IRC logs for #openttd on OFTC at 2021-11-06
⏴ go to previous day
00:11:01 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler dismissed a review for pull request #9673: Fix #9579: Object and HQ construction is Construction cost, not Property Ma… https://git.io/JPppi
00:11:04 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #9673: Fix #9579: Object and HQ construction is Construction cost, not Property Ma… https://git.io/JPppY
00:12:49 *** WormnestAndroid has joined #openttd
00:15:36 *** snail_UES_ has joined #openttd
00:26:11 *** gelignite has quit IRC (Quit: Stay safe!)
00:32:30 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on pull request #9673: Fix #9579: Object and HQ construction is Construction cost, not Property Ma… https://git.io/JPhTU
00:42:40 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #9673: Fix #9579: Object and HQ construction is Construction cost, not Property Ma… https://git.io/JPhkG
01:34:54 *** iSoSyS has quit IRC (Remote host closed the connection)
01:52:39 *** snail_UES_ has quit IRC (Quit: snail_UES_)
01:56:22 *** ChanServ sets mode: +v tokai
01:57:12 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
02:03:31 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
02:09:35 *** Wormnest has quit IRC (Quit: Leaving)
03:05:06 *** snail_UES_ has joined #openttd
03:07:18 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
03:07:40 *** WormnestAndroid has joined #openttd
04:11:49 *** pm is now known as Guest5103
04:18:31 *** Guest5053 has quit IRC (Ping timeout: 480 seconds)
04:20:55 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler opened pull request #9674: Fix: Don't show screenshot GUI in screenshots https://git.io/JPhrc
05:09:11 *** _aD has quit IRC (Quit: leaving)
05:12:17 *** snail_UES_ has quit IRC (Quit: snail_UES_)
06:42:47 <DorpsGek> [OpenTTD/OpenTTD] LordAro requested changes for pull request #9674: Fix: Don't show screenshot GUI in screenshots https://git.io/JPhAR
07:47:32 *** andythenorth has joined #openttd
09:11:01 *** HerzogDeXtEr has joined #openttd
09:54:40 <peter1138> Didn't get up in time for the club cycle ride.
10:31:22 <peter1138> Damn, I wrote some code about 9-12 months ago, but I can no longer find it :/
10:38:41 <peter1138> Surely it should just be in a branch...
10:49:17 *** Eddi|zuHause2 is now known as Eddi|zuHause
11:00:26 <Eddi|zuHause> if it was good code, you could just write it again :p
11:02:55 <peter1138> It's my code, of course it's not good.
11:03:08 <peter1138> I found it, but of course it won't rebase :-)
11:04:45 <peter1138> No, that rebased fine ;)
11:12:39 * andythenorth can actually play blitz at 60fps now
11:12:46 <andythenorth> doesn't increase my skills though
11:17:24 <TrueBrain> but you can see faster that you don't have any :P
11:19:10 <andythenorth> and so much smoother
11:19:13 <andythenorth> and all the fog effects
11:19:21 <andythenorth> so at least losing looks nice
11:24:22 <peter1138> Rebased in steps and it made more sense, so...
11:24:28 <peter1138> It doesn't work any more, but still.
11:32:16 *** WormnestAndroid has quit IRC (Remote host closed the connection)
11:35:52 <andythenorth> ouch the performance differences are so hit and miss
11:36:04 <andythenorth> python image processing is generally twice as fast
11:36:25 <andythenorth> rendering html docs seems to be 3x faster
11:36:50 <andythenorth> nmlc is unchanged
11:38:44 <andythenorth> nmlc is complicated because it's faster under pypy3 vs py3.10
11:38:50 <andythenorth> but pypy3 is x86 only currently
11:39:31 <andythenorth> does nmlc have any C++ modules?
11:43:36 <LordAro> it has compiled modules, yes
11:43:42 <LordAro> they're C, rather than C++, iirc
11:47:21 * andythenorth wonders which arch they're compiled for
11:49:13 <andythenorth> I'm compiling nml under x86 currently
11:51:22 * andythenorth compiles under arm
11:51:27 <andythenorth> no obvious difference
11:58:30 * andythenorth did a release build of openttd
11:58:31 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on pull request #9674: Fix: Don't show screenshot GUI in screenshots https://git.io/JPjDh
11:58:55 <andythenorth> ffwd will do a sustained 150x game speed factor
11:59:12 <andythenorth> interestingly, full animation on/off seems to now be negligible effect
11:59:26 <andythenorth> wonder what a big map does
11:59:44 <andythenorth> lol big maps generate slow
12:00:43 <peter1138> Fuck you fucking detours in ETS2
12:01:07 <andythenorth> I never got into ETS2
12:01:14 <andythenorth> I bought it, never really played
12:01:29 <andythenorth> seemed a bit sort of tycoon / business simulator
12:01:39 <andythenorth> I didn't want to run a company, that's my day job :P
12:01:50 <andythenorth> ETS 1 you could just open the plain text save file and grant more money
12:02:14 <andythenorth> wow a 4096 ^ 2 map is only 2.5x on ffwd, no vehicles
12:02:28 <andythenorth> 'sharding anyone?'
