB is to blame ... do I even want to know why? Ugh ..
07:26:42 <truebrain> `botocore.errorfactory.InvalidParameterException: An error occurred (InvalidParameterException) when calling the DescribeTasks operation: Tasks cannot be empty.`
07:26:49 <truebrain> well, clearly it can
07:26:53 <truebrain> wth do you want me to do!
07:28:49 <truebrain> there we go .. now the NLB knows where to find things again
07:29:08 <truebrain> and PLOP, the whole server listing is populated again
07:29:10 <peter1138> Hmm, according to Microsoft and nVidia, WSLg should support acceleration. Alas, not for me 😦
07:29:38 <truebrain> I was surprised it did work for me
07:29:44 <truebrain> didn't work for months
07:29:49 <truebrain> but after a recent update, it all of a sudden does
07:30:32 <truebrain> emperorjake: Tnx for letting us know 🙂 All fixed again!
07:31:12 <truebrain> and clearly they are just begging to be migrated to the new infra too .. but that is for next week
07:34:42 <truebrain> you running an i386? How does that even work? 😛
07:35:10 <peter1138> Are you missing the -dev part of it?
07:36:00 <truebrain> something like `libcurl4-openssl-dev`, depending on your OS
07:36:36 <truebrain> so I removed this massive amount of files from the S3 bucket yesterday, right? Now AWS measures every day how big your bucket is .. and it just says: nah, nothing changed
07:36:59 <truebrain> at least on Cloudflare you can see at any moment in time how big your bucket is .. instead of waiting for a new measurement that LIES IN YOUR FACE 😦
07:43:12 <pickpacket> libcurl4-openssl-dev is available. I'll install that. Thanks :) I have a 32 bit Atom processor
07:43:19 <pickpacket> And it works *very* well, actually
07:43:27 <truebrain> you don't see that a lot anymore
07:43:44 <pickpacket> nope! It was a budget machine when I bought it back in 2015
07:45:39 <peter1138> Didn't we determine that it is actually a 64 bit Atom, but just a 32 bit OS?
07:46:29 <peter1138> Well, #11063 escalated a bit from "oh, decode_params isn't properly initialised" to "yeet it all"
07:47:23 <truebrain> rabbitholes in OpenTTD? NEVAH
07:47:25 <pickpacket> peter1138: it doesn't boot with a 64 bit OS though 🤷
07:48:08 <pickpacket> I don't believe lscpu for it... It claims that it's a quadcore cpu, and that's simply impossible
07:48:29 <peter1138> That's common for dual-core with hyperthreading enabled.
07:48:46 <peter1138> Well, "common", no, actually normal.
07:49:45 <peter1138> Of course hyperthreading reveals a bug that can expose data, meh.
07:50:47 <truebrain> `failed to create R2 bucket: The bucket you tried to create already exists, and you own it.`
07:54:35 <peter1138> Hmm, opening the settings window has a noticable pause. I assume it's from loading all the strings for each setting to determine how big windows need to be.
08:20:12 <pickpacket> peter1138: because hunger time?
08:24:13 <pickpacket> "nlohmann_json not found; compiling OpenTTD without nlohmann_json is strongly discouraged"
08:24:22 <pickpacket> the hell nlohmann_json?
08:32:07 <pickpacket> I compiled 13.3 just the other day and can't remember seeing these warnings then. Are those deps new or was I just not paying attention?
08:36:00 <LordAro> well, not part of release/13 anyway
08:36:19 <truebrain> a 20 year old game, and every release, we add new goodies 🙂
08:40:18 <peter1138> Urgh, why are browser select lists so dumb 😦
08:42:54 <truebrain> oof, there are many links to the CDN .. this will be a bit of work to untangle that web again 🙂
09:53:45 <peter1138> So what happened to the coordinator?
10:16:27 <truebrain> you are jumping back to 3 hours ago? 😛
10:17:38 <truebrain> truebrain: this happened to the coordinator 🙂 The NLB was confused 🙂
10:17:43 *** zerokek has joined #openttd
10:18:23 <truebrain> (basically, it wasn't updating where instances were located .. so it was forwarding traffic to nothingness)
10:18:57 <truebrain> what is a bit ironic, it was broker for more than a week, but as instances rarely crash or are restarted, the old data the NLB had was fine .. it was only till all instances had crashed, that it went offline 🙂
10:19:08 <truebrain> shows how stable the software is 😄
10:32:15 <peter1138> Just not quite stable enough 🙂
10:33:02 <peter1138> So the font I am using causes something to slow down. Possibly the layouter but I thought those got cached.
