Skip Navigation

Posts
2
Comments
1,625
Joined
2 yr. ago

    1. It seems to make a LOT of calls to other servers. Its almost constantly pinging other servers asking for updates.

    The fediverse works the other way around: other instances push activities to yours. If you have a lot of subcriptions to large communities like !technology@lemmy.world it will indeed receive a lot of activities.

    1. It gets de-federated almost instantly from popular instances. Which kinda sucks.

    Mine's not been defederated from anywhere, not even Beehaw

    1. It uses up quite a bit of CPU compared to other federated applications.

    It definitely uses a fair bit of CPU but it is ingesting a fair amount of data, but still not a ton either:

    Although I do hear PieFed is a lot lighter.

    1. Subscribing to instances seems to work most of the time, but sometimes it just errors out and I have to re-do it.

    That settled for me after a week or so of running mine. My subscriptions always go through.

  • If you look at my username you'll see I do run my own instance so I've gone through the process :)

  • I would probably just skip the Lemmy Easy Deploy and just do a regular deployment so it doesn't mess with your existing. Getting it running with just Docker is not that much harder and you just need to point your NGINX to it. Easy Deploy kind of assumes it's got the whole machine for itself so it'll try to bind on the same ports as your existing NGINX, so does the official Ansible as well.

    You really just need a postgres instance, the backend, pictrs, the frontend and some NGINX glue to make it work. I recommend stealing the files from the official Ansible, as there's a few gotchas in the NGINX config as the frontend and backend share the same host and one is just layered on top.

  • Hasn't cost me a penny, hurray for unmetered bandwidth

  • Just don't copy paste the commands. Really! Just take the time to understand what the command does, read the manual, and rewrite it yourself instead of pasting it. That alone will help a fair bit and can start guessing what it should be.

    After a while of doing that it stops being a "paste this command to make the service run" and becomes "ask systemd to enable and start the service". You start associating editing files in /etc with "will probably need to slap a sudo in front of that one", you start mentally replacing nano/vi/vim/emacs/nvim/sed with your preferred way of editing the file, because you absorb the concept of "this command edits a text file".

  • I had that pretty well managed in the end. Even when it all works correctly it's just not like the first time where you feel superhuman and knock down 2 weeks of chores in a day. The body readjusts and it settles closer to a neurotypical focus.

    I'm talking about the very first time in particular, when you feel clarity for the first time and feel limitless. It's not like you lose the focus, but the euphoria that makes you feel like you can do everything fades away. It can give the impression the ADHD is gone, but you still need to stay organized, eat and sleep well to not come crashing down. And also not overcommit, it's easy to run straight into burnout because you can easily pull all nighters studying for the first couple days.

  • I don't want to break the fun, but it's pretty typical for the first couple days/weeks to have some euphoria as a side effect of the medication. It still helps a lot but you don't exactly keep the initial laser sharp focus. It's still an improvement but eventually chores become chores again, but at least it's much easier to stay focused on it anyway.

  • One thing that is very slightly harder to do on the fediverse is a bot would only see the communities at least one of the instance's users is subscribed to, so it's not trivial to make one that sees everything out of the box. Easy enough to fix with a few API calls to recursively discover most instances though.

    And there are bots: there's the Media Bias Fact Check bot, there's the PipedVideo bot that posts piped/invidious links for any YouTube link. I have both of those blocked because I don't care (and Tesseract has both built-in anyway).

    I think culturally stuff like remind me is too noisy/unnecessarily spammy, so if I were to implement such a thing I'd do it directly into the UI so you don't have to (ab)use comments for that. The haiku bot I blocked a long time ago because it's just kinda noisy, it's fun for a while but eventually it gets annoying. As an admin I also think about resource usage, it's not just wasting big VC funded companies money, it's wasting people like me's money too.

    The fediverse opens up a lot of possibilities that allows things to be done cleanly without bots and spam. As a user you're free to use any UI/frontend you want and still access all the available features, unlike Reddit, well before they killed off third party apps completely.

  • You could try having the HDMI input selected on the receiver before plugging it in, in case there's some EDID fuckery going on and it might pass it through that way. There might be some info dumped in dmesg as well that could help troubleshoot what it's not liking about. Computers try to be smart and output a supported resolution whereas a lot of AV equipment just outputs whatever it wants without checking.

