IRC logs for #openttd on OFTC at 2025-06-05
⏴ go to previous day
01:26:38 *** Wormnest has joined #openttd
02:12:29 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
02:15:55 *** WormnestAndroid has joined #openttd
02:19:56 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
02:21:24 *** WormnestAndroid has joined #openttd
02:40:48 *** Wormnest has quit IRC (Quit: Leaving)
02:44:34 *** gnu_jj_ has joined #openttd
02:48:11 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
03:20:09 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:26:18 *** D-HUND is now known as debdog
04:05:36 *** ipravd_ has joined #openttd
04:08:10 *** ipravd1 has joined #openttd
04:13:09 *** Flygon has quit IRC (Read error: Connection reset by peer)
04:13:54 *** ipravd1 has quit IRC (Quit: ipravd1)
04:14:12 *** ipravd1 has joined #openttd
04:39:37 *** ipravd1 has quit IRC (Quit: ipravd1)
06:19:20 <goddess_ishtar> you know, I've realised that people don't know the meaning of "steps to reproduce" when filing bug reports
06:51:14 <pickpacket> goddess_ishtar: I believe that people quite often don't know how to reproduce their errors
06:52:21 <goddess_ishtar> but surely, some context on "this is what I was trying to achieve" would be useful?
06:52:34 <pickpacket> I spent *hours* reproducing the error for my bug report 3 times, because it just takes so long
06:53:09 <pickpacket> goddess_ishtar: definitely, yeah... I guess there's room for improvement in most cases
06:55:07 <LordAro> PRs on the issue templates accepted!
06:55:30 <LordAro> at some point should probably convert to github's shiny new format as well
07:01:58 <peter1138[d]> goddess_ishtar: Also lack of detail on what they were doing, e.g. "using the scenario editor".
07:07:21 <peter1138[d]> Okay, in the industry window, should I just change "Production History" to "History"?
07:07:44 <peter1138[d]> Or perhaps something to suggest it'll be a graph...
07:08:18 <pickpacket> peter1138[d]: which one is the most correct? I.e. does "History" imply more or other information?
07:08:39 <peter1138[d]> History could theoretically be a text log...
07:09:22 <peter1138[d]> My branch adds cargo accepted and waiting data to the graph, so it's no longer just Production.
07:10:18 <goddess_ishtar> I think "Production History" is still the most evocative name
07:10:50 <goddess_ishtar> cargo accepted and waiting is still strongly correlated to production anyway
07:11:15 <peter1138[d]> Some industries don't produce anything.
07:12:56 <goddess_ishtar> "Cargo History"?
07:19:38 <_zephyris> It does everything, right? Delivery, stockpile, production and outgoing?
07:19:45 <_zephyris> "Logistics history"?
07:23:53 <peter1138[d]> I'd rather avoid introducing a word that isn't used anywhere else 🙂
07:24:46 <peter1138[d]> I'd also like it to be short(er) to avoid making the button wide.
07:29:39 *** f_ is now known as funderscore
07:30:15 <peter1138[d]> Okay, the button says "Production Graph" and the window says "Production History"
07:34:05 <peter1138[d]> "Cargo Graph" and "Cargo History" probably the simplest?
07:48:42 <peter1138[d]> Hmm, "Stockpiled" to match "Produced", "Transported" and "Delivered".
07:49:33 <peter1138[d]> Funnily enough, the word "stockpile" isn't used anywhere in OpenTTD except... the AI/GS API...
07:52:14 <_zephyris> I think it's clear enough.
08:45:31 *** mindlesstux has joined #openttd
09:11:12 <xarick> clicking `StationID` in ```static bool GSMarine::BuildDock ( TileIndex tile,
09:19:05 <peter1138[d]> Okay, there's a Megawad that combines levels from Doom and Doom 2. That'll be on later then.
09:44:02 <andythenorth> isn't that from 1996 or something?
09:44:06 <andythenorth> I feel like I played it
09:45:00 <andythenorth> I so should not start Doom wadding
09:56:15 <andythenorth> when we add OpenTTD Steve
09:56:36 <andythenorth> player-controllable character object that can move around map (cursors)
09:56:56 <peter1138[d]> May 2025, so not 1996.
