Fluid Unit Hell | The NeoForged Project | Page 4

astral stag Jun 20, 2024, 7:04 AM

#

I'm aware of the issues with the use of millibuckets but we have to deal with them for now. No proposal that keeps mb will solve that problem. Bringing it up again is pointless.

pure lynx Jun 20, 2024, 7:15 AM

#

if you dont need the api thats fine, the rest of us think its useful and it doesnt harm you in any way, so let us implement it

vernal bane Jun 20, 2024, 7:59 AM

#

I do not

deep hamlet Jun 20, 2024, 11:08 AM

#

No, I don't interact with any arbitrary fluid unless a pack maker also defines it via datapack (which also requires a melting temperature, specific heat capacity, etc.) So the fact a common tag exists doesn't help.

fluid mortar Jun 20, 2024, 2:30 PM

#

alright, so for your case at most a tag that is "molten iron at the 100mb standard" would be useful

#

I think thats probably the simpliest way to handle these standards, have prefixes that mean "100mb", "90mb" and "250mb" for the common standards (and 144mb if people insist on reviving the bad legacy standard). Then you just use the appropitate prefix for what you do

#

We could do that alongside the tooltip, and people wishing to do conversions between standards just use multiple tags

#

Only awkward part here is coming up with names for each standard. I've been calling 90mb "metals", which is a bit weird if you apply it to redstone for instance but not the worst thing

#

I tossed in a section about unit standard tags to the fluid tooltip proposal, a fluid constant PR could reference those tag prefixes for relevant constants (e.g. INGOT would mention "tag fluids using this unit under c:metals/<fluid name>")

solemn sand Jun 20, 2024, 2:55 PM

#

astral stag I'm aware of the issues with the use of millibuckets but we have to deal with th...

The fluid unit API does solve it though -- not for buckets, but for the case of ingots and all other units

#

It means that you don't need to worry about an ingot "standard" size that's divisible by whatever people want -- you can just have each person set that themselves

#

And in the long run -- looking at versions further down the road -- it could solve the millibucket issue too

fluid mortar Jun 20, 2024, 2:57 PM

#

yeah, its really just a question of how we want to solve this

#

everything the fluid unit API proposes apart from tooltip formatting could be done via other methods (tag conventions and mod added datamaps), however it is a lot more streamlined and potentially eaiser for pack devs to work with

#

I personally am not too partial to either approach, as long as we make no changes to the bucket divisions before we get some unit display in Neo

#

With either approach we can change the bucket divisions with minimal impact on players

pure lynx Jun 20, 2024, 5:10 PM

#

fluid mortar everything the fluid unit API proposes apart from tooltip formatting could be do...

Idk, I don’t think the “document every standard” approach is helpful. I mean it’s nice, but if you’re using anything other than the base standard it’s more helpful to use fluid units than doing special casing for each standard we write here

fluid mortar Jun 20, 2024, 6:22 PM

#

that is true, its easier to not have to special case each standard

#

it all comes down to how much do we expect conflicting standards to occur?

#

if its just 1 or 2 standards for a thing, its not a huge deal

#

if its regularly 2-4 standads, that becomes a pretty big maintenance burden on the modders side

#

The standard tags are arguably useful whether the unit API is added or not, since they grant a quick and easy way to detect fluids of a particular standard, though I'd call them non-essential with the unit API

#

without the unit API they are also arguably non-essential, though given we have some conflicting standards already its probably better to have them than not

pure lynx Jun 20, 2024, 6:33 PM

#

I think so far we have 3 reasonable standards, 90 mb, 100 mb, and 144 mb

kind stratus Jun 20, 2024, 6:46 PM

#

we don't have to have a completely separate set of tags for the standards

#

like, we could have recipes that are based on the intersection of c:molten_iron and c:ingot_is_90 [if we don't have intersection holdersets, or holderset-based ingredients, we should]

fluid mortar Jun 20, 2024, 7:43 PM

#

pure lynx I think so far we have 3 reasonable standards, 90 mb, 100 mb, and 144 mb

I'd argue 144mb is not reasonable 😛

#

250mb is the third reasonable standard

#

its mainly used for non-metals

fluid mortar Jun 20, 2024, 7:44 PM

#

kind stratus like, we could have recipes that are based on the intersection of `c:molten_iron...

thats a fair point, had not considered that

#

Neo should have intersection fluid ingredients if it does not already

pure lynx Jun 20, 2024, 8:24 PM

#

it does as of a few weeks ago

fluid mortar Jun 20, 2024, 8:58 PM

#

perfect, that solves that

#

Tag standard proposal:

c:metals/<name> or c:gems/<name> would be the standard place for metals and gems alike. What qualfies as either is the same semantics as using the item tags c:ingots/ and c:gems/ respectively
Not sure what to do with weird fluids like redstone or clay that are neither, probably just tag under root c:redstone or c:clay, though a generic c:molten/<name> might be an option
For units, we use c:units/<number> where <number> is the size of the most standard. e.g. the Tinkers ingot standard would use c:units/90, and the gem standard is under c:units/100
May be worth retroactively applying this standard to existing Neo fluid tags, e.g. c:honey and c:potion would typically use c:units/250 but you could if you wish use another standard.

#

This solves the problem of conflicting units by just telling people to publically expose which standard they are using in a way other mods can query

#

Not sure if the c:units/ proposal should be mutually exclusive with the unit API proposal. I think its probably still beneficial if that API exists as you may have a context where tags are way easier than querying the unit API

pure lynx Jun 20, 2024, 9:40 PM

#

is c:units/90 just assumed nugget, thing, and block being 10, 90, 810?

fluid mortar Jun 20, 2024, 9:41 PM

#

there is an assumption here the standard is named after the ingot, not the nugget

#

and since nugget/ingot/block is 1/9/81, you just multiply/divide the fluid unit as needed

#

non-9 ingots will break as much here as they will in item tags

#

plus, if you are using non-9, you probably don't need the 90mb standard, 4 is way better in the 250mb standard

pure lynx Jun 20, 2024, 9:49 PM

#

maybe it should be

c:unit/ninth-ninth 10 : 90 : 810
c:unit/ninth-quarter 25 : 100 : 900
c:unit/quarter 250 : 1000
c:unit/third 333 : 1000 ? this one might just be unsupported for now

solemn sand Jun 21, 2024, 1:51 AM

#

fluid mortar everything the fluid unit API proposes apart from tooltip formatting could be do...

