IRC logs for #openttd on OFTC at 2025-10-25
⏴ go to previous day
01:58:49 *** tokai|noir has joined #openttd
01:58:49 *** ChanServ sets mode: +v tokai|noir
02:05:39 *** tokai has quit IRC (Ping timeout: 480 seconds)
02:12:24 *** gnu_jj_ has quit IRC (Ping timeout: 480 seconds)
02:13:35 *** WormnestAndroid has joined #openttd
02:42:20 *** WormnestAndroid has quit IRC (Remote host closed the connection)
02:42:41 *** WormnestAndroid has joined #openttd
03:56:23 *** Zathras_11 has joined #openttd
03:59:56 *** Zathras has quit IRC (Ping timeout: 480 seconds)
04:03:17 *** gnu_jj_ has joined #openttd
04:06:44 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
04:10:37 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
04:10:38 *** WormnestAndroid has joined #openttd
04:10:46 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
04:11:04 *** WormnestAndroid has joined #openttd
04:38:39 <DorpsGek> - Update: Translations from eints (by translators)
04:46:33 *** Zathras has joined #openttd
04:49:56 *** Zathras_11 has quit IRC (Ping timeout: 480 seconds)
06:00:34 *** Extrems has quit IRC (Read error: Connection reset by peer)
06:00:39 *** Extrems has joined #openttd
07:48:12 *** Flygon has quit IRC (Remote host closed the connection)
08:41:41 <locosage> hm, chatgpt recommends lark as ply replacement...
08:48:54 <andythenorth> locosage: is this in some parsing step? I don't know your pipeline
08:49:20 <andythenorth> [context for me is that nmlc parsing is slow, and I'm curious if ply is implicated in that]
08:50:16 <locosage> it's for parsing switch/va2 code
08:50:34 <locosage> nml is unlikely to be affected as it was just bad code from me
08:50:38 <kuhnovic> For a moment I thought you were talking about a woodworking project
08:50:57 <locosage> not sure why it slowed down 3.14 that much but I shouldn't be creating a new parser for each switch anyway
08:51:25 <andythenorth> put them in a thread pool 😛
08:51:46 <andythenorth> map all your switches over the pool 😛
08:51:51 <andythenorth> "overhead" though
08:52:45 <andythenorth> I hate these 2 nmlc numbers, but it's beyond my skill level to fix 😛
08:52:51 <andythenorth> except by making a smaller nml file
08:53:50 <locosage> I hope I won't ever need threading for parsing, my idea is to make code cacheable
08:54:23 <andythenorth> I wonder if nmlc could cache the results per #include
08:55:12 <andythenorth> tell reddit, they can campaign
09:00:21 <locosage> yeah, probably could
09:01:27 <locosage> since ## is only for ids and not like a preprocessor
09:01:31 <ahyangyi> Is it normal to lose all grf-py cache upon a build failure?
09:03:29 <ahyangyi> At least it's on the "safer" side, and I should fix my problematic hash functions first 😛
09:04:55 <locosage> yeah, cache is controlled by get_fingerprint and get_resource_files as long as those stay the same it should be reused
09:04:58 <ahyangyi> Oh, and it is desirable to have some parallelism when `gorender` is involved.
09:04:58 <ahyangyi> I thought the built-in multithreading of `gorender` justified a single-thread `grf-py` to schedule around, but it turns out it was only true because I was using an ancient CPU
09:05:26 <ahyangyi> On a modern CPU, parallel building is still a few times faster than calling `gorender` sequentially
09:05:34 <locosage> yeah, I do plan to multithread sprite loading eventually
09:05:52 <ahyangyi> anyways, not a big issue since those are cached easily
09:31:23 *** Zathras_4 has joined #openttd
09:45:35 *** gelignite has joined #openttd
09:56:36 <xarick> who's Marcin Grzegorczyk?
10:06:32 <rito12_51026> That line was not modified for last 19 years
10:10:16 *** kuka_lie has joined #openttd
10:16:05 <xarick> eh, there's some html code warnings in landscape.html
10:16:14 <xarick> visual studio complains
10:25:36 <LordAro> Marcin Grzegorczyk is the guy that originally decoded the original TT(D) GRF format
10:25:47 <LordAro> or at least had a significant part in doing so
10:53:34 <xarick> is there a coding style for html?
