IRC logs for #openttd on OFTC at 2025-04-05
⏴ go to previous day
00:13:02 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:13:08 *** WormnestAndroid has joined #openttd
01:40:16 *** Wormnest has quit IRC (Quit: Leaving)
02:32:53 *** WormnestAndroid has quit IRC (Remote host closed the connection)
02:33:00 *** WormnestAndroid has joined #openttd
02:45:49 *** gnu_jj_ has joined #openttd
02:49:07 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
02:49:22 *** geizeskrank has quit IRC (Ping timeout: 480 seconds)
02:52:54 *** geizeskrank has joined #openttd
02:57:13 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
02:57:15 *** WormnestAndroid has joined #openttd
02:57:36 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
02:57:37 *** WormnestAndroid has joined #openttd
02:57:45 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
02:58:08 *** WormnestAndroid has joined #openttd
03:54:01 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:31:58 *** D-HUND is now known as debdog
04:43:50 <DorpsGek> - Update: Translations from eints (by translators)
05:13:53 <andythenorth> lol I should not reply
05:14:15 <andythenorth> I need github to test my recent coffee intake before allowing a reply
07:18:54 <kuhnovic> KuhnovicviaGitHub: xarick It would be great if you could test this one
08:35:02 <andythenorth> if you saw `length = len(my_list or [])` would that be a weird idiom?
08:50:24 <pickpacket> andythenorth: lunch *now*??
08:50:50 <pickpacket> At 9:50 in the morning?
08:51:36 * pickpacket hasn't even had breakfast yet
08:52:22 <pickpacket> Then I suggest brunch
08:52:49 <andythenorth> I was considering an end run
08:55:03 <andythenorth> cutting out the first step, going straight to the main event
09:01:30 <pickpacket> Yeah, you're right. Might as well
09:01:57 *** mindlesstux has joined #openttd
09:02:55 <pickpacket> Is a 15.0-beta2 still on the table, btw?
09:25:18 <andythenorth> is peter1138[d] on a bike?
09:34:37 *** WormnestAndroid has quit IRC (Remote host closed the connection)
09:34:41 *** WormnestAndroid has joined #openttd
09:40:40 <xarick> checking if my maths at calculating tiles are correct TestRemoveRail().TryRemove(entry_tile_2, exit_tile_2, 2 * exit_tile_1 - entry_tile_1)
09:47:12 <andythenorth> flag on badges, option to group in buy menu?
09:47:52 <andythenorth> variants implementation could be backwards supported by adding badges from variant prop
09:50:19 <xarick> it starts removing from the wrong tile
09:50:50 <xarick> this is what I expected:
09:51:26 <xarick> seems that first and last tracks are being included?
09:51:43 <xarick> gonna check openttd code
09:53:27 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
09:54:03 <xarick> the script function I use is ScriptRail::RemoveRail
09:58:45 <kuhnovic> Are you still checking the consequences of the PR I merged yesterday? You did see I have another one open since this morning right?
09:59:24 <xarick> okay, what Im doing is "wrong"
09:59:45 <xarick> or currently it got into a wrong state
10:00:36 <xarick> previously, this would fail at removing
10:01:19 <xarick> now, why did it add tile 1555 as part of the route
10:01:42 <xarick> it's part of the station
10:06:02 <xarick> hmm pff, i depend on the ALREADY BUILT error again
10:15:11 <xarick> alright, this is it, this is the problem... I was dependant on ERR_ALREADY_BUILT
10:15:38 <xarick> if already built, don't push tile 1555 into m_builtTiles
10:16:02 <xarick> but now it doesn't fail at building if it's already built, so tile 1555 is pushed into m_builtTiles
10:16:31 <xarick> when the ai notices it can't build the 2nd rail line, it removes everything that was pushed into m_builtTiles
10:16:51 <xarick> tile 1555 is present now, it was not present before Kuhnovic
10:17:47 <kuhnovic> kuhnovic: xarick for the third time...
10:18:25 <xarick> it's not a problem doing remove rail
10:18:57 <xarick> it's indeed the BuilRail not failing at ERR_ALREADY_BUILT
10:20:11 <xarick> oh, 13968, my bad I didn't pay attention
10:25:09 <xarick> yay, 13968 fixed it, just tested
10:26:04 <kuhnovic> I made the same change for roads, with a little luck it's fixed as well
10:42:46 <xarick> i have a lot of already built errors in roads hmm
10:42:59 <xarick> wondering which ones are important
10:43:05 <xarick> enough to cause issues
10:43:38 <kuhnovic> But is it any different to what it was before?