Yeah, I think the fluid unit API does it better though, cause it doesn't force different fluids to have the same sizes for a given unit -- whereas tags, whatever end up requiring mods to do that

solemn sand Jun 21, 2024, 1:53 AM

#

pure lynx maybe it should be `c:unit/ninth-ninth` 10 : 90 : 810 `c:unit/ninth-quarter` 25...

I feel like it'd almost be better to make an ingredient type that could work with the fluid unit system instead of a bunch of wacky tags like this

#

...though honestly I don't think people should be designing recipes that way to begin with. Like, your recipe shouldn't care that a fluid has 90-ingots

#

You should only care that that fluid has, say, an ingot unit and a nugget unit (and the one is 9 times the other)

pure lynx Jun 21, 2024, 1:54 AM

#

it would be better yes

solemn sand Jun 21, 2024, 1:54 AM

#

In what scenario do you actually need to know that it's a 90-ingot-fluid?

#

And how can we design an ingredient type that lets you represent that actual condition (fluid with ingot and nugget units) instead?

#

Could we have the c:ingot unit somehow define an implicit c:unit/has_ingot tag on the fluid or something?

#

And then "fluids with ingots and nuggets" is just a tag intersection

#

I'd have to think about hte practical end of that a bit more

pure lynx Jun 21, 2024, 1:55 AM

#

oh, well no, recipes still need to take in a specific liquid in order to get an outcome generally speaking

#

very rare that you need a recipe that takes all ingots

#

the recipe could take in a tag and then filter out all the fluids in the tag that dont have the particular unit

solemn sand Jun 21, 2024, 1:57 AM

#

Sure, guess I'm saying I don't really see the advantage of a ingot_90 tag. I'd argue we should be encouraging people not to rely on how "big" a given unit is for a given fluid, or hardcode stuff to only work on a certain unit size. People should instead only care which units something has (and that the ratios work as the mod expects I suppose)

solemn sand Jun 21, 2024, 1:57 AM

#

pure lynx the recipe could take in a tag and then filter out all the fluids in the tag tha...

Yeah, this makes sense

#

I want molten copper, but only if its molten copper with an ingot unit

#

etc

pure lynx Jun 21, 2024, 1:57 AM

#

i think the idea would be to use an intersection tag for that. All fluids that are both molten_iron and ingot_90, which the current fluid ingredient system supports

solemn sand Jun 21, 2024, 1:58 AM

#

pure lynx i think the idea would be to use an intersection tag for that. All fluids that a...

Right but that's the issue. People shouldn't be relying on it having an ingot unit with value 90, they should only care that it has an ingot unit

#

If you're filtering only fluids where a unit is equal to a certain value, and you have the fluid unit API available, you're doing something wrong imo

pure lynx Jun 21, 2024, 1:59 AM

#

agreed. honestly the tag stuff does seem very annoying and i dont really plan on using it for anywhere other than just writing the 1 json it takes to tell mods its in that tag

#

fluid units are more ergonomic in basically every way but people are convinced that we dont need them, not really sure what else to say on that front :/

solemn sand Jun 21, 2024, 1:59 AM

#

I don't know that people are convinced on that

#

Some people sure but I've also seen a lot of interest for it in this thread

#

Heck, I wasn't fully in favor to begin with but I think I've come around since. Just seems like the cleanest long-term solution

pure lynx Jun 21, 2024, 2:01 AM

#

yeah, i guess the question is if this PR got created what would the maintainers do with it. cuz tech is still convinced we dont need it. but shadows seems to be good with it. I think the only real concern is that the existence of this system would disincentivize people from using the tag standards

solemn sand Jun 21, 2024, 2:02 AM

#

Someone should make the PR as a draft at least just so there's something solid to work with

solemn sand Jun 21, 2024, 2:02 AM

#

pure lynx yeah, i guess the question is if this PR got created what would the maintainers ...

I mean yeah, I think it ought to. I think the idea of an ingot_90 tag is terrible just by its nature

#

Cause then you encourage people to filter things based on a units value... which you really want people not to do

#

That said, c:units/ingot tag -- for fluids with an c:ingot unit definition (of any value) -- isn't a bad idea

#

The question is, is it possible to make such a tag automagically contain any fluids that have the c:ingot unit?

pure lynx Jun 21, 2024, 2:04 AM

#

there is merit to having both imo. if you only want to work with things in 1 specific standard the tag is nice. the unit system currently doesnt really allow that. actually i guess you could do something like fluid.getUnit("ingot") == 90

solemn sand Jun 21, 2024, 2:05 AM

#

pure lynx there is merit to having both imo. if you only want to work with things in 1 spe...

My point is nobody should only ever want to work with things in 1 specific standard. Like, what's a case where you'd actually want that, when what you want isn't instead something about having a specific set of units and ratios between them? Saying "I want only fluids with with 90 units per ingot" is really an XY problem -- nobody ever wants that really, they want fluids where they can do certain unit conversions and have them work out, or grab fluid in certain unit sizes, or have certain guarantees about ratios of unit sizes or whatever.

#

Like, why would you want an ingot_90? Maybe it's that you want:

ingot unit exists
nugget unit exists
9 nuggets to an ingot
bucket to be divisible by nugget
And you should be checking that set of conditions instead, not just saying "I want things where ingot is 90"

#

Because you don't really want things where ingot is 90, you want those other conditions because something you're doing relies on those assumptions. A fluid where ingot was 180 would still work just as well in your system in that scenario, for instance.

#

So the question is: how do we design a fluid unit API that makes checking those conditions -- ideally from an ingredient or holderset thingy -- easy? Because that's the sort of stuff people actually need, not ingot_90 fluids

#

Custom holderset type maybe?

pure lynx Jun 21, 2024, 2:11 AM

#

(currently looking up what the hell a holderset type is)

#

forge has custom holder sets?

solemn sand Jun 21, 2024, 2:11 AM

#

pure lynx (currently looking up what the hell a holderset type is)

Basically: vanilla has tags. Neo lets you register custom things like intersections/unions of holdersets (tags in vanilla)

#

I think you could extend such a system with some sort of "fluid unit holder set" for filtering, say, all fluids with a given unit defined or something? Haven't poked the system enough, could be wrong

pure lynx Jun 21, 2024, 2:12 AM

#

no idea, i also need to take a look at this

#

i have not been in the "neo exclusive" game for very long lol. i have never really taken a hard look at neo's apis until like a month or 2 ago

solemn sand Jun 21, 2024, 2:13 AM

#

Regardless, I think that sort of solution is the proper approach here if its possible -- having tags for different standards or whatever just, like, bakes in those standards -- when in the long run I think we should be getting rid of a lot of those standards, and in the long run hoping to make even the size of the bucket unit fluid-specific

