• 0 Posts
  • 30 Comments
Joined 1 year ago
cake
Cake day: August 22nd, 2023

help-circle
  • podman-compose definitely got better over the past year…

    But you can also use docker-compose itself with podman instead!

    https://www.redhat.com/en/blog/podman-docker-compose

    Basically, for system level containers, you can do:

    sudo systemctl start podman.socket
    

    (or enable --now instead of start if you want it to stick around after rebooting)

    Then use docker-compose and it’ll communicate with podman instead of docker.

    For user session “rootless” containers, it’s mainly the same thing, except you’ll need to remove sudo and then add --user after start or enable in that systemctl command. And you’d need to set an environment variable (either prefixing it on the command or using export to set it in your session), like this:

    DOCKER_HOST=unix:///run/user/$UID/podman/podman
    

    (Put that in front of the docker-compose command and it’ll connect to podman as your user instead, provided the service is available. Or toss it into your .bashrc with "export " before it and new bash sessions would have it.)

    https://brandonrozek.com/blog/rootless-docker-compose-podman/

    The one big gotcha I’ve hit is that if you have SELinux on your system, you’ll want to add :z to your volume(s) mount to have it automatically deal with SELinux stuff. (Lowercase z for volumes that can be mounted for multiplayer containers and uppercase Z for volumes that are tied to a specific container.)

    But, I’ve found that using “quadlet” service files is much, much better than using podman-compose or docker-compose. There’s a program called “podlet” that can even convert compose files to service files (quadlet)… It can convert command line flags and kubes and other formats too.

    Quadlets are basically systems service files that integrate with podman, letting you easily set up a container as a system (or even user level) service, making managing a container just like managing any other service.

    Here’s the podlet command that’ll convert things to quadlets: https://github.com/containers/podlet


  • Peppers are the fruit of the plant. They’re what’s made after the flowers were pollinated and have seeds. They’re also sometimes sweet and not always so spicy.

    Of course, there’s the botanical definition and culinary definition and there’s some overlap. The most famous would be a tomato, which is also a fruit and a vegetable from different points of view.

    What’s mind-blowing to think about is that a pepper is not just a fruit but also technically a berry.

    In cooking, peppers are used as a fruit, a vegetable, and even a spice. (Depending on the pepper variety.) So, anyone classifying it as any of those things is right. 👍

    (Wikipedia mentions all this too.) https://en.m.wikipedia.org/wiki/Chili_pepper



  • Yep, ArcMenu (@ https://gitlab.com/arcmenu/ArcMenu which is the maintained one, last updated days ago instead of years ago) has a ton of different layouts which can mimic any version of Windows, and so much more.

    screenshot of ArcMenu layout settings, with the "modern" category expanded

    When using GNOME, use the “Extensions Manager” app (from Flathub) to search for “ArcMenu” and install it, then you can configure it there in the Extensions Manager app as well. In the ArcMenu configuration, go to layouts and select the modern group to see something like the screenshot above. (The previews are generic wireframe sketches; the result will look much more high fidelity.)





  • Merlin wasn’t available here when I checked at some point in time (last year?)

    whoBIRD does use BirtNET, from Cornell, so it’s basically the same backend (although it may be an older version).

    I recently tried out Merlin (which is now available here) and it’s amazing. It’s definitely more featureful than whoBIRD, although both have the core “recognize bird directly using your phone” features.

    For anyone OK with non-FOSS apps, Merlin is great. For anyone who wants a FOSS app for bird detection, whoBIRD is still pretty good.

    Either way, identifying apps using ones phone is nice. 👍 Big things to Cornell for making the ML for both of these apps.




  • darktable, hands down. It has a learning curve, but it’s a pro app and app pro apps have learning curves.

    The linear pipeline is great, masking is superb, and the app keeps getting better every release.

    The one downside is that darktable is not opinionated by default (so raw files look a little flat to begin with, without doing anything), but it’s customizable that you can even change that with auto applied presets. On the other hand, it does let you do what you want to do with an image, versus fighting with defaults (which is what it’s like to edit something in Lightroom, if you want to diverge from what it suggests by default).

    There are a bunch of great tutorials on YouTube and you’ll want to check out https://discuss.pixls.us/ too. Create an account on the Pixls forum, read some threads, try out some “play raws” (where people post their raw files under a CC license and then lots of people try their own take at editing it and post their edit).

    Rico Resolves has a half hour getting started video for darktable 4.6 at https://youtu.be/ucjAmTMIEOI

    Anything from Bruce Williams https://youtube.com/@audio2u and Boris Hajdukovic https://youtube.com/@s7habo are both great too, and more people are posting darktable videos all the time as well.

    The documentation for darktable is actually very good as well. Do not skip it. You don’t have to read it all, but try reading the intro parts and going back to it when you want some reference on how a part of darktable works. https://docs.darktable.org/usermanual/4.6/en/

    Some tips:

    • You can right click on sliders to get a special UI and you can also enter numbers (often even outside the bounds of what the slider would normally permit).

    • Modules will be applied in the best order regardless of which one you work on first.

    • There are some somewhat redundant modules, as darktable did start out as a “display referred” workflow (just like most all of the other raw editors everywhere) and moved to a “scene referred” (aka “linear rgb”) workflow, which provides better editing, improved color handling, vastly better tone mapping, and so on. If there are two similar modules, try to go with the version that has “RGB” in its title. Older modules still exist mainly for older edits. (You can also change darktable back to the old display referred workflow in the settinfs, but I strongly suggest to not do this. Scene referred is much better.)

