Game Development

@trollingchar Wish Unity had some android-like Override tag, so you can explicitly tell Unity if you indeed plan a function to be treated as a Unity function or a custom one

trollingchar

Why can't it just use an attribute like [Update]? It's already using attributes for many things

TomTsagk

I'm not familiar enough with attributes, how do they work with functions?

trollingchar

Attributes do nothing by themselves, they are used by other modules/plugins/etc via reflection or something

at least when I worked with them, I used reflection to find classes marked with attributes and process them

in this case if Unity sees for example [SerializeField] attribute it knows that this field should appear in inspector

TomTsagk

11:16 AM

Ooooh I see, I've used [SerializeField] before

From what I understand they are like annotations in Java (similar to @Override)

trollingchar

yes

they are similar

Alexandre Vaillancourt

1:45 PM

But then again, we're used to Update. But that does not describe really well what the function is doing...

TomTsagk

True, but when it comes to OOP there could always be an object that is meant to be updated manually, its nice to have the choice

Alexandre Vaillancourt

But that could be an incentive to give a better name to the function :)

We have had issues with functions named "update" that were doing "everything".

Alexandre Vaillancourt

2:03 PM

If I listen to music like metalcore and death metal when I work, when I get home at night, can I complain to my wife that I'm being yelled at all day long?

TomTsagk

I like that, does complaining to your wife come with perks? More time to spend on video games to make you feel better? Or takeaway from your favourite place?

Alexandre Vaillancourt

It doesn't come with perks, no :P

TomTsagk

It appears the movies have been lying to me

Alexandre Vaillancourt

Oh, I guess mileage may vary

TomTsagk

:P

trollingchar

2:10 PM

hello again

TomTsagk

yo

trollingchar

I can't decide how to organize states in my game state machine

TomTsagk

Is this for the worms game ? or in general?

trollingchar

yes

TomTsagk

What do you use to handle animations and states now ?

Is it a custom system or some other tool?

trollingchar

2:15 PM

this is a state machine for turns, not animations

There are actions that can affect victory conditions and those that cannot

Poison damage, water rising and nuclear missiles fall into the first category

Shield regen, wind change and crates fall into the second

TomTsagk

Wouldn't it be better to check victory condition every time a player dies? (assuming that's the only victory condition)

trollingchar

so I should synchronize clients when ckeck for victory is already made and will not be changed

@TomTsagk this is a way that prevents scheduled nukes from arriving

TomTsagk

Oh are you talking about multiplayer ?

trollingchar

okay

the game has multiplayer mode, and clients synchronize via server

how turns work

before each turn, wind changes, shields regenerate, and a crate may fall

then a player plays his turn

TomTsagk

Mmm

How much have you implemented already? is there any state machine now?

trollingchar

2:22 PM

after that, poison damage applies and water level rises in sudden death mode

my code is messy that's why I'm refactoring it

the problem is that nukes work this way: you select a target, confirm strike, everyone sees CAUTION ☢ CAUTION screen, but nuke falls after every player plays one turn, right before your turn

and if you are eliminated, it still falls

the problem is in specifying an order that I will implement

TomTsagk

You could divine the turn into phases, and dynamically add or remove phases depending on the special weapons

trollingchar

this is where i have problems

TomTsagk

so a turn goes like PREPARE PHASE (preparing the current turn, change wind, pass a turn on timed weapons etc) then PLAYER PHASE where the current player has a turn etc etc

trollingchar

I already have some phases

TomTsagk

You could add some of the effect on those phases? It helps if you "combine" similar-behaving effects, to simplify the process

trollingchar

2:29 PM

initial
before-turn
sync
ready
turn
after-turn
poison
next-turn

this is how it works now

TomTsagk

Sounds good, you could probably combine some of those phases to each other (like poison can happen during "after-turn")

trollingchar

before-turn is when harmless events happen, after-turn is when projectiles are flying but you can't move, poison is where harmful events happen

but poison should work exactly once, and nukes may arrive several times (if a player was eliminated)

my thoughts are refactoring this into this

initial
harmless events & send sync signal
wait for server
ready
turn
after turn
harmful events
loop dropping nukes until a player can become active (not eliminated)
repeat

but I don't know where to insert checks for victory and draw

will it be unfair to not drop a nuke if victory conditions are met at some moment between turns?

one of the Three Superweapons isn't something less important than an ordinary victory I think

it can turn half of the map into radioactive ashes so even if you are eliminated, your opponent may be unable to escape it, so it will be a draw

TomTsagk

What if all phases are grouped in an object oriented way, so that every phase inherits the Phase object, and every time a Phase object finishes, it checks for victory condition?

trollingchar

I can implement everything. I am not done with requirements.

TomTsagk

2:44 PM

In general I like more simplified system, so you can focus more on the actual game, instead of spending time refactoring or maintaining later on

trollingchar

if the check is done after each phase, nukes may be ignored sometimes

TomTsagk

How do nukes work exactly? I think I've misunderstood that

trollingchar

1. You select a target and confirm a nuclear strike.
2. Everyone sees a huge nuclear alarm screen and hears loud nuclear alarm sounds. They don't know where is the target.
3. Every player makes their turn.
4. Right before your turn, or instead of it (if you are eliminated), the missile arrives and deals devastating damage.

