[ Home ] [ wiz / dep / hob / lounge / jp / meta / games / music ] [ all ] [  Rules ] [  FAQ ] [  Search /  History ] [  Textboard ] [  Wiki ]

/hob/ - Hobbies

Video game related hobbies go on /games/
Password (For file deletion.)

  [Go to bottom]   [Catalog]   [Return]   [Archive]

File: 1613143570126.jpg (515.77 KB, 1196x1178, 598:589, frog.jpg) ImgOps iqdb

 No.56871[View All]

172 posts and 28 image replies omitted. Click reply to view.


File: 1647701286742.png (679.98 KB, 1024x768, 4:3, 2022-03-19.png) ImgOps iqdb

>try to look at the html of twitter. it's not even human readable. Reactjs is fucking gay. it's also these javascript websites that have weirdly autogenerated obfuscated elements and names
It's really bad. If you scroll down twitter the next block of posts is returned as a (deeply nested and garbled) json response based on scrollbar position, then that response is appended to the dom with javascript. It's designed to be slow and hard to scrape. Thankfully, like you say, people have made better public apis.

I've been porting my image/booru viewer from lisp to c. The initial port used sdl, but it's 2slow4me so I'm using plain opengl now. The first problem I encountered was managing the limited number of texture units on the gpu, which sdl does for you, because there are more image thumbnails than texture units. In the screenshot the 512:512 texture is uninitialized memory, then the unordered thumbnails are stamped in the texture using shelf next fit. Tomoko on the right is drawn using the normalized uv coordinates returned from the function.

Also, I like the opengl 1 api better than the newer versions. You should be able to open a port and stream instructions to the gpu, or do it the ogl 1 way. The newer stuff overcomplicates things. Different stages of the modern pipeline could be exposed as registers instead of the inout parameters in shaders for example, it would be simpler and faster.


File: 1648231227422.png (268.44 KB, 1024x768, 4:3, 2022-03-25.png) ImgOps iqdb

Added text stuff and asynchronous autocomplete with caching.

The main problem I had was my string pool used realloc to resize the pool, but I forgot you may lose anything that was pointing into the previous allocation, so my strings would get overwritten with server responses, and the async stuff would write into freed memory and crash the program. The fix was to expand the pool by putting new allocations in a list, so they don't change over the lifetime of the program.


File: 1648571686107.png (861.55 KB, 1024x768, 4:3, 2022-03-29.png) ImgOps iqdb

Now we can browse pages and download stuff.

There's not much new functionality. The program just extracts the necessary fields from each post (if they exist), compares them to a cache of posts, and if we don't have it download/decode the thumbnail, adding it to the texture atlas. The draw list just points into the cache and is assembled on the fly. Most of the speed improvement in this area is from using persistent connections and decoding jpegs in parallel, but we're ultimately limited by the speed of the server (200-300ms) unless the cache is hit.

The program is basically finished, but there are still lots of refinements to make:
>be able to navigate by rows (the program just blindly draws until it reaches the edge of the window, so it has no idea what a row is)
>center thumbnails vertically
>rendering is incredibly slow, we spend 3ms making 352 opengl calls just to draw a page, it can be sped up by throwing all vertices/uvs in an array
>maybe prefetch the next page to make browsing faster
>add a download queue, instead of blocking the program like a retard every time `d' is pressed
Anyway, I'm really happy now it's usable, this is much nicer to use than my first attempt.


the university should be able to offer significant help with finding an internship if it's a requirement


I got 200 rejections and was only given the chance at one job interview that I also failed. 2:10 is an incredibly ratio imo.


im figuring out the mes file format used in harvest moon and similar games. i googled it and can see other people have done this, but there doesnt seem to be anything working for river king on ps2 (harvest fishing). i already can mod the audio of this game, changing the dialogue now would be cool


ffmpeg -y -f dshow -i video="webcam" -frames:v 1 "test.jpg"

this saves a photo from the first webcam connected to pc… ffmpeg always impresses me. literally anything to do with audio or video, it is possible using ffmpeg. and for images, imagemagick. i might build a little time lapse photography box thing and record plant growth or something fun


File: 1650747502062.png (5.51 KB, 631x90, 631:90, ClipboardImage.png) ImgOps iqdb

i thought youtube-dl and youtube-dlc were just slow garbage but it seems youtube throttles them, because yt-dlp isn't throttled. never heard of yt-dlp but it works interchangeably with commands for both youtube-dl and dlc


how do i add this https://github.com/dlemstra/Magick.NET to unity? i can't find any dll or binaries or anything. it says i need to either build it myself, or use nuget. i googled nuget and it's a website and i found magick.net on there but it downloaded a nupkg file. i googled what the hell is a nupkg and it says it's used by visual studio. do i really need to use that collossal bloatware garbage? i like using np++ to write code


are there any boards out there that focus around technology/coding? something like /g/ but a whole chan for it, if there are any i'd love to hear about them.


