18358414

Sorry, forgot to spell it out. If you're already hosting binaries for 3 operating systems it's chump change to host an apk. If you're hosting 3 operating systems, plus a .deb, plus a .rpm, plus a tar.gz, then how much weight is an apk? For that matter some don't even do that, xournal++ hosts theirs on f-droid. Regardless, I wasn't talking about "small time app developers", though I realize I didn't explicitly state that in my original post, so I've inserted an edit.

Technically, yes, but it's literally clicking "build APK" from a drop-down menu in android studio. Even if you prefer to use something like Gradle, building an apk is basically a one-liner. Yeah, it's probably low-volume; but for some apps, they already have binaries for Windows, MacOS, and Linux (in addition to a code stream for mobile android and ios). It seems like building to apk is a small step. The gain would be that the people using your application aren't required to sign up for a google account, though maybe that's why it's mostly the open-source people providing the apk directly. Thinking about it though, I realized prior to you posting that the most likely reason is the analytics; if everyone downloads the apk then you can't be #1 in the 'app store'.

I have an android device that I have basically said "this never connects to google", and I've installed everything by APK thus far. I ran into some issues when I came across Floatplane and a few other packages which just don't seem to provide a direct APK download through official sources. Xournal++ offers the APK directly on their gitlab site for example (as do most of the other open source goodies I wanted to install). With floatplane specifically, I feel like this is an oversight because some of its users (myself included) want to use it to get away from Google... and yet to install it you need a Google account because you can't access the play store without an account. So is there some restriction in the developer agreement that stops people from just providing the APK directly? What about people who use amazon tablets and need an apk to install their applications? And yes, I'm aware of APK mirrors. I'd rather install something provided by the developers because there's at least some trust there. edit: I'm talking more about established projects, not small-timers who make one app for android and that's it (Xournal++ is the smallest dev group I've installed from). edit2: I don't have time for this; I've asked that this post be deleted as I've no intention to respond to any of you.

Looks like I've mostly figured it out; The keys are tight from the factory, so they take some work to get off. The longer keys which have the metal stabilizer can be removed by applying a little force to the plastic springs that hold them in place.

I'm familiar with the force needed to remove keycaps from MX Blue switches, I was expecting something similar to that. What I'm actually getting is something like 10 times that (probably about the same force needed to pull out a well-seated Molex connector) edit: Looks like it's working... apparently they're just really tight from the factory. No idea so far how I'm going to remove the long caps though.

I'm trying to restore a Dell mechanical keyboard with Alps black switches but I'm noticing I have to apply a considerable amount of force on it to get the key caps off. Is that normal? I'm worried I'll break my little plastic keycap puller.

I've got a 20 year old Volks sedan with a dead battery, needs new tires, and the timing belt should probably be replaced soon, and I have no idea how bad the trans is (but it doesn't accelerate uphill). There's a bit of rust near the glowplugs, but I can still get them out easily enough. This year the radiator cracked and a few years ago I had to do the serpentine belt and alternator. Is it worth it to keep stringing this thing along or should I bite the bullet and get a newer vehicle? For reference, this thing's been getting 39MPG (diesel) lately and gets driven about 700km/month.

Name suggestion: "Normal Linux": Linux for normal people. Or call it 'Norma' so that the full name is 'Norma Linux'. While I feel like some of your goals are a bit lofty for a free project, it's certainly a nice idea. Beware of burnout (I've done that on many a project) edit: "norma" might be taken... what about "Norme" (pronounced nor-me)?

It depends on your desktop environment and hardware, but I have Arch Linux with XFCE installed on an external IDE hard drive (Hitachi travelstar HTS421212H9AT00) and my time required to get to a working firefox browser from boot (i.e. from the second I touch the power button to the second I can complete a google search) is roughly 2m15s (that's on a I3 350m processor) A worst-case would be if you were doing something crazy like 'ramroot', which can take upwards of 10 minutes, but loads the root filesystem to ram.

I can only find black cable ties in the store; what about multiple colours so that I can get something that fits my colour-coded use case? Or maybe that would be too low-volume...