10:53:59 <xarick> majority being lower case
10:54:28 <andythenorth> it's insignificant
10:54:49 <andythenorth> whilst I personally prefer a consistent style, it doesn't make any difference to rendering
10:55:31 <andythenorth> I suspect that even trying to agreeing a html style would do more harm than good
11:32:52 <xarick> reviewing a 1600 line html file 😛
11:33:01 <xarick> for missing closed tags
11:33:28 <xarick> visual studio only points out excessive closed tags
11:34:03 <xarick> so i'm resorting to manual inspection 🙂
11:37:26 <andythenorth> the html spec has gone through multiple standards respecting tags
11:37:42 <andythenorth> and there are online reformatters and linters that will do the work
11:38:08 <andythenorth> just manually fixing things because of VS warnings you don't understand 🤯
11:41:19 <andythenorth> the DTD is html 4, but largely html 4 was a misstep
11:41:35 <andythenorth> and it's irrelevant anyway because it's just some docs
12:06:32 <ahyangyi> xarick: I think one of the most common "style guide" for virtually any language that does not come with a standard formatter is: "don't change existing code unless you're refactoring, and try to be consistent with your immediate context"
12:53:33 *** Wormnest has joined #openttd
13:08:41 <rito12_51026> That code was copied from the GetEngineColourMap
13:10:06 <peter1138> GetEngineColourMap is different because the 2cc part is conditional.
13:12:47 <andythenorth> and I tested the html file in browser, and appeared fine by eyeball
13:14:53 <rito12_51026> ```console_output
13:14:53 <rito12_51026> [new_map f6d0b9e994] Codechange: Make Tile constructor from TileIndex explicit.
13:14:53 <rito12_51026> Date: Sun Oct 12 22:21:58 2025 +0200
13:14:53 <rito12_51026> 76 files changed, 2542 insertions(+), 420 deletions(-)
13:14:53 <rito12_51026> @peter1138 should I split this commit?
13:18:46 <xarick> it is complaining about <h3> somehow
13:21:18 <xarick> something tells me this document was copy pasted from an even larger document that had the other subsections h2 and h1
13:22:47 *** alwaysampersand has joined #openttd
13:25:05 <xarick> I have this sudden urge to PR it 🙁
13:25:16 <xarick> obsessive compulsive disorder
13:25:58 *** alwaysampersand has left #openttd
13:26:42 <andythenorth> first heading should be a <h1> for preference
13:26:48 <andythenorth> depending on which html spec version
13:26:58 <andythenorth> do you actually have OCD?
13:28:55 <peter1138> Hmm, I probably have a function in my rgb company colours patch.
13:29:18 <rito12_51026> Aren't the <h*> elements obsolete?
13:29:28 <andythenorth> I mean AVPD is 25% of people in some countries
13:29:38 <andythenorth> it's only a disorder if it causes disordered behaviour
13:29:49 <andythenorth> rito12_51026: errr...no?
13:31:08 <andythenorth> hmm I should read first, AVPD is not just being Avoidant, it's the clinical disordered form of being Avoidant
13:46:34 <peter1138> Hmm, is an optional reference much different from a raw pointer?
13:46:43 <peter1138> Slight bit of semantics I guess.
13:48:54 <rito12_51026> peter1138: By the way, how compatible it is with newgrf cc1 and cc2 variables and with recolour sprites?
13:58:22 <xarick> it was supposed to list the m1 ... m8 in order
13:58:57 <xarick> but it's m1 m2 m3 m8 m4 m5
13:59:40 <xarick> where will i place my own m8 bit description now?
14:06:57 <xarick> why mention free bits?
14:08:19 <xarick> are they in use by newgrfs?
14:08:35 <_glx_> because it's important to know where the free space is
14:09:09 <andythenorth> hmm do I need more CPU cores and RAM?
14:09:32 <_glx_> and that way it's easy to check that text matches the grid
14:09:43 <andythenorth> controversial suggestion
14:09:52 <andythenorth> generate the docs from a map of bits
14:10:02 <andythenorth> as we have 2 docs
14:11:38 <andythenorth> I guess there's no single place in the actual src to parse them from?
