IRC logs for #openttd on OFTC at 2023-11-21
⏴ go to previous day
03:03:51 <johnfranklin[d]> _pruple: And anyone has complaints can make their own once you licensed it GPL or CC
03:23:19 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:32:57 *** Wormnest has quit IRC (Quit: Leaving)
04:04:26 *** tokai|noir has joined #openttd
04:04:26 *** ChanServ sets mode: +v tokai|noir
04:10:59 *** tokai has quit IRC (Ping timeout: 480 seconds)
06:00:24 <locosage> Rubidium: You can make a separate grf for options xD
06:14:24 *** keikoz has quit IRC (Ping timeout: 480 seconds)
06:34:26 <rau117> Rubidium: Well... they can simply be ignored. *As they say, if you don’t like it, don’t use it, others will use it.*
06:54:24 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
08:22:33 <Eddi|zuHause> why do i get the feeling there's half the conversation missing?
08:22:59 <locosage> because you're on irc :p
08:23:42 <Eddi|zuHause> that would be the lazy answer
08:23:58 <locosage> bridge can't into replies
08:26:01 <Eddi|zuHause> anyway, NARS probably predates all the parameter UI stuff
08:26:57 <alfagamma7> Which version of Nars
08:31:11 <emperorjake> 2.03 has old style parameters, 2.5 has new style but a lot of features were cut
08:37:39 <johnfranklin[d]> I tried parameter 1 in 2.03
08:38:05 <johnfranklin[d]> It only affects purchase menu offsets
08:55:58 <peter1138> `font medium "arial,"` is a nice crash :)
09:36:40 <peter1138> Hmm, might need some heating.
09:41:25 *** thelounge34 has joined #openttd
09:51:14 *** D-HUND is now known as debdog
10:16:54 <peter1138> Hmm, so the tile hash is changed, and theres deferred viewport hash updates
10:17:27 <peter1138> _jgr_, did you consider using a multimap instead of a the linked-list chain?
10:21:54 <_jgr_> Containers like unordered_multimap are implemented using bucket linked lists anyway, more performant hash tables don't seem to support multiple identical keys
10:22:07 <_jgr_> It's less of an issue when different tiles are no longer alised to each other
10:23:14 <peter1138> I plugged in a std::unordered_multimap and it's ~8.5% faster than vanilla on Wentbourne.
10:24:57 <peter1138> But yes, it probably mostly comes from not having only 128 buckets for the whole map.
10:25:14 <peter1138> I didn't split by vehicle type either as that's a bigger taks.
10:33:30 <peter1138> I like simpler code but I guess this is probably a case where specialization is useful.
10:37:11 <peter1138> Ah, of course this depends too much on how the multimap is implemented.
10:56:54 <peter1138> Similar performance improvement by just increasing the hash size. Hah
11:01:35 <locosage> peter1138[d]: what are you working on?
11:01:39 <locosage> sounds related to #11263
11:09:24 <locosage> well, it also involves tile hashes and viewport updates xD
11:39:11 *** thelounge34 has quit IRC (Ping timeout: 480 seconds)
11:59:35 <peter1138> Oh it's not 128 buckets, it's 128x128 buckets.
12:03:54 <orudge> truebrain: do we need to ask 1password for an open source licence extension or something similar?
12:09:38 <truebrain> yes; I was kinda surprised by the 4 day notice
12:10:48 <truebrain> `Your membership does not expire. If you signed up before November 26, 2021, please contact opensource@1password.com to convert your two-year membership into one that doesn't need to be renewed.`
12:11:20 <truebrain> so I will email them 🙂
13:06:19 <yesbad> truebrain: Btw, is it just my instance or is there no sounds? I downloaded them thru check-online-content yet I don't think it works.. Thanks :)
13:07:09 <peter1138> Did you select the sound set in the Game Options window?
13:07:29 <yesbad> Yeah I did, didn't change anything. Also did some refreshing aswell.
13:12:58 <talltyler> Is the volume on? (In the jukebox window)
13:13:40 <talltyler> Oh, this is emsccription. It had no sound for me either last time I tried.
13:28:48 *** dihedral has joined #openttd
13:30:30 <_glx_> IIRC there's no sound support in emscripten
13:32:32 <_glx_> oops I totally forgot my 1password password (but I have all other required info)
13:36:32 *** dihedral has joined #openttd
13:36:33 <yesbad> _glx_: Ah that's disappointing, got any clue why not?
13:37:38 <peter1138> OpenAL should work apparently, so that might be a place to start.
13:38:39 *** Smedles_ has joined #openttd
13:43:25 *** Smedles has quit IRC (Ping timeout: 480 seconds)
13:44:55 *** dihedral has joined #openttd
13:45:07 <_glx_> I finally remembered the password
13:54:29 <truebrain> I don't use sound (or music) in OpenTTD; so I never notice if it is there or not
13:54:35 <truebrain> I think that is the main "clue" why not 😛
13:55:19 <yesbad> truebrain: Ah, that makes sense... So would you happen to have an idea how one would get it to work? :)
13:56:27 <truebrain> enable, test, debug, try again, code a bit, get it to work
13:56:29 <truebrain> something like that?