#

This solution solves all the problems those standards are needed to solve, without the sort of limitations that they drag with them

fluid mortar Jun 21, 2024, 3:27 AM

#

the reason why you need ot know the standard is one of two reasons

#

either a: you want to know for divisibility sake or b: you want to know how much fluid is needed for a recipe

#

both are handled by a unit API of course

#

so I think based on this discussion, I'll conclude the tag standards are only needed if we don't do a unit API

#

and tbh, maybe they are what we need to convince people a unit API is good, because otherwise they will have to do that nonsense 🙂

kind stratus Jun 21, 2024, 4:14 AM

#

if we're gonna have units tags, they should probably say ingot, gem, bottle [though again that raises the question of what a unit of redstone is]

#

also, is a tag that's just a number legal?

kind stratus Jun 21, 2024, 4:15 AM

#

solemn sand The question is, is it possible to make such a tag automagically contain any flu...

i think there's been discussion about automagically populated tags in the past and the answer has been no

#

anyway, if we have a unit quantity based ingredient it can simply not match fluids that don't have the unit, if we want to check it in code we could add a .hasUnit helper method to Fluid and/or FluidStack

pure lynx Jun 21, 2024, 4:18 AM

#

or fluid resouce hehehehe

solemn sand Jun 21, 2024, 4:19 AM

#

kind stratus i think there's been discussion about automagically populated tags in the past a...

I wouldn't do it as unit tags honestly -- I'd try and make a new holderset type that grabs fluids with certain units

#

Then there's no magical tag generation

#

But yeah, regardless -- I'd just avoid the tags for units

#

Either they're redundant and a source of error (if the tag is just the same as checking if the fluid has the unit) or they include numbers which we should avoid, as you shouldn't be reasoning/filtering based off of a unit number (and a unit API would mean you don't need to)

kind stratus Jun 21, 2024, 4:21 AM

#

there is maybe something to be said for something like "is it divisible by 4, recipe quantity is 3/4 unit" but that's niche enough that it can probably be left to whoever needs it

solemn sand Jun 21, 2024, 4:22 AM

#

kind stratus there is *maybe* something to be said for something like "is it divisible by 4, ...

Sure -- and even then, that's a lot different than "exactly 90" or the like

kind stratus Jun 21, 2024, 4:22 AM

#

[i have been tempted to suggest a floating point unit quantity, but i don't actually think it's a good idea, and honestly it should really be rational, that's just ugly to try to implement in json]

fluid mortar Jun 21, 2024, 4:46 AM

#

floating point has terrible percision at the points we are trying to get by changing bucket standards

#

thta is, 3, 9, 81, etc.

#

they only wnat powers of 2 for precise values

grizzled tulip Jun 21, 2024, 6:47 AM

#

floating point precision has a much bigger issue than rounding: digits of precision.
float has 6-7 digits. that means you can't have a tank with 100 million buckets because if you extract one bucket at a time you would either lose additional fluid (if it rounds up) or duplicate fluid (if it rounds down)

kind stratus Jun 21, 2024, 6:49 AM

#

fluid mortar floating point has terrible percision at the points we are trying to get by chan...

i meant for ingredient notation ("i want 1.5 ingots"), not actual storage, but fair enough

compact lynx Jun 21, 2024, 11:52 AM

#

https://factorio.com/blog/post/fff-416

Friday Facts #416 - Fluids 2.0

Hello,
Grab your best lube, because it's time to talk about fluids!

During one of the fluids discussions, Rseding proposed an algorithm very close to one he had played with several times, an algorithm from the Minecraft mod Thermal Expansion by team CoFH:

how topical kekw

weak blade Jun 21, 2024, 11:54 AM

#

outer spade Jun 21, 2024, 1:13 PM

#

obvious issue when gamers have different opinions how realistic a game should be^^... and minecraft is as far away as you can get while some mods try to relive the real world industrialisation kek

still iris Jun 22, 2024, 11:25 PM

#

that's fun lol

desert folio Jun 24, 2024, 5:51 PM

#

oh god I missed this sidebar somehow

#

how many other terrible ideas are brewing that I'm unaware of

echo lily Jun 24, 2024, 5:51 PM

#

5, if I'm counting correctly kek

prisma skiff Jun 24, 2024, 6:50 PM

#

well I have 4 extra ideas....

#

but those are for only a few to know

fluid mortar Jun 26, 2024, 12:38 AM

#

desert folio how many other terrible ideas are brewing that I'm unaware of

The main idea brewing here is a fluid unit API

#

Either purely for rendering fluid units in tooltips (similar to what tinkers does ) or extending the idea to work with logical fluid units

#

Logical fluid units means you could for instance make an item that holds 1 ingot, but 1 ingot may have different mb on different fluids

#

Or make an ingredient that matches 1 gem and not care whether a gem is 100mb or 90mb for that fluid

#

I think it's not worth changing the bucket capacity without at minimum having a fluid unit display API, so I've been in pushing for that being done before we further bikeshed the bucket divisions

#

There is more info in the pins if you want

compact lynx Jul 5, 2024, 2:00 AM

#

hello fluid folks
just an FYI that I've made a comment on https://github.com/neoforged/NeoForge/pull/1138 (Define standard FluidConstants) proposing that PR to go forward if this discussion leads to nowhere concrete -- an issue or PR, as I've somewhat alluded to in my pinned message
just so we have something in place to discuss concretely

pure lynx Jul 5, 2024, 2:21 AM

#

alright. well honestly i dont really know where we stand on this issue

#

i understand there was still some aprehension from tech, and orion insists on a PR being written before he can decide. idk where everyone else is on this

kind stratus Jul 5, 2024, 7:01 AM

#

i'm not sure i understand what the blocker is - is it just that there's not a PR, or are there still serious objections to the proposal

drifting wraith Jul 5, 2024, 9:46 AM

#

a PR might be needed to actually arrive at some definite conclusion, it looks like.

kind stratus Jul 6, 2024, 10:07 PM

#

i'm tired of looking at or thinking about openblocks for a while, i might put together a PR

compact lynx Jul 6, 2024, 10:07 PM

#

the phrase "out of the frying pan and into the fire" comes into mind /jk

tawdry wolf Jul 7, 2024, 1:31 PM

#

aiui a PR is being requested for further discussion

#

But even if no PR arises, I don't think that #1138 should be merged, it will introduce more problems than it fixes.

kind stratus Jul 7, 2024, 7:12 PM

#

ok i think i might have some code ready to PR, though I didn't do any ingredient stuff [i don't think we have fluid ingredient stuff at all yet do we?]