    I used to shoot film and do darkroom stuff years ago. I’ve used Aperture on OS X. I used Lightroom on OS X and then on Windows. A few years ago, I switched to darktable on Linux… and darktable has gotten so, so much better each release. When I switched years ago, it was more or less a Lightroom competitor (with some advantages and disadvantages). But darktable is really amazing software now, and can give you much better results than Lightroom, when you know how to use it.


  • For the video problem, it might be codecs; try using Proton-GE if it’s in Steam or use Wine-GE if not. (IIRC, Steam will often convert the videos and give you the converted ones in the shader caching if necessary. But those outside of Steam, and sometimes a few still in Steam don’t have that workaround.)

    For the main issue you’re having, try running those games in gamescope, which itself is a compositor with a bunch of neat tricks. In this case it’d make sure to not lose the focus of the game even if the gamescope window loses focus. It can also optionally force windowed or fullscreen modes, upscale (even with FSR1), and lock the framerate.

    Changing settings in the game itself between fullscreen or borderless (borderless should usually help with the focus issues) may help too, if the game has that setting, but then you’ll probably hit the borders issue due to FVWM. (I don’t know if you’d get the fullscreen unredirection optimization in fvwm. That could be a reason to pick one or the other for you too.)

    You’re probably hitting a few edge cases by using FVWM versus a more modern environment like GNOME or KDE, but to be fair I’ve seen the focus issue happen before on a game on running through Heroic on GNOME with more than one monitor before. FWIW: I don’t remember seeing the issue in games from Steam. (It probably depends on the game itself, however.)


  • Penpot works perfectly on Linux, and you can even host it yourself in your own computer if you want. It’s web-based and works in both Firefox and Chromium browsers. (I think WebKit ones too, but it’s been a little while since I’ve tried it with Epiphany.)

    I use Penpot myself all the time on Linux, but I’m usually using the hosted version so I can collaborate with others without having to maintain a server. I have also run locally in a container using Podman, even with Podman’s rootless support.

    But to start using it, all anyone needs to do is point their browser of choice to https://design.penpot.app/ and sign in. There is no setup process or installation needed; self-hosting is completely optional.


  • Just pointing this out, as there are non-free services that the apps use:

    Frog is awesome, but note that while Frog works offline for OCR, it has TTS (text to speech) which uses an online service. As long as you avoid having it read to you, it’s all done locally.

    And Dialect always uses an online service. Some of the servers are FOSS, but some aren’t. But everything you type or paste into it is sent somewhere else. (This is the case with using translation websites too, of course.) I’m not saying you shouldn’t use it; I’m just saying that you should be aware.

    Hopefully Dialect will add Bergamot (what both Firefox by default & the “translate locally” extension use for translation) at some point. Dialect has a longstanding issue about it, but no forward motion yet. https://github.com/dialect-app/dialect/issues/183

    For something open source that runs completely on your computer for translations, you’d want Speech Note. https://flathub.org/apps/net.mkiol.SpeechNote It’s Qt based, but works well. In addition to translation, it can do text to speech and speech to text too. You do have to download models first (easily available as a click in the app), but everything, including the text you’re working with, is all done locally.

    I use both Frog and Speech Note all the time on my computer (GNOME on Fedora Linux). They’re excellent.



  • Agreed.

    Additionally, the graphic oversimplifies things as well. The resulting genetically modified crop is often not even all that close close to the same as the non-GMO, as seen by studies such as this one:

    https://enveurope.springeropen.com/articles/10.1186/s12302-023-00715-6

    Basically; GMO soybeans contain proteins which differ and also include additional proteins. This can cause allergic reactions to modified soy where non-modified soy might not cause an issue.

    Monsanto supposedly even knew about these proteins and higher risk of allergic reaction and chose to not disclose it. (I saw some research that mentioned this years ago… It’d be hard to find the exact source I read back then.) This specific paper, which talks about additional proteins and side-effects brought in by the new transgenic splicing, also explicitly states that Monsanto did studies themselves and failed to report relevant findings:

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5236067/

    Obviously, other methods can also change proteins too, but these papers show it isn’t as clear cut as the graphic in the original post claims.

    Along these lines, here’s a study that finds differences not just in soybeans grown organically versus ones treated by glyphosate (Monsanto Round-Up pesticide) but also between GMO and non-GMO crops, both treated by the pesticide.

    https://www.sciencedirect.com/science/article/pii/S0308814613019201

    But, yeah this is just a long way of agreeing with the parent post and saying that the end goal is to make the plants resistant to poison, not to make them better for humans, all to make more money. (In this case, Monsanto is even double-dipping by selling both the pesticide and the crops tailor-made for the pesticide.)

    Other GMO crops might be closer to the original crop and might also actually be beneficial for humans without drawbacks. However, Monsanto’s soybeans are problematic, and other crops might be as well, especially if they’re made by companies who have money as their primary goal.




  • Yeah, that’s a big, weird if though. Most modern apps can rely on the runtimes for their dependencies and not have to ship their own custom dependencies.

    It’s different from something like AppImage, where everything is bundled (or Snap, where a lot more needs to be bundled than a typical Flatpak, but not as much as with an AppImage).

    Additionally, there’s always some level of sandboxing in Flatpaks (and Snap packages) and none at all for RPMs, Debs, or AppImages.

    Also, Flatpak dedupicates common files shared across flatpak apps and runtimes, so there isn’t “bloat” like what you’re talking about.

    https://blogs.gnome.org/wjjt/2021/11/24/on-flatpak-disk-usage-and-deduplication/