10:41:47 <Rubidium_> yeah... that seems noticably :D
10:42:55 <Rubidium_> looks like the end of the 's'-es is missing a bit; some off-by-one error somewhere?
10:45:06 <peter1138> I don't see what you mean.
10:45:29 <peter1138> Oh in the DejaVu Sans one. Hmm.
10:46:00 <Rubidium_> it's quite noticable in the second image, not that much in the first image
10:47:20 <peter1138> Usual trick, the width of the string doesn't include space for shadows.
10:47:28 <peter1138> I'd forgotten about the one tbh 🙂
10:48:25 <peter1138> With the sprite font, the shadow is baked in and the extra width provides the spacing between characters.
10:49:11 <peter1138> With truetype, it's 1) not baked in, and 2) the width and spacing is provided by the font and layouter, so just including it would space things out too much.
10:50:21 <peter1138> But we already have compensation to provide extra vertical space when setting up cropping for text.
10:50:27 <peter1138> Maybe we need the same for horizontal.
10:50:48 <peter1138> But none of this is the performance issue I'm seeing 🙂
10:54:52 <peter1138> Fairly simple I guess.
10:55:08 <peter1138> It means it's technically overflowing its widget though, but maybe that's okay.
10:55:31 <peter1138> There may even be a GH issue about this one.
10:57:36 <truebrain> _glx_: As they didn't crash, no 🙂
11:02:29 <peter1138> This performance issue is enough to impact Wentbourne 🙂
11:04:22 <peter1138> Hmm, plan, try the same font in Windows.
11:08:29 <peter1138> Yeah, no issue there, although truetype font rendering is measurably slower than sprite font, but not anywhere near a problem.
11:09:53 <peter1138> Hmm, maybe it's breakfast time
11:15:39 <pickpacket> Happy Midsummer, everyone
11:24:10 <truebrain> peter1138: well, there hasn't been a downtime of more than 5 minutes in 4 years; so .. 99.999% or something? 😛 Guess "stable enough" 😄
11:25:08 <peter1138> More stable than Debian stable (yes, I know)
11:26:47 <truebrain> well, 5 9s is already pretty good 🙂
11:26:54 <truebrain> understatement of the year 😛
11:34:13 <pickpacket> That’s better than Fastly
11:35:46 <pickpacket> Technically their outage was only a minute or so I guess, but it cause several hours of widespread outages for services that used them
11:36:26 <pickpacket> Hasn’t gmail gone down for a few hours at least once in the last four years?
11:44:33 *** zerokek has joined #openttd
11:55:46 *** gill6151 has joined #openttd
12:17:20 <peter1138> Why can I not reply in one go... or am I missing something every time.
12:18:08 <Rubidium_> you can only do that if you review yourself
13:11:54 *** ChanServ sets mode: +v tokai
13:40:22 <LordAro> i think they want a response
13:43:10 <peter1138> Oops, committed too much 🙂
13:46:13 <truebrain> Responding to a single image without giving any context? Feels if they don't want to put in the effort, why should we? 😛
13:50:13 <truebrain> no, I appreciate you are less childish than I am 🙂
13:56:42 <peter1138> HTML select sucks... just use datalist... that also sucks... I prefer our dropdowns to this crap 😦
14:16:54 <AlphaGamma3521> _zephyris: Hi! So I tried opengfx2 out. Looks almost like the original graphics , didn't like some buildings drawn with extrazoom in mind. (Sorry but I like Opengfx1 more.) Waiting for Opengfx2+ Landscape eagerly. Btw are you going to redraw the vehicles in the style of the original graphics as well?
14:37:02 <iqTop[m]> <truebrain> "Responding to a single image..." <- Huh?
14:39:20 <iqTop[m]> You mean the currency codes? Or the two sprite sheets of roadtypes I posted?
14:39:57 <iqTop[m]> Sorry, I've got trust issues with you guys right now. I'm not sure how to deal with it.
14:45:44 *** gelignite has joined #openttd
14:46:27 *** gelignite has quit IRC (Read error: Connection reset by peer)
14:50:21 <iqTop[m]> If you did mean the roadtypes, they're supposed to be "futuristic" roadtype equivalents to monorails and maglevs.
14:52:10 <iqTop[m]> Basically, "SmartStreets" and "SmartRoads". The former is first available in 2000, the latter at some point in the New 10s, with both being based on less futuristic-looking but equivalent real life smart vehicle/highway systems.