File: 1651153094473.png (602.69 KB, 1280x720, 16:9, debil.png) ImgOps iqdb

>it says i need to either build it myself, or use nuget.
nuget is presumably a package manager for wangbows. all it will do is pull down the source and dependencies for imagemagick and build it. you can see what's happening at a high level by reading the build scripts in your link (for instance Magick.NET/src/Magick.Native/build/Build.cmd, install.ps1)

>it's used by visual studio. do i really need to use that collossal bloatware garbage?

you don't need to use visual studio's compiler and linker (cl.exe), you can use whichever c compiler you want to build imagemagick. on wintoddler 10 this will likely be mingw or clang

the instructions for what you want to do are literally in your link, it's just a remarkably badly written wrapper library (i lost my patience when i saw xml in comments), so you'll have to invest some time into examining how dlemstra builds imagemagick and creates a c# binding for it to integrate it into unity. the main difference will be instead of using whatever datatype dlemstra defines for wrapping imagemagick's representation of images you will use unibabby textures instead


man i hate the process of building from source. i know some people find that fun. i find it tedious and it pisses me off when people don't just include compiled builds with releases

i will probably just bundle imagemagick with my unity prject and use it by creating and running batch files. this breaks compatibility for linux and mac though


in the case of a batch/shell script, you'd just check to see if the user has the program installed, but there are more efficient ways without invoking a shell: you could either run the program as a sub-process, or dynamically link against the imagemagick library (.dll on wangbows, .so on loonux) and use c#'s foreign function interface


yeah i know how to run batch files on window as a subprocess and have the window and output hidden. i'm not going to bother trying to oop-ify the process of feeding parameters into an external program, the imagemagick commands are simple enough. thanks


What are some cool things that I can do with C++ available libraries. I can’t really find much other than SFML and JUCE that doesn’t require knowledge in another field like ML or Embedded.


I did a quick programming course that took 9 months and I've been working as a web developer for four years. I use an outdated technology and I sort of feel like an impostor compared to people who formally studied this in university.


File: 1654063439699-0.png (48.75 KB, 1243x682, 113:62, ClipboardImage.png) ImgOps iqdb

File: 1654063439699-1.gif (536 B, 16x16, 1:1, test side walk.gif) ImgOps iqdb

not hardcore but it is still programming. making a simple sprite character creator for myself. will have selection side and then a preview side. 18 graphics can produce an idle/walk animation for cardinal directions, with head/face/chest/legs sections. should be much more but you can flip the east/west sprites among other things. so you pick different body parts and it displays the character to the right, either idle or walking, and rotated facing whichever direction, then you can export that as a sprite sheet… to get a variety of parts to test i just hue shifted the colors from a single actual sprite. one problem i noticed though is that different skin tones don't work when you include clothing and bodyparts/skin together. i'm thinking just making skin color in the actual sprites 00FF00 and then basically chroma keying it to whatever color is selected. the alternative is doubling the graphics required and creating naked versions of everything so you have a layer of skin underneath the clothing but that sounds too annoying


File: 1654240164086.png (772 B, 64x64, 1:1, ClipboardImage.png) ImgOps iqdb

well it outputs correctly. using imagemagick to create all the graphics is fun. maybe because i did art long before learning programming, but the idea you can easily automate this sort of stuff with a few lines of code is just very cool and satisfying


File: 1654240891181-0.png (741 B, 64x64, 1:1, ClipboardImage.png) ImgOps iqdb

File: 1654240891181-1.png (824 B, 64x64, 1:1, ClipboardImage.png) ImgOps iqdb

File: 1654240891181-2.png (603 B, 64x64, 1:1, output.png) ImgOps iqdb

…well that is an interesting bug. first pic is copy/pasted from nomacs image viewer. second pic is copy/pasted from paint.NET. third is actual image upload. i can't explain how they are all different, but i can see them in the file attachment thumbnails as i type this


that's super neat


File: 1656598313969.jpeg (281.49 KB, 1200x671, 1200:671, 46a27943d330b372d0d2b033f….jpeg) ImgOps iqdb