12:02:50 <michi_cc> And still people want even larger maps
12:05:33 <andythenorth> hmm wonder why big maps are soooo slow though
12:05:47 <peter1138> That big patch pack has larger maps.
12:05:48 <andythenorth> RAM use is pretty low
12:05:58 <peter1138> It's because there's more to process.
12:06:47 <andythenorth> I removed FIRS, it's pushing 6x game speed factor
12:07:07 <andythenorth> at least twice as fast without FIRS
12:09:45 * andythenorth stops fooling with that
12:19:21 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #9673: Fix #9579: Object and HQ construction is Construction cost, not Property Ma… https://git.io/JPjQk
12:24:43 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on pull request #9674: Fix: Don't show screenshot GUI in screenshots https://git.io/JPj7n
12:29:26 <peter1138> Can I bear riding the bike in civvies?
12:30:40 <peter1138> Pretty sure it's normal but as a 'lyrca lout' it'll feel all wrong...
12:33:19 *** WormnestAndroid has joined #openttd
12:35:09 <michi_cc> Strawpoll time: How bad would it be to change TileIndex to be either an enum or a struct? (enum means sprinkling the code with casts; for struct without inlining (e.g. potentially debug build) it might get really slow).
12:35:49 <TrueBrain> Guess your poll is missing context .. what do you try to accomplish? :D
12:35:51 <michi_cc> I would like to do a template decision on if a type is a TileIndex or nor, but unfortunately typedefs don't actually introduce new types, so the current TileIndex is not unique in any way.
12:36:07 <nielsm> having it be two 16 bit fields instead of one 32 bit in a struct would be better in the long run imo
12:36:43 <peter1138> So the aim is type-safety. I guess the sprinking of casts would be the "few" cases where arithmetic is done?
12:36:46 <michi_cc> So e.g. TileIndex resolves simply do unsigned int for template argument purposes, but of course we have hundreds if types that resolve to unsigned int.
12:37:12 <michi_cc> peter1138: Mostly the places where the code freely assumes that an int will just silently convert to TileIndex.
12:37:25 <nielsm> I'd prefer a single-field struct
12:38:03 <michi_cc> With a struct you can add all the appropriate constructors and operators, I'm just not sure how bad the performance impact might be on a debug build (which will be compiler dependant).
12:40:20 <michi_cc> Of course all this would be to avoid needing both a DoCommand::Post and DoCommand::PostTile. So the third options is to just say, oh well, who needs automatic decisions anyway.
12:43:26 <TrueBrain> that makes sense :)
12:46:23 <michi_cc> But that requires that TileIndex is a distinct type :)
12:46:36 <michi_cc> And typedefs are not distinct types.
12:46:54 <TrueBrain> I am a bit surprised by that, I have to say :P C++ can be a bit weird :)
12:47:23 <michi_cc> That might be because typedefs are C of course, not only C++.
12:49:46 <TrueBrain> nielsm: maybe name the image folder based on the post, so it is easy to find them back?
12:50:03 <TrueBrain> (so including the date)
12:53:56 <andythenorth> trying to route to unreachable depots
12:54:15 <DorpsGek> [OpenTTD/website] michicc commented on pull request #229: Add: Blog post about the 12.0 signals UI change https://git.io/JPjN2
12:59:11 <michi_cc> Have to go shopping, but if somebody has any more opinions on TileIndex, I'd love the read them in the backlog :)
13:03:54 *** Etua has quit IRC (Quit: Etua)
13:06:45 <TrueBrain> for what it is worth, I have no opinion on the matter :)
13:14:52 <DorpsGek> [OpenTTD/website] nielsmh updated pull request #229: Add: Blog post about the 12.0 signals UI change https://git.io/JPs41
13:16:14 <DorpsGek> [OpenTTD/website] nielsmh commented on pull request #229: Add: Blog post about the 12.0 signals UI change https://git.io/JPjjq
13:18:25 <LordAro> TrueBrain: do you recall if there's anything stopping me from merging eints#61 ?