Here's a quick example I worked out in 10 minutes to give the basic idea. I'm want to know if any existing programs (or a function in `gpg`) provide this type of functionality: #!/bin/bash ### # USAGE: # ./script.sh [mode] # [mode] can be either 0 or 1: # mode 0 is regularization (remove random data) # mode 1 is obfuscation (add random data) ### # Copyright (C) 2022 Yuri-Fury # This script comes with ABSOLUTELY NO WARRANTY; # This is a free script, and you are welcome to redistribute it; ### mode=$1 #super-secret password MyPass="KrustyKrabPizza" Delim=`openssl passwd -6 -salt $MyPass $MyPass | cut -d'$' -f 4` #file to encrypt FNAME="MyFile.zip" FNAME2="test.wav" ############################################################################### if [[ $mode == 1 ]]; then #print 16K of random data (totally overkill; the amount can be randomized) dd if=/dev/urandom bs=4K count=4 > $FNAME2 #print password echo "$Delim" >> $FNAME2 #print file dd if=$FNAME >> $FNAME2 #print password echo "$Delim" >> $FNAME2 #print 16K of random data (still overkill) dd if=/dev/urandom bs=4K count=4 >> $FNAME2 #encrypt gpg -c $FNAME2 --output $FNAME2.gpg rm $FNAME2 elif [[ $mode == 0 ]]; then #Decrypt gpg $FNAME2.gpg #Filter out random data #Locate obfuscation delimiters POS1=`grep -abo "$Delim" $FNAME2 | head -n 1 | cut -d':' -f 1` POS2=`grep -abo "$Delim" $FNAME2 | tail -n 1 | cut -d':' -f 1` #Offset POS1 by the size of the password POS1=`expr $POS1 + ${#Delim} + 1` #Compute size of actual contained data SIZE=`expr $POS2 - $POS1` #Filter dd if=$FNAME2 bs=1 skip=$POS1 count=$SIZE of="$FNAME" else echo "MODE NOT RECOGNIZED" fi

You seem to be misunderstanding the question: I have an unencrypted .zip file I want to encrypt it with something like GPG Since .zip files have a standard header, it is possible to analyze the first 'n' bytes of the encrypted file to try and crack the encryption If there is a random few bytes of information at the start of the file, it makes that analysis much more difficult. This isn't a new concept, it's a pretty well defined problem (and some cryptography methods actually require padding to make a message long enough to be encrypted). I'm just asking if there's a way to pad files with GPG (or a homebrew solution to accomplish the same thing).

I want to pad random characters to the start and end of a .zip file to make it less likely to be cracked. From what I understand, it doesn't appear that GPG adds a random amount of padding to the encrypted data. Is this correct, or is there some way I can add random padding to gpg-encrypted files? Would doing so be over-kill? edit: for reference, I'm running `gpg --encrypt MyFile.zip`

Is the physical GPU in use at the time you try to add it to the vm? If so, and if this GPU isn't something like a quadro which allows it to be split up, you might be out of luck (unless you have a second gpu); see here: edit: just to clarify, it *is* technically possible to boot a kvm from a headless machine with only 1 gpu, but you won't be doing it from a graphical environment:

"The only winning move is not to play."

Problem: if I have 6 hard drives spun down via `hdparm -y /dev/sd*`, whenever I shut down my Linux computer there's a period of about 12 seconds where each drive that was spun down is spun up again. Since I'm shutting DOWN the system (as in power-off), those drives are spun up for a total of about 5 seconds before being spun down again. At this point, I feel like it's better if I just flip the power switch instead of the proper shut-down because at least that way I'm not spinning up the drives needlessly (and I'm in a live CD environment so I can't damage my filesystem). Question: is there a way to prevent Linux from spinning up hard drives that have been spun down to avoid unnecessary wear when I'm shutting down the system? edit: I believe this is because of a "synchronize cache" call on shutdown, but I have yet to find a way to skip that.