>there's no difference between a c string and c array besides arrays can have 0s anywhere in them
>there's no difference between a lisp string and lisp symbol besides strings can have spaces or parens in them
more programmers need to ask about the sq (the string question)


scheme@(guile-user)> (string->symbol "sp ace")
$1 = #{sp ace}#
scheme@(guile-user)> (symbol? $1)
$2 = #t
scheme@(guile-user)> (eq? $1 (string->symbol "sp ace"))
$3 = #t

solution for double escape: 08/24/21 https://wizchan.org/meta/res/60390.html#60449 (pruned) → https://github.com/towards-a-new-leftypol/leftypol_lainchan/commit/217e873e88f595a0ad7fedfeff40b08924e0701d


File: 1656706915722.jpg (38.57 KB, 235x314, 235:314, Alonzo_Church.jpg) ImgOps iqdb

Bit of harmless fun:

1. Outside code blocks:

scheme@(guile-user)> (string-any #\; "->")
$1 = #f

2. Inside code blocks:
scheme@(guile-user)> (string-any #\; "->")
$1 = #f

So as you can see, while untyped wizchan calculus might be consistent within its domain, it is inconsistent as a deductive system. You don't even need fixed-point combinators to prove this, all you need is a code block. ;)


not programming, but there is no general tech/software thread. does anyone know what would cause drop down menus in firefox to stop working?

for example, when making a post on wizchan, the email selection has a dropdown. no menu displays for me though, i have to click somewhere on it, then press left/right arrow to select something. it's been like this for maybe 5 years so i'm used to it, but every now and then i am reminded of it. like signing up for websites, picking my country/state is always annoying. i have to type 'u' to jump to countries starting with u and then click right a few times to reach usa


AdBlock? No Javascript? A function key like tilda is in an always-pressed state on your busted keyboard?


doesnt seem to be adblock or any addon, persists with everything disabled. i don't think tilde is pressed i bind it for games often. i've changed keyboards within the last year also


File: 1657025978313.png (346.81 KB, 662x662, 1:1, 7bdcc21010f5b119ab46e69be9….png) ImgOps iqdb

that's an unusual problem. you could have a look in about:config at dom.forms.select and see if anythings been changed, and some people using newer compositors suggest it may be hardware acceleration which you can try unchecking under settings > general > performance

i'd also check your userChrome.css and userContent.css (or any themes) to see if there's any styling that may interfere with select dropdowns


the profile css stuff is fine. and that particular config is ok. i just have too many configs set it has to be one of them, if i check 'show only modified' it shows almost 1000. a lot of it is to disable telemetry, sensors, popups, notifications, pocket, etc but a number of problems including this dropdown thign have appeared, like being unable to update addons or the browser itself. the next logical troubleshooting step is probably just to reset the about:config to default to see if that fixes it


it's normal to have so many configs changed, but it may just be after 5 years there's a config conflict as firefox has changed. you could save yourself some time by either trying help > troubleshoot mode to reset all settings temporarily, or go to about:profiles to try running firefox with a fresh profile (you can also do this by running firefox -p in a shell) if you're curious, about:about lists everything

or it could be something completely unrelated, like the doubleclick interval on your mouse being so short it opens/closes select menus instantly, but you would have probably noticed that long before


damn this is cool


what’s a good way to host images? I want users to be able to upload images. I’m running a django back end btw hosted on postgre.


Anyone here remember wizmud? I would like to revive it, maybe I should contact the dev? I would only take the models though.


oh nevermind I just digged and found there's a Patreon and it's alive-ish


File: 1660098546650.gif (2.72 MB, 600x338, 300:169, 1523886380191-0.gif) ImgOps iqdb

ah thats the fucking name. ive been trying to remember that thing for the past 3 years or something


I remember playing videos that I thought where cool when I was younger and annoying older wizards


anyone curl into their netgear? there is a authorization popup thing that displays within browsers for logging in, i'm having trouble getting anything to work. i originally thought you could just pass the username/password with the url and from then on supply requests with whatever cookie or authorization info it returns for interacting with the rest of the site. not so


Serious question to wizzies who know how to code: Is it possible for a nearly 30 yr old to still learn valuable coding skills?