#

i do not have a good way to test this though, there's nothing in vanilla or neoforge itself that actually uses fluid stacks

pure lynx Jul 7, 2024, 7:45 PM

#

the fluid ingredient PR was merged a while ago

kind stratus Jul 7, 2024, 7:54 PM

#

ah... maybe i should have pulled before starting work on this

#

ok how do i make a codec that does this "maybe a list or object, maybe just an integer" stuff

tawdry wolf Jul 7, 2024, 8:09 PM

#

should be an either codec?

kind stratus Jul 7, 2024, 8:19 PM

#

how does that work

#

what i want is a codec for a single type with two json representations, not a codec that can represent two types of object

#

maybe i should just make a straightforward codec and let it be ugly, [[{"unit": "mb", "amount", 10}]] isn't that bad.

tawdry wolf Jul 7, 2024, 8:29 PM

#

aiui what you want is a codec that allows both

10
```and
```json
{
  "amount": 10,
  "unit": "mb"
}
```Is that correct?

kind stratus Jul 7, 2024, 8:30 PM

#

technically i want all of 10, {"amount": 10, "unit": "mb"}, [{"amount": 10, "unit": "mb"}], and [[{"amount": 10, "unit": "mb"}]]

echo lily Jul 7, 2024, 8:31 PM

#

conecern why double list?

tawdry wolf Jul 7, 2024, 8:31 PM

#

yeah, what are the lists doing there?

kind stratus Jul 7, 2024, 8:31 PM

#

the inner lists are sums, the outer list is alternatives

#

so e.g. you can have "2 ingots and 1 nugget, or 2 gems"

tawdry wolf Jul 7, 2024, 8:32 PM

#

please for the love of god, make that "fluid_any_of" and "fluid_all_of" (or similar naming)

#

as separate ingredient types

#

(I'm assuming this is about fluid unit ingredients)

kind stratus Jul 7, 2024, 8:32 PM

#

ok it still at least needs the sum list for like "1 ingot and 3 nuggets", that's a single quantity

tawdry wolf Jul 7, 2024, 8:33 PM

#

okay, a codec that is either a single element or a list thereof is pretty simple, hold on

echo lily Jul 7, 2024, 8:34 PM

#

Codec.either(Thing.CODEC, Thing.CODEC.listOf())

#

and wrap that in another either if you want the single int too

kind stratus Jul 7, 2024, 8:34 PM

#

how do i map that to "the single thing becomes a list of a single element at runtime" though?

echo lily Jul 7, 2024, 8:37 PM

#

.xmap(either -> either.map(List::of, Function.identity()), Either::right)

kind stratus Jul 7, 2024, 8:37 PM

#

do i just have to do it by hand with xmap

tawdry wolf Jul 7, 2024, 8:37 PM

#

there's Codec.withAlternative(), which is just a wrapper over xmap

#

but otherwise, yeah what Platin suggested

echo lily Jul 7, 2024, 8:39 PM

#

yeah ok, that would make it a bit nicer, since it would just be Codec.withAlternative(Thing.CODEC.listOf(), Thing.CODEC, List::of)

tawdry wolf Jul 7, 2024, 8:39 PM

#

yeah

#

I also find that a bit nicer to read: "expect Thing.CODEC.listOf(), if it's just a Thing.CODEC then that's fine too, we'll just wrap that in List.of() and be done with it"

echo lily Jul 7, 2024, 8:44 PM

#

it does turn everything into list on encode, but I doubt that matters here

tawdry wolf Jul 7, 2024, 8:44 PM

#

yeah I figure that doesn't matter

kind stratus Jul 7, 2024, 8:45 PM

#

is there any better alternative to .filter(Optional::isPresent).map(Optional::get)?

#

hmm technically the inner lists should be maps or sets

#

probably maps, they'll look nicer as json maps too

#

is there an open map codec that returns an ordered map? the ordering would be nice to keep for the toString, but isn't critical

echo lily Jul 7, 2024, 8:49 PM

#

kind stratus is there any better alternative to .filter(Optional::isPresent).map(Optional::ge...

.flatMap(Optional::stream)

kind stratus Jul 7, 2024, 8:49 PM

#

thanks

#

ok i think i've got the ingredient type otherwise implemented, i need to go now but i'll tackle the codec later

#

...where's the open map codec at again?

tawdry wolf Jul 7, 2024, 9:02 PM

#

checked Codec, ExtraCodecs, NeoForgeExtraCodecs?

kind stratus Jul 7, 2024, 9:02 PM

#

i have checked but i don't remember what it's called either

#

oh, unbounded map

kind stratus Jul 7, 2024, 10:10 PM

#

i still don't know how to test literally any of this though

pure lynx Jul 7, 2024, 11:28 PM

#

kind stratus technically i want all of 10, {"amount": 10, "unit": "mb"}, [{"amount": 10, "uni...

??

#

you should be using the name of the unit as the key here

#

this is just really hard to read

#

it should be an either of an integer or an unbound map of string to number

kind stratus Jul 7, 2024, 11:29 PM

#

pure lynx you should be using the name of the unit as the key here

yeah i got there, i was stuck on that format because the actual datamap data, which looks similar, has more fields

#

i made it so the integer can only substitute for the whole list though, not the entries inside it

#

hmm did i? let me check again

#

oh, it accepts an integer but doesn't write it back out as an integer, but it does write single-element lists out as single values

tawdry wolf Jul 7, 2024, 11:31 PM

#

kind stratus i still don't know how to test literally any of this though

ask somebody (maybe multiple people) who would be a consumer of this system to make a branch using that PR build

pure lynx Jul 7, 2024, 11:32 PM

#

i dont have ad astra on 1.21 yet otherwise id offer

#

oh wait no, i do have a mod using fluid ingredients i could use

#

this would be really useful to me rn for liquid experience. id love to have that unit be in terms of exp points

kind stratus Jul 7, 2024, 11:33 PM

#

hmm

tawdry wolf Jul 7, 2024, 11:33 PM

#

OpenBlocks also had an XP fluid iirc, that one could be used as well

kind stratus Jul 7, 2024, 11:33 PM

#

i've been going back and forth on whether to have separate abbreviated and long translations for units, the issue is that most units don't have sensible abbreviations

pure lynx Jul 7, 2024, 11:34 PM

#

maybe keep that as a lang thing? idk

kind stratus Jul 7, 2024, 11:34 PM

#

tawdry wolf OpenBlocks also had an XP fluid iirc, that one could be used as well

yeah but it doesn't do much else with fluids. I do have plans for this once the system gets in, but it's not a very thorough test

#

I think for now I'm going to have just singular and plural, and have mb as "mb" but everything else longhand including "experience points"

tawdry wolf Jul 7, 2024, 11:34 PM

#

I think using mb is fine, everything else should use "long" names (i.e., ingots, gems, nuggets, blocks, shards)