14:52:25 *** gelignite has joined #openttd
14:57:20 *** HerzogDeXtEr has joined #openttd
15:30:20 *** gelignite has quit IRC (Read error: No route to host)
15:33:10 <iqTop[m]> <peter1138> "FFS" <- ?!
15:57:05 <andythenorth> 'ctrl-click stations' is cheating though?
15:57:38 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:03:09 <pickpacket> andythenorth: very cheaty 🤪
16:03:20 <pickpacket> Even if it’s enabled
16:03:21 <andythenorth> along with RVs and ships
16:03:26 <andythenorth> and station spread
16:03:31 <andythenorth> and 90 degree curves
16:03:55 <pickpacket> and stations longer than one tile
16:04:07 <andythenorth> I was with you until the one tile station thing 😛
16:04:22 <andythenorth> but the rule was actually TL 5 or TL 6 **everywhere**
16:11:43 *** m3henry has joined #openttd
16:18:29 *** Eddi|zuHause has quit IRC (Remote host closed the connection)
16:19:26 <m3henry> I just noticed that IH Merchandise open, hood open and regular open wagons were merged with variants, woo, but silo and cement wagons were not. Am I missing a functional difference?
16:20:37 *** Eddi|zuHause has joined #openttd
16:22:21 <andythenorth> no, it's just a default cargo and appearance difference
16:22:39 <Eddi|zuHause> trains longer than 7 tiles are definitely also cheating
16:22:39 <andythenorth> some groups are merged, some aren't, it's a bit vague as to why, it's just a gut feel thing
16:22:58 <andythenorth> isn't there some thing about trains reversing at EOL that was considered cheating?
16:23:06 <andythenorth> also PBS *was* considered cheating, but times have changed
16:23:16 <andythenorth> was YAPF cheating?
16:23:16 <Eddi|zuHause> no, that's a feature of the original game
16:23:51 <Eddi|zuHause> yapf is no more cheating than npf
16:24:30 <Eddi|zuHause> and before npf we had ntp
16:24:48 <Eddi|zuHause> which all were about equal
16:26:12 <Eddi|zuHause> i don't know if its still there, but in the old days junction examples were described with "trains have to turn right to turn left", which is really only a problem if you didn't use any of these three pathfinders
16:26:21 <andythenorth> I can't think of any other things considered cheating
16:26:44 <andythenorth> ctrl-click stations, along with use of ships, RVs and planes were the main ones
16:27:08 <andythenorth> or was that a valid meta?
16:27:10 <Eddi|zuHause> plane speed setting is cheating
16:27:19 <andythenorth> were escape depots cheating, or meta?
16:27:21 <peter1138> Now they just consider PBS "lazy"
16:27:34 <andythenorth> 'they' = one person with bizarre views on reddit
16:27:40 <andythenorth> and reddit is off so eh
16:27:55 <peter1138> "Nothing of value was lost" 🙂
16:28:42 <peter1138> That reminds me, time to manually delete more reddit comments that slowly come back.
16:30:36 *** Eddi|zuHause has quit IRC (Remote host closed the connection)
16:32:32 *** Eddi|zuHause has joined #openttd
16:38:19 <ahyangyi> andythenorth: I'm just curious, what's the piglet?
16:38:31 <ahyangyi> ( besides meaning "baby pig" in normal English )
16:41:27 <andythenorth> I can't find a reference on google
16:41:48 <andythenorth> but it's a single RV at a station which just cycles around a loop doing load / unload
16:41:57 <andythenorth> it keeps station ratings high
16:43:00 <Eddi|zuHause> didn't HEQS specifically add the forklift for that purpose?
16:43:26 <dP> andythenorth: aka rating pushing
16:43:37 <dP> first time I see it called piglet though
16:46:04 <pickpacket> Eddi|zuHause: the only thing that isn't cheating is vanilla with default settings, OpenGFX, OpenFSX, and OpenMSX. No NewGRFs.
16:46:23 <Eddi|zuHause> vanilla has plenty of cheating
16:46:30 <andythenorth> vanilla is all cheating
16:46:33 <andythenorth> especially in MP
16:46:48 <andythenorth> what was the depot crash thing?
