IRC logs for #openttd on OFTC at 2023-07-20
⏴ go to previous day
00:00:51 <kamnet> johnfranklin: Which airport patch?
00:14:45 <kamnet> Oh yes, that's a nice one. I got to design one of the layouts (Indianapolis), it works super well.
02:25:52 *** debdog has quit IRC (Ping timeout: 480 seconds)
02:45:42 *** Wormnest has quit IRC (Quit: Leaving)
04:15:04 *** HerzogDeXtEr has joined #openttd
04:27:26 *** D-HUND is now known as debdog
05:08:46 *** alfagamma_0007 has joined #openttd
05:08:46 <alfagamma_0007> Compiling old patches is a pain in the backside. Either I am screwing it up or I should have read the instructions properly.
05:52:32 <jfs> if you're trying to port them up to current versions of OTTD then yes it's going to be a big task, lots of little things have changed everywhere in the code
06:04:36 *** nielsm has quit IRC (Ping timeout: 480 seconds)
07:20:28 <shrekshellraiser> So I'm curious about why something like the polyline tool hasn't been added to base OpenTTD, is it just due to lack of effort to do so? Or is it a conscious decision
07:24:50 *** pickpacket has joined #openttd
07:27:01 <pickpacket> I’m playing a multiplayer game where I’m having some issues. I host it on my (cr|l)aptop, which is having no problem at all with it
07:29:02 <pickpacket> but the other player is using my RPi 3B+ and is *really* struggling as the game progresses. We’re now 70 years into the game and there are 1000+ vehicles on the map and the lag on that client is absolutely unbearable. It’s using 100% CPU all the time
07:30:16 <pickpacket> towns and trees are hidden in the transparency settings. I don’t know what else to do to lessen the strain on the CPU
07:31:14 <pickpacket> my dual core x86 has no issues whatsoever but the quad core ARM does 😕
07:32:12 <shrekshellraiser> I don't think openttd uses multiple cores
07:32:23 <shrekshellraiser> So you'd be bound to single core performance
07:33:33 <andythenorth> on older mac hardware, turning off full animation made a big difference
07:33:33 <dwfreed> also rpi performance is pretty crap in general
07:33:49 <andythenorth> but that was also specific to the mac build and intel integrated graphics chip
07:35:02 <truebrain> 1000+ vehicles, lol
07:36:07 <truebrain> I guess you found out why we limit it to 500 per company by default 🙂
07:36:27 <andythenorth> does the game use FIRS?
07:36:30 <andythenorth> FIRS kills performance
07:36:40 <truebrain> I blame the author btw
07:37:10 <shrekshellraiser> shrekshellraiser: Do any of you have inputs on this?
07:37:31 <andythenorth> the things that get added are things that interest people with approval rights
07:37:43 <pickpacket> truebrain: with a 500 vhcl limit I can’t expand more after 50 years 😉
07:38:02 <truebrain> pickpacket: but don't come complaining to us if you delibrately push the game over its max 😛
07:38:10 <alfagamma_0007> jfs: I agree. I was trying my hand on the shunting patch that someone on the forums had done some years back.
07:38:34 <pickpacket> I’ve set the limits to 800 trains, 500 boats, 500 eoad vehicles and 200 planes per company 😄
07:38:38 <shrekshellraiser> andythenorth: Well yeah.. but that tool has existed as a patch for years
07:38:56 <pickpacket> truebrain: I’m not complaining, just seeking advice 😉
07:38:56 <andythenorth> I believe there are 2 polyline patches
07:39:04 <truebrain> pickpacket: advice: use less trains 😄
07:39:15 <andythenorth> I think there's a polyline in JGRPP and a different one in CityMania, but this might be fake news
07:39:16 <dwfreed> alternately: get better hardware
07:39:23 <truebrain> (sorry, I am not being all that serious pickpacket 🙂 )
07:39:24 <shrekshellraiser> JGR definitely has one
07:39:25 <pickpacket> More trucks and boats instead? 😜
07:39:48 <andythenorth> are you using grf trains?