Should I allow victory if there are missiles not yet dropped?

the answer is now clear to me that I should lock victory until all missiles strike

TomTsagk

Oh I see, this sounds more like a design decision than a developing decision

trollingchar

even there are no players left, isn't it fun to witness something amusing and rare?

@TomTsagk yes

now i have better understanding how to organize states

TomTsagk

2:53 PM

I remember on the original games, when events would happen, the game would wait for them to finish, and if anyone won after that event, they would be declared victorious straight away (even if the sea level was about to rise for example)

I think its fair that if multiple events are meant to happen, if the first event kills all your enemies, that you should be victorious

trollingchar

hmm

a chess-like approach

TomTsagk

I'm not sure chess works like that :P

trollingchar

but i like the approach when you must secure your victory - no casualties from poison and flooding and nukes about to strike

I should ask players what they think, but when I release the game, you may become players, so I ask you

a working day is over, I go afk for a hour

TomTsagk

3:11 PM

Haha fair enough, at the end of the day it's your choice. I can see how both ways could work and be fun, but it also depends on the rest of the game

Has anyone done any work on detecting endian-ness on C ?

Alexandre Vaillancourt

Someone mentioned little endian?

TomTsagk

Was that drawn by you? :P

Alexandre Vaillancourt

No. Google did that :P

TomTsagk

She looks so sassy

Alexandre Vaillancourt

But no, I haven't worked with endian-ness in C. I did not work much with C.

TomTsagk

3:24 PM

What about C++? I imagine it would be similar?

I found a snippet on stack overflow, that you can create an int and a pointer to char to basically check if the current system is using little-endian or big-endian, but for some reason it feels there should be a better way?

Elva

en.cppreference.com/w/cpp/types/endian

std::endian::native == std::endian::little

And lots of weird corner cases that you should probably just do:

static_assert(
	std::endian::native == std::endian::little ||
	std::endian::native == std::endian::big &&
std::endian::big != std::endian::little)

Alexandre Vaillancourt

works for c++ 20, though, might not be implemented. This informaiton is supplied by the compiler?

Elva

Oh ++20 :o I did not notice that

TomTsagk

That's interesting, wasn't aware of this

trollingchar

3:47 PM

http://manpages.courier-mta.org/htmlman3/endian.3.html
this? it should be compatible with pure C

but it may be linux-only

Elva

There's standard to host order and to network order things ain't there?

Shotgun Ninja

Should be, though idk what they'd be

Elva

Yeah the hton and ntoh family linux.die.net/man/3/htons

Host to Network and Network to Host

TomTsagk

Thanks @trollingchar, looks interesting, but it seems to be functions to convert from one to another

I was thinking something more to detect the endian-ness of the current system

Elva

And network order is big endian: tools.ietf.org/html/draft-newman-network-byte-order-01

TomTsagk

3:50 PM

I've found this: https://stackoverflow.com/questions/2100331/c-macro-definition-to-determine-big-endian-or-little-endian-machine

but even though it works, it feels "hacky"

trollingchar

convert a number from host endianness to little or big and compare

Elva

So bool is_bigendian_machine() { return htonl( 0xff ) == 0xff }

TomTsagk

oooh I see

that makes sense

This actually seems promising, thanks a lot everyone :D

Alexandre Vaillancourt

If your at the point of needing to know what is the endianness of a machine, it's because you're deep down doing leet stuff :P

TomTsagk

It's not what it sounds like :P
I'm making a small `.ply` parser, and when parsing it, it tells you if its data is in little or big endian

Alexandre Vaillancourt

4:03 PM

Oh, ok, well reading binary stuff is a bit more involve than reading text files :P

TomTsagk

True, but reading binary data is so much faster, that makes it worth it to understand how it works, or at least try :P

trollingchar

all files are binary

you just decode them differently

Alexandre Vaillancourt

@TomTsagk I totally agree :)

@trollingchar Yeah, here "you" means "the api you use", but your right.

TomTsagk

@trollingchar isn't everything binary?

This took a philosophical turn

Alexandre Vaillancourt

@TomTsagk cough CoC update cough

trollingchar

4:06 PM

@TomTsagk If I say yes, Stack Exchange Inc will ban me

Alexandre Vaillancourt

:P

TomTsagk

Oh no you didn't :P

Alexandre Vaillancourt

I love those:

updateMouseCursorVisibility(); // Update the mouse cursor visibility

TomTsagk

Oh boy

trollingchar

omg they again

TomTsagk

4:10 PM

There was a function in Unity that had a very confusing name, and I search the docs to see what it does, and it literally said the same as the name of the function, but it had normal grammar. Which suggests someone actually wrote that instead of being automated

Alexandre Vaillancourt

Lots of people work at Unity. When you add more people, the quality can degrade :P

trollingchar

@AlexandreVaillancourt I guarantee it

Alexandre Vaillancourt

First hand experience?

Almo

5:39 PM

.

Shotgun Ninja

..

Almo