09:57:00 <andythenorth> then we can make SteveObjects
09:59:33 <andythenorth> "there is an Imp here would you like to (1) run (2) shoot the imp"
10:01:09 <andythenorth> we could black out parts of the map
10:01:13 <andythenorth> and the player has to explore them with Steve
10:01:23 <andythenorth> but Steve needs food from a base
10:05:03 <talltyler> peter1138[d]: What if you called it “waiting” to match how it’s described on the main industry window?
10:14:09 <xarick> waiting sounds good to me
10:26:33 <peter1138[d]> Town population and town buildings do not belong on the same chart, or at least not the same y-axis.
10:28:56 <xarick> SwapList is a faster operation than AddList
10:31:48 <xarick> now, the matter of the optional parameter... I honestly don't know
10:31:58 <xarick> part of me says it should be handled
10:35:35 <xarick> I think some AIs keep using their old API versions together with the newer API functions
11:21:00 <peter1138[d]> Eddi's been missing for a year.
11:26:21 <andythenorth> wonder if he's in the German forum?
11:37:57 <peter1138[d]> No, missing for the same there too.
11:41:34 <peter1138[d]> Now changed to just "Waiting"
11:51:38 <andythenorth> people going MIA in online communities 😐
11:52:13 <andythenorth> if I get dead, some of you probably can figure out how to check 😛
12:09:32 <peter1138[d]> Who will finish refactoring Horse?
12:16:26 <xarick> can you turn a tileset to using stockpiled industries
12:17:30 <andythenorth> peter1138[d]: yes
12:17:41 *** foodliker has joined #openttd
12:17:56 <andythenorth> no-one ever will
12:18:00 <andythenorth> it will just....end
12:21:14 <foodliker> you won't allow yourself to die until youre finally done fiddling around making sprites perfect
12:21:18 <foodliker> which will be never
12:21:28 <foodliker> the eternal andythenorth
12:22:30 <peter1138[d]> peter1138[d]: (Funnily enough, at least I was amused, is that internally I called the field "waiting" all along...)
12:43:30 *** akimoto has joined #openttd
12:45:57 <_glx_> xarick: Their problem, if they set version 12 they should get the behaviour of version 12
13:18:20 <jfs> If you declare a compact level, then ideally any functions added after that version are also not available
13:29:07 <_zephyris> So, default badgers.
13:29:16 <_zephyris> How's NML support for badgers?
13:30:33 <_zephyris> I'd like to add the extra-zoom ones to ogfx2 (IMO, base extra newgrf should be 1x, to match original graphics).
13:40:55 <pickpacket> _zephyris: there's a NewGRF that adds big cats on the map. Could do the same with badgers
13:42:20 <pickpacket> peter1138[d]: Eddie's been gone for a whole year? O.o Damn, time flies...
13:43:54 <peter1138[d]> jfs: That isn't the case, and enforcing that now would break compatibility. So we'd need to only apply it to new functions, and make it depend on compatibility version. It starts to get messy...
13:45:13 <_glx_> We just assume scripts using older API are old and won't use stuff added in more recent API
14:19:19 *** akimoto has quit IRC (Remote host closed the connection)
14:21:03 *** Wormnest has joined #openttd
15:58:40 <andythenorth> _zephyris: it's in a PR
15:58:47 <andythenorth> but appears to be complete for me
16:14:03 <peter1138[d]> Probably needs a rebase to get rid of the fix commits.
16:14:23 <peter1138[d]> Oh, and sorting out frosch's comments.
16:20:59 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
16:30:30 <pickpacket> Hmm. I'm trying to understand how to write a Save function in my GS. I have the MinimalGS base and the comment for the function says "[...] return a table which can contain nested tables, arrays [...] Class instances [...] cannot be stored by OpenTTD."
16:30:59 <_glx_> some class instances can
16:31:22 <pickpacket> An instance is pretty much a table... And in my case I have class instances containing lists of class instances
16:31:47 <_glx_> but an instance has underlying c++ object
16:32:16 <_glx_> at least for openttd provided objects
16:33:32 <pickpacket> hmm. Alright I have a class ExampleItem which contains a few keys with values, and a class ExampleStorage which contains a table where each value is an ExampleItem and a list where each element is an ExampleItem
16:34:12 <pickpacket> Or do I have to convert it into tables with tables or lists with tables, etc?
