networks

[elder cobalt]

I guess it might be because of this line

so I decided to load up a server with AI where each AI has a session ID + 2 real clients which are controlled from me

and when I put my session id which is 18 and not the AI ones

its following one of the AI's instead of me

it's following session ID 25

[cloud spruce]

can you add this print statement just (other lines are just for context) before the w.send(packet) of the first if block:

                packet.payload = bytes(data)
                print(f"target pos(x: {to_float(X_pos)}, y: {to_float(Y_pos)}, z: {to_float(Z_pos)}), rot (x: {to_float(X_rot)}, y: {to_float(Y_rot)}, z: {to_float(Z_rot)})")
            w.send(packet)
```and this `print` statement almost at the end of the `elif` block like this:
```python
            packet.payload = bytes(data)
            print(f"    my pos(x: {to_float(X_pos) + X_rel_pos}, y: {to_float(Y_pos)}, z: {to_float(Z_pos) + Z_rel_pos}), rot (x: {to_float(X_rot) + math.pi}, y: {to_float(Y_rot)}, z: {to_float(Z_rot)})")
            w.send(packet)

[elder cobalt]

like this?

or like this?

[elder cobalt]

if its like this then

yeah

its definitely following the wrong person

💀

coz my coords are this

[cloud spruce]

let me check the session_id conversion code, haven't really read it, just copied it

[elder cobalt]

if I do understand this logic properly

data[1:2] will be comparing it with char of 18

[cloud spruce]

are you sure the session id is at [1]?

[elder cobalt]

yeah

when im using my old code

its working

[cloud spruce]

ok, so that is the position of right car then?

i think i know what it is... hold on

no, it wasn't what i first though

[elder cobalt]

yeah\

[cloud spruce]

i think i have all the changes in this copy of the code: https://paste.pythondiscord.com/musecaqere

[elder cobalt]

hmm looks like even using my statement in yours doesnt work

[cloud spruce]

try with the one i sent just now

i hope i have not forgotten a fix that you have implemented in between, but i think i got them all

[elder cobalt]

👍

[cloud spruce]

rewriting packets can be hard when i can't experiment with it my self 😆

[elder cobalt]

yeah I get you, it's a huge guess work since you can't really know what type of packets come across or go lol

also

[elder cobalt]

tried the new one and for some reason it's not sending print statements?

[cloud spruce]

have you changed the server_port and Session_ID to be correct for the code i sent you? (since i don't know what should be there for you)

[elder cobalt]

yeah did

[elder cobalt]

okay changing it to this

from

works

properly

and even captures the right car

just edited this in

btw was this supposed to be Z_rot or X_rot?

when I do relative_distance=5

when its relative_distance=0

relative_distance=3

[cloud spruce]

it should be Z_pos there as that indicates what way the car is facing, isn't it?

[elder cobalt]

Not sure, Z pos is basically

The car position in a XYZ Cartesian graph

If I recollect these are the only packets which are received by the client from server

[cloud spruce]

this should only be some basic trigonometry to position your car relative to the target car
we solve a right triangle (where the angle of C is 90 degrees) for the "adjacent side" (b) and the "opposite side" (a)
using relative_distance, which is really the length of the "hypotenuse" (c, the longest side) and "theta" (angle of A) but in radians instead of degrees and of us i think that is Z_rot

[elder cobalt]

Hmm adding relative distance to 0 basically means assigning the X Y and Z coords of Car B to Car A basically resulting to this

Increasing it doesn't seem like to increase the distance between them but instead be positioned to the left of the car

[cloud spruce]

yeah, that should just copy the other car, but then we have a 180 degree rotation in there too

[elder cobalt]

hmm so you think rotation is messing it up?

[elder cobalt]

well I mean even if its without rotation it would just be this but the car is facing front

dont you think?

[cloud spruce]

i might have messed up the order of the sin and cos calculations in the function for the X and Z, just try this

def relative_position_2d(distance: float, rad_angle: float) -> tuple[float, float]:
    return distance * math.sin(rad_angle), distance * math.cos(rad_angle)
    #return distance * math.cos(rad_angle), distance * math.sin(rad_angle)
```the old return value is just commented out so that we can restore it easily if this messes things up even more

[elder cobalt]

I feel like I might've found a relation between worldPosition and velocity

when my X velocity is negative my world position becomes negative as well

when X velocity is positive my world position becomes positive as well

problem starts when X velocity is at 0 lol

not really 0 but yeah

[cloud spruce]

oh, the velocity is probably just like what you would get with a 3D gyroscope in a phone or other device that has them

[elder cobalt]

using this modification when relative distance is at 5

3

0

[cloud spruce]

then it's not correct, i need to revisit the calculation again, it should be find at 5, it's something else that's up with the positioning

what happens if you turn the target car like 90 degrees to one or the other side?

[elder cobalt]

with relative distance = 3

[cloud spruce]

with 5 that should look better

[elder cobalt]

with 5

[elder cobalt]

you think anything can be done with the help of this relation?

eh cant

[cloud spruce]

don't think the velocity should change anything here, position and angle should be enough

[cloud spruce]

is the drivers seat on the right side of the car in those cars?

if it is it would look like the middle of the car is centered on the driver and not the middle of the car

[elder cobalt]

for the bmw it's left

[cloud spruce]

okay, then that theory is out the window

[elder cobalt]

as well as for the abarth its left

[elder cobalt]

yeah its cantered for the bmw lol

[cloud spruce]

i still wonder if the car position in the game is at the center of the car or the center of the driver, i hope the car or else there will be problems getting it right

[elder cobalt]

hmm I feel like it might be at the center of the driver?

not sure

well lets say if its at the centre of the car

[cloud spruce]

maybe, then it will be hard to adjust for that without knowing the drivers position in relation to the car body for both the target car and your car

[elder cobalt]

if I were to increase X and Z by + 1 it should move + 1 in front of XZ direction right?

[cloud spruce]

i hope it's the center of the car, then it's no problem

i think it will move in relation to coordinates of the world map

[elder cobalt]

let me check up real quick with someone who might know the game's physics

[cloud spruce]

if you substitute X_rel_pos and Z_rel_pos with 5 in the calculation where you assign the value to data, like:

            data[6:10] = from_float(to_float(X_pos) + 5)
            #data[6:10] = from_float(to_float(X_pos) + X_rel_pos)
            data[10:14] = Y_pos
            data[14:18] = from_float(to_float(Z_pos) + 5)
            #data[14:18] = from_float(to_float(Z_pos) + Z_rel_pos)
```where the old calculations are commented out

your car will probably just be stuck on one side on the diagonal according to the map orientation and move with the target but not correctly when the target rotates
it will be like a static offset relative to the cars position but not take into account the direction of the car (other then our rotation) for positioning, that will be stuck with the world map

this is just to debug things

[elder cobalt]

I feel like I've tried this before

[elder cobalt]

and this was exactly the outcome

[cloud spruce]

and the distance will be more then 5 since it's on the diagonal, it will be around 7.07107 for distance

[elder cobalt]

wait so if I do know the drivers position in relative to the car's body that could be one way to find the center of the car?

@cloud spruce okay if I exclude the center of the car and maybe it's centered around the driver is there anyway to like mirror?

That way I always stay in front of the driver even if I turn or something?

[cloud spruce]

it you are okay with putting your drivers position right in front of the targets driver position we don't have such a big problem
it will just look a little bit weird since the car bodies will not align

[elder cobalt]

Well gotta sacrifice sometimes

But yeah for now I'm fine with it

[elder cobalt]

I guess the code mentioned here should be fine or should there be any changes?

eh guess I'm done for today

[cloud spruce]

i'm trying to mock things over here, but it's hard without any example data

[cloud spruce]

haha, okay

[elder cobalt]

gotta hit the bed

😭

but atleast found out what's possible and what's not

[cloud spruce]

to be continued then

[elder cobalt]

maybe once I wake up tomorrow if I have some time left I could look into the game's memory to find if it even loads up other car's heading

surely a client should know where the other car is facing if isn't sending packets maybe it's doing it within game itself but that'd take alot of years

luckily I have a debug file but I'm so bad at reverse engineering got to see what I can do

[cloud spruce]

guess i'll be seeing you here tomorrow, until then 👋

[elder cobalt]

@cloud spruce iirc the game has some sort of prediction system filled in place as well so gonna try fiddling with velocity while my xyz pos are the same

[elder cobalt]

@cloud spruce asked a friend of mine he replies

[cloud spruce]

yeah, that's kind of what it though too, velocity shouldn't have anything to do with it other then that we move with the other car
that is, we should just copy all the velocity values from the target to our own car without any manipulation, i think the velocity is in relation to the map (just like position) and not in relation to the rotation of the car

[elder cobalt]

I've decided to try out hispos + (hisvel * distance) it somewhat worked but it was really janky

Where I asked again how to get hisfwd and he replied you get it based on the rotation

I've checked on the internet many say you normalize the Z rotation which finally made sense when you added in Z_rot but I didn't get this calculation

https://stackoverflow.com/questions/61520151/normalise-all-vectors-on-the-z-axis

[elder cobalt]

@cloud spruce any idea if we can use math function to normalize Z rotation between -1 or +1?

wait it's already being done

but it looks like it's only like adding + 5 to the current position

when I use

def relative_position_2d(distance: float, rad_angle: float) -> tuple[float, float]:
    #return distance * math.sin(rad_angle), distance * math.cos(rad_angle)
    return distance * math.cos(rad_angle), distance * math.cos(rad_angle)

[cloud spruce]

your not really using that function right now if you still have this debug code in use:

            data[6:10] = from_float(to_float(X_pos) + 5)
            #data[6:10] = from_float(to_float(X_pos) + X_rel_pos)
            data[10:14] = Y_pos
            data[14:18] = from_float(to_float(Z_pos) + 5)
            #data[14:18] = from_float(to_float(Z_pos) + Z_rel_pos)
```to use it again switch which statement is commented out and not

    return distance * math.cos(rad_angle), distance * math.cos(rad_angle)
```is wrong, it should be:
```python
    return distance * math.cos(rad_angle), distance * math.sin(rad_angle)

[elder cobalt]

when I corrected the statement output shows

    my pos(x: 653.6578938854869, y: 89.28119659423828, z: 1410.950487438022), rot (x: 4.261828215914317, y: 0.023133646696805954, z: 0.06406708806753159)
target pos(x: 648.6681518554688, y: 89.28119659423828, z: 1410.63037109375), rot (x: 1.1201612949371338, y: 0.023069046437740326, z: 0.06374324858188629)
    my pos(x: 653.6579972901552, y: 89.28119659423828, z: 1410.9488715457721), rot (x: 4.261753948526927, y: 0.023069046437740326, z: 0.06374324858188629)
target pos(x: 648.6681518554688, y: 89.28119659423828, z: 1410.63037109375), rot (x: 1.1201030015945435, y: 0.02300095371901989, z: 0.06352806091308594)
    my pos(x: 653.6580657119953, y: 89.28119659423828, z: 1410.9477977851993), rot (x: 4.261695655184337, y: 0.02300095371901989, z: 0.06352806091308594)
target pos(x: 648.6681518554688, y: 89.28118896484375, z: 1410.63037109375), rot (x: 1.119991421699524, y: 0.022939220070838928, z: 0.06317082047462463)
    my pos(x: 653.6581787912351, y: 89.28118896484375, z: 1410.9460151659682), rot (x: 4.261584075289317, y: 0.022939220070838928, z: 0.06317082047462463)
target pos(x: 648.6681518554688, y: 89.28118896484375, z: 1410.63037109375), rot (x: 1.1199955940246582, y: 0.02289349026978016, z: 0.06316276639699936)
    my pos(x: 653.6581813332951, y: 89.28118896484375, z: 1410.9459749758937), rot (x: 4.261588247614451, y: 0.02289349026978016, z: 0.06316276639699936)
target pos(x: 648.6681518554688, y: 89.28118896484375, z: 1410.63037109375), rot (x: 1.1200497150421143, y: 0.022851286455988884, z: 0.06337600946426392)
    my pos(x: 653.6581139195009, y: 89.28118896484375, z: 1410.9470390579015), rot (x: 4.261642368631907, y: 0.022851286455988884, z: 0.06337600946426392)```

yikes

there a bit readable

target pos is at 648, but because of the way its facing it should be at 648-5 which is 653 but instead it's at 648+5

and Z position is not really changing as well, it should be 1410+5 = 1415 from the way its facing

[cloud spruce]

are you using this code now?

            data[6:10] = from_float(to_float(X_pos) + X_rel_pos)
            data[10:14] = Y_pos
            data[14:18] = from_float(to_float(Z_pos) + Z_rel_pos)

[elder cobalt]

yep

also just in case

[cloud spruce]

could you also update the two print statements to this:

print(f"target pos(x: {to_float(X_pos):>12.3f}, y: {to_float(Y_pos):>12.3f}, z: {to_float(Z_pos):>12.3f}), rot (x: {to_float(X_rot):>12.3f}, y: {to_float(Y_rot):>12.3f}, z: {to_float(Z_rot):>12.3f})")
```and for the last one:
```python
print(f"    my pos(x: {to_float(X_pos) + X_rel_pos:>12.3f}, y: {to_float(Y_pos):>12.3f}, z: {to_float(Z_pos) + Z_rel_pos:>12.3f}), rot (x: {to_float(X_rot) + math.pi:>12.3f}, y: {to_float(Y_rot):>12.3f}, z: {to_float(Z_rot):>12.3f}), rel_pos(x: {X_rel_pos:>12.3f}, z: {Z_rel_pos:>12.3f})")
```it should make the output even more readable independent of the position on the map (edit: i shortened the name of `relative_pos` to `rel_pos` in the last snippet)

