Yeah, remembered that wrong, thanks for the correction! Corrected it in my comment.

winecfg can add drives too, I always put an U: to /home/username , that makes it easier to pick a better location to install the game files (in my case, to ~/Games). A lot of things can be tweaked with winetricks and winecfg hahaha.

I think WINEPREFIX is all those tools need. Lutris can use that too (if i remember correctly).

For the setup:

Final Edit: The wine prefix is the same for the games and the installer, just a windows subsystem that has all the dependencies.

First I install the wine-cachyos package, create a WINEPREFIX variable for the folder i want to put the prefix in, then i run winecfg to start the prefix, then I install the dotnet packages and vcredist packages through (before, when the repos existed) these links for the vcredist and dotnet (maybe archive.org or somewhere else still have the executables), or using winetricks too, but these are nicer to use as they install everything in one go.

Then I install the dxvk and vkd3d (you can install through lutris, so this is kinda useless hahaha).

Then I run the fitgirl repack through lutris and click to install the vcredist or .net when asked.

Then I just run the game with a runner or wine-cachyos.

I’ll update this with valid links when I find them.

Edit: last time I used this and this, installing just the .net runtimes, not directx, java, silverlight, etc.

I use it all the time on linux, what problems did you have?

Another tip is the YouTube mobile extension. It hides the shorts and is a lot faster than even chrome, but some people might not like it like that. If on a linux tablet just use this link.

To revert to desktop, use this link.

Edit: On ironfox this worked, on librewolf on an intel tablet this made it go like 10 fps.

deleted by creator

What is your purpose?

Do you just want a job or do you want to program for fun? If you want backend, that’ll be very different than frontend, do you want corporate or startup?

I didn’t care about that, but I think java, for backend would be a first choice, as it has the biggest amount of “it just works with that”.

I think scala is cool too, very modern and with A LOT of less boilerplate, all the JVM support, so it can work with almost anything in the backend, but for working for jobs, it’s not very used.

The most used language has the most of the worst code, quarkus is a new framework (and very good), ejbs are enterprise java beans (the worst thing ever made).

I think a lot of the code was written by a lot of people in a rush, very backend of the backend, but EXTREMELY needed. Java is hated (I hate it too), but the new things in it? It’s amazing too (I only love the new stuff).

Gson in the corner murdered:

it wasn’t funny in any way, but JSHITBOSS “microservice” (it was a jboss service with one microservice inside, in a kubernetes pod, with only one core, tell me about redundancy). Service classes with over 2000 lines of code, it shouldn’t even be called spaghetti code, more like lasagna code, the pasta came in layers, separation of concerns was a mere suggestion, code was not thread safe (and it needed to be), but there was only 40 Ejbs for each “stateless” service inside de EAR, so number go up, code goes better.

I refactored it, it’s now in the glorious quarkus 3.27, on virtual threads and java 21, not hyperbole, but 5x more throughput and you don’t need 7 phds in italian cousine.

Edit: I also saw, in Angular, the infamous:

if (variable === true || variable === 'true' || variable === "true") {

Their drivers are SHIT, for wifi there is CNSS, ICNSS, then QMI, all do the same thing, but differently, but NO, it’s the same thing, but what does this do??? Is this really a different event or is this just rewritten in that event? Idk still, no one knows.

Edit: I tried to port the not working kernel drivers for the wifi on the Redmi Note 9s to postmarketos (wifi is not working), didn’t work and it’s now on LineageOS

Yeah, it’s a joke, but I disagree on the void, the other languages on the meme doesn’t need a return type (when they are returning nothing), java needs it.

And the best:

public static void () {}

I just put a video of a depressed phd student trying to make some money on, like dig, that explains some of the bronze age archeology and history.

Are you tring to study for some exam or just learning about something?

The Gripen is quite interesting, who would have guessed that licensing fighter jets for manufacturing on other countries (Embraer is building them too, Colombia bought some, other countries seem to want it too) would be a good strategy?

I’ll wait 50 years for a Gripen to be sold on Aliexpress for 150 thousand dollars, 250 thousand with replacement parts with a cool paint job.

Yeah, I need to kill the witness to that crime.

Thanks Facebook!

Edit: Corrected the integration

This is a simulation I made for a ball falling (the principles can be used everywhere):

import numpy as np
import matplotlib.pyplot as plt


def calc_acceleration(pos, v):
    return np.array([0, 0, -9.8])


def integrate(pos, v, a_old, delta_t):
    pos_new = pos + v * delta_t + 0.5 * a_old * delta_t * delta_t

    a_new = calc_acceleration(pos_new, v)

    v_new = v + 0.5 * (a_old + a_new) * delta_t

    if pos_new[2] < 0.0:
        v_new[2] = -v_new[2]

    return pos_new, v_new, a_new


# Constants
delta_t = 0.01
t_max = 10

# Time evolution
times: np.ndarray = np.arange(0, t_max, delta_t)

# Initial state
v = np.array([15.0, -3.0, 15.0])
pos = np.array([0.0, 0.0, 7.0])
positions = np.zeros([times.size, 3])
a_old = calc_acceleration(pos, v)

# Simulating
for i in range(times.size):
    pos, v, a_old = integrate(pos, v, a_old, delta_t)
    positions[i, :] = pos

# Plotting
fig = plt.figure()
ax = fig.add_subplot(projection="3d")
ax.scatter(positions[:, 0], positions[:, 1], positions[:, 2])
plt.show()

Very simple, you can see a ball falling and getting up again from that starting point, so you just see at least some starting point on how to do this stuff, the basics are not that difficult. But every simulation is like this, just more complicated hahaha.

Oh yeah, someone finally asked something about my field.

So, maybe COMSOL is what you’re looking for, but if you want the cool, smooth and amazing stuff of the pros:

Time to learn numerical calculus, like verlet integration, maybe some calculus for turning potential fields into force (maybe some forward differentiation with dual numbers, just like sympy does hahaha, you can use that too). If it’s a hairy equation, that takes some hypergeometric, bessel, etc, scipy has all the special equations.

I think verlet + sympy for turning the potential fields into the force fields at the particle spots will do amazing, if you don’t want paper ready simulations. Like, pos[n, 3], v[n, 3], a[n, 3] (maybe some torque if you want) would be almost all you need.

For visualization, matplotlib has a lot of stuff, you can make animations and videos.

If you want to simulate the fields itself, you can use finite differences or finite element methods.

If you need a lot of performance, I recommend learning Julia or the GODLY Fortran 2008 (fprettify, fortls, gfortran will help to build a modern Fortran program that you can call from python using cdll and ctypes with the ieee c libs for fortran).

I tried KDE mobile (and a lot of other DEs) on postmarketos, seems everyone has the same experience, it drains the battery a lot faster than other phone DEs.

Yeah, the pin being only 6 digits is quite bad I liked the mobile version, but there are still some issues to iron out.

I like the vanilla gnome on my tablet, but smartphones are quite different. As for phosh, it seemed to be kinda buggy on my smartphone, I couldn’t lock it and it had some other problems. Linux on phones have some road ahead to be good, but we’re getting there. Maybe kde mobile might be better, but it sucks all the battery (idk why).

I use CachyOS on my X220 with btrfs and lzo as disk compression (lzo is very good on old cpus and makes the SSD go really fast). But I think any distro could be good on that hardware.

As a side note, I would really like an x86_64-v2 distro, people jumped from no additional instructions to v3 in no time, but these thinkpads and older pcs could really shine with that kind of optimization.