Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
I've recently gotten this idea to take a Chromebook and run Linux on it.
Why a Chromebook, you ask? Well ChromeOS is a Linux kernel, boots using coreboot, and Google now requires that OEMs support LVFS (fwupd), i.e. the Chromebook is a very Linux friendly laptop.
Some other reasons I would use a Chromebook are: it's cheap, basically useless otherwise, amongst other things.
So, what do you guys think, good idea or not?15
Argh! (I feel like I start a fair amount of my rants with a shout of fustration)
Tl;Dr How long do we need to wait for a new version of xorg!?
I've recently discovered that Nvidia driver 435.17 (for Linux of course) supports PRIME GPU offloading, which -for the unfamiliar- is where you're able render only specific things on a laptops discreet GPU (vs. all or nothing). This makes it significantly easier (and power efficient) to use the GPU in practice.
There used to be something called bumblebee (which was actually more power efficient), but it became so slow that one could actually get better performance out of Intel's integrated GPU than that of the Nvidia GPU.
This feature is also already included in the nouveau graphics driver, but (at least to my understanding) it doesn't have very good (or none) support for Turing GPUs, so here I am.
Now, being very excited for this feature, I wanted to use it. I have Arch, so I installed the nvidia-beta drivers, and compiled xorg-server from master, because there are certain commits that are necessary to make use of this feature.
But after following the Nvidia instructions, it doesn't work. Oops I realize, xrog probably didn't pick up the Nvidia card, let's restart xorg. and boom! Xorg doesn't boot, because obviously the modesetting driver isn't meant for the Nvidia card it's meant for the Intel one, but xorg is to stupid for that...
So here I am back to using optimus-manager and the ordinary versions of Nvidia and xorg because of some crap...
If you have some (good idea) of what to do to make it work, I'm welcome to hear it.6
Just came across the NERSC Docs (https://docs.nersc.gov), absolutely wonderful open source docs by the National Energy Research Scientific Computing Center.
I believe they were written as a guide for people that would be using their super computers, but it's a very good linux beginners guide.
Plus, it looks nice (no visible dark mode tho...).
What's a good password manager for Linux?
A few (optional) conditions (in order of preference):
1. It's free
2. It supports ssh, gpg, etc.
3. It has a GUI (a nice one with gtk/qt support)
4. It's (properly) secure
5. It has FIDO U2FA support (i.e. supports physical security keys like Yubikey or Solo)
6. It has a browser extension
7. It's compatible/non-conflicting with gnome-keyring17
Everybody saw this coming! A privacy breaching telegram vulnerability!
I find myself using explainxkcd more often on my phone...
Because Firefox doesn't show the whole title text...
I was casually browsing some issues for a project on github, and I certainly across an issue where someone wanted support for for URI handling (there's a good reason for his request).
I was bored, and I just wrote a script to do exactly that. I'm gonna polish it on Sunday, and then upload it, (plus add a PR).
I just started maintaining a few AUR packages, and I got to say it's rather fun and rewarding, just to know that your responsible for making sure that something is up to date, and that no one using the package is getting anything bad.2
Any one ever heard of the Solo? It's basically an open source FIDO compliant U2FA usb (with planned support for PGP/SSH key storage!).
The guys who made it are now miniaturizing it into the "Somu" (Secure Tomu).
Please support it! It's a great project and a great (and cheap) addition to basic system security.
What would all you guys say is a good (preferably easy) language for writing CLI applications? Something that runs fast, the less dependencies at runtime the better, and (this goes lower on the list)of thess logic required for argument handling the better.26
Tl;Dr This guy thinks apple is poised to switch the Macs to a custom arm based chip over x86! He's now on my idiot list.
"They've made a custom GPU", great! That's as helpful as "The iPad is a computer now", and guess what Arm Mali GPUs exist! Just because they made their own GPU doesn't make it suitable for desktop graphics (or ML)!
"They released compilation tools right when they released their new platform, so developers could compile for it right away", who would be an idiot not too...
"Because Android apps run in so many platforms, it's not optimized for any. But apple can optimize their apps for a sepesific users device", what!? What did I miss? What do you optimize? Sure, you can optimize this, you can optimize that... But the reason why IOS software is "optimized", and runs better/smoother (only on the newest devices of course) is because it's a closed loop, proprietary system (quality control), and because they happen to have done a better job writing some of their code (yes Android desperately needs optimization in numerous places...).
I could go on... "WinTel's market share has lowly plataued", "tHeY iNtRoDuCeD a FiElD pRoGrAmMaBlE aRrAy"
For apple to switch Macs to arm would be a horrible idea, face it: arm is slower than x86, and was never meant to be faster, it was meant to be for mobile usage, a good power to Wh ratio favoring the Wh side.
Laides and Gentelmen! It my pleasure to present to you the next level in the Linux desktop! MATERIAL SHELL!!!
Demo video: https://i.imgur.com/2UVZTnk.mp426
I've been waiting all week for the Manjaro maintainers to release the next update wave.
Ahh... finally something to do in the terminal :)6
I just upgraded to an SSD (finally! I know right?), and holy crap, SSDM showed up in less than 2 seconds!
Oh I've been missing out on life!
IT'S SOOO FAST!!!!4
YES! YES! YEEEEES! FINALLY!
I've been trying to get some systems to run successfully on docker swarm. Now forget a regular swarm, there have issues just getting it to pass CI builds, and Finally, YES FINALLY it passed!!!!
I now need to pass my self on the back...
We're getting there!
These stats are based on User Agent, check https://w3counter.com/globalstats.p...9
Wow. I feel kinda bad!
I just raged at someone for removing a feature, that disappeared because I CHANGED my config file.
I just feel stupid... Wow.22
Man I'm annoyed!
TL;Dr what does it mean "we're trying to reduce options to a minimum", why don't you go closed source!? why don't you remove themes!?
For anyone who uses rofi, they would know that a few months ago an update made it more compliant with the free-desktop spec, that it only uses the first .desktop file for the given Name tag.
I only found out about this recently as I was only able to update Manjaro recently, and it really annoyed me, cause it took me a while to figure out why tons of my desktop entries disappeared.
Turns out someone made an issue about this, and the given answer was: "that's against the spec". Ok, fine. But when I asked if they could add an option to still ignore that aspect of the spec (i.e. --show-duplicated), the response I got was: "going against the spec is a no-go". WHAT!?
There are so many things that have behavior that goes against the spec (ex. gnu-utils), why can't they add an option to do this!? An OPTION!?
When I decided to try (I don't know C yet) and make a PR, the first and last (it got locked afterwards!) comment I got was:
" As explained on #941, this is a no-go. We want to reduce the number of options to the minimum, and non-compliance to a well-defined and widely implemented spec is definitely not something we want."
Why are you so closed minded!? Yes compliance is amazing, but it's not a safety standard, it's okay if you *give an option* to go against the spec!!!!
WHAT THE HECK!?!?!? WHY!?!?!?
Why is a open source project closed to new features that are part if the scope of the project, and require minimal maintenance!?11
I feel like an idiot... I just realized why my builds were failing.
My entrypoint was setup as to not overwrite a specific file if it was already there. The problem is, the said file is auto-generated, and I need a different configuration for it to work properly in docker. But I forgot this, and the result was a failing build, plus me scratching my head for a few hours. :(3
Boy I'm proud of myself! I just managed to produce a nice little awk script, which makes printing a file in a log look good!