[elder cobalt]

getting this error

for the second print statement

the first works fine

[cloud spruce]

ah, i forgot about one thing

there, i edited the last one now, the X_rel_pos and Z_rel_pos are already floats and don't need to be wrapped in to_float()

[elder cobalt]

took 3 snips

from start, in middle and end @cloud spruce

[cloud spruce]

looks like you are almost stationary, is that correct?
just moving a few meters and rotating a bit

[elder cobalt]

Yeah I'm stationary at the first snip

[cloud spruce]

i think i know what's wrong! 💡

[elder cobalt]

After that I started rotating in a circle

[cloud spruce]

i see it in the numbers now

we are using the angle from the wrong rotational axis, change:

X_rel_pos, Z_rel_pos = relative_position_2d(relative_distnace, to_float(Z_rot))
```to:
```python
X_rel_pos, Z_rel_pos = relative_position_2d(relative_distnace, to_float(X_rot))

we are turning around the X axis to change the heading of the car, not the Z axis

that's also the variable that we flip 180 degrees or 1 pi in radians to reverse the car relative to the target

@elder cobalt how is it working now?

[elder cobalt]

Loading the game up, took a short break but will get back to you in a bit 👍

[elder cobalt]

But any idea behind why X rotation? With the return distance * math ... Function won't it just multiply my distance into cos(X_rotation)? Then add it to Car B position?

So for the sake of simplicity let's keep relative distance at 1 so it'll be cos(3.14159) when my cars at 180 degrees which is 0.99849715 and then add that to my car B position?

[cloud spruce]

X_rot is the heading of the target, so it will change all the time when the target turn left and right
distance * cos(X_rot) and distance * sin(X_rot) will give you the offset X and Z coordinates that your car need to be at in relation to the target to always be that distance in front of the target regardless of where the target is pointing around the X_rot axis

i think you will see exactly how behaves with the debug print statements on the console as you turn the target car around

[elder cobalt]

hmm no luck with changing to X_rot as well

car's facing like this

[cloud spruce]

oh, now i think we need to do one more change

[cloud spruce]

can you paste the code you have right now for the definition of the relative_position_2d() function?

[elder cobalt]

def relative_position_2d(distance: float, rad_angle: float) -> tuple[float, float]:
    return distance * math.cos(rad_angle), distance * math.sin(rad_angle)```

[cloud spruce]

try to just switch cos and sin

[cloud spruce]

and if you take a picture of the cars in relation to each other, how does it look?
i'm wondering if you ended up in front or behind the other car (depending on the world coordinate system in the game)

[elder cobalt]

looks like it ended up similar to before

[cloud spruce]

could you paste your code that you have right now to https://paste.pythondiscord.com/ ?

[elder cobalt]

sure

[elder cobalt]

https://paste.pythondiscord.com/axitojejiy

[cloud spruce]

after you went to sleep i found out why data[1:2] works and not data[1], it's because we are working with bytearrays, if you access just one element python returns that elements decimal value instead of the byte but when using slices it returns a new bytearray but holding just one byte in this case 🤦

so, you will need to use slices everywhere when working with the bytearrays, so that needs correcting again as i had missed this detail with bytearrays when i edited the code

[elder cobalt]

ah I see, not an issue

hmm so it looks like when I move in a diagonal (I assume this because both my X and Z coordinates from world position are either both decreasing or increasing when I move) my X rotation is positive but my X and Z position coordinates are negative(or decreasing as I move)

[cloud spruce]

does the world coordinates and rotation that the game print out match what you see in the console debug output from the script for the "my pos(..." lines?

[elder cobalt]

yeah

current world position

this is of the red bmw

[elder cobalt]

now if I were to move to the coordinates of the teleported car

[cloud spruce]

this was the world position for the target car, what about your car (teleported one)?

they match the console output as well?

is the red car the teleported car or the target car?
i'm guessing it's the teleported car that is red, right?

[elder cobalt]

red car is the target car, the orange car is teleported

[cloud spruce]

oh, then i had that totally backwards 💡

[elder cobalt]

also yeah they do match the world position

[cloud spruce]

if you were to turn off your script and position the orange car approximately where you would expect it to be in relation to the red car without moving the red car from where it is right now and then give me the world coordinates and rotation of both cars i think i can work out what's wrong with the calculation for the relative positioning

[elder cobalt]

regarding this can't we just do a check where if X_rot is positive keep subtracting 1 from the target car * relative distance? and vice versa?

[elder cobalt]

sure

[elder cobalt]

let me start from first

[cloud spruce]

i don't think that would work when the car moves around and turns on the map

[elder cobalt]

position and XZ rotation of the current red car bmw

[cloud spruce]

does the game also report the rotations like that?

[elder cobalt]

nope need to check it through wireshark since it sends those to server

[cloud spruce]

okay, but if you position the orange car what you think is about five meters in front of the red car and facing each other, what world position does the orange car report then?

[elder cobalt]

orange teleported car

[cloud spruce]

i mean if you have your script turned off and just drive there with the orange car and see what the orange car has as it's world position

[elder cobalt]

5m in front of this should be

[elder cobalt]

yeah I turned off the script and just drove there

only for rotation I enabled a separate script which just reads my rotation thats it

not manipulating it

[elder cobalt]

regarding this...

[cloud spruce]

okay, but in the picture it doesn't look like the orange car is positioned in front of the red car

[elder cobalt]

5m in front of the red car

[cloud spruce]

perfect, then i would also need the rotation reported right now from the red car, can your other script that doesn't modify the position print that value?

[elder cobalt]

its this @cloud spruce

my red car is in the pits and the pits location is the same if I teleport back to it

and I haven't even moved from it

[cloud spruce]

so, which value is the X rotation for the red car when it's positioned just as it is right now?

[elder cobalt]

this is X rotation

of the red car

as it is right now

[cloud spruce]

perfect, then i'll just crunch some numbers to try to work my way backwards to what we need for the calculation

[cloud spruce]

is there any compass or heading in the game when driving the car or a picture of the map with the car on it in relation to the map or other indication of where north on the map would be in relation to the car?

[elder cobalt]

unfortunately no 😔 got to access the shared memory for that

[cloud spruce]

alright, i'll have to try to reverse engineer how the world position coordinate system and cardinal direction and car rotation relates to each other then

[elder cobalt]

would sharing you the .pdb file of the game be of any help?

[cloud spruce]

don't know if i could use it on linux as i don't have windows and don't have the games exe and dll files anyways

[elder cobalt]

ah yeah mb, need the executable of the game first

[elder cobalt]

uhh @cloud spruce have you heard about or tried kaitai?

[cloud spruce]

no, but looks really cool 👍

[elder cobalt]

O well I do most of my packet decoding with the online IDE but looks like I was in for a surprise when I was checking how things were working with modded server packets

I found out that I haven't decompiled the entire packet, it looks like modded server packets send bigger udp packets which contain information of other players in one single packet as a way to optimise

Thankfully it's repeatable I'm not sure tho how I could possibly do that with python but I've deconstructed the packets using kaitai into something more meaningful

[cloud spruce]

do you have a spec for the packet format of both the standard and the modded servers?

[elder cobalt]

Yeah

[cloud spruce]

could i have the url?

[elder cobalt]

Sure hold on

[errant bay]

Hey @elder cobalt!

It looks like you tried to attach file type(s) that we do not allow (.zip). We currently allow the following file types: .gif, .jpg, .jpeg, .mov, .mp4, .mpg, .png, .mp3, .wav, .ogg, .webm, .webp, .flac, .m4a, .csv, .json.

Feel free to ask in #community-meta if you think this is a mistake.

[elder cobalt]

oops

https://files.catbox.moe/4y8fxv.zip @cloud spruce just drag and drop the contents to https://ide.kaitai.io/#

This is just to get you started if you ever plan on using this in the future https://doc.kaitai.io/user_guide.html

should look somewhat like this(I've deleted some files but you could have a look at them)

client.ksy for client2server.bin
serverSTOCK.ksy for server2clientSTOCK.bin then,
serverMOD.ksy for either server2clientMOD.bin or need2decode blablabla

[cloud spruce]

thanks for making me aware of Kaitai Struct, i think i'm going to find uses for it now and then 🙂

[honest drift]

I'm sorry for making such a dumb question, but I couldn't find the answer I was looking for on the internet.
What is the use for network programming?
Why is it so important?
What would be an example irl that you'd need it?

[cloud spruce]

network programming is what makes devices/computers able to talk to each other and exchange information
without it the internet (and the information age and the world as we know it today) wouldn't exist and you wouldn't be able to surf the web, search google, watch youtube, stream music, movies and series, read your e-mail, use social media and chat, make phone calls (other then as they did back in the day when you first talked to an operator who physically connected phone lines with wires for you to get to whom you wanted to talk to) or video calls
network programming enables all that, even if you have not written the code someone else has
you will need to learn it if you want to write your own code/program/application that exchanges information between devices

[cloud spruce]

i think i solved it now, and if i'm right it was stupidly easy 🙃

[elder cobalt]

o what was the fix this time?

coz I just opened up a debugger for checking if I can find the direction of my car

[cloud spruce]

we have to subtract a rotation constant because first the X rotation is at a 90 degree angle, then we need to adjust a bit more because the center seem to be centered around the driver instead of the middle of the car body

[elder cobalt]

yeah

[elder cobalt]

found something interesting

[cloud spruce]

how about this?
https://paste.pythondiscord.com/ezajegateh

[elder cobalt]

trying it out will let you know in a few mins 👍

[elder cobalt]

should I just

[cloud spruce]

haha, the problem with not being able to run the code at all 😆

[elder cobalt]

X_rel_pos, Z_rel_pos = relative_position_2d(relative_distance, to_float(X_rot) - X_rot_fix)

[cloud spruce]

yeah, that is the fix for that

[elder cobalt]

hmm tried it out no luck looks like it's still retaining the same old position as before?

[cloud spruce]

that is so odd, same map and same position of the red target car?

[cloud spruce]

aha, i think the rotation goes the other way around, try adding instead of subtracting

[elder cobalt]

almost there

[cloud spruce]

and in that case, that probably goes for the subtraction in line 31, might need addition there instead

[elder cobalt]

line 31?

o X_rot_fix?

[elder cobalt]

replaced it with an addition

[cloud spruce]

try value 0 for X_rotation_adjustment

[elder cobalt]

yep

car's always in front it's a good sign

[elder cobalt]

yep looks like it

[cloud spruce]

we might be a little bit off still, so you can adjust that angle in that variable to fine tune it (it's in degrees and you can use decimals and negative numbers too)

[elder cobalt]

I suppose you use relative_distance to define how far apart the car is?

[cloud spruce]

yeah

[elder cobalt]

what about x_rotation_adjustment whats that used for?

sorry

[cloud spruce]

that is to do the small adjustments to center the care head on, since the center of both cars seem to be the driver and not the car body

[elder cobalt]

perhaps rndpkt with the calculations you did won't it be possible to know where the car is facing?

[cloud spruce]

it's in degrees and you can use like -1.5 or 3.2 or what you need to fine tune the relative position

[elder cobalt]

like cardinal direction?

