IRC logs for #openttd on OFTC at 2024-05-26
⏴ go to previous day
01:14:42 <yiffgirl> what would be the best way to split this into two patches? i guess an interactive rebase?
01:18:12 *** talltyler has joined #openttd
01:18:12 <talltyler> Yes, you could make a new branch based on the existing branch and remove the commits you don’t want, or start with a clean branch and cherry-pick.
01:42:21 <johnfranklin> Windows always lets RAM half occupied when turning on the device... From the 4GB device 10 years ago, to the 32GB device now, I don't know why.
02:11:51 *** Wormnest has quit IRC (Quit: Leaving)
02:37:04 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
03:04:13 <yiffgirl> now do i update the treespreading branch right now or let yall examine this patch first
03:06:39 <talltyler> If it can stand alone, I’d update it separately. I don’t know what Zorg means by the claim that improved and original are the same — and I’m not at a computer to check.
03:34:52 <yiffgirl> git diff HEAD~1 my beloved
03:49:46 <yiffgirl> i'll look into CircularTileSearch stuff tomorrow
03:50:16 <yiffgirl> or. when i can i guess. probably going to be busy tomorrow
04:10:00 *** reldred has joined #openttd
04:10:00 <reldred> talltyler: I mean, they aren't, but Zorg will Zorg...
04:11:19 <yiffgirl> the screenshots they posted are from like 2021 too
04:16:16 <reldred> though i tend to behave myself these days
04:18:05 <yiffgirl> i'm curious as to what your crack smoking moment was
04:21:16 <reldred> Hah, was a fair few years ago now, I think I was pissy that nobody was reviewing PR's.
04:42:25 <DorpsGek> - Update: Translations from eints (by translators)
05:50:35 <andythenorth> can FIRS just do wall clock values for supplies delivery windows etc? Town growth is now in seconds....that would do the job
05:54:11 *** XYZ has quit IRC (Ping timeout: 480 seconds)
05:58:18 <andythenorth> oh industries already do production in last minute
05:58:46 <andythenorth> so it's probably just 'deliver supplies every 3 minutes' (adjusted by wallclock setting value)
06:17:44 <merni> johnfranklin: It's Windows.
06:20:26 <wensimehrp> merni: Memreduct helps a lot I'd say
06:23:02 <wensimehrp> I just installed it hours earlier, and ram consumption decreased by 7 gigabytes (before is 13)
06:23:02 <wensimehrp> To be fair with or without the software isn't a very big deal for me as I have 64 gigs of ram; nonetheless it is a great software.
07:26:19 <andythenorth> to find the town for an industry, do I want GSTile.GetTownAuthority or GSTile.GetClosestTown?
07:26:34 <andythenorth> some tiles have no authority, so I assumed I need GetClosestTown
07:26:43 <andythenorth> but that might be using grid distance
07:28:12 <andythenorth> GetClosestTown is resolved by ClosestTownFromTile in OpenTTD
07:28:29 <andythenorth> which is also used for the grf var for industry town, so I guess that's fine?
07:50:19 *** SigHunter has joined #openttd
08:16:59 *** ChanServ sets mode: +v tokai
08:23:44 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
08:42:09 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
08:44:46 *** SigHunter has joined #openttd
09:02:07 *** aperezdc has quit IRC (Remote host closed the connection)
09:03:54 <peter1138> Okay. Whoever it was who asked for more waypoint classes...
09:04:55 <peter1138> Ah, it was georgevb
09:06:04 <peter1138> Crap, forgot about my toast. Cold toast 😦
09:11:44 *** HerzogDeXtEr has joined #openttd
09:21:09 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
09:26:02 <kuhnovic> I'm gonna try out that memreduct tool, let's see what it can do for my 12 year old 8 gb windows laptop 😛
09:36:23 *** SigHunter has joined #openttd
09:36:30 *** aperezdc has joined #openttd
09:40:10 <kuhnovic> It's not my main development machine, luckily
09:40:27 <Rubidium> so, what's that going to do? Flush all the caches making everything that used to be already in memory loaded from disk?