16:35:48 <pickpacket> my script is dependent on other things that are new in 15. Backwards compatibility is a non-issue
16:38:55 <andythenorth> you have to serialise / deserialise
16:38:59 <andythenorth> or don't use classes
16:39:12 <andythenorth> I keep all persistent state in a persistent storage manager (table)
16:39:37 <andythenorth> think I just do one root table, easier to serialise/deserialise
16:40:38 <andythenorth> looks like I wrote a storage manager class that handles writing to a root table
16:41:26 <andythenorth> then save / load uses that table
16:41:58 <andythenorth> in some respects I wish OpenTTD handled more of this, but leaving it to authors is more flexible
16:43:47 <_glx_> indeed, if you wrote the class you can decide how to save/load it
16:45:06 <peter1138[d]> SaveloadDescriptions? 😄
16:45:15 <pickpacket> yeah, I guess serialisation/deserialisation is the way to go
16:45:25 <_glx_> ScriptObject save/load handling is super low level, it happens during parsing of the table returned by Save() call
16:46:14 <pickpacket> hehe. Tempted to just not do classes now :P
16:47:55 *** scrubbles3587 has joined #openttd
16:48:12 <_glx_> might be possible to add support for SaveObject() for squirrel instances, but I'm not sure about how loading would work
16:48:18 <pickpacket> Do I need to declare a "class Something" in order to have "function Something::Foo()" ?
16:49:05 <_glx_> Something.foo <- function() ...
16:49:31 <_glx_> squirrel happily supports extending classes from outside the declaration
16:50:05 <_glx_> but you can't do it once an instance has been created
16:50:25 <_glx_> compat scripts do that all the time 🙂
16:51:02 <_glx_> copy class member into a new member, redefine old member
16:51:55 <pickpacket> yeah, but can I make a library with functions rather than a class with methods?
16:52:26 <andythenorth> scrubbles3587: not enough smoke detectors on that TGV though
16:53:11 <andythenorth> pickpacket: it's squirrel, you can do it however your want
16:53:13 <andythenorth> all rules are off
16:53:21 <andythenorth> either the language doesn't do it, or you do what you want
16:54:45 <andythenorth> I can't find any evidence that you can just declare functions though
16:55:17 <andythenorth> I don't know what the scope would be
16:55:37 <andythenorth> do you need to do saveload with classes?
16:56:22 <pickpacket> andythenorth: yeah
16:56:41 <pickpacket> but I figure that maybe I should just not use classes
16:56:53 <pickpacket> instead of having to deal with serialisation
16:57:26 <andythenorth> it's not wildly hard to treat it like a database, assuming your classes are only singletons
16:57:48 <andythenorth> `::persistent_storage[foo]` vs `_this.foo`
16:57:57 <andythenorth> or something like that, I forgot all the squirrel
16:58:08 <andythenorth> treat it like an external datastore
16:58:29 <andythenorth> if you have multiple instances then...yeah
17:04:58 <_glx_> hmm actually I may have an idea to also support save/load of script provided instances
17:06:01 <andythenorth> isn't there some issue where the object hash might change across platforms / client versions?
17:06:07 <andythenorth> might be mixing that up with something else
17:17:29 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
17:18:45 *** WormnestAndroid has joined #openttd
17:20:51 <_glx_> classes could implement an optional `SaveObject()` function returning a 2 element array [name of the class to construct it on load, data to use in LoadObject()], and a `LoadObject(data)` function
17:21:50 <_glx_> then it's just a matter of plugging that in the existing handler
17:23:02 <andythenorth> would they automatically get suscribed to save, if that SaveObject() method exists?
17:23:10 <andythenorth> or would all instances have to be walked, calling save?
17:24:17 <_glx_> the idea is to check method exists, if not try the native (which most likely will fail and return `You tried to save an unsupported type. No data saved.`)
17:24:37 <andythenorth> ok so OpenTTD would try this?
17:24:45 <andythenorth> it's not 100% squirrel author side?