[elder cobalt]

ah I see

[cloud spruce]

probably, i think it's might just be something like math.degrees(to_float(X_rot) + math.radians(90)) to get the heading of the car in degrees, it depends on if rotation in the game is clockwise or counter-clockwise
if you want it, but i don't know what for
i think 0 should be north, 90 east, 180 south, 290 west after that adjustment
or if we are 180 degrees wrong, then you subtract instead of add in that calculation to fix it
and if the game rotation goes counter-clockwise we would need to flip that number around to fix it

[elder cobalt]

kinda sucks that there's no way to know where your facing in game until you access shared memory

[cloud spruce]

but does it matter to you which cardinal direction you are facing in the game, and you use that information in any way if nothing depends on it?

[cloud spruce]

is there a map in the game?
in that case, is it fixed to north or does it rotate freely with the heading of the car?

[cloud spruce]

by the way, i added teleport = True that you can set to False if you just want to see everything printed (like what it would do) but not affect the car in the game

[elder cobalt]

maybe in future I might have some use for it as it would be really handy to know where I am facing

something like a compass

[cloud spruce]

okay, but only as console output on the side then?

[elder cobalt]

yeah

in case something breaks down

[cloud spruce]

i wonder if they have done some real world modeling in the game when it comes to light direction 🤔

[elder cobalt]

iirc you could set the angle of the sun in game

[cloud spruce]

i mean, if they set the correct angle of the sun depending on were on earth the city that you drive in is located (say the map should be located in Monaco) and the date of the year and time of day that the events in game should play out in

[elder cobalt]

ah nope

[cloud spruce]

okey, i see

[elder cobalt]

not sure if these could be of any use?

[cloud spruce]

it is, since you can set the location of the sun in degrees in relation to the map (heading in there), the height angle doesn't matter much as long as it's not to much overhead, and angle of 20-35 degrees over the horizon is probably fine
we should be able to point the front of the car towards the sun a few times when the sun is in different places and check the X_rot value to see how they relate to each other in game and work out if X_rot is 180 degrees reversed or not and if the whole map is also 90 degrees on it's side and work out the cardinal directions from there

but first we would need to know something more basic, like if X_rot is clockwise or counter-clockwise in the game
and that's easy to check by you just getting the X_rot of the car and then turning the car like 45 degrees to the right (or left, as long as i know which one you did) and then getting the X_rot again

[elder cobalt]

so far I guess the games finding the cardinal direction based on your X-axis rotation?

hmm

[cloud spruce]

yes, the direction of the car that is
but if you want to translate it to real cardinal directions like N, NNE, NE, ENE, E and so on there is some more work to be done

[elder cobalt]

will get back to you

[elder cobalt]

@cloud spruce good news found a car in game which has an inbuilt compass

took a bit of searching around for old mods

[elder cobalt]

This is from

Hmm what type of correlation tho

O

60 degrees difference

orange car

[cloud spruce]

haha, nice find! 👍

[elder cobalt]

this seems sorta confusing tho, let's say the difference between heading and rotation of X axis is 60 degrees

so when heading's negative rotation is positive and vice versa

[cloud spruce]

what happens with the X_rot value when you start turning right, does it increase or decrease?

[elder cobalt]

hmm let me try real quick

[cloud spruce]

also, somewhere it will flip from positive to negative and when you are close to 0 degrees in-game (which i think should translate to N) what is the X_rot value?

[elder cobalt]

a mb for taking too long, it looks like the server is having a seizure gonna ping you once I can come back with some readings

[cloud spruce]

i also wonder if different maps can have different rotations in regards to cardinal direction and if there is something in any network packets to indicate the map rotation in that case
or if all maps have the same rotation, i think that would make most sense

[elder cobalt]

I'll load in a server with nothing but just a flat plane that should help getting things started

[elder cobalt]

Hmm not sure

[cloud spruce]

if the offset between X_rot and north would be 1 radian it would translate to about 57.2958 degrees, can that be what you are seeing?

[elder cobalt]

Yeah

Basically rounding it off to 60 degrees

[elder cobalt]

So it looks like

Car's facing straight?

ok it actually corresponds to this nice

so east is -90 degrees, -180 and +180 might be south, +90 must be west then 0 and -0 should be North

[nova cloud]

Well yes but also it's easier to work with only degrees between 0 and 360 so I recommend adding 360 to the number until it's positive that way you don't have to keep in mind the number might be negative

[elder cobalt]

hmm yeah fair

thinking of doing that once I get to know NSWE

so it looks like

-90 = East = 90 in map
0 = South = -180 to +180 in map
90 = West = -90 in map
-179.9 to +179.9 (rounding it to 180) is = North = 0 degrees in map

hmm time to convert it to an actual compass

o wtf

the game does not go to 360 heading as well

I guess it'd be pointless to convert it to a full 360 might confuse the game

I guess implementing a cardinal direction : it's heading degrees will be better

[tired river]

Can anyone tell me is CCNA certificate important out there or not?

in recruitments u k

[cloud spruce]

depends on the company recruiting, i have never given much stock to any certifications as our company don't need them, for a consultancy company which offers support contracts to customers or a reseller it can be a requirement either that you have them or get them soon after employment
we had a similar discussion about certifications in #cybersecurity a while back which then was generalized to most higher certifications within IT, it's around this point in the chat history of that channel: #cybersecurity message

[honest drift]

thank you 🙂

[wanton dock]

do guys have an idea how for an program in the field of networking?

[cloud spruce]

what do you mean?
the network field is huge/broad, you would need to be a little bit more specific then that
what is it you would like to do?

[wanton dock]

so i coded a litle file server (more or less a cloud) wich was a great project but only lastet for 2 days. an this hooked me really to servers. but now i dont have any idea of an project wich gets me deeper into the topic and is also fun to program

[cloud spruce]

it's impossible for us to say what you will think will be a fun project or a project on the right level (not too hard and not too easy either), which makes it very hard to come with suggestions 🤷
it's best if you pick something you are passionate about or something you feel you are lacking (that you are in need of), that will make it easier to find motivation to work on the project

[wanton dock]

huh... ok thanks!

[elder cobalt]

def compass_direction_name(angle: float):
    angle = angle % 360
    octant = round(angle / 45)
    return {
        0: "S",
        1: "SW",
        2: "W",
        3: "NW",
        4: "N",
        5: "NE",
        6: "E",
        7: "SE",
        8: "S"
    }[octant]

print(compass_direction_name(math.degrees(struct.unpack('<f', X_rot)[0])))

You think these are enough to get a compass?

here's a rough sketch

Hmm now I need to figure out it's heading 😅 took a while to make the compass and correct it @cloud spruce

[cloud spruce]

is east really -1.57 as the X_rot of the car?
i though it would be west that would be - 🙃

[elder cobalt]

Yeah

[cloud spruce]

so, the X_rot is counter-clockwise then?

[elder cobalt]

-90 degrees of X_rot is 90 degrees in game heading

[cloud spruce]

haha, okay, that feels backwards, but okay 🤷😆

[elder cobalt]

let me just recheck for the rest

[cloud spruce]

what is X_rot from the packets in radians if you are heading as close to north as you can?

[elder cobalt]

hold on

[cloud spruce]

like, without any adjustments, the raw X_rot of the car in radians

[elder cobalt]

true north is supposed to be -3.14 or 3.14

if it goes beyond -3.14 it jumps to +3.14 and vice versa

its just that because the car keeps moving I can't get the exact

3.14 radians

but very close

[cloud spruce]

and east is then close to -1.57?

[elder cobalt]

Yeah

[cloud spruce]

now i see, it's because everything is flipped around 180 degrees or 1 pi radians

[elder cobalt]

Exact -1.57

Hmm you think we would be able to get heading in degrees with this?

Problem is heading is supposed to be calculated till 360 maybe I'll put an indication saying that if it's 360 it's (0) and if it's 270 then it indicates (-90)?

[cloud spruce]

yeah, no problem, now when we know how the X_rot in the data packets works in relation to the game world

well, 0 = 360, usually it's 0 .. 359,99999... and then you wrap around to 0 again

[elder cobalt]

Hmm any idea on how to implement? My rough guess is we take X_rot and add or multiply to it? @cloud spruce

[cloud spruce]

try:

math.degrees(to_float(X_rot) + math.pi)

[elder cobalt]

yep working perfectly!

[cloud spruce]

taking + math.pi removes any negative numbers and turns the heading so that 0 radians is north, 0.5 pi is east, 1 pi is south and 1.5 pi is west

[elder cobalt]

also any idea how to maybe format it such that output shows Facing: {cardinal_direction}, Heading: {heading from 0-360} Then some sort of a suggestion pointing out that in-game it's -90 degrees/+90 degrees?

using these two print statements atm

        print(compass_direction_name(math.degrees(struct.unpack('<f', X_rot)[0])))
        print(round(math.degrees((struct.unpack('<f', X_rot)[0] + math.pi))))

really bad when it comes to formatting stuff into something readable 😅

[cloud spruce]

where you use struct.unpack('<f', X_rot)[0] you can do to_float(X_rot) instead to make it easier on the eyes

if you still have that function defined in your file

and you can define your function as def compass_direction_name(angle: float) -> str: if you want full type hints

[elder cobalt]

does this look clean?

[cloud spruce]

sure, but does it work all 360 degrees around?

[elder cobalt]

Yep 👍

now I guess it's time for me to start decoding modded server packets 😩

Thanks for helping me out soo far @cloud spruce

really mean it 🙂 wouldn't have got this far without any of your help

[cloud spruce]

you're welcome 🙂

[elder cobalt]

Oh btw have you had a look with kaitai?

it looks like for modded packets there's an identifier after the type which indicates how many updates are there in that one single packet

only problem is that now I need to say python to look for the numberOfUpdates byte packet then based on that search through the packets with the sessionID which you are trying to find then splice the data accordingly

problem is that how would I say python to make sure it searches for the right data 🤔 then say python that entire struct is belonging to this sessionID?

isn't as simple as the + offset but not sure

fortunately it follows the same structure

[cloud spruce]

i have, but not in this detail for that specific issue, hold on and i'll have a look

but i must again say that i really liked kaitai 🙂

[elder cobalt]

ah I see, I just got to know about kaitai a year or two ago but haven't really used it until now

was a breeze to setup things and start working with the packets though

[cloud spruce]

with pydivert, are you allowed to:
use packet.payload[9:13] directly?
assign things directly to it like packet.payload[9:13] = X_pos?

or are you required to make it into a bytearray first?

[elder cobalt]

need to assign it as a bytearray first

data = bytearray(packet.payload)

[cloud spruce]

are you a 100% sure you can't even access them the other way without first making it in to a bytearray?

i'm just checking if there is another way to do a few things which would be an advantage when processing the modded server packets

[elder cobalt]

you mean manipulating the data without the bytearray?

[cloud spruce]

or at least read the data when it's data from the server

[topaz dune]

Heyyy
Anybody knows a way to send files from my pc to a friend(different network) through terminal?

we both have linux

[elder cobalt]

Hmm not sure but maybe something like this?

import pydivert

with pydivert.WinDivert("tcp.DstPort == 80 and tcp.PayloadLength > 0") as w:
    for packet in w:
        print(packet)
        w.send(packet)
        break

[cloud spruce]

there are lots of ways, but it would require on of you to act as a server and have either a public ip address or do port forwarding in the home router of the person running the server side

[cloud spruce]

i can't test you know, as i don't run windows i can't run pydivert as it's windows only, that's why you have to do all the testing for me 😄

[elder cobalt]

Ah I see what exactly do you want? Maybe first starting the filter then? Should I like try see if there's a way to print out the payload of the packet?