14:15:00 *** Rubidium has joined #openttd
14:16:35 *** brickblock19280 has joined #openttd
14:16:35 <brickblock19280> xarick: Lift position?? I don't see anything about that in the grf specs
14:29:09 <_glx_> in newgrf it's replaced with animation state
14:29:58 <_glx_> and newgrf can use it for whatever they want
14:41:34 *** Wolf01 is now known as Guest29884
14:47:04 *** Guest29884 has quit IRC (Ping timeout: 480 seconds)
15:02:34 <xarick> I am worried about newgrfs populating bit 14
15:02:44 <xarick> if they can even do that
15:06:04 <peter1138> NewGRFs don't have any write access to map data.
15:15:00 <xarick> water misalignment is right where I need to place the m8 bit 14 info
15:16:21 <peter1138> andythenorth, I have a patch...
15:17:12 <peter1138> Or rather, had one. Not sure if I scrapped it.
15:20:03 <xarick> void tiles cannot have underground tunnels, I suppose
15:20:13 <xarick> but it's still checked
15:24:49 <xarick> any particular reason `t.m8() = 0;` is not there?
15:34:07 <rito12_51026> Propably all `t.mx() = 0;` are redundant
15:35:05 <peter1138> Absolutely not redundant.
15:35:30 <peter1138> For MP_VOID it doesn't particular matter, but it probably should be there.
15:49:28 <andythenorth> peter1138: a patch for? 😛
15:49:38 <andythenorth> that you have a patch is a default expectation 🙂
16:22:39 <kuhnovic> If void tiles don't use m8 then this should be fine, but I would rather set it anyway. I think there was a PR that wrote tests for this, the one that stirred up the whole AI discussion
16:25:48 <peter1138> I'll make a PR to set it.
16:59:46 <peter1138> I think I ballsed up something :)
17:14:04 <mmtunligit> im seeing variebles that are provided by table/strings.h but i can't seem to find strings.h in table? i fetched from upstream earlier today so i should have everything right
17:14:51 <michi_cc> It's a generated file
17:15:59 <mmtunligit> i need to add something to it for the thing im working on, how would i go about doing that
17:16:13 <mmtunligit> or do i just declare the new string and then give it an entry in the lang file
17:16:21 <michi_cc> you modify english.txt
17:18:29 <mmtunligit> im trying to make it so you can move signs and station/waypoint names, signs will be able to be moved anywhere, gonna require stations and waypoints to be on a stationtile
17:18:59 <mmtunligit> and theres no "move" string
17:21:31 <xarick> I'm not sure about void tiles. I'm not restricting setting a bit there too in the eventual case a tunnel can be placed there
17:23:14 *** kuka_lie has quit IRC (Remote host closed the connection)
17:25:23 <xarick> and the MakeVoid doesn't even zeroes the m8
17:30:44 <peter1138> I do wonder if the current method is "over optimising" changing tile types.
17:47:10 <peter1138> andythenorth, do you have fully-functioning FIRS 5 game with industries in use etc?
17:50:06 <peter1138> Do you not play with your own set? ;-)
17:50:29 <andythenorth> any save I have will be some local build of OpenTTD with remote PRs applied
17:50:39 <andythenorth> and possibly broken grfs
17:50:59 <peter1138> Hmm, do coal piles change?
17:51:06 <peter1138> I see they are animated on & off.
17:51:42 <andythenorth> if it's FIRS, they are animated by accident
17:51:55 <andythenorth> maybe I'll fix that next week 😛
17:52:20 <andythenorth> mindless no-visible-change commits, might be good for a week off 🙂
18:51:59 *** SigHunter has joined #openttd
20:01:49 *** Smedles has joined #openttd
20:03:11 <xarick> i'm getting this crash
20:04:12 <kuhnovic> An asset is not a crash
20:07:28 <xarick> gonna try a fresh build
20:10:03 <xarick> did something change in world generation?
20:13:23 <kuhnovic> Someone optimized river generation? 😛
20:13:43 <xarick> i suspect it's the scale town/industry commit
20:15:13 <xarick> I am generating 1 town and industry set to funding only, i get a crash trying a 512x512 map
20:41:14 *** gelignite has joined #openttd
22:08:09 <xarick> 56 seconds to flatten a 4096x4096 from height 1 to 0
22:08:24 <xarick> in scenario editor, empty map
22:29:09 *** keikoz has quit IRC (Ping timeout: 480 seconds)
23:10:14 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:29:10 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
continue to next day ⏵
