IRC logs for #openttd on OFTC at 2024-12-31
⏴ go to previous day
00:10:31 <xarick> tile + GetDockDirection(tile) == t
00:10:31 <xarick> what's evaluated first?
00:19:32 <talltyler> Aw, Copilot got it right 😦
00:22:20 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
01:58:29 *** nielsm has quit IRC (Ping timeout: 480 seconds)
02:03:59 *** gelignite is now known as Guest4580
02:04:02 *** gelignite has joined #openttd
02:11:18 *** Guest4580 has quit IRC (Ping timeout: 480 seconds)
03:18:14 *** gnu_jj_ has joined #openttd
03:21:36 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
03:25:31 *** Zathras has joined #openttd
03:28:48 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
03:28:53 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:34:05 *** firmat has quit IRC (Remote host closed the connection)
03:58:57 *** firmat has quit IRC (Remote host closed the connection)
04:42:27 <DorpsGek> - Update: Translations from eints (by translators)
07:54:14 *** Flygon has quit IRC (Remote host closed the connection)
09:45:13 <peter1138> I mean it works, but the UI side is basically duplicated.
09:47:27 <peter1138> And that we haven't migrated the old setting-style cheats to settings means it's not too coherent.
09:54:45 <peter1138> Is this too verbose?
09:55:27 *** tabytac has joined #openttd
09:55:27 <tabytac> Would a more long term solution be to just have 1 settings page with added tabs for something like "client", "Company", "Game" and "Sandbox" added to the existing Game Options windows, maybe needing a slight design to accommodate the extras. If splitting up certain categories more to like "current game/all games" and "current company/all companies"
09:55:47 <kuhnovic> peter1138: Why is there separator line? What makes these settings different from a user perspective?
09:56:56 <truebrain> peter1138: Think about the savegame size ! 😛
09:57:30 <peter1138> truebrain: Yeah. We currently only log changes to things that can't be changed in network games.
09:59:10 <peter1138> Logging everything is clearly too much.
10:01:01 <peter1138> kuhnovic: Exactly. There are some cheats that are settings-style, and then other cheats that are actions. If all the settings-style cheats are moved to settings, the distinction is that the line separates actions from settings.
10:01:11 <peter1138> But yes, that isn't the case here right now.
10:01:56 <peter1138> Amazing, I can actually see sky today.
10:08:18 <kuhnovic> I still don't really see how that makes a difference from a user perspective 😛
10:13:15 <kuhnovic> The cheese is brown here
11:35:45 *** Flygon has quit IRC (Ping timeout: 480 seconds)
11:40:50 *** Smedles has joined #openttd
11:54:06 <peter1138> Why is it ARMOURED, and why is the description so long?
11:56:31 <peter1138> This one is also very long, but at least it is formatted to be readable.
12:01:19 *** Flygon_ has joined #openttd
12:06:44 *** Flygon has quit IRC (Ping timeout: 480 seconds)
12:30:20 <xarick> 69.8 seconds to 70.7 seconds...
12:30:37 <xarick> it's not much of a slowdown
12:46:50 <andythenorth> what are we doing today?
12:46:57 <andythenorth> compression fit waste plumbing?
12:50:07 *** gelignite has joined #openttd
13:20:25 <andythenorth> I think the plumbing
13:20:28 <andythenorth> owing to the leaks
13:21:49 <_glx_> yeah plumbing feels it should always be first
13:22:38 <_glx_> but it's often something that needs to be done on sunday, when everything is closed
13:26:08 <xarick> could you even tell the difference?
13:30:48 <xarick> there's a big emphasis on the "performant" path
13:31:02 <_glx_> right finished after left
13:31:25 <xarick> but it's kinda a "meh" difference between them
13:31:50 <_glx_> yeah 1s or 2s diff on a very long run doesn't really matter
13:36:32 <xarick> thinking what to do next
13:38:08 <xarick> should I chase those consts that don't allow me to use tile + dir directly... should I revert all the tile + dir smart code, ...
13:39:39 <peter1138> Colour's all wrong.
13:41:49 <xarick> I'm worried about tile + dir being very prone to errors
13:42:08 <xarick> not sure it's worth the risk
13:43:31 <andythenorth> peter1138: a bit
13:43:31 <xarick> maintainability and all that considered, it feels bad this just won't work for all the cases
13:43:41 <andythenorth> "settings are hard"
13:43:47 <andythenorth> windows control panel anyone?
13:54:03 <xarick> how to fix this stack overflow?