[topaz dune]

nice. No idea yet how to do that, but I will try

[elder cobalt]

@cloud spruce ```python
import pydivert

with pydivert.WinDivert("ip.DstAddr = 87.106.198.64 and udp.PayloadLength = 62") as w:
for packet in w:
print(packet.payload)
w.send(packet)
break

Get the output as

```b"FBQ\x90\x02\x00'\xb8\x13D\xcd\xe9\xb5B.\xd8\xb3D\x14\x10\x06\xc0\xaf\xb6n\xbd\xb8\x08\x1b\xbd@Z.\xbb\xbfh\x8b:\x13\xd5q:dddd\x7f\x7f\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\xd6\x94m?"```

looks good?

[elder cobalt]

Oh bruh was doing just that here

sorry didn't look through your message properly

[cloud spruce]

and that works if you print X_rot?

[elder cobalt]

Yes

That's how I got the compass to work

[cloud spruce]

you did that without going through the bytearray?

[elder cobalt]

Yep

[ember ledge]

hi what to learn here. what about networking if I may ask

[cloud spruce]

nice, then we can use another pretty cool python feature to access each packet segment 🙂

[ember ledge]

is it like "hackers" here?

[cloud spruce]

nope, unless you go back to the original meaning of the word "hacker" which meant something along the lines of "tinkerer" and would today probably be more in line of what is now called a "maker" 😉
at the core what a real hacker is hasn't changed, it's a person that takes a thing that is made to do something(s) and make it do something else that it wasn't originally made to do but you would like it to do instead
that is basically what today's definition of hackers is still about if you think about it and filter out the pop culture around the word

[elder cobalt]

Oh damn that's pretty interesting, unfortunately need to head to bed again so cya take care

[ember ledge]

well worded

[elder cobalt]

I still recollect a quote somewhere stating that it takes only one programming language to master and you can kick start with anything network related.

To be honest, I'm glad I've taken python partially because of how it was taught at school but more over on how easy it is to understand and how versatile it is, idk I've been thinking about getting into lua or c/c++ down the line but python just keeps pulling me back, maybe because of how I'm somewhat familiar with it and if I ever wanted to integrate with any other field maybe not solely into computer software but hardware related it'd be easy to do so with a bit of tinkering

To be short I just like tinkering with stuff on the hardware level, packets and networks grasped my attention after learning how easy it is to get started with python

[undone harbor]

does anyone know how to create sftp server with python

[elder cobalt]

yikes I guess I found an issue with my compass

game heading does not correspond at all

game_heading = round(math.degrees((to_float(X_rot) * -1)))

might be because of the * -1

[cloud spruce]

i think we can do something about that

[elder cobalt]

for instance

car rotation in rads

the compass heading works fine tho

[cloud spruce]

remember that game heading is flipped 180 degrees

[elder cobalt]

yeah

so * -180?

woops

[cloud spruce]

wasn't it:

game_heading = round(math.degrees(to_float(X_rot) + math.pi))

[elder cobalt]

thats for a full 360

but game calculates between 180 to -180

[cloud spruce]

no, math.tau = math.pi * 2 = 360 degrees in radians

[elder cobalt]

real_heading = round(math.degrees((to_float(X_rot) + math.pi)))

I use this variable to show from 0 to 360

[cloud spruce]

the + math.pi will make the compass go from 0 - 359 degrees instead of -180 - +180

[elder cobalt]

a is there anyway to show between -180 to +180?

[cloud spruce]

oh, you want the negative heading as well?

[elder cobalt]

yeah

[cloud spruce]

but you want it flipped so that 0 is north instead of south?

[elder cobalt]

180 is north and 0 is south

[elder cobalt]

here

well I made a dict not sure if I can use the same here

def compass_direction_name(angle: float) -> str:
    angle = angle % 360
    octant = round(angle / 45)
    return {
        0: "S",
        1: "SW",
        2: "W",
        3: "NW",
        4: "N",
        5: "NE",
        6: "E",
        7: "SE",
        8: "S"
    }[octant]```

[cloud spruce]

yeah, but do you want it the other way around?

[elder cobalt]

nah want it with respect to game so that's -90 which is east, 90 is west, 0 is south and 180 is north

ig I drew the diagram wrong when I was editing it

north should be up and south should be down

there

[elder cobalt]

Hmm any luck? @cloud spruce maybe should I like create some conditions for + and - events?

[cloud spruce]

i would probably do it in another way, and not use a hash and do the calculations directly in radians as that is what the data is originally to save some precision and cpu time not having to do the conversion from radians to degrees

[halcyon lily]

Hi, I'm not sure if I'm posting in the good channel. I tried positing my question in a help channel but got no answer. I have something I don't understand, mostly because I'm a beginner with sockets. I wanted to know if there is any possibility that a recv could be blocking (or stuck) with a socket put in non-blocking state, when having check before with a select if there was data

#help-avocado message

This is the link to my question if by any chance someone have an idea

[elder cobalt]

Huh so I guess doing a comparison with radians itself and then converting them to degreea

[cloud spruce]

a non-blocking socket should never block
are you on windows?
select() isn't very good on windows and the same goes for non-blocking sockets in windows if i remember correctly

[halcyon lily]

I'm on macOS

I am checking in a loop with select if there is any data in reading for my socket. then if there is, I'm doing a recv. My problem happens like 1 or 2 times in a day

So it's not something happening each time and it's very hard to reproduce

If the socket is stopping from the other side, while doing a recv, I'm expecting to have an error that I would catch, but it seems I'm stuck in the recv while having set sock.setblocking(0) so not sure what could be happening here

[cloud spruce]

can't really test, but a complete compass.py might look something like this:

import math

import pydivert

server_address = "146.19.191.145"
server_port = 9948
Session_ID = 12
# 0 = N, 90 = E and so on, others may also go in between as long as they are in order
# cardinal_directions = ('N', 'E', 'S', 'W')
cardinal_directions = ("N", "NE", "E", "SE", "S", "SW", "W", "NW")
# cardinal_directions = ('N', 'NNE', 'NE', 'ENE', 'E', 'ESE', 'SE', 'SSE', 'S', 'SSW', 'SW', 'WSW', 'W', 'WNW', 'NW', 'NNW')

def cardinal_direction(rad_angle: float) -> (str, float):
    return cardinal_directions[math.floor((rad_angle % math.tau) / rad_cardinal_precision + 0.5) % len(cardinal_directions)], math.degrees(rad_angle % math.tau)

def to_float(data: bytes) -> float:
    return struct.unpack("<f", data)[0]

session_id = struct.pack("<B", Session_ID)
rad_cardinal_precision = math.tau / len(cardinal_directions)

with pydivert.WinDivert(f"(inbound and ip.SrcAddr = {server_address} and udp.SrcPort = {server_port} and udp.PayloadLength > 60)") as divert:
    for packet in divert:
        if packet.direction == pydivert.Direction.INBOUND:
            if packet.payload[1:2] == session_id:
                X_rot = packet.payload[21:25]
                cardinal_heading, degrees_heading = cardinal_direction(to_float(X_rot) + math.pi)
                print(f"{round(degrees_heading):3} degrees, {cardinal_heading}.")

        divert.send(packet)

and you can pick which granularity you want for the directions, any of the three should work just fine

math.floor(some_float + 0.5) in there is almost like round(some_float) but deals with numbers like 18.5 and 20.5 differently

[elder cobalt]

ah yikes, I just made my own but will save your copy as well if any case mine breaks out

used

def g_compass_convert(radians: float) -> float:
    if to_float(radians) < 0:
        return to_float(radians) + math.pi
    else:
        return to_float(radians) - math.pi

g_compass = round(math.degrees(g_compass_convert(X_rot)),1)

[cloud spruce]

the relevant bits are in:

cardinal_directions = ("N", "NE", "E", "SE", "S", "SW", "W", "NW")

def cardinal_direction(rad_angle: float) -> (str, float):
    return cardinal_directions[math.floor((rad_angle % math.tau) / rad_cardinal_precision + 0.5) % len(cardinal_directions)], math.degrees(rad_angle % math.tau)

rad_cardinal_precision = math.tau / len(cardinal_directions)

# loop here
    cardinal_heading, degrees_heading = cardinal_direction(to_float(X_rot) + math.pi)
```and you could also go for a 4 or 16 segment compass with that instead of an 8 segment compass

[elder cobalt]

yeah wanted the cardinal directions so bad lol

thanks

[cloud spruce]

that code will give you both the cardinal direction and degrees
degrees are simplest calculated with:

round(math.degrees((X_rot + math.pi) % math.tau))

tau is the constant of a full circle in radians, 2*pi, what would be 360 degrees

[wise jungle]

glad to see rndpkt is still going strong in #networks

[cloud spruce]

haha, it was far from yesterday we chatted, i think it's been well over a month, maybe even two
did you sort out the problems with the latency between your server and clients?
perhaps you move over to async programming instead for your software to cut down on the latency?

[wise jungle]

Haha yea i havent looked into it in a month i got a job its taking up my time

[cloud spruce]

sounds like the development of that voice chat has stalled then

[cloud spruce]

i think your best bet would be to convert the code to use async programming instead of threads or a select() loop if you ever get the time to continue with that project
either way, happy to see you back here 🙂

[neon horizon]

Hello, I'm trying to get into socket programming.. I have trouble getting the server and client to send each other messages smoothly..

here is my code...
https://www.toptal.com/developers/hastebin/ripivezebu.properties

my problem is that when i run the client script, either the client doesn't receive the initial message from the server, or the server doesn't receive what the client has inputed

could anyone please help me figure out what's wrong

[cloud spruce]

first of all, see: #networks message
it's about the construct socket.gethostbyname(socket.gethostname()) that you too are using
i'll continue reading your code, this just jumped out at me

[neon horizon]

ok, so i changed the server ip to 127.0.0.1
i also copy pasted someone else's code to see if it works, and if so, how it's different from my code..

looks like people don't send the length of a message before sending the actual message..

[cloud spruce]

you probably want to do that for the client too, to make sure it connects to the same address that the server is listening on

[cloud spruce]

i have checked you code now
no, you are making some assumptions that just isn't true for the socket api
the recv() function can return 1 or more bytes up to the number of bytes you have specified, it will not block and wait for data until the buffer is full
so you can't do recv(64) and expect it to always be returning 64 bytes, remember that you are reading from an unpredictable stream

i think the problem you are facing is that you get out of sync with the data you expect and the data you get and communications breaks down

you would need to have a read loop that fills a buffer variable in your program until you have 64 bytes and then process the header if you are going to use a fixed 64 byte header

[neon horizon]

i was using the fixed buffer because i thought the recv can only receive a certain amount of bytes vs up to a certain amount of bytes

would i just send and receive as many bytes as i need without the buffer?

[cloud spruce]

that is the thing, the socket api doesn't work like that, you only request the maximum amount of data that you want to receive
but there is no guaranties about how many bytes you will get, as soon there is data you will get the data that is available at that point in time when you call recv(max_bytes)

[cloud spruce]

but you probably want a delimiter (character or string) between messages
or some fixed header, like you have attempted, but implemented in another way

this goes equally for the server and client programs of course

[neon horizon]

why do i need the fixed header in that case?

*what would the header be doing?

[cloud spruce]

you don't have to have a header at all, but you would need something to tell your program when a full message is completely received or else it's just a stream of octets (almost like bytes), but certainly not characters
as you are using utf8 which can be multi-byte characters, one utf8 character can be split between two different recv() calls
how would you know it's safe to do decode() on the received bytes, so that a multi-byte character doesn't end up in between two different buffers that you run decode() on separately?

[wise jungle]

yessir i was thinking about starting async programming as it just saves a lot of time

[opal orbit]

I'm trying to implement a full-duplex connection in python with sockets and I'm struggling with this concept, how am I supposed to know if the socket is sending or recving the data? Because calling the wrong method will cause timeout err

[cloud spruce]

it depends on how you are writing the code: select, threads, asyncio, blocking or non-blocking io, etc
it sounds like you are using normal blocking io without asyncio, but with or without threads and are you using select()?

[elder cobalt]

@cloud spruce just a headsup regarding modded server packets, I'm thinking whether I can say python to maybe search 61 bytes each time that way it can see if it has the required information or not, and if that entire packet does not contain it'll just send it to my client and proceed with the next packet? maybe once it finds the required session_ID change the offsets in a way where it splices 61-62 bytes?

[opal orbit]

I think the select is the easiest method? I've read that there are other more efficient methods but some are OS dependent and basically its just harder to implement

[cloud spruce]

one thread per connection is usually simpler and tend to lead to cleaner code as you only have to deal with one connection in the code path
but if you want to exchange data between clients through the server thread can make it a little more complicated instead (as you then need to transfer data between thread in a thread-safe manner)

old fashioned select() is the least efficient, but on Linux and BSD it's nowadays backed by more modern kernel facilities like epoll and kqueue which makes them more efficient then threads

[cloud spruce]

and by the way, if you're on windows you have IOCP (I/O Completion Ports) which is what should be used there instead if you need performance and/or scalability, even if select() works there it's very inefficient on windows

[opal orbit]

@cloud spruce How would you do it with only threading? Everything I can think off here is quite opposite of clear, like some recursions and cross-connecting classes etc, my head just hurt from thinking about it

[cloud spruce]

do you need any information to be transferred between clients, like a chat, shared world game server?
or is it just data to and from the server, like a web server, file server, mail server, etc?

also, what kind of data is it and how does the flow of data look (a lot of back and forth, just a request and a response, just a stream, or somewhere in between)?