10:52:31 <_glx_> Write only discord can be annoying 😉
10:55:05 <xarick> I'm not sure, i'm investigating
11:08:02 <xarick> oh, i just remembered i have a GS already ready to test it, I'm stupid
11:09:44 <xarick> hmm looks like it's not working
11:20:27 <xarick> just a typo on line 1026
11:27:16 <kuhnovic> That code was there before 13885, I simply reverted that change and started over
11:29:15 <xarick> needs further investigation, but it's lunch time now, afk
11:58:46 <kuhnovic> xarick: FYI no functional changes, just fixed commit messages
12:16:49 <xarick> gonna investigate why it used to work
12:21:55 <xarick> ah, it's the had_success
12:27:07 *** kuka_lie has joined #openttd
12:30:24 <xarick> okay, gonna change my suggestion, or make a new one perhaps
12:31:51 <kuhnovic> Oh, I now see that the return value on road_cmd.cpp line 1055 depends on had_success. You might be on to something after all.
12:32:10 <kuhnovic> This shit is delicate
12:34:43 <xarick> `if (ret.GetErrorMessage() == STR_ERROR_ALREADY_BUILT) had_success = !is_ai;`
12:36:04 <xarick> `if (!is_ai && ret.GetErrorMessage() == STR_ERROR_ALREADY_BUILT) had_success = true;`
12:36:43 <xarick> maybe the 2nd is better
12:37:57 <xarick> what if the ai builds a partial road
12:40:33 <kuhnovic> The first one is simply incorrect as it would set had_success to false even if it was already set to true
12:40:57 <kuhnovic> So the second one is better. i can change it.
12:43:24 <peter1138[d]> You should find a rock, it would be more useful.
12:47:41 <xarick> yeh, 2nd is correct, 1st does something different
12:50:35 <_glx_> hehe test was not needed for 1st one, it was clearly wrong
12:51:45 <kuhnovic> xarick: Then it would have to be had_success_at_least_once, which is way too verbose
12:52:24 <_glx_> had_success is fine, you don't care how many times it had success
12:52:51 <_glx_> could even be just `success` (quite common in source code)
12:57:27 <xarick> kuhnovic: coming from the guy with `node_water_patch_on_high_level_path`
12:58:02 <kuhnovic> xarick: Just buy an ultrawide ok 😛
12:58:47 <kuhnovic> In all seriousness, although a bit verbose, that code is quite complicated. A bit code self-documenting doesn't hurt in such cases.
12:58:49 <_glx_> we should go back to BASIC time
12:59:09 <kuhnovic> 640x480 screens only
13:02:23 <kuhnovic> Some colleagues of mine work on a different (ancient) product that has function names like ed15 in them. Turns out the author's initials were E.D. and he was already at number 15 with this numbering scheme...
13:03:04 <kuhnovic> Then you start do wonder, should I review this code or just punch the guy?
13:06:43 <_glx_> I remember the start on opendune, function names were based on memory addresses
13:07:09 <_glx_> same for variables of course
13:07:47 <_glx_> hard to do different with decompiled stuff
13:08:03 <peter1138[d]> andythenorth: is it lunch?
13:09:12 <andythenorth> ran out of coffee
13:09:17 <kuhnovic> A former colleague of mine was a judge in an international robotics competition. He once had to take a look at a piece of code that was entirely written in japanese. So all your regular syntax was in english, but everything else was essentially obfuscated (for him) 😛
13:09:45 <_glx_> hehe like simutrans with german everywhere
13:11:57 <kuhnovic> Haha yeah, although as a dutch speaker I can usually work out what it means. But it is a bit silly for an open source project IMO.
13:23:24 <andythenorth> wonder if I should just have one level of variant nesting
13:23:30 <andythenorth> less string faff 😛
14:26:03 *** blathijs has quit IRC (Quit: brb - reboot)
14:28:08 <xarick> changing AIs in scenario editor is no longer crashable, right?
14:29:41 <xarick> I can safely permit AI/GS parameter changes inside editor
14:40:17 *** blathijs has joined #openttd
14:40:17 *** ChanServ sets mode: +o blathijs
14:46:11 <xarick> the reload AI button is available in scenario editor
15:00:12 <_glx_> it will fully reset the company
15:06:38 <kuhnovic> xarick: did you pull my latest changes?
15:06:48 <andythenorth> don't set variant_group to self 😛
15:08:26 <xarick> i tested road, it works now, did you change anything for rail?