13:54:37 <peter1138> Make your operator-() not call itself.
13:55:28 <peter1138> By implementing the operator properly.
13:56:01 <peter1138> Adding/subtracting an axis doesn't seem like it should be an operator anyway.
13:56:55 <andythenorth> would we like me to learn C++ for general entertainment?
14:03:50 <Rubidium> andythenorth: yes, please. Entertain us ;)
14:13:29 <xarick> i just "implemented it properly"!
14:13:49 <xarick> too much copy paste code
14:15:43 <truebrain> andythenorth: trying to learn me NewGRF was entertaining, so I don't see why not!
14:18:41 <peter1138> Hmm, crazy devs, removing features that never existed.
14:20:01 <Rubidium> oh, what did we remove this time?
14:38:53 *** frosch123 has joined #openttd
14:38:53 <frosch123> Do we still need the Apple Developer Program? It's about to expire, and I don't know who reads info@
14:40:50 <_glx_> learning squirrel stack is entertaining too
14:49:01 <truebrain> frosch123: when I read those emails, I just forward them to orudge 🙂
14:49:29 <truebrain> and yes, we still need it 😛 The MacOS builds are signed because of it 😄
14:50:30 <truebrain> but so far it has been the case of: it expires, because we didn't sign their new agreements (there is a new one every, what, 3 months? It is insane), we notice the nightly fails, I poke orudge, he fixes it 😛
14:51:17 <truebrain> I really do not know why they insist on changing their terms this often ... they really are a special snowflake 🙂
14:53:42 <truebrain> (I just forwarded the mail to orudge)
15:03:16 <xarick> `DrawTrainDepotSprite(x, y, widget - WID_BRAD_DEPOT_NE + DIAGDIR_NE, _cur_railtype);`
15:03:16 <xarick> WID_BRAD_DEPOT_NE + DIAGDIR_NE is complaining
15:05:32 <peter1138> const-ness again, in case you can't figure it out like last time.
15:08:59 <xarick> please explain, i don't get it
15:09:47 <xarick> i addec const to TileIndex
15:11:17 <peter1138> Hehehe, still cursed.
15:11:31 <peter1138> Black on mauve is a bit...
15:12:22 <xarick> > error C2666: '+': overloaded functions have similar conversions
15:12:22 <xarick> > note: could be 'TileIndex operator +(const TileIndex &,DiagDirection)'
15:12:22 <xarick> > note: or 'built-in C++ operator+(int, DiagDirection)'
15:12:45 <xarick> it's the 2nd one in this case
15:13:33 <peter1138> Ah of course, you've overloaded things and now there are multiple options.
15:13:33 *** Wormnest has joined #openttd
15:13:55 <xarick> enum BuildRailDepotWidgets : WidgetID {
15:14:06 <xarick> WidgetID + DiagDirection
15:14:13 <peter1138> Convert DIAGDIR_NE to integer first.
15:15:32 <xarick> nice, that fixed it, but now this leaves me wondering what else could be broken
15:20:27 <xarick> AddSortableSpriteToDraw(overlay + RTO_SLOPE_NE + static_cast<int>(tunnelbridge_direction)
15:21:38 <xarick> what i find it boggling is that it's not even a TileIndex
15:21:48 <xarick> why would it catch that
15:26:39 <mnhebi> andythenorth: I mean if you like cthulhu picking apart your brain, sure C++ is entertainment
15:28:02 <peter1138> Because StrongType has a non-explicit default constructor from uint32_t.
15:28:33 <peter1138> And also an assignment operator.
15:29:41 <peter1138> Which kinda makes a mockery of StrongType. But at some point...
16:02:33 <talltyler> peter1138: White on mauve is better, but might be a bit harsh to look at
16:03:31 <talltyler> That said, remember when we had a disagreement several years ago about too many window colours, and I wanted more to be grey? I’ve since changed my mind. I want more colours! 😄
16:04:32 <talltyler> I’m even wondering how the game would look with the green vehicle windows from TTD… 😛 (or TT? I’ve never actually played)
16:06:37 <talltyler> Speaking of the settings menu, I wonder if it’s worth splitting Settings into client settings (news notifications, signal side, gui stuff, etch and game settings. Client settings would be a natural fit for the Options menu even if the game settings didn’t get merged into that window.
16:07:15 <peter1138> talltyler: This is what someone said above.
16:07:38 <talltyler> Oh, I guess I didn’t read the whole scrollback 😄
16:07:55 <peter1138> So actually I did something completely different to that suggestion, which wasn't about merging options and settings at all.