I'm really getting into Html and python atm and I think I'm able to pick up on stuff when I'm not being too lazy.


learning to code isnt too hard. the hard part is joining another team and contributing and interacting with others and using the same tools as them, having a boss, doing reviews, fixing other peoples stuff. im guessing. i read about the sctra stuff that programmers and coders do for a their jobs and it doesnt sound fun. anyone can just learn programming though, as long as BASIC variants still exist


Yes, but I recommend getting into javascript because browser integration, runs it in browser console, runs in extension, runs in bookmarklets and runs in userscript managers like Tampermonkey, allowing to get used faster because is visual and can be used to solve your daily browser problems. E.g: redirecting Youtube to CloudTube.
var url = window.location.hostname
window.location.hostname = url.replace('www.youtube.com', 'tube.cadence.moe')

Some sites you can get userscripts:


Javascript is a really ugly language. The only good thing about your advice is there is a lot of software in node community. But it would be far better to learn Python (as a first language) and then learn a niche tech stack like Nim, Go, or Rust if you want a job. Nim is actually a fucking beautiful language, too.



Yeah, I used to love coding but then I started doing it as a job and now I hate it. Sometimes I think about quitting my job and moving back in with my parents but it's hard to say no to 300k.

Maybe I can find another job as a "junior" dev where I have to do less bullshit and can just chill for less money.


How the fuck am I supposed to feel confident in applying for jobs. I have an IT background but I fucking hate it,and find it suicide inducing. I've been learning webdev on and off for about 2 years. My js is pretty solid, and I'm learning React with FullStack Open. Like, is there a minimum amount I need to know to even have a chance at a job?


Just apply for the job.


Aso net core Mvc

Spent like 5 hours yesterday screaming at the screen because my viewmodel was empty when i tried to update it on it respective post controller. After a lot of stress near the deadline i figured it out. The viewmodel wasnt getting anything because i put a name attribute for my select so it was appearing on the Request.Form array but not on the view model.
Good damn Microsoft fuck net core its a terrible language. Why in the name of good you got select tag helper, dropdown list and other stuff that deserves the same porpuse but it just have different names with minors differences. its confusing and annoying as fuck when you try to google a error but half the solutions doesnt work for you because they are in razor pages or in .net 4.7 or whatever the hell those idiots just build out of nowhere just to disccount it after 2 years.
Im so dick of .net and i dont even wanna try dealing with xamarin, universal forma IOT or any stupid useless shit the envioriment has to offer.



File: 1662451725916.png (239.59 KB, 1351x857, 1351:857, title only view.png) ImgOps iqdb

I added a togglable title only view to wizchan.
It shows the number of newly added replies to each thread since you last visited.


>Tab characters are not allowed in Ren'Py scripts.
>#if c == u'\t':
> #raise ParseError(filename, number, "Tab characters are not allowed in Ren'Py scripts.")
u cant tell me what to do, fartsniffer


how would you go about reversing the reflection from a curved surface? for example a sphere… i think you'd have to raytrace, but in reverse? like project the reflection's pixels values to a curved 3d surface representing the mirror in the image

is there software for this or is it a good opportunity to just make my own tool


File: 1663347484770.png (6.17 MB, 2048x1384, 256:173, ClipboardImage.png) ImgOps iqdb

like this for example. if all you had was a 3d model of the blob mirror, and the reflection from its surface


File: 1663359353754.jpg (2.65 MB, 2417x2396, 2417:2396, Circular_fisheye_view_of_O….jpg) ImgOps iqdb



>Absolute data is specified as the mouse's actual x-coordinate; relative data is specified as the number of mickeys moved. A mickey is the amount that a mouse has to move for it to report that it has moved.

i saw references to 'mickey' and thought wtf is a mickey? then as i typed in "windows mickie mouse api" in google i realized the joke and it made me laugh

[View All]
[Go to top] [Catalog] [Return][Post a Reply]
Delete Post [ ]
[ Home ] [ wiz / dep / hob / lounge / jp / meta / games / music ] [ all ] [  Rules ] [  FAQ ] [  Search /  History ] [  Textboard ] [  Wiki ]