• 0 Posts
  • 117 Comments
Joined 1 year ago
cake
Cake day: June 17th, 2023

help-circle
  • A normal copy consists of a program reading from one file and writing to another. There is no way for the filesystem to do a reflink in that case, it just sees that the program is reading and writing stuff. In order to do a reflink, the program must tell the filesystem what data should be “copied” to where using FICLONE or FICLONERANGE. Though some programs will do that by default if possible nowadays when copying files or when moving files between different subvolumes on the same partition, including the Coreutils cp, mv and install commands and some GUI file managers.






  • None of which are in this picture. The person in the picture talks only favorably of immutable systems yet is apparently against them, thus making for an easy target by arguing against themselves, so a straw man.

    I’m actually positive to immutable systems, I just thought the argument wasn’t great. I realize that’s about what Skinner does in the meme, but it feels weak.

    On second thought, I think the reason it was so jarring is because normally points against Skinner are in top picture, and the bottom picture has him abandon that line of thoughts in favor of something simplistic, thus changing his mind from one side to the other. Whereas here, the points against Skinner are at the end point of the meme, and thus he argues in both directions simultaneously.




  • It’s better in one way, in that updates are applied on reboot rather than pulling the rug put from under running applications. But I agree that it doesn’t go all the way, as it doesn’t provide a verifiable base system with clearly separated modifications. OSTree would be great.

    Another possibility would be to distribute a base image as a btrfs send stream (possibly differential against previous versions) containing a compose-fs image and associated files. And then OS extensions could be installed with systemd-sysext.






  • LaggyKar@programming.devtoLinux@lemmy.mlVLC Player
    link
    fedilink
    arrow-up
    15
    ·
    6 months ago

    I used to use it, but then I switched to MPV, as it works a lot better with hardware acceleration. MPV supports more methods for hardware decoding (e.g. nvdec), and also MPV will keep the frames in VRAM when doing hardware decoding, and do additional processing and presentation using the GPU, while VLC copies everything back to system RAM and processes the frame on the CPU.

    At the time I switched hardware decoding with copy-back would actually result in twice the CPU usage compared to software decoding, but that was a long time ago. Also, I would get tearing in VLC and not in MPV.