What do you need help with? v9 - Developers

Well shit I just locked the last one. Old ones: What do you need help with? Version 5 What Do You Need Help With? V6 What Do You Need Help With? V7 What Do You Need Help With? V8 Versions 1-4 are dead links now

Does the WAYWO highlight generator not work any more?

Probably not. I wrote a quick script to digest the last thread but nobody on the Discord wants to help me convert it to a post because they say "Facepunch is dead". Here's the JSON file. Maybe you can do something with it? results.json Quick rundown of the structure/how it was calculated: Users: This field has each user who posted in the last WAYWO, and the grand total of all of their ratings. The Ratings take the format [ YourUsernameHere: { X: Y }, SomeOtherUsername: { X: Y } ] Where X is the rating ID and Y is the number they received, throughout the entire thread. HighestPosts: This field is the top 5 highest rated posts for each rating category. HighestScores: This is what you get when you total all the ratings a user got across the entire thread, and choose the user with the most ratings in EACH category. For example, rating #3 is "Funny", and throughout the entire thread, I received 88 funnies, putting me at #1. --- These are all a little skewed: the ratings from before the update got wiped out so the first page had zero ratings. A manual curation of top posts and most interested works should be undertaken as well. And, of course, here's the code I wrote to do all this. It's quick and dirty and has to be used in a kind of precise way. Im not proud of this code, I wrote it really quickly: index.js as a gist: https://gist.github.com/collinoswalt/03dcbeaf7b4944617649db6b43e762c7 Feel free to ask questions or make suggestions. theres no license to the code so consider it public domain. You use it like this: ./node index.js [URL] > file.json where URL is of the format: https://forum.facepunch.com/f/nerds/btaaj What is important is that the last parameter there (btaaj in this case) is the LAST thing in the URL, since I manually concatenate the /Page# and add the ?json=1 at the end as well.

More of a math based question but I'll post it anyways. Currently I'm working with someones legacy c++ code double derOriw = pCamera->GetOrientation().w; double derOrix = pCamera->GetOrientation().x; double derOriy = pCamera->GetOrientation().y; double derOriz = pCamera->GetOrientation().z; double dcamYaw = atan2(2.0 * ((derOriw * derOriz) + (derOrix * derOriy)), 1.0 - (2.0*(derOriy * derOriy + derOriz * derOriz))); double dcamPitch = asin(2.0*(derOriw * derOriy - derOrix * derOrix)); double dcamRoll = atan2(2.0 * (derOriw * derOrix + derOriy * derOriz), 1.0 - 2.0*(derOrix * derOrix + derOriy * derOriy)); Anvel::Vector3 camDir = Normalize(Anvel::Vector3(cos(dcamYaw), sin(dcamYaw), -sin(dcamPitch))); Anvel::Vector3 camDir2 = Normalize(Anvel::Vector3(cos(dcamYaw), sin(dcamYaw)-.5, -sin(dcamPitch))); Anvel::Vector3 camDir3 = Normalize(Anvel::Vector3(cos(dcamYaw), sin(dcamYaw)+.5, -sin(dcamPitch))); Essentially, pCamera is a sensor and derOriw/x/y/z is the orientation of the said camera using a quaternion, and it gets w x y z from it. I have no idea what the math is behind the yaw pitch and roll, but I'm assuming ultimately that will give me the numbers that will become normalized into a normal vector in Normalize(). This vector is what the front of the sensor points to, and sends a ray out the front. Now what I needed to do was try to get instead of a line, a 60 degree cone in front of the sensor. I assumed it would be easy, just as an example camDir2 and camdir3 have -0.5 and +0.5 and it creates a vector that points to the left and to the right of the original camDir, but that is if the sensor lies flat on the ground. It took me a while to realize but if the sensor is pointing into the sky or angled off the ground, the +0.5 and -0.5 won't be rotating on the same plane but instead just spin around the same axis, and if it is pointing straight up all it would do is spin the sensor around but point in the same direction. Does anyone have any insight on if there's some sort of way to calculate the normal for x y and z such that it would spin while staying perpendicular to say the XY plane instead of just spinning around the Z axis

I'm trying to figure out how to resize this -__- <video autoplay="" loop="" muted=""><source src="Banner.mp4"></source></video></p> Its going into a blog post it wont resize! I've tried width="x" height="x"> behind the link but it did nothing. Can someone help me? I cant use <div> because it isnt the structure of a template, but a blog post. It works in my post but I cant get it to not be gigantic.. I would be very grateful for some help. Please and thank you

Not sure if this fits here but Since i am a beginner coder (currently learning python, have some experience with Lua), I was wondering how exactly do people make model/asset extractors for games? Say I wanted to extract some sounds from GTA: San Andreas. How would I go about figuring out how the files are compiled/how they work in order to build a program that would decompile/extract the necessary data that I would want? I hoped I've framed this articulately enough.