Android is Linux (it runs on the Linux kernel). There is a considerable number of people who own an android device and play video games on it (in fact, it's one of the two main contenders in the mobile market; the other one being BSD-based). Ipso facto, Linux is perfectly fine for gaming. On the desktop (and for AAA titles), I'd argue that the lack of a standardized environment makes it challenging (though not impossible) to support. The reason Windows gets so much better support is because, while Windows is kind of crap, they've standardized the crap so everyone has the same mud on their boots. Linux distributions unfortunately suffer from the fact that users can choose what they want to put on top of the kernel, and that freedom means uncommon combinations are likely. We are human, so we can never expect a consensus on the best combination of software packages. If you give people the freedom to do what they want, you'll get that. You know, if everyone exclusively used Edge for their web browser, everything on the internet would 'just work'. But is that a world we want to live in? Instead, we have choices. Of course, if Microsoft were to release their proprietary libraries to the public such that software like WINE could achieve 100% compatibility with Windows programs, would that really cause a mass migration to Linux? I think not, personally. I use Linux pretty much exclusively, and when I have to use Windows to test code it's a pain. It doesn't even come with a decent text editor. But I feel like the reason I don't like Windows is because I never use it. When you've been using an operating system environment for 15 years, suddenly changing to something completely different is just a pain that you don't want to have to deal with.

I get a similar issue sometimes on Arch Linux with my 750TI, but usually only when I reboot from Windows back to Linux. For a brief moment while the X server is starting up some icons and visual elements from my last Windows session are displayed on screen because I guess NVidia doesn't care to clear out the video memory when you reboot your machine. It freaked me out the first time it happened to me, because I started up Linux and right before the login screen rendered a distorted picture of the Terminator jumped up on my screen. It's been off-and-on for about 4 years now I guess. I strongly suspect you are having the same issue because your image #1 appears to show Windows UI elements that are rendered and not stored as actual files. If you're getting coherent images when you do a fresh boot (i.e. going from completely shut down to completely powered on), something weird is happening; if you get coherent images when you reboot, it's probably just whatever was stored in video ram from your last session. There shouldn't be a way for Linux to coherently read files from a Windows partition until it's been mounted, and even then Linux devs are very picky about avoiding random filesystem reads. I am not currently aware of any method to wipe video memory or re-initialize it with all zeros aside from writing a custom program that just stuffs your video ram with zeros.

Commodore PET 4016: can't recall the specs off the top of my head... but it had an external tape drive Commodore 64: can't recall the specs off the top of my head, but it had a 5.25" external floppy drive IBM 386 with Windows 3.1, 386MB hard drive space, 16MB of RAM, 2x CDRom drive, 3.5" floppy drive, 800x600 display output, SoundBlaster 16 sound card NEC Versa V/50 laptop, 16MB of RAM, 860MB hard drive, Windows 3.1, PCMCIA, removable floppy, colour screen (TFT), but I forget what CPU it had... No-brand Pentium II with 256MB of RAM, 8MB video ram (SIS integrated graphics), 5GB hard drive, 52X cdrom, 3.5" floppy, 250W PSU, and later swapped in the SoundBlaster16 from the IBM PC when the onboard audio died. Compaq Presario S5010NX, 256MB RAM, 40GB hard drive, Intel i810 integrated graphics, Celeron 2.5GHz (don't remember the specific model), 230W PSU which later turned into a toaster. No-brand tower PC with an Athlon 1800+, 256MB RAM (later upgraded to 1GB), 40GB hard drive, Integrated GeForce 2 MX graphics, DVD-ROM drive, later upgraded to have a DVD+-RW drive, again later upgraded to increase hard drive capacity, then used it as a router for a couple of years before replacing it with a cheap laptop. (at some point I had another HP computer which I pulled off the side of the road, 40GB HDD, 256MB ram, some kind of celeron processor, but it ran away one day when I was at school). ACER WLCI(somethingorother) laptop 1.2GHz celeron processor, 256MB Ram, I don't know much else about it other than it was enough to play Midtown Madness. ASUS EEEPC 701 with 512MB RAM, 4GB soldiered hard drive (no expansion option), 7" screen. I believe this had a Celeron processor underclocked at 600MHz ZaReason Terra HD laptop, 14" screen, 1GB ram, some kind of 64-bit capable dual core ATOM processor, 320GB hard drive. Later the frame cracked and the battery died, so I bought a USB NIC and started using this laptop as a network gateway/router. Acer Aspire laptop with I3 530 processor, 6GB RAM, 500GB hard drive, 15-inch(ish? Maybe 17?) screen. Currently my daily driver laptop. Home-built Phenom II X4 965 BE with ATI HD4200 integrated graphics, 4GB RAM (was 8 at one point, but one of the memory channels died), currently 4TB of storage in RAIDZ1+0 serving as a storage server, but was my main daily-driver from 2010 to 2016/17. Upgraded around 2015-ish to add a GTX750TI. Home-built Ryzen 3 1300x with 8GB RAM, 512GB NVME SSD+1TB HDD, re-using the Compaq Presario case, Blu-Ray drive, GTX 750TI (from the Phenom build). I know I'm forgetting a few. Computers 1 through 9 and 12 were hand-me-downs/computers that I saved (or scrapped together) from the municipal garbage dump, 10,11,13,14 were ones I spent my own money on. Except for 3, 7, and 8, all of these computers still work; #7 I think I can fix, but it's not worth it. The No-brand Pentium II was basically scraped together from whatever I could find. There's something of a story behind each computer, but I've already written too much. The real story is that I need to get rid of some of these.