17:25:28 <_glx_> SaveObject() is on author side
17:25:51 *** WormnestAndroid has quit IRC (Remote host closed the connection)
17:25:58 <andythenorth> it wouldn't be very hard to do the whole thing author side
17:26:13 <andythenorth> instances could register into some root table of persistable classes
17:26:23 <andythenorth> then just walk that on save
17:26:42 <andythenorth> I have got used to doing silly things in squirrel 😛
17:27:10 <_glx_> if you pass an instance in the to be saved array it has to implement SaveObject()
17:28:01 <_glx_> so still "manual" serialisation, but called automatically
17:28:39 <_glx_> with benefit of the instance being automatically recreated on load
17:30:27 *** WormnestAndroid has joined #openttd
17:40:15 <pickpacket> I don't understand the syntax in the squirrel docs... this is apparently a function declaration: "stat:= 'function' id ['::' id]+ '(' args ')'[':' '(' args ')'] stat"
17:46:46 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
17:48:07 *** WormnestAndroid has joined #openttd
17:57:17 <LordAro> pickpacket: grammar description
17:57:40 <_glx_> the `'[':' '(' args ')']` is for free variables (kinda like c++ lamba captures)
17:59:41 <_glx_> hmm I can see a possible issue with my `SaveObject()` idea, infinite recursion
18:22:36 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
18:30:30 *** WormnestAndroid has joined #openttd
18:37:31 <pickpacket> can things like GSStoryPage and their contents be saved as is?
18:47:14 <peter1138[d]> Story Pages are part of the game state.
19:02:44 <_glx_> hmm actually infinite recursion will probably trigger max depth before becoming an issue
19:33:36 <peter1138[d]> Hmm, didn't expect such swings
19:34:39 <talltyler> What is that a graph of? I can see it's a town, and know the colours are passengers and mail, but is it transported? Delivered?
19:34:53 <peter1138[d]> It's actually stations.
19:35:14 <talltyler> Amount waiting at a station?
19:35:14 <peter1138[d]> It's the amount waiting to be loaded.
19:35:32 <talltyler> Looks like you need some unbunching 😛
19:36:44 <talltyler> Lots of trains? Throughput looks like 2000 passengers per month, not accounting for any cargo disappearances due to low station rating.
19:37:47 <peter1138[d]> Maybe I can graph cargo lost.
19:38:04 <talltyler> A graph of station rating 😛
19:38:09 <talltyler> (probably overkill)
19:38:21 <talltyler> Also different units, of course
19:38:44 <peter1138[d]> (Yes, the y-axis is wrong when they're all enabled)
19:38:46 <talltyler> Oh, that's interesting
19:39:40 <talltyler> What are supply and demand in this context? Supply = cargo produced by houses and industry tiles in the catchment area, regardless of whether it actually moves to the station?
19:41:04 <peter1138[d]> Supply is that figure.
19:41:31 <peter1138[d]> Demand is something that shouldn't be graphed I think.
19:41:37 <peter1138[d]> They're both linkgraph related, so unsure.
19:41:39 <talltyler> I have honestly never looked at that number 🙂
19:41:51 <peter1138[d]> I don't remember it being there.
19:41:52 <talltyler> Is that before or after station rating?
19:42:01 <peter1138[d]> Which means it was probably added 15 years ago.
19:42:39 <talltyler> Hmm, I assume you're in calendar mode? I'm sure I did find-and-replace for wallclock mode, even if I've never paid the number any thought 😄
19:43:07 <peter1138[d]> It says Supply per minute in wallclock mode, so...
19:43:19 <peter1138[d]> This is just an old save.
19:43:51 <talltyler> I never knew I'd be so happy to see OpenTTD 15 be the Statistics Update 😛
19:44:18 <talltyler> (legitimately, this looks like useful info I never knew I cared about when it wasn't easily knowable)
19:46:27 <jfkuayue> peter1138[d]: Famous person party game?
19:50:39 <_glx_> ah yes the test game I first tried to join with a debug build
19:50:56 <peter1138[d]> Yeah, demand seems to be something quite internal to the linkgraph.
19:51:07 <peter1138[d]> Replace with lost cargo, I guess.
19:52:05 <talltyler> Something something cargodest
19:52:23 <peter1138[d]> That's something someone else was looking at...
19:52:47 <talltyler> Lost cargo could be helpful, although I wonder how much it's an impossible goal to chase because 100% station rating is almost impossible
19:53:36 <peter1138[d]> talltyler: It would be more of a taunt, wouldn't it?