Apply generic knowledge of game programming to take educated guesses at formats, combined with reverse engineering the game to see where it loads etc. Alternatively dumping VRAM

for the current project im working on i feel like i need to create a javascript engine, including the interpreter and everything but i got no idea where to start i know the ecmascript standard including the documentation for it exists but ive never written something like this before and im kinda freaking the fuck out over here because i feel like im in over my damn head, its gotta be written in c, im not asking for anyone to make me but im just panicking over here because i feel like i cant do it and im not knowledgeable enough oh god

unless this is a school project, you don't need to write a javascript interpreter yourself. What language are you working in?

not a school project, it's for my job but I really have no idea where to start. working in C on a small MCU, i don't know exactly how big the sram is but considering it's embedded, it really should be as light-weight as possible

Why do you need javascript in an MCU??? Here's one library, I don't know what it's compatibility would be: http://duktape.org/ Here's a table: https://files.facepunch.com/forum/upload/132997/1e40fb43-190b-42f9-b944-7a2785a64478/Screen Shot 2018-06-21 at 12.00.48 PM.png What compiler are you using?

I'm using GCC. The reason I wanna use Javascript on an MCU is because I've already written a little thing that uses JSON, to configure a device using a small website that we use for various other things.

I'm confused - why not just rewrite your "little thing that uses JSON" in C/C++ then?

This might just be a question about just Visual Studio, or might require Xamarin knowledge. Right now I'm using Xamarin in Visual Studio to make an android app using .NET. I've been adding all of my asset files to the default Assets folder, but none of the files within it get compiled into the APK. When I specifically add each file to the solution, they get properly compiled. However, I have hundreds of asset files, all organized into many different subdirectories. It takes a lot of time to add them to the solution using Add Existing Item, and I have to frequently redo it as I add new files pretty frequently. Is there any way to tell Visual Studio to create the APK using all files in the Assets folder, whether included in the project or not? Or a way to automatically add the whole Assets folder and its files into the project? I know how to do it when Show All Files is toggled, but I don't have the option of doing that when using Xamarin, only for Windows projects.

I just took a programming quiz on TripleByte which is a service that is somehow supposed to match you with a tech job based on your strengths and weaknesses. Except the four timed questions they made me code for all were algorithmic string manipulation, which I absolutely suck at. I was working on the code in PyCharm because it's a lot easier, I went back to paste the code into the box and on each question I ran out of time and it submitted some preliminary garbage. So basically they're gonna review my submissions, which weren't the final code that I had, and think I'm some sort of script kiddy who just learned Python on youtube.

Every interview I've ever had (admittedly, on the East coast and I haven't had TOO many) has not been abstract Algorithm/Data Structure questions, they've been very practical questions about (for example), writing a SQL query, simple shit like "reverse a string in javascript", maybe some other JS questions (I'm in web tech). Personally, the #1 tool that I've found useful for finding a job has been Linkedin. Beef out your profile, start connecting with everyone you're even TANGENTIALLY related to (college friends are a good way to get connections). Follow up on every InMail message, even if you think it's just a robo-message; 9/10 times I follow up with InMail, the recruiter gets back to me. Brush up on the core tech you want to work with; if you want to do C# Web Dev, then make a project in .Net MVC or something. If you wanna be a Unity dev, make a game in Unity, etc. Of course, your mileage may vary depending on the company you apply to or the area you're in, but these guidelines have worked for me pretty wells

That's what I thought. It seemed stupid that nearly all of the questions on the quiz were algorithmic. And the company claims to "find your strengths and weaknesses and match you to the right job"

Like I said, your mileage may vary. From what I understand, the algorithmic type questions are more popular in Silicon Valley, but I live in an area with a lot of defense contractors and the more popular measure of skill is experience rather than arcane knowledge.

it's not entirely programming related... but does anyone here have experience in getting a publisher, or doing fundraising? most of our experience tends to be companies jerking us around and then just ghosting us. we end up wasting time going back and forth and getting a polished build for them, and then leave without any feedback -- even after follow-up emails. we've done a lot of work since we shipped our early access on itch.io but I'm not sure if that's enough to change impressions we've already given to some publishers?