Don't know what error message Linus specifically got, 'nor do I know what packages he had installed when he tried running 'apt get' so I can't say whether there was an autocompletion (eg: pkgfile) tool trying to "help", but you can install apt on arch via the AUR: https://aur.archlinux.org/packages/apt But it doesn't even have to search aur; here's what pkgfile's 'command-not-found.bash' hook returns when you just type 'apt': apt may be found in the following packages: extra/jdk7-openjdk 7.u261_2.6.22-1 /usr/lib/jvm/java-7-openjdk/bin/apt If I didn't already know that this wasn't the debian package manager, I might reasonably assume that installing this "dependency" would provide 'apt'.

I've had nothing but issues with Gnome. Randomly locking up due to its dependence on communicating with dbus (having your whole DE lock up until the gnome-settings-daemon times out is annoying, also annoying is the fact that my journalctl gets stuffed with 'gsd-binary timed out' even on a fresh install), sometimes it randomly breaks every GTK GUI because of a 'tracker-miner' process that goes commando (requiring you to run a tracker reset), as well as the whole desktop restarting every time you save a file because your /home/user/.config/user-dirs.dirs broke during an update (this will play tricks with other desktop environments, but Gnome handled it the worst that I've seen). I think they dodged a bullet not using Gnome. It may be the default for many distros, but that doesn't mean they've implemented it correctly. In terms of being able to efficiently use your machine, I'll say Gnome is pretty good (it gets the job done)... but if you don't know how to fix these under-the-hood things, it can be a very bad experience.

I'm kind of torn between the two. Since I'm involved in research that involves heavy computation, I need something that I can dump half a terabyte of information on and leave running for several hours processing. Desktops with 32GB of RAM + a Quadro GPU aren't too expensive when you compare them to equivalent laptops. Better yet, if I can afford a better GPU in the future, I have the option... and that option doesn't involve buying a new everything. I also like the added benefit that I can use an old computer as a compute node to offload some low-priority tasks. Using an old laptop for this looks kind of goofy and I find it awkward (not to mention loud when I hit it with an all-core load). Part of me wants a cheap laptop because it would make it easier when I go to visit family; then I can just use VNC for the heavy lifting... but I hate the fact that I'll have this thing with a screen and a keyboard and the most likely end-result is I will never be able to use that screen or keyboard for anything else -- they're stuck to the device. So what happens when it goes out of date? More e-waste because the screen and the keyboard are a part of the device. The laptop I use is a i3 530 from 2010. Realistically it's pretty closely matched for being able to do VNC, but the IGPU is knackered and it freezes for a couple of seconds every few minutes; not to mention the keyboard is grotesque. Seems like a frame.work laptop would likely match my needs, but then again, so would a Rasperry Pi with a decent screen attached. So overall, when it's time to throw down, desktop wins over laptop. I could be convinced otherwise with some compelling arguments, but they would have to be VERY compelling.