[opal orbit]

@cloud spruce Im trying to do a basic proxy, one connection is just local with binded port, the other is remote addr, i know there can be issues with some of the protocols hence they have to get some things like banners first before the data but I will try to handle it later after the basic mode works

Its pretty ez without the full duplex mode, but some real troubles starts here for me

[cloud spruce]

aha, then you do need multiple sockets per thread even in a threaded server, since you need one on the client side and one on the server side

[opal orbit]

class ConnRemote(Thread):
    def __init__(self, host:str, port:int): #TODO: Multiple target support
        self.host, self.port = host, port
        self.sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        self.sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
        self.sock.connect((host, port))
        print('[*] Connection established:', self.sock.getpeername())
        self.data = io.StringIO() 
        self.bytes = 0

    def run(self):
        #   Read only if the previous data where already read,
        #   after the fileptr is being set to 0 again:
        while not self.data.tell: 
            data = self.sock.recv(4096)
            if data:
                self.bytes += data.write(data)


class ConnLocal(Thread):
    def __init__(self, host:str, port:int):
        self.host, self.port = host, port
        self.sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        self.sock.bind((host, port))
        self.data = io.StringIO()
        self.bytes = 0

    def run(self):
            self.conn, self.addr = self.sock.accept()
            while not self.data.tell():
                self.bytes = self.data.write(self.sock.recv(4096))


class ConnArt(Thread):
    def __init__(self, lhost:str, lport:int, rhost:str, rport:int):
        self.lhost, self.lport = lhost, lport
        self.rhost, self.rhost = rhost, rport

        self.conn1 = ConnLocal(lhost, lport)
        self.conn2 = ConnRemote(rhost, rport)

    def run(self):
        self.conn2.sendall(self.conn1.data.getvalue())
        self.conn2.data.seek(0)

        self.conn1.sendall(self.conn2.data.getvalue())
        self.conn1.data.seek(0)
    
    def cleanup(self):
        self.conn1.shutdown()
        self.conn1.close()
        self.conn2.shutdown()
        self.conn2.close()
    
    def stats(self):
        return self.conn1.bytes, self.conn2.bytes

Idk I guess im just too dumb, it feels bit easier to work with sockets in old c though, more explicit

[cloud spruce]

i think you would get problems if you use separate threads for the local and remote connection, then you need a queue to transfer data between them in a thread-safe way
https://docs.python.org/3/library/queue.html
this is what i meant that threads can be kind of a problem if you need to share data between connections/threads
you might want to keep the pair of connections in the same thread and do select() between them instead
or just drop threads all together and switch over to asyncio, i think that would be the best option for you really if you are willing to take the plunge and get in on async programming (if you haven't already)
besides the standard library module asyncio you could check third party modules like Trio which also has awesome documentation for learning async programming and concepts
asynchronous programming is the most scalable option (and doesn't have the overhead of threads and no problems with thread safety) you have if your program is mostly io bound

[cloud spruce]

as you can see there are lots of options to choose from, i think you'll like the asynchronous programming paradigm when you get the hang of it and is very well suited for your use case with the proxy server

[opal orbit]

Okay, thanks, there is a lot to look into, I think I'm gonna try building both threading and async versions for learning purposes. I was taught that threading is superior, but I guess asyncio has its strong points too

[cloud spruce]

to be blunt, threading is quite inferior for I/O bound workloads (more or less regardless of language) compared to async io implementation due to the overhead incurred by starting threads and context switching done by the OS, add to that the total lack of control of when context switching and execution of a thread happens
threading in python is also unsuitable for cpu bound loads due to pythons GIL, for such loads you need to go with multi-process when using python
unless you are calling C/C++ libraries that explicitly release the GIL for some of it's work, which they can't do while exchanging data with the python parts of the code

[elder cobalt]

@cloud spruce guess I'm pretty much stuck with decoding modded packets lol, can't really find a way to implement it the way I imagine

can't seem to splice particular parts of the data payload

[cloud spruce]

i'm on my mobile right now and without access to a computer for a few more hours

[elder cobalt]

Ah no issues, just thinking if maybe you might've found a much better solution than mine lol

Mine just doesn't work 😔

[shy lion]

hello, can i ask here somthing about netmiko and linux ?

[cloud spruce]

just ask and we'll see if anyone can help

[edgy scroll]

Is there anyone work as a freelancer? I wanna start a freelance career. I would appreciate it if you guys help me out.

[ripe spear]

Go and find customers who wants to buy your service. E.g. if you can design them a logo, tell them how much you charge and when it can be delivered. If they like your service, they'll probably hire you again or refer you to their contacts for new businesses.

[edgy scroll]

Actually im gonna work in data science and ml besides i dont have a degree. Do they care?

[green siren]

Kind of a broad question unfortunately,

I am creating a server that will be accepting TCP connections from multiple users, they will be logging in, and then accessing/querying/changing data inside of a Postgres database I have setup.

I plan on setting each TCP Connection inside its own thread so I don't have to close them, as I plan on users regularly querying new data (I shouldn't have to close it and reopen it constantly, should I?)

Can each thread push a query to the database without me worrying about creating some kind of queue, or locking a stand-alone thread that passes these queries to the database?

Does the database handle all concurrency issues by default?
Will I have an issue on pythons end by pushing multiple queries out at the same time?

[cloud spruce]

it will very much depend on each customer, but this discussion really belongs in #career-advice

[green siren]

Not sure if making fun of my question or a serious question 🤔

[cloud spruce]

starting what within networking exactly?

[cloud spruce]

i would go with async io with one of the libraries that enables that instead of threads in that specific situation to handle all the connections, together with asyncpg and a connection pool towards the database

[green siren]

Don't think I've ever used async io or asyncpg how is it different?

And what do you mean connection pool towards the database

[cloud spruce]

with any kind of async io you typically run everything in only one thread (or at least a low number of threads) and you may write functions what you declare as async
then you call your own and other async functions from different libraries with await which tells your program "this call might take a while before it returns, while we wait the program can go on and do there tasks in the meantime, then when this call eventually is finished we'll continue with the execution where we left off"
for asyncpg specifically, see: https://github.com/MagicStack/asyncpg and https://magicstack.github.io/asyncpg/current/
and to use it with connection pooling towards the database, see: https://magicstack.github.io/asyncpg/current/usage.html#connection-pools (remember that this is just an example, there are probably better libraries for you then aiohttp that is used in that example, as connection pooling has nothing to do with aiohttp specifically)

i think this is a very good tutorial to start with: https://realpython.com/python-sockets/
but where they use recv(1024) i would use a slightly higher multiple of 1024, at least 2048 in general and maybe higher numbers still, such as 4096, 8192 or 16384 in LAN environments that support jumbo frames (which isn't very common)

[cloud spruce]

what kind of API would your server provide to the clients to access the database (as in: sockets, REST API, websockets, etc)?

[elder cobalt]

btw @cloud spruce you think it's better to use sockets or pydivert in the long run? So far I'm doing just fine with pydivert but not sure if I should branch out and learn sockets as well?

[cloud spruce]

sockets will not do what you want here, as you want to divert, hold/pause the matching packets and process them before you send them on their way
pydivert is using a windows facility known as windivert to accomplish this for you, linux and bsd has other facilities to accomplish similar tasks, it can't be done with just sockets

[wise jungle]

ok so how would u actually use UDP in async?

i assume you dont just slap async into the function ofcourse

[wise jungle]

@cloud spruce

[green siren]

So I can use this to listen to many different sockets at once?

[green siren]

No idea what REST API or websockets are, ect. I'm only just barely familiar with normal sockets right now

At the moment my client and server are connected via a stream/ TCP socket

[wise jungle]

ya

for tcp sockets its fine to use multi threading

but then later on its ideal to get into async

i still dont know how to but i really think u just turn the functions into async and run the app as a loop

[cloud spruce]

the use case was a proxy server, i it will be better to go async from the start

[cloud spruce]

i haven't found the time to write some example code for you yet, working on something else right now

[wise jungle]

i dont want example code for my case i just want a generalized one

i maybe able to do it myself but i havent worked with async that much

[brave raptor]

When I do a request like this requests.get(...), how does the site in question know that a script is making the request, not a genuine user? Currently it's returning an access denied error code.

[prisma cobalt]

the user-agent header might yield some infomation

[brave raptor]

oh, do you know if there's a way to print out the header?

[prisma cobalt]

import requests
                 
r = requests.post(API_URL)
# get request headers
print(r.request.headers)

from: https://stackoverflow.com/questions/52570933/what-is-the-request-header-by-default-in-python-requests

[brave raptor]

ah, thanks

[karmic tangle]

Is this code only or networking as general ? 🤔 I am trying to find someone that would help me with management switch

[cloud spruce]

generally it's code for almost all channels on this server besides maybe #microcontrollers
but the topic on this channel might suggest otherwise, so try asking and we'll see if anyone can help

[karmic tangle]

Our IPTV has 2 IPs. One local and one from WAN . Normal switch don't forward both IPs and IPTV doesn't like that.

I borrowd managment switch and i wanted to set it but i dont know how.

anyone help please?

[cloud spruce]

what is the make and model of the managed switch?

[karmic tangle]

mikrotik hEX
ID: RB750Gr3

[cloud spruce]

are you using windows on your computer or something else?

[karmic tangle]

yes windows. but i am trying to set it throu web

[cloud spruce]

do you already have a ip address setup on it that you know so that you can login to it?
otherwise mikrotik has a program called winbox that you could download and use to connect to the device to manage it

[karmic tangle]

when i reset it and conect to pc, i get address

[cloud spruce]

so, you know the address to the switch so that you can login to it?
or do you mean that you get an address for your computer?

[karmic tangle]

bouth, i am curently in the switch managmet UI (winbox )

[cloud spruce]

doesn't your provider use a separate VLAN for IPTV?

[karmic tangle]

yeah probably

1_INTERNET_R_VID_836 836/0 IPv4
5_TR069_R_VID_851 851/6 IPv4

personaly i think, the first is normal internet and the second is for iptv

it's quite late here, good night. i will be back in morning

[main spruce]

Does anyone here have experience converting UDP from matlab to python?]

[cloud spruce]

agreed, sounds very plausible

[cloud spruce]

i haven't used matlab and know what matlab has to do with UDP
but that doesn't mean i won't try to help anyways if i can be of any help 😉

[main spruce]

hey @cloud spruce, thanks for responding. I've posted my question in #help-chocolate

[wise jungle]

asyncio udp 😄

need assistance

no clue how to do

or just handling multiple clients

i dont know what i did wrong

@cloud spruce back to square one with https://github.com/e1pupper/VoiceChat

im gonna rewrite it

[cloud spruce]

follow along a bit in the #help-chocolate then 😉

[wise jungle]

ya i just saw ur ping im gonna read everything

[cloud spruce]

i think you'll be very happy with working with async when you've got the hang of it
as you would be writing functions or methods to handle events (like when you receive new data on the socket)
then you just have to keep some state like the list of all sockets so that you know where to the incoming data to reach the other participants on the server

[wise jungle]

this is absolutely crazy what u just linked me

i love you !

[cloud spruce]

happy to help where i can 🙂

[wise jungle]

so how is it not non blocking

[cloud spruce]

the idea with async programming is that you should call functions that are declared with the async key word with await (meaning "asynchronously wait")

[wise jungle]

ya dats why i was asking about it

cause i dont think this really solves the issue

but is part of the solution

[cloud spruce]

functions that can either block or take time to execute because we would have to wait for something like opening, reading or writing a file are declared with async

when you call them you use the await key word in the call

that tells the program that "while we are waiting for this to finish we can go and do some other work instead"

in that way you are in control of at which points in the code context switching can happens, as it only happens when you run await

[wise jungle]

im finna look into it tmrw morning 3 am rn

[cloud spruce]

i would recommend the official asyncio documentation at https://docs.python.org/3/library/asyncio.html as it's the most correct and up to date with the current practices for async programming in python
but first you might want an introduction and overview to get to grips with the terminology to find your way around the official documentation
for that i would recommend https://realpython.com/async-io-python/
or another library (not part of the standard library) named Trio https://trio.readthedocs.io/en/stable/ which has very good documentation and tutorials on the subject even if it's understandably kind of bias towards their own library, but don't let that discourage you from reading it as they still teach the general basics of asynchronous programming, and in the process you might also come to appreciate or even like their library 😉

[elder cobalt]

Btw @cloud spruce you think making some sort of a counter where it keeps adding +61 to check the sessionID since all of them are evenly spaced? That way if it does hit the correct sessionID it'll apply offsets automatically to other variables, and if the counter reaches the max counter of the packet it implies that the sessionID does not exist in this packet and it'll simply send and then look into the next packet sent

[wise jungle]

okay before i get into async and all dat i managed to clean up the code from github

i will do some tests now

[light brook]

I need list pip python libraries for network analyzer

[quick moat]

hello
should https stuff be setup in gunicorn or in nginx?

i have flask --> gunicorn --> nginx

[karmic tangle]

so, i solved my Router problem. 😄 now to the code... 😄

mikrotik > RouterOS... anyone? 😄

[cloud spruce]

what/which network analyzer are you talking about?