pure lynx Jul 7, 2024, 11:35 PM

#

maybe the lang would be like 1 ingot(s)?

kind stratus Jul 7, 2024, 11:35 PM

#

i already have plurals working

tawdry wolf Jul 7, 2024, 11:35 PM

#

yeah, experience_points is fine because there is no user-facing xp anymore, and removing the _points makes it unclear

kind stratus Jul 7, 2024, 11:35 PM

#

well, theoretically working as i have not tested anything

#

er i was using xp for the key but "Experience Points" to display to end users

tawdry wolf Jul 7, 2024, 11:36 PM

#

pure lynx maybe the lang would be like `1 ingot(s)`?

nah, I'm assuming count == 1 ? singularComponent : pluralComponent

kind stratus Jul 7, 2024, 11:36 PM

#

[the translated names are capitalized because it's easier for a mod that wants to display in lowercase to lowercase it than to titlecase]

tawdry wolf Jul 7, 2024, 11:36 PM

#

kind stratus er i was using xp for the key but "Experience Points" to display to end users

I'd rather that be experience_points internally tbh

#

or at the very least, xp_points

#

because otherwise it could be interpreted as xp_levels

kind stratus Jul 7, 2024, 11:37 PM

#

experience_point then? every other unit is singular internally

#

also

pure lynx Jul 7, 2024, 11:37 PM

#

the game calls it totalExperience and experienceLevel, so it should be full "experience"

kind stratus Jul 7, 2024, 11:37 PM

#

is there a resourcelocation codec that can use a different default namespace?

tawdry wolf Jul 7, 2024, 11:37 PM

#

ah, yeah make it singular, that's fine

kind stratus Jul 7, 2024, 11:37 PM

#

i'm half tempted to make one

#

tired of staring at "c:"

pure lynx Jul 7, 2024, 11:38 PM

#

seems beyond scope, the way it works rn is fine imo

kind stratus Jul 7, 2024, 11:38 PM

#

fair enough

#

so, experience_point

#

also, i set it up so you can put "bucket" or "mb" in the list for display purposes, but they're required to have the correct values

tawdry wolf Jul 7, 2024, 11:39 PM

#

1 bucket is still a constant 1000 mb?

kind stratus Jul 7, 2024, 11:39 PM

#

[i needed them in the map internally for the recipe stuff]

#

yeah, changing that is beyond scope

tawdry wolf Jul 7, 2024, 11:40 PM

#

yeah that's fine then

pure lynx Jul 7, 2024, 11:40 PM

#

maybe just make them properties of the datamap thing? cuz you'll have more data than just the list in the datamap no?

kind stratus Jul 7, 2024, 11:40 PM

#

not sure what you mean

#

[why the hell doesn't applyAllFormatting actually fix any of the issues it complains about? isn't that the purpose of having check and apply separate?]

pure lynx Jul 7, 2024, 11:42 PM

#

nvm, i just typed out what i was thinking and it was dumb

kind stratus Jul 7, 2024, 11:42 PM

#

i guess it only applies whitespace

pure lynx Jul 7, 2024, 11:42 PM

#

what other values are we putting in the individual entries for the datamap values

kind stratus Jul 7, 2024, 11:43 PM

#

display_in_tooltip flag

pure lynx Jul 7, 2024, 11:44 PM

#

"#tconstruct:fluid_units/metal": [
      { "name": "c:block",  "value": 810 },
      { "name": "c:ingot",  "value":  90 },
      { "name": "c:nugget", "value":  10 },
      { "name": "tconstruct:copper_can", "value": 90, "show_in_tooltips": false }
    ],

"#tconstruct:fluid_units/metal": {
  "c:block": 810,
  "c:ingot": 90,
  "c:nugget": 10,
  "tconstruct:coppper_can": {
    "amount": 90,
    "show_in_tooltips": false
  }
}

should be possible and how id prefer it to be written

#

i think this would just change the value codec to be an either

kind stratus Jul 7, 2024, 11:46 PM

#

... can you do that?

pure lynx Jul 7, 2024, 11:46 PM

#

i think so?

#

i can try it

kind stratus Jul 7, 2024, 11:47 PM

#

oh

#

i can just make the entry codec not include the name

#

and use that as a value in the unboundedmapcodec

pure lynx Jul 7, 2024, 11:47 PM

#

yeah, itd be an unbound map of string to either an int or "configuredint" or something

kind stratus Jul 7, 2024, 11:48 PM

#

wait no how would i get the name into the entry when reconstructing the

#

do map codecs guarantee order?

#

actually, i don't need them ordered, nevermind

pure lynx Jul 7, 2024, 11:48 PM

#

it doesnt really matter the order

#

its ordered by the value anyways

#

but in my experience yes, the order does usually stay the same

tawdry wolf Jul 7, 2024, 11:49 PM

#

yeah, type should generally be int here

kind stratus Jul 7, 2024, 11:49 PM

#

i just realized i don't actually need a class that includes both the key and the value

kind stratus Jul 8, 2024, 12:14 AM

#

gah i'm getting an exception on this codec and i don't know why

#

public static final Codec<FluidUnitSpec> CODEC = RecordCodecBuilder.create(instance -> instance.group(
            Codec.unboundedMap(FluidUnit.CODEC, UnitInfo.CODEC).fieldOf("units").forGetter(s -> s.map),
            FluidUnit.CODEC.optionalFieldOf("default_unit", FluidUnit.BUCKET).forGetter(FluidUnitSpec::defaultUnit),
            Codec.INT.optionalFieldOf("smallest_decimals").forGetter(s -> s.smallestUnitDecimals)).apply(instance, FluidUnitSpec::new));```

#

[on actually statically initializing the codec, as far as i can tell]

#

...oh

#

it was caused by me putting the static public constants in the FluidUnit class above the map that holds them [toolaction style]

#

i guess that's why tool actions have a separate class for holding the constants, i guess i could do that

kind stratus Jul 8, 2024, 1:22 AM

#

one last push, gonna leave it up overnight for people to look at before doing the PR

#

lol i just realized all my discussion from the last hour was in #neoforge-github

pastel herald Jul 8, 2024, 3:31 AM

#

kind stratus i guess that's why tool actions have a separate class for holding the constants,...

theres also Codec.lazy

kind stratus Jul 8, 2024, 4:02 AM

#

it wasn't a codec issue at all, i screwed up the static initializers in the FluidUnit class so it crashed on classload

pastel herald Jul 8, 2024, 5:01 AM

#

ah

kind stratus Jul 9, 2024, 12:10 AM

#

I posted the PR

compact lynx Jul 9, 2024, 12:15 AM

#

enabled PR publishing

kind stratus Jul 9, 2024, 12:17 AM

#

@fluid mortar @pure lynx

compact lynx Jul 9, 2024, 12:17 AM

#

could you link the PR here so I can pin it

kind stratus Jul 9, 2024, 12:19 AM

#

https://github.com/neoforged/NeoForge/pull/1264

GitHub