09:44:53 <andythenorth> unload all the fonts
09:47:41 <Rubidium> ah, no... it's using the page deduplication feature that was added to Windows and subsequently disabled by default as it allows essentially any local process to peek into memory from arbitrary processes...
09:48:45 <Rubidium> and yes, it's trashing (file) caches. That'll definitely save some memory
09:49:04 <Rubidium> not sure whether it makes your compiles any faster though
09:53:19 <andythenorth> grfid: "\F1%\00\08";
09:53:26 <andythenorth> that's an nml grfid
09:53:39 <andythenorth> how do I convert that to a python raw string that I can write back as that?
09:53:44 <andythenorth> and what's % doing there?
09:54:20 <andythenorth> the grfid as shown on bananas is "f1250008"
09:56:09 <Rubidium> I'd guess that 0x25 in ASCII is '%'
10:01:42 <andythenorth> I hate everything about grfids 🙂
10:03:27 <peter1138> It's just a 32bit value that NML decided should be a string.
10:28:40 <andythenorth> it was convention in early nml grfs to put it in via string()
10:28:53 <andythenorth> because it's not translatable
10:29:06 <andythenorth> it was something to do with nmls magic tokens and all the CPP templating crap
10:45:47 <merni> PeterNviaGitHub: > The only language file missing a layout is hindi.txt which is non-latin and should probably have a specific layout anyway.
10:45:47 <merni> Making a keyboard layout for complex scripts like Hindi isn't very simple... for the same reason languages like Chinese, Japanese and Tamil in openttd just use the English keyboard layout for the OSK :)
10:50:04 <merni> Hindi could probably go with the same for now
10:50:17 <merni> Though the Hindi translation in openttd is anyway woefully incomplete
10:56:04 <merni> Also, interesting that the openttd OSK kind of has the UK layout hard-coded in, as there are 2 symbol keys on the first row (after P) but 3 after L, unlike the US layout which has 3 keys after P and 2 after L
10:59:48 <merni> hm, that appears to vary even within the US layout 🤷
11:02:18 *** ahyangyi has quit IRC (Quit: User went offline on Discord a while ago)
11:24:59 <peter1138> Yeah, looks like the OSK puts that key on the first row, where backspace is normally.
11:25:19 <peter1138> I do actually have a patch to fix that 😄
11:25:38 <peter1138> merni: What's not simple about it>
11:26:28 <peter1138> Main thing missing is some more modifiers than just Shift would be useful.
11:59:42 <peter1138> Rubidium: The old "free memory is wasted memory"
12:05:24 <Rubidium> kinda yes. If you really like free memory, just disable all file caching ;)
12:07:55 <Rubidium> it's also a question of how the specific operating system shows used/free memory. Windows shows cached mmap-ed files as used memory, which gives me discussions with out sysadmins about an application taking a huge amount of memory
12:08:52 <Rubidium> ... but no applications crashing due to being actually out of memory. Windows just trashes a few caches and it got heaps of memory again
12:19:50 <_glx_> and all the standby memory
12:40:10 <talltyler> andythenorth: No, scaled by “industry cargo scale”
12:41:22 <talltyler> That setting also changed the required frequency even in calendar mode 🙂
13:46:52 <peter1138> The best sort of town generation
13:46:57 <peter1138> One long high street 🙂
13:47:16 <peter1138> Shun those silly unrealistic 2x2 and 3x3 grids.
13:49:52 <peter1138> Hmm, not sure why some of 2x2 layouts show the icons.
13:52:06 <peter1138> Well, it's like the NewGRF is deliberately not drawing the building graphics for the front tile.
13:52:33 <peter1138> Maybe it's intentional 🙂
14:04:01 <_glx_> or the draw order of the tiles, front might actually be an overlay from back
14:04:58 <peter1138> Winner winner chicken dinner.
14:05:02 <peter1138> It's incorrectly coded.
14:05:25 <_glx_> big 2x2 sprite not split ?
14:08:36 *** locosage has joined #openttd
14:08:36 <locosage> peter1138: If they use basset sprites they sometimes don't have building on front tile
14:08:55 <locosage> Or just copy vanilla layouts
14:08:57 <_glx_> clearly not a baseset 🙂
14:09:04 <peter1138> It's a NewGRF house so that's not relevant.
