The Random Thread

ChatGPT is literally the worst of them all. Sam Altman is an awful everything.

Google Gemini is scoring the best of all, with Grok being a firm second. Claude is meh. I have seen Gemini produce insanely advanced stuff that no other model came even close to. Grok is also strong is some very useful arenas, but I feel like Grok probably has no advantage of Google's immensely advanced internet search background.

I don't trust AI benchmarks at all. For me, that's not a valid argument. I respect your opinion, though, but my experience is what I've explained.
 
I just found out by accident that if you "shake" the mouse (move it quickly from side to side), Plasma makes the cursor frighteningly huge for a couple of seconds so you can find it.

1773086494050.png


And it seems it can make it even bigger:

1773086877256.png


You can disable this, as the above image shows.
 
I normally use 5 workspaces ("virtual desktops," according to Plasma) and maintain two extra ones at the end just in case. I always use the same "workflow" (I open the same applications for the same purpose in each workspace).

I switch between workspaces using Super+1, Super+2, and so on, and also Super+Repag and Super+AvPag to go to the previous or next one respectively.

1773144259950.png


What about you?
 
Who needs more than one workspace? 🥸
Esp. when considering the amount of RAM needed to keep the relevant apps open anyway? Esp. with browser tabs? And apps like Konsole offering ways to split up your screen for the workflow?
😲

1) I have never run out of memory. My system is also very responsive.
2) I use Kronkhite tiling extension, so I have even another way to split my screen.
3) On the konsole instances of the first and second workspaces I usually have several tabs open.
4) My workflow has a purpose and serves it very well. I can do very quickly and mostly from the keyboard several things that I usually do simultaneously (I rarely do only one thing at a time—while I'm doing something, I'm normally doing one or two secondary things at the same time).
5) I have a second screen that I only use for monitoring purposes. I use the workspaces mainly on the main screen. The secondary screen is normally fixed on something.
6) I guess I'm not the only person on the planet happily using several workspaces, or else such a feature wouldn't even exist.

🤠
 
Keep in mind that you can mark certain windows "sticky", so that that particular window follows you around over all workspaces.

I use that quite a bit when (also) watching video on one workspace.
 
Keep in mind that you can mark certain windows "sticky", so that that particular window follows you around over all workspaces.

I use that quite a bit when (also) watching video on one workspace.
I know that, thank you. I use it as well.
What does your sysutils/neofetch dump show? What about the dump shown by /usr/bin/top?
It says that my system works very well, thank you. I didn't know I was being subjected to an interrogation. It's perfectly fine that you work how you work and I work differently. It's called diversity. Modern software is very pliable and serves many purposes and workflows.

👾
 
I know that, thank you. I use it as well.

It says that my system works very well, thank you. I didn't know I was being subjected to an interrogation. It's perfectly fine that you work how you work and I work differently. It's called diversity. Modern software is very pliable and serves many purposes and workflows.

👾
Some workflows like video editing or 3D modeling or analysis of large datasets take up a truckload of RAM. Try compiling www/firefox from source, you'll run out of RAM in a hurry.

I assume you're familiar with the screenshots thread (Thread freebsd-screen-shots.8877) . There you can see neofetch dumps in all colors of the rainbow.
 
I'm moderately disillusioned with the forum, but it's certainly a source of good ideas for my crumbs. Many times, instead of arguing in a thread, I write a crumb and toss it into the void that is the general internet.
 
Some workflows like video editing or 3D modeling or analysis of large datasets take up a truckload of RAM. Try compiling www/firefox from source, you'll run out of RAM in a hurry.
I found out about OOM killer on Linux trying to compile LineageOS :p

But that was interesting: Wayland was a larger priority than session restore, OOM killer didn't mind taking out DEs, along with somehow still a thing with 1TB swap involved.
 