16:08:03 <talltyler> Huh, read the posts above and below, and missed that one…
16:08:26 <peter1138> IMHO I don't see a reason to merge only some of the options.
16:08:29 <talltyler> That’s what I get reading Discord first thing in the morning
16:08:39 <peter1138> But I do see value in splitting out the client-side preferences from the gameplay preferences.
16:09:48 <peter1138> The big tree was certainly an idea, but it's complex and a bit confusing.
16:10:04 <peter1138> It was originally tabs, too.
16:10:55 <talltyler> I do like the search feature, I use that more than exploring the tree
16:11:26 <peter1138> _glx_: Having looked, I think quite a lot of it wrong and only works because of the implicit conversion from basetype to strongtype.
16:57:03 *** nielsm has quit IRC (Ping timeout: 480 seconds)
17:20:11 <xarick> what's the difference vs inline and debug_inline
17:21:01 <xarick> #define debug_inline inline
17:23:01 <xarick> is there an advantage in not using inline for release?
17:28:15 *** kuka_lie has joined #openttd
17:32:15 <peter1138> debug_inline means it should/could be inlined even with a debug build.
17:34:20 * Rubidium wonders why I put all that documentation in stdafx
17:40:25 <peter1138> Write-only documentation?
17:45:22 <xarick> btw Intellisense is always complaining about script_ai
17:45:36 <xarick> seemingly for no reason
17:48:16 <xarick> it wasn't captured by visual studio
17:48:43 <peter1138> It's there. In the source code. You just need to look at it.
17:55:54 <xarick> okay, so i need to test performance on debug builds, does that also incluse RelWithDebInfo?
17:56:28 <xarick> not quite sure what a (pure) debug build is
17:57:43 <peter1138> You probably do not need to test performance on debug builds.
18:02:45 <xarick> i hear a ship sound and then silence
18:19:36 <xarick> wentbourne, let me check where i have it
18:20:03 <xarick> how do i run for 100 ticks
18:36:34 <kurzov> Not sure whether to ask this in Discord channel #add-on-development or here because it has to do with the addon hosting service: But is there any plans to make the "Check Online Content" / BANANAS addon browser less of a "giant monolith" (for lack of a better term) to browse - maybe some sorting options, such as by last update date?
18:41:24 <_glx_> one day someone® may do it
18:43:55 <xarick> last of my very scientific performance tests
18:46:44 *** tokai|noir has joined #openttd
18:46:44 *** ChanServ sets mode: +v tokai|noir
18:52:03 <xarick> let me remove all inlines and debug_inlines, see what happens
18:53:36 *** tokai has quit IRC (Ping timeout: 480 seconds)
18:54:07 <xarick> oh, link errors 😦 many of them
18:58:29 <xarick> whatever, can't even make it build, so...
19:10:34 <johnfranklin> £3 single bus ticket cap… it is still not sure whether a single ticket would rise up to £3.
19:11:25 <johnfranklin> New year gift, but that “gift” is in german language.
20:37:36 *** Flygon_ has quit IRC (Read error: Connection reset by peer)
21:05:38 <xarick> 3 minutes to recompile after a comment change 😦
21:10:52 <xarick> 2 minutes and 10 seconds...
21:26:28 <xarick> wish I could ask questions to rubidium without angering him
21:27:13 <xarick> `TileIndex operator+=` or `TileIndex &operator+=` ?
21:31:48 <peter1138> Have you tried thinking about it? Are you returning a new TileIndex or the existing one?
21:33:40 <xarick> I have this weird situation `*tile += DIR_N;`
21:34:08 <peter1138> Okay, which one of those will return a new TileIndex, and which one will return an existing TileIndex?
21:35:29 <xarick> `if (CircularTileSearch(&tile, 40, SearchLumberMillTrees, nullptr))`, this tile is passed around and is updated, needs to return back updated
21:36:07 <xarick> I think it's `TileIndex &operator+=`
21:36:39 <xarick> but I also have cases where it's just tile += other_dir
21:37:55 <xarick> question is, do I need both `TileIndex operator+=` or `TileIndex &operator+=`?
21:54:09 <peter1138> I said something and this happens 😉
21:55:24 <xarick> woah, only 22 changes files? I changed 54
21:55:27 <johnfranklin> Say someone, someone arrives
21:55:33 <peter1139> Rubidium, how did you identify all occurances?
21:55:57 <Rubidium> make the constructor explicit
21:56:07 <peter1139> Ah, tested that way.