19:53:59 <talltyler> I suppose, might bother the OpenTTDCoop folks 😛
19:54:56 <talltyler> My brain is mush, $real_job is in crunch time 😦
19:54:57 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
19:55:46 *** WormnestAndroid has joined #openttd
19:56:29 <locosage> 100% is feasible with regular advertisement
19:58:36 <locosage> lost cargo would be an interesting stat to have as cargo loss mechanics are well-hidden normally
20:10:18 <peter1138[d]> It does correlate with the rating going bad, although that really needs a second y-axis 😦
20:12:48 <locosage> it's more useful than rating
20:12:57 <locosage> rating is usually more or less stable
20:13:09 <locosage> but cargo loss have some hidden thresholds
20:13:37 <locosage> like 4096 units wating
20:15:28 <peter1138[d]> Wondering about loaded and unloaded data.
20:16:14 <peter1138[d]> "Time since last vehicle"
20:16:32 <peter1138[d]> (As it's monthly, that's not all that useful)
20:16:45 <locosage> loaded/unloaded is a curious stat but not particularly important imo
20:17:11 <locosage> time since last vehicle is basically debug info for station rating xd
20:18:07 <peter1138[d]> Yeah, rating either needs to be a separate chart, or a separate y-axis...
20:18:59 <locosage> separate axis is better for showing correlation
20:19:29 <peter1138[d]> Okay, I probably need to handle aircraft crashes separately.
20:23:25 <locosage> btw, would be nice to be able to add charts from gs
20:23:32 <locosage> at the very least for towns in city-builder
20:24:02 <peter1138[d]> Yes, it would be.
20:25:12 <peter1138[d]> Potential issues with updating the data.
20:26:52 <peter1138[d]> Assuming the data is date related then it should automatically rotate every month. But it needs the GS to issue commands to fill in the data.
20:28:21 <locosage> yeah, gs is a bit wonky by design
20:28:31 <locosage> but it mostly works ok so maybe it's fine xD
20:29:06 <locosage> it can rotate months on command from gs
20:29:46 <peter1138[d]> Rotate on command would avoid data being missed, or being in the wrong month. But that means the months might not actually be months, and might not align with all the other graphs.
20:30:53 <locosage> yep, and wallclock adds more more complexity too
20:31:22 <locosage> but unless gs is overloaded it shouldn't be a noticeable problem
20:32:37 <peter1138[d]> hmm, 1) get data for graph 2) send data for graph for specific date 3) profit.
20:33:10 <peter1138[d]> get *date* for graph, that should be.
20:33:45 <peter1138[d]> So rather than a command that assumes "now", it knows which current or previous slot should be filled.
20:34:47 <peter1138[d]> Being able to read graphs?
20:35:05 <peter1138[d]> I think there's some current stats that can be read, but not historical data.
20:35:22 <locosage> get date is surely more reliable but it'd say just spamming data monthly is good enough
20:35:56 <peter1138[d]> "good enough" often isn't.
20:36:53 <peter1138[d]> Also depends it how the command looks... e,g, sending one command to do lots of updates vs loads of commands..
20:38:11 <locosage> reading graph history is potentially useful to avoid saving data twice
20:39:00 <locosage> but graph data is usually tailored for better presentation so may not be useful for anything else
20:49:23 *** k-man has quit IRC (Ping timeout: 480 seconds)
21:01:34 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:16:04 <peter1138[d]> Hmm, do I have any... big but not slow game...
21:20:37 <peter1138[d]> Not sure Supply should be included.
21:20:56 <peter1138[d]> It's kinda the same as an industry's Transported, but not quite. And a bit of lag.
21:23:14 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:25:10 <locosage> it's nice to have with multiple industries
21:25:25 <locosage> also stations are few so extra stat doesn't cost much
22:01:05 *** ChanServ sets mode: +v tokai
22:06:06 <_glx_> indeed infinite recursion is already handled 🙂
22:08:04 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
22:52:51 <peter1138[d]> Hmm, interesting with high waiting station.
22:54:15 <peter1138[d]> Like constantly losing ~1000 units every month.
23:04:55 <locosage> well, it's supposed to lose a certain % of cargo above 4k units so seems consistent
23:10:54 <peter1138[d]> Yes. It's a station with lots of vehicles waiting, and cargo is going up and down.
23:11:07 <peter1138[d]> So you might not actually notice it's being lost.
23:12:33 <peter1138[d]> Feels like maybe it's revealing too much.
23:30:54 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
23:31:41 *** WormnestAndroid has joined #openttd
23:52:17 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
23:56:46 *** WormnestAndroid has joined #openttd
continue to next day ⏵