I use Redcore-Linux (Gentoo-Clone). Compile everything from source.
With poudriere I have to be carefull.
Never been OOM in Linux. I don't even need swap.
PS : I always use 32GB of swap , with 16GB of memory. The sweet spot.
 
I'm compiling some code from an Area51 repo just to get to KDE 6.6.0. Official FreeBSD repos are still on KDE 6.5.5 at this time.
Update to that:

There's a reason that bleeding-edge stuff is labeled "Area51", "Not for general-purpose use", "Not for daily driving"... I compiled my way to 6.6.0, but it was not a usable desktop. Konqueror would not start (which is an old problem that I complained about earlier, in another thread). Even Konsole would not start, and along the way, I had to do pkg-fetch(8) on a quite a few packages that did not compile at all, like multimedia/ffmpeg. I had to resort to multimedia/ffmpeg4 just to finish.

With 6.6.2 in official ports at this point, I'm using the official ports tarball, trying again to compile my way into KDE Wayland. And the compilation is going a LOT more smoothly, stuff compiles fine (as long as I don't specify circular dependencies by accident). Lesson learned.
😅
 
official ports tarball ... Never used tarballs except for kernel.txz & base.txz. With poudriere one does normally does not need tarballs ? No ?

All from source ; pkg info | grep ffmpeg
Code:
ffmpeg-8.0.1_1,1               Realtime audio/video encoder/converter and streaming server
ffmpeg3-3.0.2_1                Realtime audio/video encoder/converter and streaming server (legacy 3.* series)
ffmpeg4-4.4.4_28               Realtime audio/video encoder/converter and streaming server (legacy 4.* series)
ffmpegthumbnailer-2.3.0_1      Lightweight video thumbnailer that can be used by file managers
kdemultimedia-ffmpegthumbs-25.12.1 FFmpeg-based thumbnail generator for video files
py311-imageio-ffmpeg-0.5.1     FFMPEG wrapper for Python
 
official ports tarball ... Never used tarballs except for kernel.txz & base.txz. With poudriere one does normally does not need tarballs ? No ?

All from source ; pkg info | grep ffmpeg
Code:
ffmpeg-8.0.1_1,1               Realtime audio/video encoder/converter and streaming server
ffmpeg3-3.0.2_1                Realtime audio/video encoder/converter and streaming server (legacy 3.* series)
ffmpeg4-4.4.4_28               Realtime audio/video encoder/converter and streaming server (legacy 4.* series)
ffmpegthumbnailer-2.3.0_1      Lightweight video thumbnailer that can be used by file managers
kdemultimedia-ffmpegthumbs-25.12.1 FFmpeg-based thumbnail generator for video files
py311-imageio-ffmpeg-0.5.1     FFMPEG wrapper for Python
You only really need multimedia/ffmpeg... Did you notice that others are labeled as "Legacy"? Ever wonder why?

And tarballs are just an easy way to download the ports snapshot into a clean install where you don't even have libgit installed (but fetch(1) is in base)
😏
 
Maybe he noticed and he has a reason to have all of them installed. Or he just likes it this way. That is, I find it quite belittling to assume that another person doesn't understand such an obvious thing.
The Ports Collection does include tools for dependency analysis. Those tools do allow you to see which ports actually depend on a given port. Same information can be found on the port's FreshPorts page. And offhand, I can say with confidence that there's no ports that depend specifically on having ffmpeg3/4, only plain ffmpeg. Please prove me technically incorrect here.

A reason to have ffmpeg3/4 installed is for obscure features not found in plain ffmpeg. But you'd have to be an A/V geek (with a deep understanding of how to use ffmpeg in command-line, and to have an appreciation of the difference in results produced by the legacy versions vs the plain, up-to-date version) to actually need it all. If he has the expertise (or wants to play with different versions), more power to him.

For compilation from source into a functional DE, ffmpeg3/4 is not technically essential. Plain ffmpeg is, though. Dependency analysis will show you that much.
 
Back
Top