13:19:05 <TrueBrain> I don't do eints :P Better ask frosch123 :P
13:19:19 <TrueBrain> (I really do not know the impact etc)
13:22:33 <TrueBrain> learning new dialetcs?
13:22:49 <LordAro> only the differences between seconds and milliseconds :p
13:23:19 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #9672: Fix #9669, cbe00ec: Only try valid reverse directions https://git.io/JXeeD
13:24:20 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #9653: Update: add setting to hide news about competitors vehicle crash https://git.io/JXeep
13:26:24 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #9652: Fix #8797: Use logical rail length when placing signals https://git.io/JXev0
13:27:12 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #9639: Feature: Ctrl-click to remove fully autoreplaced vehicles from list https://git.io/JXevr
13:28:29 <LordAro> right, that's my OTTD duties done for another week :p
13:36:52 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
13:37:58 *** WormnestAndroid has joined #openttd
14:05:36 <LordAro> TrueBrain: you're the one with the scripts :p
14:06:03 <TrueBrain> I gisted it! But it also requires you first to merge those PRs ;)
14:12:51 *** jottyfan has joined #openttd
14:25:30 <DorpsGek> [OpenTTD/website] michicc approved pull request #229: Add: Blog post about the 12.0 signals UI change https://git.io/JXeOT
14:35:18 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
14:37:34 <DorpsGek> [OpenTTD/website] nielsmh merged pull request #229: Add: Blog post about the 12.0 signals UI change https://git.io/JPs41
14:50:06 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #9674: Fix: Don't show screenshot GUI in screenshots https://git.io/JPhrc
14:51:57 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #9674: Fix: Don't show screenshot GUI in screenshots https://git.io/JXenV
14:54:01 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #9674: Fix: Don't show screenshot GUI in screenshots https://git.io/JXenA
14:57:11 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler commented on pull request #9673: Fix #9579: Object and HQ construction is Construction cost, not Property Ma… https://git.io/JXecV
15:16:06 <michi_cc> Well, this is basically my struct variant. What I'm afraid of is that all these tiny functions might not get inlined in a debug build, making it even more slow.
15:16:32 <michi_cc> I have not measured that though, but probably will do it, as the enum variant requires little changes all over the code and is still less flexible than a struct.
15:18:32 <michi_cc> The problem with TileIndex is that the underlying base type is uint32, but there are lots of places that try stuff a (signed) int into it, which the strong typedef examples also try to prohibit (or at least make not implicit).
15:19:07 <andythenorth> anyone want to play coop NoCarGoal on pikka's server?
15:21:09 *** esselfe has joined #openttd
15:35:34 *** WormnestAndroid has joined #openttd
15:35:48 *** Wormnest has joined #openttd
16:32:41 <DorpsGek> [OpenTTD/OpenTTD] glx22 merged pull request #9672: Fix #9669, cbe00ec: Only try valid reverse directions https://git.io/JPb8b
16:32:44 <DorpsGek> [OpenTTD/OpenTTD] glx22 closed issue #9669: [Bug]: Ships can exit depots and locks through the sides if the exit is a dead end https://git.io/JP7NY
16:46:06 *** jottyfan has joined #openttd
16:46:21 <michi_cc> nielsm: Your website post will be live whenever the GH action has run through.
16:54:54 *** frosch123 has joined #openttd
16:55:13 *** jottyfan has quit IRC (Quit: jottyfan)
16:56:32 <LordAro> string fixed, now for the hard bit - working out how to leave the team...
16:58:08 <LordAro> i can remove myself, but i'm pretty sure that's the admin interface
16:58:52 <frosch123> michi_cc: struct is better than enum. it allows to add a default constructor, and possibly members x(), y() in the future. you can add the actual integer as public member and access it directly in places where it matter. i would hope there would be not that many: mostly map-accessors and tile_func...
16:58:56 <DorpsGek> [OpenTTD/OpenTTD] LordAro commented on issue #9657: [Bug]: In English (AU) language, "Show tooltips" setting says the unit of time is seconds when it should be miliseconds https://git.io/JiW3T
16:58:59 <DorpsGek> [OpenTTD/OpenTTD] LordAro closed issue #9657: [Bug]: In English (AU) language, "Show tooltips" setting says the unit of time is seconds when it should be miliseconds https://git.io/JiW3T
16:59:18 <frosch123> LordAro: there is a "leave team" button on the team page
16:59:21 <frosch123> andy tested that before
16:59:41 <nielsm> frosch123: there's some concerns about POD types if you add constructors, iirc
17:00:27 <frosch123> nielsm: move-constructor stuff, i don't think that matters for an int
17:00:59 <michi_cc> frosch123: Oh, conceptually a struct is definitely better, just have to test if there are any performance problems from it. Might also just stick a cast operator to uint into it for starters to avoid having to touch too much at once.
