I was visiting a friend, and their whole neighborhood was immigrants. One thing that was clearly different was their lawn barely had grass, at most a patch of 6x6, and everything else was vegetable garden and most trees were fruit trees. 6-7 different types of fruits in the front lawn instead of ornamental plans and seasonal vegetables on backyard.

I really don’t understand the obsession with wasting good land to put useless things. And put so much resources in maintaining it

This. After leaving Reddit I feel like I lost a place I could post about my little foss app and get people to try it. Specially in the niche topics.

I am trying my best to get people I know try but they don’t understand the domain, and I don’t have a reach to people that will understand

I have two folders for git though, my projects, and other people’s packages and such

You’re right. But I guess we could say they know/learn what they need to survive at that moment of infancy. That’s how natural selection works.

Sea turtles learn how to walk in sand, but I’m “guessing” most vital skill for them is swimming.

Considering they just hold back packages, but do not do additional testing to release them, yeah, they should not do that.

Arch already has testing repo, normal repo packages on arch are already stable enough

Good luck with everything. Go at your own pace, hosting is just leaving a computer on. Add more things as you need and that’s all there is in my opinion. But of course you’ll learn security, backups, and other things as you need.

I recommend you gnu parallel. It does similar things, but runs the commands in parallel. And it’s way easier to pipe than xargs. If you really need it to run one command at a time you can give number of cores to 1. And it also has progress bars, colors to differentiate stdout fo different commands, etc.

Basic example: to echo each line

parallel echo < somefile.txt

To download all links, number of jobs 4, show progress

parallel -j 4 --bar ''curl -O" < links.txt

You can do lot more stuffs with inputs, like placing them wherever with {}, numbers ({1} is first) that allow multiple unique arguments, transformers like remove extension, remove parent path, etc. worth learning

deleted by creator

Sure, if we can share config like that it could work for many people. I still want to see something that dynamically downloads the shared libraries/dlls for tools you add, but that requires complete rewrite. Maybe we can do it with imagemagick scripts

Now I’m thinking why don’t we make an image editor that we can customize the simple UI. Like users choose which sliders and tools to put on the ui, it’ll be simple UI with like just 5-10 buttons/sliders/tools, but you van customize it to have basically anything. That way you can simply drag drop tools make new UI and then use it for specific use cases.

Maybe already existing tools have that options. Or maybe we need to make a new one, in that case, it’d be nice if we could just add all different tools in dlls or sth, so that you can only download/keep the tools you use.

I guess it can go unnoticed, I use Arch so maybe that’s why I got more involved. I remember searching why auto completion didn’t work, then finding out I need to install bash-completions package. After knowing that it makes one curious about how it works. Then the next stage is writing it for my own programs because it obviously won’t come with bash-completions package.

I once wrote a shell (terminal) to watch anime, and I wrote auto completion for different commands on it, it was really nice to just type play then prefix and then tab for auto completion on anime names, and even for episodes I wrote auto completion give me last episode I watched + 1.

Whenever someone says they don’t really like terminal because they don’t like to type or remember commands. This is what I think “they didn’t use auto complete”.

Auto complete works for file names and paths by default, but the development can write it to only complete certain extensions. Like auto complete for image program only completes image files. Then you have completion for commands, subcommands and flags.

Auto complete is done through calling a bash script with currently typed line, and the bash script can call other commands. So developer can write a really complicated auto complete and make it available as a binary if they want, and just use that in bash. Or you can use many tools that will generate auto complete script for you based on your commandline args.

If you write your own scripts/cli binaries I recommend learning how to write auto complete for it. Makes it incredibly easy to use the tools.

Inkscape can generate QR code

But you need tdf installed on the server for that right?

I realize I rarely have to do it so I tend to just download and open the pdf, or use X forwarding. Or while using emacs I just open the remote file (which basically downloads and opens I guess).

I was thinking I’d appreciate a website/community where people talk about potential new open source software ideas and interested people can collaborate. Could be small things. And able to filter by category, language, etc.

But I don’t know how the logistics would work. But at least minor things that could be fun. But it could be abused. It’s just my thoughts when I’m like “I wanna make something this week”. Or when I start something and I don’t have the skills for all components.

Arch also kinda allows that if you write custom PKGBUILD file. It’s easy to write for simple stuffs that are based on make/cargo etc.

It’s time consuming if some program gives you 100s of lines of code in bash script to install their program though.

Edit:

Another disadvantage of building from source is dependency management. You might accidentally uninstall some dependencies, the standard library versions might change and break your packages, etc.

Using package manager mitigates that.

Can’t you just keybind the switch to that key? I use arch and I have keyboard layout switch between three languages (one is Japanese which might have similar tech/typing style), and the program I use (ibus anthy) allows me to define my keybind.

Wait, are there repo that just has dating info? You just make PR for your profile. Honestly with GitHub free pages we could definitely do that lol

Then you’re just running bash scripts with bash. You’re not running bash scripts with fish.

The android auto equivalent for cars would be something I’d be interested in, that’s the only reason I had to reenable google on my phone. I don’t see any open source software that do it.