21:56:23 <peter1139> Should we use TileIndex{...} so that it doesn't look like a function call?
21:57:28 <peter1139> Oops, I should comment on the PR. Silly me.
21:58:12 <Rubidium> peter1139: could do that, but window_number is int32_t and TileIndex wants uint32_t, so it'd become TileIndex{static_cast<uint32_t>(...)}
22:00:12 <peter1139> Ah right, it's basically only this->window_number then.
22:00:42 <peter1139> At least for this set.
22:00:46 <LordAro> feels like they should not be different underlying types
22:02:14 <peter1139> Technically TileIndex{static_cast<uint32_t>(...)}
22:02:26 <peter1139> ... makes it clearer what is going on. But it's also noisier.
22:03:18 <peter1139> LordAro, you mean this->window_number should be uint32_t?
22:03:47 <peter1139> Or perhaps we should drop the window_number system and use real types. Not sure how feasible that is.
22:04:10 <peter1139> this->window_number refering to different things is a bit annoying sometimes.
22:04:19 <peter1139> But here we have scope creep.
22:05:17 <xarick> what about places that check for tile 0 ? 😦
22:05:46 <xarick> TileIndex{} = undefined value
22:05:59 <peter1139> That's not implicit construction.
22:06:47 <peter1139> Hmm, last time I thought about using constructor-casts like that somebody disagreed. Hmm.
22:09:11 <peter1139> I guess this is where other languages have (expensive) unboxing.
22:19:00 <xarick> it's new year time soon
22:24:09 <johnfranklin> How long will I get used to say 2025 instead of 2024…
22:24:41 *** belajalilija has joined #openttd
22:24:49 <belajalilija> i still think 2013
22:25:18 <belajalilija> i think i wrote 2014 on something recently and it wasnt a typo
22:43:53 <peter1139> So many rabbit holes.
22:50:09 *** gelignite has quit IRC (Read error: Connection reset by peer)
22:57:09 *** gelignite has joined #openttd
22:59:33 <Rubidium> yup, rabbit holes ;)
23:15:47 <xarick> Rubidium stuff gets merged quickly
23:19:10 <peter1139> LordAro, stop partying :D Any better ideas than TileIndex(this->window_number)?
23:19:58 <xarick> TileIndex{this->window_number}
23:20:46 <peter1139> Would need to be TileIndex{static_cast<uint32_t>(this->window_number)}, which is a mouthful.
23:25:49 *** keikoz has quit IRC (Ping timeout: 480 seconds)
23:34:40 <xarick> wow, i just discovered a fix
23:35:21 <xarick> ```TileIndex operator+(const TileIndex &tile, DiagDirection &diagdir)
23:35:21 <xarick> TileIndex operator+(const TileIndex &tile, DiagDirection diagdir)```
23:36:05 <xarick> the first one solves the issue of `DrawTrainDepotSprite(x, y, widget - WID_BRAD_DEPOT_NE + DIAGDIR_NE, _cur_railtype);`
23:36:14 <xarick> i don't need static_cast anymore
23:36:22 <_glx_> multiple overloads for similar things is never a good idea
23:36:33 <xarick> I can get rid of the 2nd
23:38:34 <_glx_> but nothing is a TileIndex there, so it should not use any of these
23:39:12 <peter1139> It does because of what Rubidium is now working on fixing.
23:39:13 <_glx_> (I know it's the implicit conversion)
23:39:41 *** kuka_lie has quit IRC (Quit: hmm should i go to sleep? its just 1.39 am)
23:43:45 <peter1139> Not really helping but tidies that up.
23:48:13 <xarick> oops, didn't work for SpriteID
23:48:26 <xarick> SpriteID + DiagDirection 😦
23:48:58 <xarick> > 23:48:11:439 E:\OpenTTD Visual Studio\SamuXarick\OpenTTD\src\tunnelbridge_cmd.cpp(1316): error C2666: '+': overloaded functions have similar conversions
23:48:58 <xarick> > 23:48:11:720 E:\OpenTTD Visual Studio\SamuXarick\OpenTTD\src\map_func.h(503): note: could be 'TileIndex operator +(const TileIndex &,DiagDirection &)'
23:48:58 <xarick> > 23:48:11:721 E:\OpenTTD Visual Studio\SamuXarick\OpenTTD\src\tunnelbridge_cmd.cpp(1316): note: or 'built-in C++ operator+(SpriteID, DiagDirection)'
23:50:30 <_glx_> because any number can be implicitly converted to TileIndex
continue to next day ⏵