15:09:44 <kuhnovic> Yes I made the same change for rail. It only sets had_success if it's called in non-ai mode
15:10:08 <xarick> oh, then i suppose it works
15:10:14 <xarick> let me just test quick
15:10:32 <xarick> oh, im in the middle of a rebase
15:11:49 <kuhnovic> It was probably working before, but this check doesn't hurt. Just to make sure there is no use case that I've overlooked.
15:22:54 <xarick> can you wait a few hours?
15:23:17 <_glx_> andythenorth: I hope OpenTTD validates that
15:23:37 <xarick> i'm working on something else
15:23:54 <andythenorth> _glx_: crashes 🙂
15:23:59 <andythenorth> probably an issue somewhere
15:25:16 <andythenorth> I can make a test case
15:25:36 <andythenorth> but currently Horse needs the badges branch 😐
15:29:04 <_glx_> ```/* Check engine variants don't point back on themselves (either directly or via a loop) then set appropriate flags
15:29:04 <_glx_> * on variant engine. This is performed separately as all variant engines need to have been resolved. */``` it should be handled
15:29:44 <_glx_> open an issue if you have reproduceable test case
15:41:28 <_zephyris> What needs to be done for default badges?
15:46:24 <kuhnovic> xarick: Are you talking to me?
15:46:53 <xarick> yes, you got me in the middle of a tough rebase, and i can't cancel it atm
15:47:05 <kuhnovic> I'll be AFK the rest of the night so knock yourself out 😛
15:51:02 <xarick> GetBridgeType does not exist on 15.0-beta1? my AI crashed
15:52:50 <_glx_> ah yeah might be from after beta1
15:55:11 <_glx_> beta1 is not a stable version, so API 15 is only valid for current master
15:55:38 <_glx_> beta1 is a nightly without a moustache
16:03:32 <xarick> nop, can't notice anything
16:03:53 <xarick> ah, it closes dropdown from the list
16:03:59 <xarick> maybe it does something
16:07:47 *** kuka_lie has quit IRC (Remote host closed the connection)
16:29:06 <peter1138> What should a script list do/return if you call Next() when it's already at the end?
16:30:16 <xarick> i think regression tests that. I doesn't crash the script
16:30:24 <xarick> what it returns... no idea
16:32:41 <xarick> trees in scenario editor in fast forward are planted but don't grow?
16:36:42 <peter1138> The regression tests it, but IMHO gives a weird result.
16:37:18 <peter1138> It returns an item that is no longer in the list, but then again it
16:37:32 <peter1138> does also return true for IsEnd()
16:38:32 <peter1138> As any integer value is valid, there's nothing it could return that could mean 'invalid'.
16:47:07 <peter1138> Maybe it should error :)
16:47:58 <xarick> does it return a warning message?
16:53:58 <xarick> _glx_: have you tested what happens if we change the game script to none in the scenario editor?
16:54:24 <xarick> is the instance really gone?
16:54:52 <peter1138> Even the errors for calling Next() or IsEnd() before Begin() is called are actually just debug messages, so don't show in the script debug window.
17:01:04 <_glx_> xarick: of course it's gone
17:13:29 <xarick> IsEditableItem greatly simplified
17:15:01 <peter1138> Hmm, and getting an item from a list returns 0 if the item doesn't exist.
17:15:08 <peter1138> But 0 could be a valid value. Hmm.
17:18:45 <_glx_> that's why there's also HasItem
17:20:30 <_glx_> and GetValue() also only reads the first matching item
17:21:11 <_glx_> though the list prevents adding duplicates
17:21:51 <xarick> what does AddItem do if item already exists?
17:22:11 <xarick> changes value if it's different?
17:23:02 <xarick> wondering about AddList
17:23:47 <xarick> if there's items on both lists does it keep the values or updates the values?
17:24:11 <_glx_> AddList is AddItem then SetValue
17:25:19 <xarick> okay time to test 13968 for rails
17:25:59 <_glx_> btw documentation mention all of that 🙂
17:34:17 <peter1138> What are we ruining now?
17:34:46 <xarick> there was a bug in my code though while I was figuring out where things were going wrong
17:35:01 <xarick> turns out it wasn't it
17:35:07 <xarick> but still, good that I looked
17:36:32 <andythenorth> peter1138: "yes"
17:39:17 <andythenorth> I should de-pluralise these though eh?
17:39:36 <andythenorth> even though the group contains multiple variant types
17:43:19 <peter1138> Pfft, variants aren't group.s
17:44:41 <andythenorth> I could de-group the ones that have high variation
17:45:15 <andythenorth> those vehicle transporters look quite different
17:45:25 <andythenorth> grey area falls on ones like this
17:47:24 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
17:47:28 <_glx_> You can't degroup a non existing group
17:48:03 *** WormnestAndroid has joined #openttd
17:48:22 <xarick> 4x zoom sprites when? 🙂
17:49:59 <andythenorth> who named the nml property variant_group? 😛
17:58:12 <peter1138> Vehicle Groups. Engine Groups. What ever next?