[cloud spruce]

i would defensively configure the TLS termination in nginx
and if that is on another machine or even container and you need higher/stronger security you would need to encrypt it again and run TLS between nginx and gunicorn (which runs your flask app directly)

[quick moat]

oki

thank your

[cloud spruce]

but most people can do without using TLS between the nginx and app server instance even if they run on different servers or containers as long as they are on a private network and don't have very strict security requirements impost on them

[quick moat]

its on same machine

no need https if between two containers?

for backend api

[cloud spruce]

if it's on the same machine and you don't let other people in to that machine, then no
unless you are really paranoid or work for a company with very strict security requirements
just see to it that your gunicorn instance isn't accessible from outside in any way

[quick moat]

ok

[wise jungle]

have u heard of cyberpanel before @cloud spruce

[cloud spruce]

no, why?
is it like cPanel?

[wise jungle]

not sure but i just started using it to make a mail server

pretty handy

[cloud spruce]

okay, when i checked their website (if it's even the same thing as you are talking about) it looked like it was mostly hosting for wordpress and php (which i'm really not a fan of)

[wise jungle]

kinda old but quiet convenient

and its constantly updated

[cloud spruce]

what is?

[wise jungle]

cyberpanel

[elder cobalt]

@cloud spruce for some reason this code does not work 🙁 https://paste.pythondiscord.com/ozesuxevih any idea?

I'm trying to create a counter where if it's not equal to packet counter it will jump to the next if statement where if the data spliced location(with offsets) is equal to the session_ID specified it will print out the XYZ and etc., else it'll keep adding +1 to the counter so that the next time it runs this loop it'll do 1 + 2 + 60 * 1 such that it'll read rest of the packet data before fully discarding it.

But let's say if that packet does not contain any info from the specified session_ID, and my counter is greater or equal to the packet counter it will set the counter back to 0 and continue to send packets that way it can proceed to the next packet. Is this flawed or are there any other ways to improve this? because my count loop seems to keep increasing to 100s and 200s which I am sure has most likely crossed my packet counter

[light brook]

@mdpkt

trafic

⚗️

[main spruce]

Hey guys, I currently have some networking code that is connecting to the host server and the host server confirms connection; however when i run s.recvfrom(1024) my python interpreter hangs up and I have to kill the process. I've tried adding a timeout and a try/except block but regardless the interpreter still hangs up

Anyone ever experienced this before, it's a udp port

[steady horizon]

socket.socket.recvfrom intentionally hangs until it receives 1024 bytes of information

Actually I don't remember whether it waits for the entire 1024 bytes of information to be received or it returns as soon as a packet with ≤ 1024 bytes of information

[cloud spruce]

s.recvfrom(1024) will return as soon as there is any data available, with as much data that is available at that time, up to and including the requested number of bytes (in this example 1024 bytes inclusive)

so, there shouldn't be a problem

[cloud spruce]

which OS (and version) are you running it on?

[cloud spruce]

i was quite busy during the weekend and didn't have much time to write and test code
but today i have had some time to look into this and i got something that should work
i'm just not super happy with the performance of it yet, so i'm trying a "few" different things 😉
spoiler alert: i tried to use kaitai in the python code, it was unfortunately the slowest of all the things i have tried yet in my benchmarks for this specific problem

[elder cobalt]

Ah I see, yeah I don't really rely on kaitai's python parser only use it for knowing where to splice

I took your implementation and tried changing it in a way that offsets get multiplied by a counter * 60

[cloud spruce]

or, well. not always the slowest, but for the modded server it was really slow as it always decodes everything from all instances

[elder cobalt]

o I see

[cloud spruce]

i wanted to try it out as i have never used it in python before 😁

[elder cobalt]

I guess it'll be useful for small projects

or places where you don't need to decode stuff very rapidly

[cloud spruce]

yeah, it was quite nice to work with in python as well (not just the webui)

just had to find out how to use it with python as that specific part of the documentation was quite lacking

but after checking some source code it was quite obvious, so not really a big problem

[elder cobalt]

it helped me for decoding packets ngl, but didn't really find any use for using it on python since the bytes were enough alone to do comparing and etc...

I just did some debugging from my code and it looks like the counter keeps going towards 200+ which I am sure has most likely satisfied this statement,

            elif counter >= struct.unpack("<B", packet_counter)[0]:
                counter = 0

but for some reason it's not resetting the counter so it looks like it's likely failing this else if check and just keeps adding, perhaps should I send you the full code?

[cloud spruce]

i think i'll going to write a custom class to deal with the data instead where i can tailor for the specific task to optimize it for speed, as this otherwise will add unnecessary latency to the communications in both direction between server and client

the extra latency between the server and the client is due to that we halt the packets on their way in both directions to be able to do modifications on them before we release them to let them continue on there way

that's why i think we should try to optimize this a bit more for speed than the current implementation

[elder cobalt]

In theory this sounds so simple and logical but in reality it's taking me over a week lol

[cloud spruce]

well, it would have been quicker if i have had more time to dedicate to this

[elder cobalt]

yeah that's what I'm afraid it is, sometimes I like to think that's not the case until someone points it out lol

[elder cobalt]

no issues lol, understand you were busy throughout the weekend

[cloud spruce]

there is one thing i have been thinking of trying, but i think i'll just skip it for now since i don't even know if the idea is viable at all, but if it works it might be the most performant

[elder cobalt]

Kinda wished the modded server packets were like the stock lol

But ig makes sense if the server wants to offload most of the packets to client

[main spruce]

i'm running windows 10, python 3.9.7

[cloud spruce]

should work much the same, just know that non-blocking io is wonky on windows and while select() does work it's highly inefficient and doesn't scale well on windows
if my memory is correct i've also seen problems if attempting to use a timeout of exactly zero as that is kind of non-blocking territory, instead use something just above zero, like 0.000001 instead of zero if a zero timeout would be desired on windows, but many times select is better to use with a longer timeout or even without a timeout if the program doesn't need to do periodic or other tasks in between

[minor creek]

i’m good with radios

[steady horizon]

If you want asynchronous networking you can use asyncio

It's not as low level as socket though

[random dock]

is there some stron boy from networks in debian?

[ember ledge]

Hi, I am a complete newb to UDP and sockets. I have a python script that attaches itself to a UDP on initialization. It has multiple methods that each send data to that UDP.

The problem I’m having is that after I call one method, that method will take control of the UDP. So when I try to call the same method again or another method that wants to send data over that same UDP, it wont let me do that.

My question is: how can I attach all my methods to the same UDP so they all can send data to the UDP?

@steady horizon

[cloud spruce]

you should be able to bind to the UDP port in one place in your script and then reuse that socket object everywhere as you see fit, or you can pass the socket object/variable into the methods

[elder cobalt]

@cloud spruce any luck? or maybe any idea where I could start?

[deft sonnet]

raw_data, addr = conn.recvfrom(65535) # Conn receives from '65535'
OSError: [WinError 10022] An invalid argument was supplied

[proud swift]

Is your socket bound?

[deft sonnet]

?

[mighty sand]

Hey ! did you ever have an issue with aiohttp failing during an assertion port is not None ? i have no idea why because when i test the query manually it works fine

[proud swift]

Do you have a traceback?

[mighty sand]

  File "/usr/app/[...].py", line 77, in execute
    async with session.post(
  File "/usr/local/lib/python3.10/site-packages/aiohttp/client.py", line 1138, in __aenter__
    self._resp = await self._coro
  File "/usr/local/lib/python3.10/site-packages/aiohttp/client.py", line 535, in _request
    conn = await self._connector.connect(
  File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 542, in connect
    proto = await self._create_connection(req, traces, timeout)
  File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 907, in _create_connection
    _, proto = await self._create_direct_connection(req, traces, timeout)
  File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1146, in _create_direct_connection
    assert port is not None
AssertionError```

i have higher trace back for my function stack but the error comes from an aiohttp.ClientSession.post

basic aiohttp query code that looks like this

async with aiohttp.ClientSession() as session:
    async with session.post(
        self.url,
        headers=self.headers,
        json={"query": BULK_QUERY_TEMPLATE, "variables": {"query": self.query}},
    ) as res:
        response = await res.json()

which i've tested the query manually

[proud swift]

Manually?

[mighty sand]

i made the post query myself in insomnia

it's a graphql query

[proud swift]

How did you make the socket?

[proud swift]

Can you make a run-able example?

[mighty sand]

na cause it's sensitive information :/
bulk query to a shopify endpoint, if you know the API

i can give you the exact endpoint if needed

[proud swift]

Try a breakpoint at the line that fails?

[mighty sand]

yes i've already done that, and as i said, previous code isn't important as the query is right, and it works

problem seems to come from aiohttp or how i pass the query to aiohttp

looking at the stacktrace, the issue seems to come from the connection part so maybe the headers prefetch fails, i'm gonna retry and see

[deft sonnet]

raw_data, addr = conn.recvfrom(65535) # 'Conn' receives from '65535'
OSError: [WinError 10022] An invalid argument was supplied

[cloud spruce]

just before the inner async with print self.url and see if it contains what you expect it to

[mighty sand]

Ye it does

[proud swift]

What does it look like with the breakpoint there?

[proud swift]

Does it work with httpx?

[mighty sand]

i have a meeting but ill tell you after

[ember ledge]

can someone help me with a slow loris ddos attack

[prisma cobalt]

no

sorry

!rule 5

[errant bay]

Rules

5. Do not provide or request help on projects that may break laws, breach terms of services, or are malicious or inappropriate.

[cloud spruce]

are you under attack and need help protecting your self from it?
otherwise: no, see above rule

[ember ledge]

No i got oermission from my friend to try take down his website

Permission

Iam not breaking any laws

[cloud spruce]

no difference, we can't know if you are telling the truth or if you are going to reuse it later for something else
so, the rules still apply

[ember ledge]

Thats true

[stuck pivot]

Ima go

[ember ledge]

Ok i nevermind then

Have a good day

[main spruce]

When I open a UDP connection in python via socket
is it possible to open/write/read using that socket? Like how t is defined

[dusty shoal]

How to send inputs using socket

TCP

[sweet breach]

like input()?

[sweet breach]

you freak me out lol

[dusty shoal]

yes

[elder cobalt]

hmm @cloud spruce found something interesting, it looks like my counter is really slow at catching up to my packet_counter? 🤔 (I've did a print statement to check if counter>=packet_counter)

[cloud spruce]

packet_counter is pretty self explanatory, but does it count all packets in both directions in one counter?
and what is counter counting exactly?

[elder cobalt]

not sure what you meant by both directions, but it sends a batch of SessionIDs included with their information like position, rotation and etc. in one single UDP packet to me (the client) from server

https://paste.pythondiscord.com/sifobomoxe

this is so far my latest modification

I'm using counter to iterate everything being received from the server

like every byte being received

[cloud spruce]

oh, i first thought packet_counter what the number of network packets that you have processed
but is it rather the number of updates contained in the packet?