13:56:57 <truebrain> possibly requires implementing an OpenAL backend
13:57:19 <truebrain> the SDL2 driver might work
13:57:49 <truebrain> but you can figure all that out yourself too 🙂 😛
13:58:19 <LordAro> that's not the right terminal
13:58:28 <truebrain> You can't comment me out!
13:58:29 <LordAro> nor what i meant to type
13:58:35 <truebrain> You can try, but it doesn't work like that! 😛
14:06:12 <locosage> hm, do I read it right that palette animation colors are hardcoded?
14:06:19 <locosage> I thought they're taken from the palette
14:08:04 <_glx_> well index doesn't change but the index->color map is rotated IIRC
14:09:00 <locosage> but the colors it takes doesn't seem to come from the palette
14:09:07 <peter1138> They are separately recorded, yes.
14:09:19 <peter1138> table/palettes.h:115
14:09:41 <locosage> palette.h have different colors for water cycle for example
14:10:04 <LordAro> documentation? incorrect? unheard of!
14:10:50 <peter1138> Treat those as placeholders.
14:11:13 <peter1138> (They always were placeholders)
14:46:23 <peter1138> nielsm, is the viewport sign Kdtree meant to be sorted?
14:53:21 <peter1138> Hmm, replacing the vector with sets does at least fix the glitching, but it's probably sorted by pointer 'value'.
15:00:01 <jfs> I think it's only supposed to be consistent enough that things don't flicker and all signs are properly registered, updated and removed when they need to
15:01:03 <jfs> I don't remember trying to add any kind of Z-sorting or similar
15:01:58 <jfs> (the purpose is to make it fast to cull signs that can't be visible on screen, not to determine draw order)
15:03:25 <peter1138> Okay. Well the problem is things do flicker :)
15:04:11 <peter1138> But yes, after looking I figured any sorting should be outside of the Kdtree.
15:07:28 <peter1138> I wonder if I broke something. It's usually me :-)
15:08:28 <jfs> I don't remember if the floating income signs are part of the signs tree
15:09:00 <peter1138> It's the station signs underneath that I'm talking about.
15:09:01 <jfs> or if it's just their redrawing that causes the signs to be invalidated
15:09:27 <jfs> I mean, if the income signs were being inserted, updated and removed from the tree that would jiggle things around
15:12:15 <peter1138> I think they're not.
15:12:30 <peter1138> But there are AIs running that are rapidly creating and deleting temporary signs
16:36:14 *** gelignite has joined #openttd
17:00:45 *** thelounge34 has joined #openttd
17:32:02 *** HerzogDeXtEr has joined #openttd
17:35:20 <goddess_ishtar> > Use highly advanced modern global search and replaces tool to expand the macros.
17:36:34 <peter1138> Seriously, back in the day I was using vim which did not make things like that easy.
17:36:37 <_glx_> I think originaly the macro was just a constant number
17:36:42 <peter1138> Easy enough within the file.
17:38:29 <peter1138> Possibly some micro-optimization to be done, e.g. calling it once per loop in various drawing routines.
17:49:40 <peter1138> But at least now it's clearer that it is calling a function.
18:40:09 <DorpsGek> - Update: Translations from eints (by translators)
18:44:11 <peter1138> Now I found another bug :(
18:44:17 <peter1138> (Not with that change.)
18:58:33 <talltyler> Obviously you’ll have to rebase one of those…
19:03:23 <goddess_ishtar> this is maybe more of a general git question than an OpenTTD question, but I'm still a little confused why you rebase instead of merge
19:03:56 <goddess_ishtar> I know it keeps the commit history cleaner, but isn't it bad form to rewrite the git history of a branch that is already available to others?
19:04:14 <peter1138> I do it because 'Create a merge commit' is disabled.
19:05:28 <goddess_ishtar> are the guidelines for rewriting histories different for PRs?
19:06:44 <peter1138> Basically... someone made a decision when we switched to git, and it has stuck.
19:07:48 <peter1138> It definitely makes it easier to track when breakages are made.
19:07:49 <_jgr_> goddess_ishtar: This only really matters when you expect other people to make their own branches starting form the branch that you publish, which isn't the case for a published PR
19:16:07 <peter1138> talltyler, thanks ;D
20:14:46 *** virtualrandomnumber has joined #openttd
20:16:17 *** virtualrandomnumber has quit IRC ()
20:34:03 *** Wormnest has joined #openttd
21:11:23 *** gelignite has quit IRC (Quit: Stay safe!)
21:41:25 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:11:10 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:39:55 <peter1138> Hmm, this window is awful.
22:45:47 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:05:31 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
23:23:32 <goddess_ishtar> PowerShell fails at PowerShell
23:34:35 <peter1138> Hmm, maybe the button should be on the left.
23:35:24 <peter1138> Or potentially in the 16th slot, although that requires a bit of effort.
continue to next day ⏵