Fluid Unit system by Random832 · Pull Request #1264 · neoforged/Neo...

This implements the system described at https://gist.github.com/KnightMiner/ec4f21d56d466e32ac5d14a1b1124923, with a handful of extra features [a default display unit can be specified for contexts ...

compact lynx Jul 9, 2024, 12:20 AM

#

kind stratus https://github.com/neoforged/NeoForge/pull/1264

pure lynx Jul 9, 2024, 12:25 AM

#

why optional of an int instead of just returning 0?

kind stratus Jul 9, 2024, 12:26 AM

#

0 means round, empty means it's calculated based on the size of the unit being displayed fractionally

#

er, i assume you mean in the decimal places parameter, optional is a couple other places but it's used with map/orelse, you can't really do that with an int with a magic sentinel value

#

if we really don't like Optional<Integer> all uses of it can be replaced in principle but it's more and less elegant code

fluid mortar Jul 9, 2024, 1:17 AM

#

We could also use -1 if we need a third state there

#

Though that gets awkward

#

I'll take a look when I have some free time and give my thoughts

pastel herald Jul 9, 2024, 2:46 AM

#

Isn't there also an OptionalInt you could xmap with?

kind stratus Jul 9, 2024, 2:54 AM

#

I tried OptionalInt first, it doesn't have the map methods

#

It's also not just about the code in the ingredient handler - getValue is arguably a user-facing API too, and returning 0 or -1 there is just asking for trouble, people will call it not expecting it and end up actually using that value

#

though looking again, i'm actually returning null there instead of an optional

#

the optional for decimals is mostly internal and fine to change i guess

#

i actually should probably do a pass through the codecs adding validations to prevent zero and negative values where inappropriate in general

#

i think i was stuck on the codec needing an optional, i'll go ahead and change the decimals field since it's internal

#

ok yeah most of my concern was actually for getAmount [which i misrembered as using an optional for, i think i had one at one time and then switched it to a nullable], not the decimals field

#

incidentally, i was half tempted to implement a system for allowing fractions to be printed as fractions too, but that seems out of scope.

#

hmm

#

should i put some of this codec stuff in NeoForgeExtraCodecs?

#

lol oops i completely forgot to make a stream codec

#

is there a built-in stream codec for maps?

kind stratus Jul 9, 2024, 4:00 AM

#

found it - i do not understand why it cannot infer these types though ```
ByteBufCodecs.<RegistryFriendlyByteBuf, FluidUnit, Integer, Map<FluidUnit, Integer>>map(HashMap::new, FluidUnit.STREAM_CODEC, ByteBufCodecs.INT).apply(ByteBufCodecs.list()),

#

ok i've got a bunch of the codec stuff fixed, and that Optional removed, gonna push as soon as i can verify it still compiles and runs datagen

pure lynx Jul 9, 2024, 4:13 AM

#

are there really languages where the way its currently setup wouldnt work? I know mandarin doesnt really make a distinction between plural and singular but that could just be solved by using the same text twice. Maybe the plural ibm thing would be helpful, i havent ever used it tho so idk how thatd work

kind stratus Jul 9, 2024, 4:13 AM

#

gonna also add simple tests of the ingredient

#

the plural thing is something i was concerned about, but gave up since, well, i have no understanding of how PluralFormat is actually available using the Component system. I had actually assumed it wasn't available, and that adding support to it would be a different PR entirely.

#

@chrome dirge ???

chrome dirge Jul 9, 2024, 4:17 AM

#

After I spent an hour looking into the Java side, I was a bit fed up with the topic and didn't look into the Minecraft side of things. However, at its base, what's behind the formatter is a lookup which of 6 translation forms to use. So even fi the formatter itself is useless, that lookup still should be in there and we may need to go traditional with 6 seperate lang keys instead of the modern 1-key-with-inlines format

kind stratus Jul 9, 2024, 4:18 AM

#

ok but the function may be called on the server, and the user's locale is unavailable there

#

i think pluralization for languages that don't use a simple singular-plural form is outside of scope if there's no actual way to do it

#

it's worth looking into in general though

#

i wrote some gametests for the ingredient types, and my matching is failing for some reason

chrome dirge Jul 9, 2024, 4:21 AM

#

oh, yes. there it is: com.ibm.icu.text.PluralRules

solemn sand Jul 9, 2024, 4:22 AM

#

pure lynx are there really languages where the way its currently setup wouldnt work? I kno...

Yes -- consider, say, the pluralising of "0 buckets" in different languages. You can find some iffy ones with that pretty easily

chrome dirge Jul 9, 2024, 4:22 AM

#

um, getTooltipComponents() called on the server? seems odd

solemn sand Jul 9, 2024, 4:23 AM

#

chrome dirge um, getTooltipComponents() called on the server? seems odd

Getting components on the server is common

pure lynx Jul 9, 2024, 4:23 AM

#

dont commands do that?

kind stratus Jul 9, 2024, 4:23 AM

#

chrome dirge um, getTooltipComponents() called on the server? seems odd

The basic formatting function could be called to format a command output. I also think AE2 and the like might build tooltips on the server for search purposes

solemn sand Jul 9, 2024, 4:23 AM

#

You often grab them there, then sync them to the client for whatever reason

chrome dirge Jul 9, 2024, 4:23 AM

#

https://www.unicode.org/cldr/charts/45/supplemental/language_plural_rules.html here are some examples...

solemn sand Jul 9, 2024, 4:26 AM

#

Does MC deal with this anywhere?

kind stratus Jul 9, 2024, 4:26 AM

#

ah shit i forgot to run apply formatting before pushing

kind stratus Jul 9, 2024, 4:26 AM

#

solemn sand Does MC deal with this anywhere?

MC mostly just does always plural or (s)

solemn sand Jul 9, 2024, 4:28 AM

#

kind stratus MC mostly just does always plural or (s)

Why don't we follow that then?

kind stratus Jul 9, 2024, 4:28 AM

#

because it's ugly

pure lynx Jul 9, 2024, 4:28 AM

#

(s) i think is fine?

solemn sand Jul 9, 2024, 4:28 AM

#

1 bucket(s) looks fine to me

kind stratus Jul 9, 2024, 4:28 AM

#

and it's mostly in command outputs, which is less user-facing