18:03:48 <andythenorth> dunno, but I might remove the pluralisation
18:04:24 <andythenorth> maybe just use a badge, no name string 😛
18:29:15 <andythenorth> all "Coil Carrier" and then a badge for "Enclosed"? 👀
18:36:39 *** talltyler has joined #openttd
18:36:39 <talltyler> Nah, I can use my eyes
18:41:30 <frosch123> Did you even try LOAD_TEMP for badges?
18:42:34 *** ChanServ sets mode: +v tokai
18:45:18 <frosch123> No idea what that topic has to do with badges
18:47:12 <andythenorth> badge everything 😛
18:49:27 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
18:49:53 <peter1138> Ah, usually it's "I want to combine strings for the extra text callback"
18:50:13 <peter1138> Which can be replaced by badges.
18:51:21 <andythenorth> wonder how much compile time I save if variants level 0 and level 1 have same name string
18:51:34 <andythenorth> 2000 instances of stuffing text stack for level 0
19:33:07 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
19:33:08 *** WormnestAndroid has joined #openttd
19:57:37 *** WormnestAndroid has quit IRC (Remote host closed the connection)
19:58:52 *** WormnestAndroid has joined #openttd
20:05:50 <_glx_> oh it's unrelated to this one
20:11:30 <xarick> also crashes on master, confirmed
20:13:18 <_glx_> yeah I confirmed it too
20:14:22 <_glx_> ok it confirms what I suspected
20:20:16 <andythenorth> name callback (161) can return 0x400 for default name, but I don't see an nml constant for that
20:24:36 <frosch123> CB_RESULT_NO_TEXT :p
20:26:36 <_glx_> funny how a "simple" change can reveal stuff
20:27:22 <andythenorth> frosch123: dunno why I looked at that, then ignored it
20:36:29 <xarick> assert(MayHaveRoad) isn't listed in view call hierarchy
20:38:10 <xarick> `GetRoadTypeRoad`, `GetRoadTypeTram`, `GetRoadOwner`, `SetRoadTypeRoad`, `SetRoadTypeTram` need to check hierarchy of these
20:47:45 *** Wormnest has joined #openttd
20:52:41 <xarick> towns can't own road bridges?
20:54:01 <xarick> oh, trams can't own trams, my bad
21:04:32 <xarick> suspicious bad code style in smallmap_gui.cpp
21:08:15 <xarick> it's called twice, could be put into a variable
21:11:37 <_glx_> you mean something like ```if (RoadType rt_road = GetRoadTypeRoad(tile) != INVALID_ROADTYPE) {
21:11:37 <_glx_> rti = GetRoadTypeInfo(rt_road);
21:12:54 <xarick> if it makes drawing the map faster, yeah i guess
21:13:19 <_glx_> it will reuse the value
21:23:03 <_glx_> oh actually it does the call 2 times, but should not affect perf
21:34:23 <_glx_> but only when asserts are enabled, else the code is inlined
21:34:43 <xarick> how can a road tile have no roadtype...
21:35:41 <_glx_> tram and road types are stored on different location
21:36:21 <_glx_> so a MP_ROAD can have only road, or only tram, or both
21:36:46 <xarick> wondering if GetPresentRoadTypes could be used
21:37:02 <_glx_> and for smallmap road has priority if present
21:39:06 *** nielsm has quit IRC (Ping timeout: 482 seconds)
21:39:21 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:45:37 <xarick> ```const RoadType rt_road = GetRoadTypeRoad(tile);
21:45:37 <xarick> rti = GetRoadTypeInfo(rt_road != INVALID_ROADTYPE ? rt_road : GetRoadTypeTram(tile));```
21:46:39 <_glx_> change probably unneeded
21:47:59 <xarick> GetPresentRoadTypes is also meh... i just noticed
21:48:45 <_glx_> but `rti != nullptr` check is useless I think (but doesn't hurt)
21:53:58 <_glx_> anyway I don't think it will be a huge or even visible performance gain
22:00:59 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:08:37 <xarick> 2 weaks ago this started
22:09:06 <xarick> SingleRail pathfinder gets 4000+ trains, DoubleRail pathfinder gets 2000+
22:09:20 <xarick> map isn't good enough for ships 😦
22:22:17 *** gelignite has joined #openttd
continue to next day ⏵