Hello I am trying to stream processed OpenCV video from a remote Raspberry Pi to my computer running ubuntu. In particular I am trying to utilize GStreamer. A snippet of my Pi code (Python): import cv2 from PiVideoStream import PiVideoStream vs = PiVideoStream().start() out = cv2.VideoWriter('appsrc ! videoconvert ! ' 'x264enc noise-reduction=10000 speed-preset=ultrafast tune=zerolatency ! ' 'rtph264pay config-interval=1 pt=96 !' 'tcpserversink host=INSERT.IP.HERE port=5000 sync=false', 0, 32.0, (320, 240)) while True: frame = vs.read() """OpenCV algorithm that is irrelevant""" out.write(frame) key = cv2.waitKey(1) & 0xFF if key == lord("q"): break out.release() vs.release() cv2.destoryAllWindows() To start my GStreamer pipeline in terminal: gst-launch-1.0 -v fdsrc ! h264parse ! rtph264pay config-interval=1 pt=96 ! gdppay ! tcpserversink host = INSERT.IP.HERE port=5000 And then I try to view it in VLC by trying to open a network stream at, with no luck: tcp://INSERT.IP.HERE:5000/ Another question I have along with this is, does the GStreamer module for openCV need to be enabled for this to work? From my understanding it was separate but now I'm not so sure anymore

How would I find out the structure of a binary file where I do not know the format?

If you're lucky and it's not encrypted, binwalk will try to find any recognizable formats in there or you can try changing known values in whatever writes to the file and comparing the differences. If that fails, attach a debugger to the program that writes to it

most files have a header of somekind. od -c [filename] and look at the ASCII printout of the first few bytes

It's propriatary stuff from a scientific instrument. I tried binwalk but it's really damn complicated.

Binwalk is interpretting compression but, without knowing much about any LZMA, it look incorrect because the dictionary sizes for a lot of these is either 0 bytes or -812384256 bytes

The first bytes (026 \0 \0 \0 001 \0 \0 \0 001 \0 \0 \0) are almost definitely little endian 32 bit values (integers?) because of the zero bytes following them

I have ASCII files that are basically dumps of old binaries of the same format. I just realized I could probably train a neural network to crack this file format.

I did the triplebyte test for fun a few months back. Once you pass the algorithm part (you probably will if you haven't already) they do an hour-or-two-long skype interview and if that checks out you go for the third round, which is where they assign you a person to set up interviews with the actual companies. You also fill out a survey about your interests/what type of companies you want to work at, and they throw a bunch of random companies your way you can apply to. I never bothered setting up the company interviews though.

Any ideas why string connectionString = @"Driver={Microsoft Access Driver (*.mdb)}; Dbq=" + Settings.DatabaseLocalLocation + ";"; //Settings.DatabaseLocalLocation is Settings.DatabaseLocalLocation = Path.Combine(Path.GetDirectoryName(Settings.SettingsLocation), "PROJECT.mdb"); is throwing errors but string connectionString = @"Driver={Microsoft Access Driver (*.mdb)}; Dbq=C:\Users\MYUSER\AppData\Roaming\PROJECT\PROJECT.mdb;"; is fine? Even hardcoding the string into Settings.DatabaseLocalLocation throws the errors. System.Data.Odbc.OdbcException (0x80131937): ERROR [IM006] [Microsoft][ODBC Driver Manager] Driver's SQLSetConnectAttr failed ERROR [01000] [Microsoft][ODBC Microsoft Access Driver]General Warning Unable to open registry key 'Temporary (volatile) Jet DSN for process 0x1c88 Thread 0x2828 DBC 0x751af24 Jet'. Repeated over and over, and it still reads the DB, it just doesn't read all of the DB. Language is C#

How can I handle a 404/403/401, etc. response from an AngularJS route? For example, if you get a 401 error, then it should show a 403 View. If you get a 404, it should show a 404 page. As it is, there's only an "Otherwise" route which handles just 404 (maybe 401 and 403), but there's no way to route specific errors