#

"1 block(s), 4 ingot(s), 4.5 nugget(s)"?

solemn sand Jul 9, 2024, 4:29 AM

#

There's no good way to handle plural vs singular via translation components as they currently work

solemn sand Jul 9, 2024, 4:29 AM

#

kind stratus "1 block(s), 4 ingot(s), 4.5 nugget(s)"?

Then do always plural. 1 blocks, 4 ingots, 4.5 nuggets. Yes, it'll be technically wrong sometimes, but, eh

pure lynx Jul 9, 2024, 4:29 AM

#

lets just make everyone learn mandarin or something, no distinction between singular and plural 😄

#

the optimal solution™️

kind stratus Jul 9, 2024, 4:30 AM

#

how about this: people translating for languages that the english rules don't work for can put their equivalent of bucket(s) in both keys if they want, but we'll keep the plural system for those that do

#

[it'll work for french too, these never get printed with zero, at least not by the helper function]

solemn sand Jul 9, 2024, 4:31 AM

#

kind stratus [it'll work for french too, these never get printed with zero, at least not by t...

thonk what?

#

How does 0 print then?

kind stratus Jul 9, 2024, 4:31 AM

#

Empty

pure lynx Jul 9, 2024, 4:31 AM

#

im also personally iffy on the keys being singular. for ingredients it doesnt look correct

{ buckets: 1 } > { bucket: 1 }

solemn sand Jul 9, 2024, 4:31 AM

#

Ah, okay

solemn sand Jul 9, 2024, 4:31 AM

#

pure lynx im also personally iffy on the keys being singular. for ingredients it doesnt lo...

The keys should be singular because it's the name of the unit itself

#

And in the data map for setting those, the plural versoin would look weird

pure lynx Jul 9, 2024, 4:33 AM

#

well either way its going to look weird somewhere. the datamap one is more for the mod dev while the usage in recipes is more used for consumers and users, i say make the key plural :P

#

you write a data map once, you could end up writing dozens of recipes, it affects one more than the other

kind stratus Jul 9, 2024, 4:34 AM

#

~~what if i made each one have two names~~

pure lynx Jul 9, 2024, 4:34 AM

#

oh god no

#

aliases.. hm

kind stratus Jul 9, 2024, 4:34 AM

#

have that and a special codec that omits the c: namespace, for maximum friendliness

#

[to be clear, having two names is not actually viable, they'd have to be registered in code instead of auto-interning like tool actions]

#

the no-namespace codec is viable though

solemn sand Jul 9, 2024, 4:36 AM

#

No aliases. I'd rather have only-plural than that

#

no-namespace codec though -- that's a good idea. Stops you from accidentally using minecraft: units too

chrome dirge Jul 9, 2024, 4:38 AM

#

aside from the decision for this specific case, what would be the general opinion in the room on patching proper pluralisation support into the translatable component? Naturally, using it for Components sent to vanilla clients wouldn't work, but aside from that I see no compatibility issues when hijacking the % placeholder as any other character than % and s after it is a fatal error atm.
(i.e. would it be a waste of time if I were to look into doing that)

pure lynx Jul 9, 2024, 4:39 AM

#

as a separate PR it might be a nice feature

#

im sure this is not the first or last time someone will run into this issue

kind stratus Jul 9, 2024, 4:40 AM

#

i like the idea

#

but i wonder if inline formatting might not be the best solution

#

actually nevermind, fuck messing with the component codecs

chrome dirge Jul 9, 2024, 4:41 AM

#

my idea if to hijack the % and use e.g. "%(standard ICU plural format)" for this

kind stratus Jul 9, 2024, 4:41 AM

#

[yeah inline formatting is fine]

#

yeah, i had a vague idea of using some sort of custom json format, then i just realized what a nightmare that would be to implement

#

i think there are some places in FML that use some kind of %{} thing, so it's not without precedent, though i think they only work for logging not for the game itself

#

what exactly is the ICU plural format? like, is %() part of it, or is that your own invention? all i can find just uses {}

#

and named variables instead of numbered ones

#

i guess as long as it works it doesn't matter what the particulars of the format are

chrome dirge Jul 9, 2024, 4:44 AM

#

so in the json it'd be "key": "I see %s %(one{dog} other{dogs}) here!",

kind stratus Jul 9, 2024, 4:44 AM

#

%1$(...) surely

#

[not passing the argument twice means we can in principle fix vanilla's (s)s with a resource pack]

chrome dirge Jul 9, 2024, 4:45 AM

#

%s is the only allowed placeholder in vanilla, and while it would be great for localisation to provide positional ones, I'm not sure if that should be mixed in with pluralisation

#

oh, that

kind stratus Jul 9, 2024, 4:46 AM

#

i'm pretty sure vanilla does support %1$s

#

it's been a while since i looked at it though

chrome dirge Jul 9, 2024, 4:47 AM

#

yes, I think "key": "I see %(one{%s dog} other{%s dogs}) here!", could work
But I'll have a look ath what the ICU classes provide first before I spin my own thing here

elfin yacht Jul 9, 2024, 4:48 AM

#

pluralization support is... complicated

chrome dirge Jul 9, 2024, 4:48 AM

#

kind stratus i'm pretty sure vanilla does support %1$s

I don't see it in net.minecraft.network.chat.contents.TranslatableContents.decomposeTemplate(), but there may be other places that split up that string?

elfin yacht Jul 9, 2024, 4:48 AM

#

its simple for english

kind stratus Jul 9, 2024, 4:48 AM

#

it could, but not being able to reference the argument twice means you're limited in cases where a different variable needs to be printed in between the number and the pluralized form

elfin yacht Jul 9, 2024, 4:48 AM

#

but some languages go up to... 6? different forms based on plurality

chrome dirge Jul 9, 2024, 4:49 AM

#

Shadows, see above, I posted a reference link

kind stratus Jul 9, 2024, 4:49 AM

#

yup, we're aware - i only used simple pluralization because fixing the component system to do more is outside the scope of this particular PR, but that doesn't mean it's not worth doing

#

(incidentally, it's not like complexity is foreign to english - our plurals are simple, but we have the same problem with ordinals)

#

some of the rules on that page for plurals actually remind me a lot of english ordinals, with the 1 vs 11 vs 101 thing

chrome dirge Jul 9, 2024, 4:51 AM