[elder cobalt]

correct

I'm using my own counter and based on that do some conditions like,

If my counter does not match the packet_counter keep adding +1 in hope that I find a sessionID matching the one I want, if it does it'll change the offsets accordingly with the counter variable and print out X, Y and Z of the car's position then reset it back to 0 so that it can start from 0 for the next packet

If let's say my counter variable is equal to my packet_counter or greater than that reset the counter to 0, it implies that this specific UDP packet didn't contain the sessionID I wanted but maybe the next UDP packet will so put it back to 0 and start reading from first

if you notice

if data[1 + offset : 2 + offset] == Session_ID:

Here in the offset variable, I did

offset = 2 + reader

and the reader variable is

reader = 60 * counter

so let's say my counter is 0, offset will be 2+60*0 which is 2

then it'll do

if data[1 + 2 : 2 + 2] == Session_ID

which is basically

if data[3:4] == Session_ID

but if it does not match my Session_ID, then I say the counter to add +1

counter=1
reader = 60 * 1 # Which is 60
offset = 2 + 60 * 1 #Which is 62
if data[1 + 62 : 2 + 62] == Session_ID  # Basically data[63:64] == Session_ID

@cloud spruce hopefully this makes sense 🤔

only issue is now my counter is so slow that it can't iterate through all the data before the server sends the next UDP packet with a new packet_counter

and that might be why python might be confused and just keeps adding +1

but I'm not sure if there's any other way to refine this

this is already fast enough because it's only comparing sessionID sent from server to the sessionID I want

I could maybe make a delay between the server sending information to the client but that would noticeably create both lag from my side and I won't get information as quick as what the servers sending

[arctic meadow]

hi i am working on a socket program i need to have a list of clients connect to my socket and pass that data to api as a json form when i try to convert the clients socket object to a json object i got an error TypeError: Object of type socket is not JSON serializable so how can i convert my socket object to a json object

[main forum]

idk if this is the right place to ask but its a networking problem 😄

can someone help me configure the /etc/network/interfaces

i seem to do something wrong and it pisses me off

[cloud spruce]

sounds like a debian based distro, but for people to be able to help you easier it's good to announce which distro and version you are running

[cloud spruce]

i'm still struggling with a bug or two in my mocking code that i have built so that i can run the rest of the code and verify if it's working or not
for the fake packets i'm using the .bin files from the archive you posted previously
if i can just get this sorted I think i'll have something to send you really soon
unfortunately i'm away from the computer for a few more hours until i can continue

[elder cobalt]

Ah no issues looking forward to it

Also glad you found the .bin files useful I should've provided them from the start lol

[elder cobalt]

@cloud spruce sorry for pinging you but I'm not sure if you've caught up to this but I just realised that the SessionIDs are in +1 order so if it starts with 46, and the packet_counter is 6, it implies 46+6 46,47,48,49,50,51,52 are all the sessionIDs in the packet

This could be a way for python to just add the starting number and check if the sessionID I want even exists and if it does not it'll just skip

But in the event it does exist let's say my sessionID I want is 48, I'd need to ask python to add 46 to some number which will give 48 (which is 46+2) that way it's way quicker

[elder cobalt]

woops looks like this didn't work out in the end as well

😅

looking forward to your code, ngl I feel like I've tried out everything but for some reason it just does not work

https://paste.pythondiscord.com/usumejabem made a new one

Really frustrating ngl, thought this could've worked but nope

[main spruce]

Does their exist anything like the BytesAvailableFcn for python?

ref: https://www.mathworks.com/help/instrument/bytesavailablefcn.html

[main spruce]

Does anyone have experience defining a udp socket using asyncio

[cloud spruce]

yes, and I think asyncio or other async framework/library will be your best bet as it's event based

[main spruce]

do you have any examples?

[cloud spruce]

we can continue in there 👍
but didn't understand what you wanted to do exactly
do you want to setup a firewall?

[cloud spruce]

https://docs.python.org/3/library/asyncio-protocol.html#udp-echo-server
maybe not the best, but at least something and it's part of the asyncio documentation

[ember ledge]

i hope this is where you ask iptable related questions.. if not please reply with the correct channel, ty

just wondering if theres a way to track udp packets that are being sent to various ports by a single host that are malicious?
(kind of like tcp-connection rate-limiting)

[cloud spruce]

sure, you can use iprables for that by setting up logging and rate-limit or blocking connections from a specific host

i think in here or in #unix or maybe #cybersecurity are good places for that question

[ember ledge]

lets say a collection of random ips send the data, then what? the rate-limiting is useless then.. but thanks

[cloud spruce]

yes, or ip spoofed packets, then you need DDoS protection services
from your previous question I thought that you knew the ip addresses that you wanted to act on

[ember ledge]

i mean, i do.. i discovered the vuln and exploited it on my own servers but i cant fix it :P
if you know anyone or any articles that could possibly help that would be awesome, thanks

[cloud spruce]

sorry, but i fail to see the vulnerability other then the classic DDoS vector where someone trys to exhaust your bandwidth by filling your uplink with unsolicited traffic, or am i missing anything here?

[ember ledge]

can i show you what im doing? i think it would be alot better to see what im talking about

[cloud spruce]

sure, but show me how, screenshot?
a detailed description or through explanation is often helpful

[ember ledge]

this is an OVH VPS that should block basic attacks and ive added my own iptables so its pretty hard to push unwanted traffic through,

the server im sending the data with is pushing all of its bandwidth through the firewall(s) (which none of it should go through...)
the data thats being sent is malicious but somewhat disguised because its going to its designated port (ex... telnet goes to port 23)
source of packet data: https://github.com/Phenomite/AMP-Research

im not well educated on networking terms, i hope you get what i mean by packet data

im just trying to track this... somehow...

[cloud spruce]

do you mean payload?

[ember ledge]

yeah my bad

[cloud spruce]

what I don't understand is what you mean by "track", is it just to keep track of or to trace it to its source?

[ember ledge]

like, able to control... rate-limit or drop.. etc...

[cloud spruce]

if you know the port or the source ip addresses you could, but only to a point since they might be able to fill your uplink if you filer the traffic after that point
in such cases you need external helt from a service provider with more bandwidth then the collective bandwidth of the attackers
there are lots of provider that offers these DDoS protection service at a premium rates
cloudflare even has a decent basic service for free

[ember ledge]

well ugh, this bypasses ovh, path, 100up... etc, the list goes on so i might just be in a pickle...

thank you for your time

[spare rose]

Any easy way to calculate time taken to send certain mb of data between 2 cities?

[cloud spruce]

no, it will depend on available bandwidth of the whole network path between the two and possible congestion which can fluctuate wildly during the time of the transfer, the best you can do is try to observe the throughput and calculate it based on that

[cloud spruce]

unfortunately a very common problem 😞

[spare rose]

I dont need exact timings. Just approx. I found https://wondernetwork.com/pings . Will this be any useful?

I mean will ping stats be approx same as sending even 1kb/1mb/1gb/1tb of data proportionately?

[cloud spruce]

no, and you would need to measure the data transfer rate between the two clients specifically

[dusty shoal]

Hello I am having this error in socket.io

import socket

s = socket.socket()
print("Socket created")

s.bind(("localhost", 9999))

s.listen(3)
print("Waiting for connections")

while True:
    c, addr = s.accept()
    name = c.recv(1024).decode()
    print("Connected with ", addr, name)
    while True:
        c.send(bytes(f"Hello, {name}, welcome to the world of coding", "utf-8"))
        output = c.recv(1024).decode()
        print(output)
        if int(output) < 12:
            print(f"Client {name} not eligible")
            c.send(bytes("Folks it looks like your age is a bit lower than expected.", "utf-8"))
        else:
            print(f"Client {name} with age: {output} accepted")
            c.send(bytes("Folks your age is enough to countinue."))
        c.close()

Server Side

import socket
import time
c = socket.socket()
command = ""
c.connect(("localhost", 9999))
name = input("Enter your name: ")
c.send(bytes(name, 'utf-8'))
print(c.recv(1024).decode())
age = int(input("Age: "))
c.send(bytes(age))

Client Side

Socket created
Waiting for connections
Connected with  ('127.0.0.1', 63519) Loss
           
Traceback (most recent call last):
  File "C:\Users\Pytho\IdeaProjects\socks\server.py", line 19, in <module>
    if int(output) < 12:
ValueError: invalid literal for int() with base 10: '\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'

Error

@humble island can you help?

[humble island]

i'm trying to think of something, but this isn't really a networking issue

just converting numbers from one base to another

[dusty shoal]

oh so the base of the numbers are different?

[humble island]

this is in hex, isn't it?

[dusty shoal]

oh yes yes i converted it into bytes because that is the only way to send from client to server and vice versa

[humble island]

right

the issue is converting from bytes back to int

[dusty shoal]

the this i converted bytes to string and then to int

i actua;lly printed the output

as you can seee the characters are not identified

[humble island]

i'm aware, i'm trying to see how to convert that back to its

[dusty shoal]

oh wait i already converted it into int

age = int(input("Age: ")) , it is converted to int already

[humble island]

you want something like int.from_bytes(mystr.encode('utf-8'), 'big'), maybe? though idk which endiannness you used

[dusty shoal]

i also forgot to mention the encoding

yes it worked

[humble island]

i'm not sure about the part where you do bytes(age) though

[dusty shoal]

oh oh the another error

Traceback (most recent call last):
  File "C:\Users\Pytho\IdeaProjects\socks\server.py", line 16, in <module>
    c.send(bytes(f"Hello, {name}, welcome to the world of coding", "utf-8"))
OSError: [WinError 10038] An operation was attempted on something that is not a socket

@humble island any alternatives for socket?

[viscid cipher]

Generic question: someone knows how to calc the limit of data that i can send as a post request to a server before it gives me gateway timeout error?

[cloud spruce]

is this the second pass through the code?
because you are doing c.close() in the inner while True loop
then the second time through the loop the socket will have been discarded and would lead to such an error
instead remove one level of indentation from the c.close() line of code so that it is only within the scope of the outer while True loop
also, you would need some error handling to take care of the case where the client disconnects by doing something like:

    name = c.recv(1024).decode()
    if output is None:
        break
    print("Connected with ", addr, name)
    while True:
        c.send(bytes(f"Hello, {name}, welcome to the world of coding", "utf-8"))
        output = c.recv(1024).decode()
        if output is None:
            break
        print(output)

[cloud spruce]

you will always use sockets in one way or another when doing [normal] network programming or else you would have to circumvent the OS network stack
but you may use high level libraries and functions that hides it from you by abstracting it way (out of sight, out of mind 😉)

[dusty shoal]

@cloud spruce Thanks for the suggestion and code!

[cloud spruce]

you won't know other then by trail and error, it depends a lot on the server and the code running there as well as configuration of the same

[cloud spruce]

another suggestion would be to send the questions from the server instead of having them hard coded in the client software

[dusty shoal]

so should i make a list of questions would tahtb efficient?

that be*

if yes then can you show how, (again i'm very very new to this stuff, though it is quite interesting)

@cloud spruce ?

[cloud spruce]

in the server you could send the question just prior to the line where you do s.recv() and wait for the answer
but the client would have to be reworked a bit

[dusty shoal]

also how can i send inputs to the user as send?

and yes I hereby confirm that I am total moron

[cloud spruce]

you wouldn't send the inputs to the user, just the text string with the question as bytes
the client would just s.recv() the question use it in the input() on the client

[dusty shoal]

how to send multiple inputs?

ir. text strings

[cloud spruce]

just loop in the client as well

[dusty shoal]

so a for loop in the while true loop in the client and a for loop in the client am i right?

[cloud spruce]

no need for any for loops and you can keep your server mostly as it is, just put a s.send() with the question before each s.recv() and handle all input() as strings (not int) and send it over to the server as the string converted to bytes and do the conversion on the server slightly different
in the client, just do something like:

while True:
    question = c.recv(2048).decode()
    if question is None:
        print("The server closed the connection.")
        break
    answer = input(question)
    if answer == "":
        break
    c.send(bytes(answer, 'utf-8'))
    reply = c.recv(2048).decode()
    if reply is None:
        print("The server closed the connection.")
        break
    print(reply)
c.close()

[burnt quartz]

anybody here?

[cedar forum]

yep

[main spruce]

is it possible to check if data was sent to a UDP port

via cmd

[main spruce]

or i guess received

right now im using netstat -sp udp

but i want to specify one port

[cloud spruce]

you need the process to be bound to the UDP port at the time your machine receive the packets or the packets will be discarded by the OS
the corner case is if you use raw sockets, it can be used like a sniffer for both inbound and outbound traffic on an interface but requires you to have superuser privileges

on linux you can use iptables to count the packets to a specific port if you make a specific rule for that port

[cloud spruce]

the you need to bind to that port with your program or setup a firewall rule with counters that you can check or use a raw socket och other way to sniff traffic preferably with a filter and then count it

[main spruce]

i have this condition satisfied

when I run netstat -sp udp

and my program

I see packets are sent/received

but I also have a bunch of receive errors

but netstat -sp udp is for every udp port

I want it for just one

[cloud spruce]

yeah, you would have to count the packets with your program rather then with netstat

or use a firewall rule for one port and make the firewall rule count the packets for you, then you don't need anything listening on the port

or have an application that sniff packets and count the matching packets that way

[main spruce]

so that's the issue

my server registers my client's send request and sends packets to the client

but the client isnt registering the data for some reason

basically my socket.recv() isn't returning anything except a timeout

I know the program is sending it to the port

I want to know if the port is receiving it

[cloud spruce]

ah, i see, then you probably need to break out a packet capture program, also called a sniffer, that you can run if you have superuser privileges

[main spruce]

I’m on Windows

Is there no way to do it without external program

[cloud spruce]

that's fine, as long as you are admin on your computer you can install and run wireshark on it

[cloud spruce]

i don't know windows well enough to know if there is some special facility that you would be able to use for this on that platform

[main spruce]

I need something either in code, in a terminal, or cli

[main spruce]

No worries I’ll keep Searching around online

[cloud spruce]

why not an application like wireshark?

wireshark comes with tshark, it's more or less the same thing for the cli (in both windows, linux and mac)

[elder cobalt]

a @cloud spruce lmk if you need more packets to test with, perhaps you could use a wireshark file to maybe replay all the packets instead of .bin files?