17:01:05 <frosch123> i think the ABI uses "fits into register" as rule, not "is an integer"
17:01:59 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
17:02:58 <nielsm> frosch123 it's more about things that turn out to technically by UB when you manipulate the type with memcpy(), memset() and friends
17:03:09 <nielsm> technically already violated on the Money type
17:08:03 <frosch123> yes, there is trivially-constructable, trivially-copyable and trivialy-moveable
17:09:01 <frosch123> i don't remember the current state, there were issues...
17:09:55 <frosch123> like "new T;" not calling default constructor, "new T()" calling it, and who knows what "new T{}" does
17:11:34 <frosch123> i think the solution was to remove the term "POD" and "std::is_pod" from the standard, and replaceing it with more explicit terms
17:12:26 <nielsm> yeah, but that's C++20, as long as we're on an older version it's still a concern
17:12:29 *** snail_UES_ has joined #openttd
17:12:53 <nielsm> although probably not one that really matters
17:31:45 <Tryard> Hi, how can I enable buy shares of company into my dedicated server?
17:54:02 <nielsm> set economy.allow_shares true
17:54:06 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #9675: Fix 65cbde4b: Writing to uninitialized string when loading a currency NewGRF. https://git.io/JXeAA
17:55:48 <DorpsGek> [OpenTTD/OpenTTD] glx22 approved pull request #9675: Fix 65cbde4b: Writing to uninitialized string when loading a currency NewGRF. https://git.io/JXexl
17:57:49 <nielsm> and remember that allowing buying shares in multiplayer enables some really silly exploits, because the economy around it is way too simple, and also that it does not allow you to buy out human players (unlike AI players in singleplayer, where you can buy 100% to take over their company)
18:01:46 <DorpsGek> [OpenTTD/OpenTTD] michicc opened pull request #9676: Fix 55a11710: Curly braces need to be double-escaped for (std::)fmt. https://git.io/JXepp
18:03:51 <DorpsGek> [OpenTTD/OpenTTD] LordAro approved pull request #9676: Fix 55a11710: Curly braces need to be double-escaped for (std::)fmt. https://git.io/JXehc
18:09:05 <Tryard> Ye, I can buy up to 75% of the company right?
18:10:01 <Tryard> And if I buy a shares company there is a dividends?
18:15:33 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #9675: Fix 65cbde4b: Writing to uninitialized string when loading a currency NewGRF. https://git.io/JXeAA
18:19:53 <glx> but you can make profit by selling shares later
18:23:28 <andythenorth> nml CHIPS when? :P
18:23:37 * andythenorth should get the PR
18:26:23 * andythenorth tests, nml example station compiles
18:27:51 <glx> michi_cc: checkout failed but retried
18:30:18 <Tryard> It is possible to enable AI in a server?
18:43:09 *** snail_UES_ has quit IRC (Quit: snail_UES_)
18:47:21 <DorpsGek> - Update: Translations from eints (by translators)
18:50:26 <andythenorth> could we make the game run slower somehow? :P
18:54:28 <frosch123> drink more coffee, then it appears slower
19:16:07 <DorpsGek> [OpenTTD/OpenTTD] Rau117 commented on pull request #9002: Change cargo age to have more effect on cargo payment. https://git.io/JXvmo
19:21:11 <DorpsGek> [OpenTTD/OpenTTD] michicc merged pull request #9676: Fix 55a11710: Curly braces need to be double-escaped for (std::)fmt. https://git.io/JXepp
19:34:00 <TrueBrain> (and michi_cc ofc for maintaining the OSX parts despite not owning a Mac :P)
19:35:06 <michi_cc> TrueBrain: You did some stuff on the last big round, too.
19:35:44 <TrueBrain> I am pretty happy to see OpenTTD in that list :D
19:35:46 <frosch123> hey, our sister project 'battle for wesnoth' is also on the list
19:41:07 <andythenorth> lol I was going to say 'Blitz works fine'
19:41:13 <andythenorth> but it seems to be using Rosetta
19:41:59 * andythenorth wonders if DOSBox will work
19:42:06 <andythenorth> if I disappear, it kernel panicked
19:42:56 * andythenorth not really understanding the debates on 9002 :P
19:43:37 <michi_cc> This is someone on a crusade after we took away their signals.
