DDraceNetwork - developer

Can anyone give me some thought on how to find out what is in the camera so visible to the player with the zoom level. Would then use the information to compare to a rect and the camera automatically has that field or rect in the center.

@Daniel wdym

01:28

are u talking about that : https://github.com/ddnet/ddnet/blob/master/src/game/server/entity.cpp#L90

ddnet/entity.cpp at master · ddnet/ddnet

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

i want to calculate the view of the camera like im on camera position 10 10 and the zoom level is 12 then i want 2x vec2 to get the rectangle of view

01:33

example image for better understanding.. hard to explain for me

01:34

I can calculate the green box but not the red box of the current camera view

do you have the current position of the tee? or let me rephrase. is the tee in the center of that red box? (edited)

Ravie

@Jupstar ✪ here's an idea with a scrollbar, we could also do something about the spectator box below to have more space for players

so like, will it be like you have to hold tab and use the scroll wheel?

default

do you have the current position of the tee? or let me rephrase. is the tee in the center of that red box? (edited)

no the tee is not at the center

the input parameters and specified output would help :)

how to fix it oop

1

Daniel

no the tee is not at the center

just tried to find x/y position using CE and the hud. they seem to hide it. can't help then. sorry. not interested to unrandomize the values at 02:50 in the night. (edited)

Voxel

so like, will it be like you have to hold tab and use the scroll wheel?

either wheel or drag the scrollbar with the cursor like normally

Ravie

either wheel or drag the scrollbar with the cursor like normally

i think its better to use scroll wheel, unless you want a seperate cursor just for that menu

01:49

like in emotes

01:50

even then you can still look around holding tab

cursor is more intuitive for new people who don't know the shortcuts

1

fair

i use cursor for zoozing

i think

        float ScreenX0, ScreenY0, ScreenX1, ScreenY1;
        Graphics()->GetScreen(&ScreenX0, &ScreenY0, &ScreenX1, &ScreenY1);

i think this is my solution

ddnet-maps

[ddnet-maps:master] 1 new commit

418af5c M Quantum, A FarLands - ddnet-maps

Jupstar ✪

do you still need it?

nice. would be nice to fix this error, though it could be tricky if this undefined behavior has always been part of the shapshot compression.

Robyt3

nice. would be nice to fix this error, though it could be tricky if this undefined behavior has always been part of the shapshot compression.

i named myself ᴷᵉᵏᶳ over a bind then in F1 typed player_name nae as soon as the server confirmed the name change it happened

Jupstar ✪

i named myself ᴷᵉᵏᶳ over a bind then in F1 typed player_name nae as soon as the server confirmed the name change it happened

doesn't seem to work for me locally :/

u need the client version that still prints the errors

10:44

or add it back, i tested it on multiple ddnet servers

I'm on current master with the errors enabled again

mhh let me test

10:48

for me it still happens: 37.230.210.231:8344 even renaming to keks

10:49

i uncommented this lines: https://github.com/ddnet/ddnet/commit/b62e7f1299fc672e210a761caba7780d6ee9d06a

Quick fix for old client on new server · ddnet/ddnet@b62e7f1

Introduced in https://github.com/ddnet/ddnet/pull/6214/ Not a proper solution yet

10:52

well weird it behaves different between us i use clang 16 on debian linux i dunno if this could be simly because the compiler compiles it differently but i am sure it also happened on gcc so maybe bcs my pc runs faster, or bcs linux/windows difference

10:53

or network/ping

10:54

btw in_range<long> is that even a safe template parameter? can't long be 32-bit on a 32-bit computer? (edited)

ah, should probably use int64_t

if there is smth that helps you understanding the problem just ping me, i can reprod it even in debug build

debug: sizeof(int)=4 sizeof(long)=4 sizeof(int64_t)=8

10:59

yeah, I guess the check never worked on my Windows system

oh ok

now I get the error immediately when renaming

Robyt3

[ddnet/ddnet] Pull request opened: #6284 Implement integer wrapping instead of preventing overflow/underflow

As the integer overflow/underflow in UndiffItem can happen during normal gameplay, we should in this case neither ignore the snapshot delta nor show an error message. Instead of depending on the particular compiler doing integer wrapping, when integer overflows or underflows occur, we make it part of the design, by implementing integer wrapping with add_int_wrap.

Checklist

[X] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in comb...

Uf, is that really the prettiest way to write add_int_wrap? I'm really not a huge fan of the random 64 assumption in there

Our assumption that int's are 4 bytes also annoys me

16:56

But I guess that one is deeply rooted

Is there a smarter way then doing a memcpy to convert between type and unsigned type?

template<typename T, typename U>
constexpr inline T bit_cast(U u)
{
    static_assert(sizeof(T) == sizeof(U) && std::is_trivially_copyable_v<T> && std::is_trivially_copyable_v<U>);
    T t;
    std::memcpy(&t, &u, sizeof(T));
    return t;
}

template<typename T>
constexpr inline T add_wrap(T a, T b)
{
    typedef typename std::make_unsigned<T>::type unsigned_type;
    return bit_cast<T>(bit_cast<unsigned_type>(a) + bit_cast<unsigned_type>(b));
}

When I use reinterpret_cast I get

error: invalid cast from type 'int' to type 'unsigned_type' {aka 'unsigned int'}
  148 |         return reinterpret_cast<T>(reinterpret_cast<unsigned_type>(a) + reinterpret_cast<unsigned_type>(b));

reinterpret_cast takes pointer

17:25

use *reinterpret_cast<unsigned int*>(&a)

I see, thanks

I actually wonder if that reinterpret cast is actually legal. Though not enough to dig out the standard

18:00

Mh, the bit_cast one is definitely legal. reinterpret_cast might not be, I don't really remember now

With gcc -O2 both implementations have the same assembly https://godbolt.org/z/McW3EWaqE

bit_cast - the code is practically the same as in their possible implementation segment. They also note

reinterpret_cast (or equivalent explicit cast) between pointer or reference types shall not be used to reinterpret object representation in most cases because of the type aliasing rule.

https://stackoverflow.com/questions/98650/what-is-the-strict-aliasing-rule https://stackoverflow.com/questions/14623266/why-cant-i-reinterpret-cast-uint-to-int (edited)

cpp/numeric/bit cast

What is the strict aliasing rule?

When asking about common undefined behavior in C, people sometimes refer to the strict aliasing rule. What are they talking about?

Why can't I reinterpret_cast uint to int?

Here's what I want to do: const int64_t randomIntNumber = reinterpret_cast<int64_t> (randomUintNumber); Where randomUintNumber is of type uint64_t. The error is (MSVC 2010): error C244...

Maybe do the bit_cast just in case. We had issues with strict aliasing once and its really no fun to debug

Learath2

I actually wonder if that reinterpret cast is actually legal. Though not enough to dig out the standard

you can safely reinterpret_cast between types of same size

discord troll admins

18:48

fat fingers

18:48

woden hands

18:49

i mean its not only #developer

18:49

even showroom

wtf, someone reordered channels order?

discord really needs a layout lock xD

Chairn

you can safely reinterpret_cast between types of same size

Actually even the size doesn't seem to matter

C++11 5.2.10p7 allows converting any object pointer to an object pointer of different type

19:00

Let me see if you are allowed to actually access it without invoking UB though

probably not

pls

19:00

move #developer back

19:00

under #general

19:01

@Learath2 this is a matter of life and death

I think that was intentional we've been moved down

nonono