07:40:01 <andythenorth> Iron Horse has about twice the CPU demand of vanilla trains (for example)
07:40:14 <alfagamma_0007> My rpi 4B doesn't work well with long length jgr games either
07:40:28 <alfagamma_0007> andythenorth: Hmm
07:40:28 <alfagamma_0007> Makes sense now
07:40:58 <pickpacket> andythenorth: Yes. AsiaStar Redux and the tea carriages from Tea Tea Deluxe
07:41:34 <pickpacket> though the problems arose long before the AsiaStar Redux trains came into play
07:42:42 <shrekshellraiser> shrekshellraiser: Citymania mentions terra forming with the polyline tool, so I think they differ
07:45:30 <shrekshellraiser> And Novapolis, its apparently previous name, only comes up with a server host that hasn't been updated since 2015
07:45:52 <shrekshellraiser> Oh I see, it was closed source before being opened and taken over in 2015
07:47:22 <alfagamma_0007> Never used it tbh
07:47:27 <shrekshellraiser> Okay. That's what I was asking.
07:49:02 <dwfreed> heh, self-replies on discord look funny on irc
07:49:06 <truebrain> Rubidium: `uint` still exists as type; I somewhat assumed that would also be translated to `uint32_t` .. would that be possible / wanted, you think?
07:51:37 <truebrain> I had to build OpenTTD on a single core as compilers suck at giving errors back when you use multiple cores and multiple compile-failures happen 😛
07:52:55 <alfagamma_0007> At least openttd works
07:53:01 <alfagamma_0007> Rather than freecol
07:56:30 <dwfreed> truebrain: you might like adding '-O' to your make invocation along with your -j
07:57:25 <dwfreed> makes parallel compiles much less messy in terms of output
07:57:36 <truebrain> but removes colours! But yeah, good point
07:58:09 <dwfreed> CXXFLAGS=-fdiagnostics-color=always
07:58:26 <truebrain> and why is all that not the default, you wonder 😛
07:58:41 <dwfreed> because less does not process colors by default
07:58:53 <truebrain> I know; my point exactly 🙂
07:58:57 <dwfreed> nor do most other things on the receiving end of a pipe
07:59:14 <truebrain> all this stuff is designed to act like 1970 by default 😛
07:59:21 <dwfreed> piping ANSI SGR sequences to a pastebin, for example, is especially gross
08:00:59 <peter1138> Hmm, the hunt for a spoke begins.
08:01:27 <truebrain> the spoke was not to be spoken of
08:01:49 <dwfreed> but yeah, imo, -O should be default-on when -j is specified
08:02:22 <peter1138> They come in different lengths and of course I've no idea how long I need.
08:02:33 <truebrain> twice the half-length
08:02:52 <truebrain> I will be here in this corner being very silent now 🙂
08:02:58 <peter1138> And the step between lengths is often 1 or 2 mm, so super easy to get the wrong one.
08:05:00 <peter1138> Hmm, do I trust "line - IL_RATE1" to be valid...
08:08:50 <truebrain> whoho, I broke GCC again! (clang works fine)
08:17:50 <peter1138> LordAro, what spoke length do I need?
08:19:59 <peter1138> Hmm, a piece of string says around 265mm.
08:20:23 <peter1138> "How long is a piece of string" turns out to be useful here.
08:21:03 <peter1138> Maybe take off a bit due to stretch...
08:21:39 <peter1138> Trying to work out how to implement better resource filters, i.e. at a database level instead of application level.
09:12:58 <Rubidium> truebrain: sorry. I left uint along because it's the sibling of int, and I'm not sure whether we want to change all int to int32 too
09:13:46 <truebrain> I am not sure either .. saw a few weird uints, which made me wonder
09:17:22 <Rubidium> and another question would be should we go away from short/int/long then as well in favour of specifically sized versions
09:18:12 <truebrain> Short and long should be banned 😛
09:21:13 <peter1138> we could change 'uint' to 'unsigned' so that it's at least standard.
09:21:18 <peter1138> not a very good standard, but still...
09:21:49 <peter1138> byte could just be uint8_t, and in many places it should be that.
09:22:09 <peter1138> but in some places it is... a byte, not an 8 bit unsigned integer.
09:22:37 <peter1138> I think that distinction probably doesn't matter much.
09:29:04 <truebrain> But especially `long` should be banned 😛
09:29:16 <truebrain> So annoyingly undefined
09:29:33 <Rubidium> they are all essentially undefined
09:29:37 <peter1138> ParseIntList uses long...
09:30:28 <Rubidium> (they as in: char, short, int, long)
09:32:16 <peter1138> and then we have "word" and "dword" which are treated as 16 and 32 bits, but technically that isn't what word means at all 😄
09:36:38 *** ChanServ sets mode: +v tokai
09:42:19 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
09:45:14 <LordAro> int to int32_t does seem excessive
09:49:52 <peter1138> As does uint to uint32_t in a way, but int is at least standard.
