That's a bummer. I still don't know what it's useful for, except for not having to type SSH passphrases. I think it doubles as a password manager? I don't need that, I use Bitwarden.

Is that why my gf keeps getting that wallet thing on every reboot and I don't? She only wanted to put a password for updates (well not really but she has no choice) there is nothing on her pc but games, no personal info

Yeah, it is safe.

But last I checked, enabling autologin means your GNOME Keyring / KDE Wallet won't unlock automatically. Something to keep in mind.

I don't know anything about anything, but as far as I know anyone with physical access to an unencrypted Linux computer can get in easily. I agree with you that auto login should be fine for your threat model.

It does, the link in the other reply is describing the problem exactly but for krunner

Odd, I don't run into this myself. Does krunner (the dialog that pops up when you hit alt+space) also have the same behavior?

Yes, sorry! I forgot there's two Fedoras.

The list should be updated before the [application] execution happens

This is how I expected it to function. I was really hoping it'd be fixed when I upgraded from 43.

I guess my temporary solution could be to change the shortcut for Notesnook to just "N" lol

Are you using KDE? I've recently switched to KDE to try it out after using gnome for nearly two decades and have come across this annoyance.

It searches the first letter then waits for a pause in typing before searching again which has slowed down my workflow significantly.

This seems to be the bug report regarding the issue: https://bugs.kde.org/show_bug.cgi?id=455239

@klisurovi4@midwest.social

I tested building from the source repo the AUR PKGBUILD uses. Vendored libupnp needed -Wno-incompatible-pointer-types to build with modern GCC. Otherwise, the build succeeded without a problem (see the end of this comment).

I also tested the executable itself, and it seems to work too, although the tool doesn't expose an option to pick the device to listen to, which may not work for users with multiple active network interfaces, if the wrong one is picked.

Here is the executable if you don't mind running binaries from a random person on the internet 😉. libfuse.so.2 should be the only non-system dependency.

git clone https://github.com/amiri82/djmount
cd djmount
# apply the diff below, then
./bootstrap
./configure
make
# the executable should be at djmount/djmount

 libupnp/configure.ac | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libupnp/configure.ac b/libupnp/configure.ac
index 8c073e7..889ed9f 100644
***
a/libupnp/configure.ac
+++ b/libupnp/configure.ac
@@ -56,7 +56,7 @@ RT_BOOL_ARG_ENABLE([tools], [yes], [helper APIs in upnptools.h])
 #
 # Checks for programs
 #
-CFLAGS="-fcommon"
+CFLAGS="-fcommon -Wno-incompatible-pointer-types"
 AC_PROG_CC
 AM_PROG_CC_C_O
 AM_PROG_AR

If the AUR package works and nobody commented otherwise you might want to take another look, esp. at the PKGBUILD which can help you build it on other distros, too.

Or just install it from your distro's repos?

How old is your router?

Yeah, as I said I currently access the server via FTP. I was just hoping there would be a better way to do it, because, while working, this isn't as reliable as I'd like it to be.

I guess I didn't really phrase myself correctly. I found a bunch of stuff but nothing that really helps. Most of the threads were either ancient or not related to my use case. I found one promising utility (djmount) but it was last updated like 2005 and I had no luck getting it to run on my pc.

That being said, I didn't think of searching for service discovery protocols

Yeah, I found it too at one point. Ancient is an understatement lol. I gave up trying to make it work after a few hours. It's honestly crazy that nobody has made another utility like it

I searched for "upnp fuse" and found a project called "djmount". The code is ancient (last updated 20 years ago), but it may still work (AUR package uploaded in 2023).

If if it indeed works, then you can just mount your shares, and access them system-wide, from any file manager or utility.

My google-fu appears to have failed me, because I have genuinely found nothing

This is strange. My search for "fedora upnp" returns plenty of possibilities. "linux upnp filesystem integration" as well.

Apparently simple service discovery protocol is another important keyword.

I don't use this myself...

Not a solution to your exact problem, but when you are connected to a remote server via Dolphin you can right click anywhere in the file view and select "Add to places" (exact wording might be different) to create a shortcut.

That way I have easy access to several different drives. Webdav for my Nextcloud server, sshfs for most of the Linux machines in the house. And when we still had Windows PCs I'd access those via Samba. FTP is possible as well.

But yeah, UPNP is not aa universal as one might hope. Every time I try it with something it's buggy even on devices that supposedly support it fully. So I usually end up creating alternative access methods, like you did.

The article might've be interesting as purely historical reflection, but the actual commands and code comparisons following are absolutely pointless:
On any distro that isn't explicitely non-systemd, one should use systemd commands & syntax these days.

Also, init.d is a directory, not a command.

The article also isn't dated. And just from skimming, there is at least one objective error in it.

That blog also vies for Google money, and OP is one of those "power posters".

@kkarhan @linux4noobs@programming.dev @linux4noobs@lemmy.world does systemd submit to #ageverification ??

@codewizard @linux4noobs@programming.dev @linux4noobs@lemmy.world np.I think it's the best "non-partisan" explainer that doesn't expect you to read through #POSIX, #LinuxFromScratch and #LinuxStandardBase #documentation and just focuses on the "How?" "When?" and "Why?"…

@kkarhan @linux4noobs@programming.dev @linux4noobs@lemmy.world thanks for sharing this video.

@codewizard @linux4noobs@programming.dev @linux4noobs@lemmy.world #systemd is the standard and it integrates well woth modern applications.#SysVinit is the preceding standard inherited from old #Unix OSes.I recommend #BennoRice's Video as an introduction…

To make binaries trusted, reproducible builds are a thing. People build the same software from source independently and get the same binaries. Then it can be said that the binaries haven't added code missing from the source. It is a difficult goal to achieve because how many complex moving parts are needed to build modern software. Much work relies on distributions' standardized packaging infrastructure. Read more about monthly achievements in this field: https://reproducible-builds.org/

Hash, as anamethatisnt said, is just to confirm nothing's corrupted.

Without knowing how to read code, best you can do is check the issues section, the number of "stars" in the repo (similar to likes), if the AppImage is provided by someone other than the original author (common when the original project doesn't include an AppImage), and other indirect signs.

If you know a bit of troubleshooting, not to be confused with knowing how to code, you could also download the AppImage to a virtual machine, cut all direct communications of the virtual machine to the internet and to the host machine, and unpack the AppImage to see if any files are amiss.

Alternatively, but that also requiring some knowledge of troubleshooting, you could ask a LLM to make an "AppImage recipe" for the program you want and it should explain step by step how to make the needed AppImage. And troubleshooting comes into play because you better check at each step if it isn't hallucinating or linking you to shady sites (latter extremely rare in my experience but better than to trust blindly).

Yeah, somewhere along the line you end up with a question of trust. "Do I trust the developer of this AppImage?", "Do I trust the result of this automated tool that checks the code for malware?" or "Do I trust my IDE and myself when I downloaded the source and tried to verify it in my sandboxed VM?".

My main point was that the hash doesn't really tell you anything about the source, except whether you got an exact copy of it or not.

You can't exactly read code and determine it's not malicious.

This is exactly why the rise of 0 width characters being used in malware is scary: Human readable source is not 1:1 with human verifiable behavior.

We've entered an arms race of "use automated tool. Review automated tools work. Used automated tool to review automated tools work. Review automated tool's automated tool's work..."

I am personally not going to start reading assembly.