14:10:04 <locosage> peter1138: Some of these sprites look like they're taken from baseset
14:10:57 <peter1138> They're not taken from the baseset, they ARE the baseset.
14:11:56 <peter1138> I think it's pretty clear I was only referring to the 2x2 builds that are showing an icon, not other houses that are not 2x2 or not showing an icon.
14:13:15 <peter1138> Well, I don't really fancy using a fake viewport for the house previous so that sprite sorting can be done.
14:14:22 <peter1138> This building does draw the same sprite 3 times.
14:14:35 <_glx_> just draw back to front and assume grf is correct
14:14:49 <peter1138> That's what it does 🙂
14:15:15 <peter1138> Tell-tale sign is double transparency.
14:15:58 <peter1138> Maybe @George can fix it, 12 years later 🙂
14:16:34 <peter1138> Maybe Discord can highlight when I use @
14:17:09 <peter1138> I would've cropped custom layout bounds to 0-16 myself.
14:17:12 <peter1138> But it's a bit late for that.
14:17:32 <peter1138> As enforcing it now breaks lots of things that "work" (even though they generally glitch out at some point)
14:21:11 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
15:12:40 <peter1138> Hmm, so should I automatically load a wide unicode-covering font automatically, even if those characters aren't needed by the language?
15:16:22 <Rubidium> given that some NewGRF authors use those characters in their descriptions... that might be useful, but not consistent
15:18:25 *** georgevb has joined #openttd
15:18:25 <georgevb> peter1138: Why don't you like this style? 😉
15:19:02 <peter1138> Tiles should only draw on their own tile, not neighbour tiles.
15:19:39 <Rubidium> peter1138: because depending on the user a different font might be chosen, which might or might not cover the characters used in the specific NewGRF
15:20:31 <georgevb> peter1138: Yes, so, ffxxxxxx class would be waypoints?
15:22:04 <peter1138> (And WAYP still, of course)
15:22:23 <georgevb> peter1138: I do not know when I will start ECS 2.0
15:23:23 <peter1138> Oops, something broken with dimensions 😒
15:27:29 <peter1138> Rubidium: If a different font doesn't cover it, then the automatically loaded unicode-covering font should.
16:30:07 <peter1138> Hmm, I thought I could dismiss CodeQL warnings.
16:43:19 <_glx_> of course x86 is not happy
16:46:01 <peter1138> Hadn't noticed that.
16:46:59 *** Wormnest has joined #openttd
16:48:36 <peter1138> D:\a\OpenTTD\OpenTTD\src\network\core\packet.cpp(133): note: see reference to function template instantiation '_Ty &std::vector<_Ty,std::allocator<_Ty>>::emplace_back<uint>(uint &&)' being compiled
16:48:41 <peter1138> Seems to be... unrelated? o_O
17:00:55 <Rubidium> is that maybe a new MSVC release?
17:02:38 <Rubidium> but indeed unrelated *and* inconsistent, as it ought to be warning a lot more in that file
17:34:30 *** gelignite has joined #openttd
18:00:17 <_glx_> funny, it doesn't warn for master, but it does for the PR
18:06:38 <Rubidium> is master compiled with the same version of MSVC?
19:27:14 <peter1138> Can you try just the first commit?
19:27:30 <peter1138> That one fiddles with stdafx.h which seems to be the only relevant thing for packet.cpp
19:35:14 <_glx_> with only first commit warning are still there
20:04:35 *** gelignite has quit IRC (Quit: Stay safe!)
21:25:30 <peter1138> Hmm, what was #12646 fixing... I forgot 😦
21:28:49 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:58:23 <peter1138> Oh, some NewGRF moved cargo around because it sorted them alphabeticall.y
22:00:07 <peter1138> FIRS 5 and old bus sets. Hmm.
22:15:31 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:32:11 <peter1138> Okay, I don't think I need to worry about pre cargolabel sets being broken.
22:34:02 <peter1138> Urgh, just loaded up an old version, interface scaling is terrible 🙂
22:36:20 <peter1138> Yeah, the old cargo sets don't even work in 1.11.2
22:46:45 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
23:43:25 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
continue to next day ⏵