#

yes, funnily enough, given how many different variants of languages there are, they all are a mix of very few building blocks

kind stratus Jul 9, 2024, 4:51 AM

#

[i don't know for sure if there are places in vanilla or mods that do "1th" ordinals, but i don't particularly doubt it either]

chrome dirge Jul 9, 2024, 4:52 AM

#

I'm pretty sure vanilla only uses number inserts in their strings very, very, rarely

elfin yacht Jul 9, 2024, 4:52 AM

#

yeah vanilla doesn't do it very much

#

mostly in the options menu or debug

#

where they don't need to display X things, but rather Value: X

kind stratus Jul 9, 2024, 4:53 AM

#

of course

#

there is another alternative

chrome dirge Jul 9, 2024, 4:53 AM

#

"translation.test.complex": "Prefix, %s%2$s again %s and %1$s lastly %s and also %1$s again!", ok, there is a complex pattern. I just need to find the code that handles that

kind stratus Jul 9, 2024, 4:54 AM

#

we could always print singular

#

eh scratch that

#

for some reason i had it in my head that vanilla does "1 [singular item name]" in tooltips, but the count isn't in the tooltip

chrome dirge Jul 9, 2024, 4:55 AM

#

um, duh. found it. the code checks the second part before the fist part. I stopped reading before I got to that 🤦

kind stratus Jul 9, 2024, 4:55 AM

#

anyway, we should probably split off another thread for this

elfin yacht Jul 9, 2024, 4:56 AM

#

It looks like we can also just add new ComponentContents

#

So if we need a different "plural-sensitive-contents" that's a thing we can do, I guess

chrome dirge Jul 9, 2024, 4:57 AM

#

finally found one in the vanilla lang file: "Chat message was too long (%s > maximum %s characters)", and chat.queue": "[+%s pending lines]",, "clear.failed.multiple": "No items were found on %s players",, "commands.banip.info": "This ban affects %s player(s): %s", "commands.banlist.list": "There are %s ban(s):", and a couple more command result messages

kind stratus Jul 9, 2024, 5:01 AM

#

elfin yacht It looks like we can also just add new ComponentContents

huh, how hard is that? i'm not convinced it's a good solution on its own, but in combination with something someone else was showing off [a port of some fabric thing called owo-lib that lets translation strings be components themselves] a while back it could be nice

elfin yacht Jul 9, 2024, 5:01 AM

#

yeah we have that thing

#

it's why we have a custom contents in the first place LUL

#

https://github.com/neoforged/NeoForge/pull/1134

kind stratus Jul 9, 2024, 5:02 AM

#

nice

#

that's probably a better path then, i just assumed it would be too difficult

elfin yacht Jul 9, 2024, 5:03 AM

#

iirc there were some vanilla compat concerns but I guess they got that worked out

chrome dirge Jul 9, 2024, 5:03 AM

#

grrr, not a single line of javadoc. how will anyone know how to use that? (re 1134)

elfin yacht Jul 9, 2024, 5:04 AM

#

I don't think you are intended to interact with the contents object

kind stratus Jul 9, 2024, 5:04 AM

#

i tried to document most of my public methods, i might have missed a few though

#

eh

#

i mean, javadoc isn't a totally non-viable venue for documentation of a json format

#

my ingredient type has json info in the javadocs on its codec fields [copied from sizedfluidingredient which also does it]

elfin yacht Jul 9, 2024, 5:05 AM

#

I have taken to describing json syntax with markdown using js highlighting: https://github.com/Shadows-of-Fire/Placebo/blob/1.20/schema/ItemStack.md

chrome dirge Jul 9, 2024, 5:12 AM

#

grrr, even after spending ten minutes reading that PR's diff, I still have no idea how I would use that thing other than copy&pasting the test that comes with it and trying out what works .-(

elfin yacht Jul 9, 2024, 5:13 AM

#

there's a formal doc here https://docs.wispforest.io/owo/rich-translations/

chrome dirge Jul 9, 2024, 5:15 AM

#

"When owo is installed, " Alt-F4.

;) ;)

chrome dirge Jul 9, 2024, 8:55 AM

#

just a quick&dirty proof-of-concept hack. that patch is 5 lines plus a tool method to safely convert the parameter to double

tawdry wolf Jul 9, 2024, 1:06 PM

#

A problem you will run into as well is existing resource packs

#

especially if you extend it to vanilla plurals

chrome dirge Jul 9, 2024, 1:30 PM

#

should be no issue, unless someone uses "%(" somewhere...and I just noticed that that was an ignored, not an illegal combination. I may need to change it to e.g. "%n(", an actually illegal combination.

compact lynx Jul 9, 2024, 2:18 PM

#

chrome dirge just a quick&dirty proof-of-concept hack. that patch is 5 lines plus a tool meth...

have you tried 0 yet

chrome dirge Jul 9, 2024, 2:50 PM

#

compact lynx have you tried 0 yet

you mean as a value? sure, that has been the test case for my last "does it still work" tests...

#

very convenient one, already there and comes up when starting the server ;)

kind stratus Jul 15, 2024, 7:10 AM

#

ugh i forgot to check notifications - i'll go through the PR review comments in the next few days

kind stratus Jul 15, 2024, 4:02 PM

#

Its probably worth having a method on FluidStack to get the full tooltip including display name to standardize the order and color between that and the amount string. Plus would call appendHoverText

I thought about this and actually almost started writing one, but then it occurred to me that about half of the vanilla itemstack components that affect tooltips [e.g. name, lore, rarity, etc - maybe potion effects] could be applied to fluidstacks as well and figured it's something that deserves some discussion on its own.

kind stratus Jul 15, 2024, 5:08 PM

#

is there a tool i can run to find public methods that don't have javadocs? i think i got all of them but it'd be nice to be sure

kind stratus Jul 16, 2024, 4:50 AM

#

ugh i thought i did run datagen, apparently it failed

fluid mortar Jul 16, 2024, 3:59 PM

#

kind stratus > Its probably worth having a method on FluidStack to get the full tooltip inclu...

I partly wonder if such things shouldn't be left up to future work

#

we could start with a bare bones "getHoverText" then expand it in the future

#

alternatively, we could do that PR first and do the units PR after

#

only concern with that idea is the unit display would have to be fixed to millibuckets which makes it more difficult for mods with their own unit display in the meantime

#Fluid Unit Hell

Friday Facts #416 - Fluids 2.0