[cloud spruce]

thanks, that would have been an idea, but now i have built some mocking code for use with the .bin files where there is only the payload of one packet per .bin file
otherwise i would need to write something custom again for reading in the pcap file and extracting the payload and doing something similar with it
would probably not be a lot of work now when i already have an mvp for creating fake packets with payload compatible with pydivert

[cloud spruce]

right now i'm looking in to why i sometimes get an error that i need to take care of

[main spruce]

Can’t display

[cedar forum]

what do you mean can't display

tcpdump is another good bet

[cloud spruce]

i haven't really used tcpdump on windows and don't know if it's any good
as @main spruce is on windows if i understand it right

[cloud spruce]

that seems unusual if your on windows

[cedar forum]

right that's true, in that case I'd probably do netsh tracing and then import into wireshark/tshark

don't have to necessarily capture with wireshark if that's playing up

[cloud spruce]

sounds like you know windows a whole lot better then i do
all @main spruce wants to do is to count udp packets that matches one single port to debug why the application isn't receiving any packets on the port that it's listening to

[cedar forum]

i've done packet tracing on windows server once so my experience isn't super complete 😄

but that does sound like a job for wireshark/tshark, but in the past you can do something like netsh trace start capture=yes, do your packet stuff, netsh trace stop, then use this https://github.com/microsoft/etl2pcapng

though really wireshark should just be doing it, so i'm curious as to what the problem there is, though i've never used wireshark on windows

[cloud spruce]

haha, i've only used wireshark (gui) and tshark (cli) for that on windows

[cedar forum]

going to spin up a test bench on a windows box i have somewhere

hmm

yeah fresh install of wireshark on latest windows 10 is working fine here

i forgot how much UDP multicast runs over this network lol

[cloud spruce]

interesting, i've never used or even knew about netsh trace
i'm happy to learn of its existence if i need it for troubleshooting in the future 👍

[cedar forum]

yeah, it's been nice to use in the past, mainly because it means I don't need to install stuff on the remote

[cloud spruce]

yeah, i've been using wireshark/tshark [even on windows] for years and it works really well

[cedar forum]

crikey, only after a couple of minutes lol

[cloud spruce]

yeah, that's a very good point, just transfer the file to another system and convert it to pcap format and run it there ✅

[cloud spruce]

haha, must be updates 😆

[cedar forum]

wireshark does some nice graphs

[cedar forum]

radio station, so just a whole lot of multicast audio streams 😄

[cloud spruce]

yeah, love wireshark

[cloud spruce]

oh, i see 💡 📻

[cedar forum]

all our studios are AoIP, so each microphone is an individual multicast AoIP stream, as is music, etc.

we moved from analog a year or two ago and AoIP ecosystem is so so nice, I can plug equipment in to any ethernet port and pick up/insert any AoIP stream I want into the network

the only bit of analogue we've still got is in inside our coffee table, we thought it'd be fun to hook the old studio selector up so you can see the needle graphs flicking 😄

[elder cobalt]

Ah I see, sorta surprised tho that it's this hard 😅 normally decoding the stock server packets was a breeze but this seems to be a lot more problematic

[dusty shoal]

can you help me with socket, I will send you the code: Server Code

import socket

s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
print("Socket Created")
s.bind((socket.gethostname(), 1234))
s.listen(5)
print("Socket is now listening")
while True:
    client_name, addr = s.accept()
    questions = ["Name: ", "Age: ", "Country: "]

    def send_message(message:str):
        client_name.send(bytes(message, "utf-8"))
        print(f"sent message to client: {client_name, addr}")

    for question in questions:
        send_message(question)

Client Code :

import socket
import time
c = socket.socket()
command = ""
c.connect((socket.gethostname(), 1234))
print("Connected successfully")
while True:
    question = c.recv(20).decode()

I expect it too well say questions one by one

[drowsy nebula]

you didn't put any statement to wait for an user input for example.

[cloud spruce]

well, my problems was with a few other things when i refactored the code a lot without being able to test it, but now i've built the mocking and are able to test the code at last

[elder cobalt]

yeah, atleast it's a bit easier now than just guessing

[cloud spruce]

yeah, a lot i would say

[cloud spruce]

i just hope that the mocking part of the code actually is accurate enough to emulate the parts of pydivertthat i need it to emulate

[vernal smelt]

Hey is it actually okay here if you share a project here? I don't really intend to advertise but I'm just so enthustiastic about what I create.

[elder cobalt]

I can send you in more lol if you want to refine it a bit more

[fervent thorn]

I have a server capable of serving multiple clients and I want to determine packet rate at which clients are sending packets to my server

The idea I came up with is keeping a frequency map for IP addresses of all the clients and storing the packets count for some arbitrary amount k seconds. Now after k seconds we traverse the map and see what IP addresses have more than 100*k packets, now we block these IP addresses. After every k seconds we empty the map and start again.

If somebody has clear idea about what I'm trying to ask and achieve through code, please DM me I need help

I've gone nuts trying to figure this out from last 2 days😭😭

[cloud spruce]

last i checked this server and channel has more to do with python then c++ 😉
here are three different techniques for rate limiting implemented in python: https://dev.to/satrobit/rate-limiting-using-the-sliding-window-algorithm-5fjn
but if you know both python and c++ to can probably transfer the general techniques to between the languages
and there is even more of them if you just google for "rate limiting algorithm" and go from there

[ember ledge]

is it possible to make a permanent ip blacklist using iptables?

when my server gets attacked my iptables drop most of the malicious packets but my servers resources are overwhelmed in the start of attacks which causes issues, I was wondering if it was possible to somehow drop all traffic from a host if they've previously been -j DROP'ed?

[cunning garden]

you may be interested in fail2ban

but to permanently ban an ip, you can use a command line to drop all the traffic coming from a specific ip

[ember ledge]

i was away from my pc when you responded, i think this is what im looking for

thanks

[ember ledge]

really weird but im actually also having issues with the same thing

xD

[ember ledge]

import time
from DDPClient import DDPClient
client = DDPClient('wss://atshop.io/websocket')
client.connect()

while True:
    try:
        time.sleep(1)
    except KeyboardInterrupt:
        break```
ssl.SSLError: [SSL: SSLV3_ALERT_HANDSHAKE_FAILURE] sslv3 alert handshake failure (_ssl.c:1129)

[halcyon arrow]

Your website has SSL correct? You need to catch that SSL certificate in python and use it to communicate with the sever. Or in turn disable ssl on your website (this is not advised). Here is a stack overflow page with an explanation of the error. https://stackoverflow.com/questions/56719290/how-to-solve-the-handshake-failure-using-ssl-in-python

[ember ledge]

it's not mine, but it has ssl

its the same website u were talkin with a guy a year ago

u can connect to it but not with the ddpclient module in py for some reason since it returns ssl errors

[halcyon arrow]

I am not sure about the ddpclient I’ll have a look, but usually you need to wrap the connection in some kinda ssl layer.

[ember ledge]

alr

@halcyon arrow btw i even used the meteorclient module which is kind of based on ddpclient

[halcyon arrow]

I don’t think the module u use is the problem. It a socket level issue with an ssl certificate I think you need to look into python ssl handling.

[fervent thorn]

How can I blacklist a client if I have it's IP address. I want them to not be able to send udp packets

[cloud spruce]

they will still be able to send them, it's you who will be blocking them so that you don't receive the packets
what OS does the machine have that you want to block packets to?

[fervent thorn]

I mean yeah, I might have framed my query wrong but all I want is to not receive their packets at my server.

And server is running on windows machine

[cloud spruce]

oh, i don't know windows that well, but i believe you should be able to set up the windows firewall to block certain ip addresses

[fervent thorn]

I have not worked with firewall at all and rn I'm reading documentation of windows firewall with advanced security api

Kinda overwhelmed by the docs

I just need to add the ip addresses in the block list of firewall

[full dust]

you should be able to do it with a powershell one-liner:

Net-NewFirewallRule -DisplayName "Block IPv4 Ingress" - Direction Inbound -Action Block -RemoteAddress X.X.X.X

https://docs.microsoft.com/en-us/powershell/module/netsecurity/new-netfirewallrule?view=windowsserver2022-ps

that method isn't terribly scalable though, I would hate managing production ip blocklists in the windows firewall lol.

I'd at least look at hosting whatever application this is with IIS and setting up IP and Domain Restrictions there.

[cloud spruce]

not applicable as @fervent thorn wanted to block udp packets to the server, i'm assuming the application is listening on an udp port and isn't performant enough to take care of the filtering it self

[full dust]

IIS can't frontend a udp listener?

I'm not a sysadmin and primarily use nginx - not trying to start an arguement just curious

[fervent thorn]

Well I'm going through the documentation on msdn and am trying to achieve this kinda thing by their api programmatically

[jade aspen]

Anyone have experience witih PySNMP and/or EasySNMP that can chat for a minute?

[fervent thorn]

Yes my server socket is a raw socket listening listening to mainly udp packets

Is that a problem?

Heyyyy I need a favor

I'm trying to fill up the fields of the rule according to my needs but I have just one machine with one ip address so I can't test things out

Can somebody help me figure out the desired fields

Dm please?

Will hardly take 5 minutes

[fervent thorn]

I've added an inbound rule to block udp connections on the port I'm running the server

But when I send packets using scapy to my server, it is still receiving it

[cloud spruce]

you said you are using raw sockets, that usually means it's the same as setting up a sniffer, which will see traffic before the firewall can block it
your options are either to not use raw sockets and just bind to one or more udp ports that you are interested in or filter the traffic in a physical firewall or server in front of your server running the program

[fervent thorn]

On what layer does the firewall work?

[cloud spruce]

just above the raw socket layer, raw sockets bypass all such things
it's even required for packet capture/sniffer applications like wireshark to work as expected/designed

[ember ledge]

does sending requests count as networking?

[cloud spruce]

i guess... ask your question and we'll see if anyone answers

[ember ledge]

nvm i figured it out

[ember ledge]

hey guys I'm trying to learn how to get requests to access info and urllib to automate stuff on websites. I feel completely ignorant on the topic, does anyone know where should I start to learn the network tab, understand how to make requests etc.. Most of youtube videos just show how to do it but not the actual science behind it so If someone could tell me where to look would be greatly appreciated.

[teal karma]

Hey I am wondering I would like to host my website with out paying a fee for a service my current method is hosting it through a server I already have using apache2 is this the best practice for what i am doing or is there a better/safer way?

[cunning garden]

gitlab/github pages are good options

[errant birch]

Hello, does http proxy work when requested https link?

[cloud spruce]

yes, if your client and proxy server can handle the connect method (most can) used to proxy https or any ssl/tls over a http proxy

[elder cobalt]

@cloud spruce ```python
wordlist = ["POST", "scripts", "Plugin"]

for x in wordlist:
conv2hex = x.encode("utf-8").hex()

with pydivert.WinDivert("ip.DstAddr = 147.135.137.85 && tcp.PayloadLength > 50") as w:
for packet in w:
data = bytearray(packet.payload)
if conv2hex in data.hex():
print("Condition met!")
w.send(packet)
else:
w.send(packet)

I'm trying to make some sort of a filter list where if it contains one of those words(wordlist) in data.hex() it'll print the Condition met! statement but for some reason it only works with the Plugin filter, it's not filtering out POST and scripts any idea why so?

[cloud spruce]

i'll check out your new code later today
i've just finished with https://paste.pythondiscord.com/ugojunedep for you
much more then half of the code is just for debugging (if/when needed) and/or mocking, i left it all in there
run the script with the -h or --help options to see how to start it
as an example run it with the options -s <server-ip> -p <server-port> -i >session-id>

[elder cobalt]

Oh holy shit thanks alot

Damn that's definitely so huge

[cloud spruce]

everything that has to do with mocking you can ignore
that whole MockDivert class and a few more bits and pieces is just placeholders for me because i can't run pydivert as i'm not on windows (and the fact that i don't have the game)
the code that actually do the important stuff isn't really that much