16:47:17 <Eddi|zuHause> you put the depot at the end of a competitors line, and build a train that crashes into the competitor train
16:47:29 <pickpacket> Eddi|zuHause: you're right. Playing the game is cheating
16:47:31 <andythenorth> so "infrastructure sharing"
16:47:57 <Eddi|zuHause> somewhat... you can't actually enter their tile, but the overlap is enough to cause the crash
16:48:11 <Eddi|zuHause> (this has been disabled now)
16:49:05 <peter1138> There's still some tile-end issues though 🙂
16:50:18 <Eddi|zuHause> anyway, the original AI never actually replaced the crashed trains
16:59:43 <jfs> are you manually assigning vehicle IDs to them in your NML or NFO?
17:01:37 <andythenorth> the nml auto-id is a misfeature 🙂
17:01:44 <andythenorth> it cannot end well
17:02:59 <Eddi|zuHause> pickpacket: there's a sort "function" in nml, where you can put in the default trains and your trains
17:19:07 <pickpacket> Eddi|zuHause: thanks :) I’ll look it up
17:19:32 <pickpacket> Btw the topic in the irc channel still says ”13.1”
17:19:39 <Eddi|zuHause> it works a bit weird, because the underlying NFO feature is even weirder
17:19:55 *** Kitrana has quit IRC (Quit: Leaving.)
17:20:41 *** Kitrana has joined #openttd
17:45:16 *** DorpsGek changes topic to "13.3"
17:45:39 <glx> @topic set 13.3 | Website: *.openttd.org (source: github, translator: translator, server list: servers, wiki: wiki) | Don't ask to ask, just ask | 'Latest' is not a valid version, 'Most recent' neither | English only
17:45:39 *** DorpsGek changes topic to "13.3 | Website: *.openttd.org (source: github, translator: translator, server list: servers, wiki: wiki) | Don't ask to ask, just ask | 'Latest' is not a valid version, 'Most recent' neither | English only"
18:24:03 *** gelignite has joined #openttd
18:41:38 <DorpsGek> - Update: Translations from eints (by translators)
18:46:54 <andythenorth> played my game for 130 years
18:46:58 <andythenorth> might be time for a new one
19:13:11 <peter1138> Hmm, need to undo a couple of translations
19:15:36 <peter1138> Or even just fix some.
19:29:02 <andythenorth> let's try Async GS
19:29:21 <andythenorth> without it, on a 512x512 map with 'high towns' monthly town processing appears to never complete
19:29:33 <andythenorth> which is expected ish
19:30:03 *** Wormnest has joined #openttd
19:35:27 <andythenorth> not sure how to use GSAsyncMode
19:36:33 <andythenorth> GS appears to stay in Async after the function exits, even though the mode object is local to the function
19:37:00 <andythenorth> if I try setting `local asyncmode = GSAsyncMode(false);` at the end of the function, GS errors
19:37:54 <andythenorth> presumably I don't want to just put the script in async permanently?
19:52:15 <andythenorth> can't see a way to read the current mode
19:59:48 <andythenorth> hmm maybe this Async is fine
19:59:54 <andythenorth> maybe it's just _that much_ faster
20:01:14 <andythenorth> no it really is fast
20:01:26 <andythenorth> I was confused, it's so much faster I thought my script was broken
20:06:09 <andythenorth> probably don't want to build industries in Async mode
20:12:35 <peter1138> Isn't it still limited or is that just as a network server?
20:13:47 <andythenorth> seems quite unlimited in SP
20:14:01 <andythenorth> I don't have a reliable way to tell when the async jobs actually complete
20:14:07 <andythenorth> but the storybook is updating quickly
20:14:34 <Eddi|zuHause> one would assume the async commands complete on next tick
20:17:17 <truebrain> lol, do we need to slow down SP for async too? 😛
20:25:29 <peter1138> Well it'll be limited by the op-code limit anyway.
20:28:44 <andythenorth> some kind of timing tool perhaps?
20:52:13 <jfs> I haven't looked much into the async GS thing, but is/would it be possible to ask the async mode object for a kind of cookie for the last invoked command, and later on ask for the command result associated with that cookie?
20:52:46 <jfs> (perhaps have some kind of timeout on how long a command result is stored)
21:13:53 <andythenorth> things with no map effect...
21:59:57 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:06:17 *** gelignite has quit IRC (Quit: Stay safe!)
22:16:59 <Eddi|zuHause> well, if you try to do things with map effect, they might fail because the map changed inbetween
22:17:19 <Eddi|zuHause> and you'll never find out about it failing or working
22:19:02 <andythenorth> trees...forest...sound
22:33:52 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
continue to next day ⏵