19:45:47 <LordAro> they're one more misstep away from just getting blocked from the GH org
19:46:02 <LordAro> (they've been good for the last few comments though)
19:56:09 <TrueBrain> I think it is safe to say that if you piss off LordAro , you made some poor life choices :)
19:57:44 <peter1138> took away... or hid behind a setting... hmmm
20:02:27 <dP> imo 9002 should just be closed at this point, it's 3 urelated commits and it started on the wrong foot
20:03:25 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
20:09:45 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
20:09:49 *** WormnestAndroid has joined #openttd
20:13:15 <dP> follow_track.hpp:124: bool CFollowTrackT<Ttr_type_, VehicleType, T90deg_turns_allowed_, Tmask_reserved_tracks>::Follow(TileIndex, Trackdir) [with TransportType Ttr_type_ = TRANSPORT_WATER; VehicleType = Ship; bool T90deg_turns_allowed_ = true; bool Tmask_reserved_tracks = false; TileIndex = unsigned int]: Assertion `((TrackStatusToTrackdirBits( GetTileTrackStatus(m_old_tile, TT(), (IsRoadTT() && m_veh != nullptr) ? (this->IsTram() ? RTT_TRAM : RTT_ROAD) : 0) ) &
20:13:15 <dP> TrackdirToTrackdirBits(m_old_td)) != 0) || (IsTram() && GetSingleTramBit(m_old_tile) != INVALID_DIAGDIR)' failed.
20:14:06 <dP> can't find anything like that
20:14:20 <dP> but it kinda just happens, surely someone must've run into it already
20:15:37 <dP> oh, 4 hours ago, guess my master wasn't recent enough xD
20:21:37 <dP> ok, can confirm, doesn't happen on latest master
20:22:33 <dP> would be nice to backport though I guess ;)
20:25:01 <dP> oh, right, was looking at the wrong issue
20:25:22 <glx> I mentionned the issue, not the PR :)
20:26:02 <dP> yeah, and I couldn't find pr for some reason
20:35:58 *** Wormnest has joined #openttd
20:43:49 *** gelignite has joined #openttd
21:01:52 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:19:51 <andythenorth> if we knew a caboose was a caboose
21:20:04 <andythenorth> we could build new wagons before the caboose, if it's the last vehicle in the train
21:20:10 <andythenorth> this would reduce some yak-shaving
21:20:57 <glx> or let the newgrf handle that
21:21:16 <glx> if caboose anywhere in the chain, draw it at the end
21:23:10 <andythenorth> surprisingly plausible :P
21:23:31 <andythenorth> don't think I can change the vehicle name string on cb36 though
21:23:50 <andythenorth> means 'Information' would be wrong
21:26:47 <frosch123> did you get reports about missing "dining car" infos?
21:27:42 <glx> I remember dbset and "MORE" wagons :)
21:28:26 <frosch123> and then you joined ottd? :p
21:28:46 <andythenorth> hmm town authorities are annoying eh :)
21:28:53 <andythenorth> there's a GS that nerfs them I think
21:59:03 * andythenorth wonders if industry sub-layouts is revivable
22:00:04 <andythenorth> nielsm had a patch for it
22:00:14 * andythenorth about to replicate a similar thing in grf
22:14:55 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
22:23:39 <DorpsGek> [OpenTTD/OpenTTD] J0anJosep opened pull request #9677: Fix: Do not redraw vehicle lists when skip order command is tested. https://git.io/JXJMt
23:00:03 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
23:00:29 *** WormnestAndroid has joined #openttd
23:00:34 <dP> why does new client list have its own widget system?
23:00:42 <dP> openttd already does dynamic buttons in other places
23:18:36 *** andythenorth has quit IRC (Quit: andythenorth)
23:23:37 <glx> MakeCompanyButtonRows() is not really dynamic, it's used only for window creation
23:27:56 <dP> yet it does update dynamically just fine
23:28:14 <dP> in detailed perf rating window for example
23:28:22 <glx> but it's only a company list
23:28:49 <glx> can be made generic I guess
23:29:46 <dP> it already looks pretty generic as it adds containers and spacers and stuff
23:29:54 <glx> current name doesn't tell it can work for something els :)
23:30:05 <dP> watch patch actually does even more complex ui the same way
23:30:34 <glx> but yeah it's already almost generic (except the sprite for size)
23:31:02 <dP> proper generic name to look at is NWidgetFunction ;)
23:45:08 *** Kitrana has joined #openttd
23:45:12 *** Kitrana1 has quit IRC (Read error: Connection reset by peer)
23:49:04 *** Kitrana1 has joined #openttd
23:53:10 *** Kitrana has quit IRC (Ping timeout: 480 seconds)
continue to next day ⏵