09:57:28 *** pickpacket5 has joined #openttd
09:59:12 <truebrain> Int for loops etc is useful
10:02:54 <Rubidium> until something is unsigned in the loop comparison
10:04:47 *** pickpacket has quit IRC (Ping timeout: 480 seconds)
10:04:47 *** pickpacket5 is now known as pickpacket
10:11:07 <truebrain> Pff, than it is a size_t obviously 😛
10:13:18 <LordAro> needs more -Wconversion
10:13:43 <peter1138> Can't believe how sidetracked I got.
10:14:44 <peter1138> Such out of stock :/
10:15:05 <peter1138> Always the best discounts on things you can't buy.
11:17:59 <FLHerne> _glx_: it's conceivable I broke it when refactoring a couple of years ago
11:18:15 <FLHerne> haven't made time to look at the code yet
12:03:42 <_glx_> I think it was already partly broken before the refactoring
12:04:58 <_glx_> The `register_names` stuff delays the list init
12:13:51 <peter1138> OCL_SVAR( OC_TTD | OC_UINT16, Industry, produced[0].waiting ),
12:14:03 <peter1138> Oh, well that isn't going to work. How old is the title game? 😄
12:14:15 <peter1138> Not OC_TTD for sure.
12:14:40 <peter1138> Oh, that's a version type, but with special handling, I guess.
12:31:14 <peter1138> I found some old TTD scenarios which will do 🙂
12:39:00 *** talltyler has joined #openttd
12:39:00 <talltyler> shrekshellraiser: A lot of patches aren’t merged because nobody proposes to do so. There’s a lot from JGRPP that could be upstreamed to vanilla, if someone is willing to align it with vanilla and open a PR 🙂
12:39:51 <talltyler> Nobody is excluding them, it just hasn’t been done yet 🙂
12:44:24 *** pickpacket has quit IRC (Ping timeout: 480 seconds)
12:49:28 *** pickpacket has joined #openttd
12:57:05 *** pickpacket5 has joined #openttd
13:02:49 *** pickpacket has quit IRC (Ping timeout: 480 seconds)
13:02:49 *** pickpacket5 is now known as pickpacket
16:54:43 *** gelignite has joined #openttd
18:21:18 *** k-man has quit IRC (Ping timeout: 480 seconds)
18:21:18 *** k-man_ is now known as k-man
18:32:38 *** k-man has quit IRC (Ping timeout: 480 seconds)
18:32:38 *** k-man_ is now known as k-man
18:37:12 *** locosage has joined #openttd
18:37:12 <locosage> shrekshellraiser: citymania polyline originates from the same patch but adds some features on top
18:39:21 <DorpsGek> - Update: Translations from eints (by translators)
18:52:49 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
18:55:04 <peter1138> Phew, old cassette seems fine
19:02:45 *** Kitrana has joined #openttd
19:06:51 <peter1138> Just not as clean as a new one 😄
19:08:54 *** Kitrana2 has quit IRC (Ping timeout: 480 seconds)
19:26:33 *** Wormnest has joined #openttd
20:07:13 <pickpacket> What’s ”polyline”?
20:12:57 <FLHerne> it makes building track marginally faster
20:17:54 <andythenorth> when it places it where you want it
20:17:57 <frosch123> on touchscreens it makes building tracks 10x faster :p
20:18:42 *** gelignite has quit IRC (Quit: Stay safe!)
20:18:49 * andythenorth wonders about 'place n parallel lines' track builder :P
20:18:57 <andythenorth> probably too many edge cases to handle
20:37:31 <michi_cc[d]> andythenorth: Someone reviewing it.
20:38:20 <locosage> andythenorth: it's fully deterministic so you decide where it places
20:49:07 <locosage> problem with polyline is that it doesn't pathfind so navigating slopes and objects is a bit of a pita
20:55:10 <locosage> But it's still a lot faster than vanilla tool once you get used to it
21:09:14 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:21:43 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:30:54 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:04:32 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
23:49:51 *** tokai|noir has joined #openttd
23:49:51 *** ChanServ sets mode: +v tokai|noir
23:56:54 *** tokai has quit IRC (Ping timeout: 480 seconds)
continue to next day ⏵