    If it is an EDID problem, there's ways to override it for the port so you could dump your TV's, then force it to use that one even when plugged into the receiver. Or you can use a generic 1080p/4K one.

    Failing that, if your receiver and TV support eARC, I guess plugging directly into the TV and have the TV feed the receiver is an option.

  • If going the emulation route and willing to use a different kind of gun, you just need to feed the XY coordinates to the emulator. Duck Hunt with a Wiimote is pretty trivial to implement. Once the emulator knows the XY coordinates it can just feed the correct inputs to the game no matter the latency of the gun: if the emulated screen is a white pixel at that coordinate, then the controller port gets a one, and circumvents the entire problem. It can probably be done on original hardware too with an RP2040 to intercept the video signal and inject the correct input, as long as it knows the coordinates before sending the fire input to the console.

    To make work on the original gun, the emulator could rewind and replay with the correct inputs fast enough to be completely invisible to the user, we already do that for netplay over the Internet. Play once, process the gun input, rewind and replay with the correct inputs (in the background) and then continue at the exact frame we rewound and it's completely invisible except a tiny bit of rubberbanding.

    Original console and original gun though, it's tricky but if we could frame quadruple the thing to 240Hz and use an OLED with zero input lag, we could theorethically have it displayed in time for the console to be able to read it with the light gun by vblank. The tolerances of that would be insane.

  • The light gun doesn't need a CRT per-se, but rather the lack of input latency. Games usually flash the targets one per frame, and then it knows based on light level if you were pointing at the target and which one. If it sees light on the third frame then the target must be the third one it flashed. That requires the console to be able to read the light level basically immediately after the console's done scanning out the image but before the phorphor fades out, so the timing is very tight.

    If we made OLEDs with direct scanout/zero latency, the guns would work just fine. But because of scalers and filters there's usually at least one frame of latency which means at best you're one target off, or it thinks you're cheating and registers a miss (games usually do a full frame of black first to see if the gun's pointed at a light source, which if you have a frame of latency on the screen it'll register the last frame which will be bright and thus register a cheat/miss).

    Add just a frame of latency to a CRT and it'll stop working there too. Later progressive scan CRTs that buffer two frames to deinterlace the signal also don't work with the light guns.

    Here it is in action: https://youtu.be/V6XnSvB34y8

  • It's a machine that used to be well oiled but management's been deferring maintenance for decades, the oil's gross, it's leaking everywhere and overheating, it's barely hanging on, and the manufacturer's long been out of business.

  • Unless it's running as your own user as part of your session. /etc for system-wide and ~/.config for your own user makes a lot of sense.

  • How does that even work for those hosting their own? Do I just give myself Bluesky+? Because all those features I already have by virtue of hosting my own data.

  • Maybe it can be hacked together with Syncthing: have your phone's camera sync with an inbox folder on the desktop, have the desktop pick up the files and transcode them with handbrake, then move the original out of the inbox. This will cause Syncthing to sync the deletion back to your phone, and sync the transcoded version back on your phone.

    I'd also check if you can just change the bitrate in your camera app's settings in case there's a way to lower the quality there. Could be noticeable, could be just as good as handbrake, never know with hardware encoding.

  • I believe I may have been able to mostly reproduce your setup directly in Bottles, or at least found another way to achieve the same.

    I made a completely fresh bottle, went to dependencies and then installed cnc-ddraw. Then I installed the game, and it works. The intro videos play fine, but then the menu doesn't draw properly as you said, but the whole menu is drawn at least which makes it just janky but totally fine to navigate. Then once in-game it runs just fine.

  • It doesn't, moving it to the end of the disk is a fairly common workaround for this specific issue. UEFI only looks for a GPT partition table and a partition within it with the UUID that corresponds to the EFI System Partition (ESP) type with a supported filesystem on it. The filesystem in question is implementation dependent, but FAT32 is guaranteed to be supported so most go with that. Apple's firmwares can also do HFS+ (and APFS?). More advanced firmwares also let the user add their own drivers, in which case as long as you can find a driver for it you can use whatever filesystem you want.

    It is common however to do so, out of convenience. Usually it's other partitions you want to resize, and when imagine to a new bigger disk (or cloud environments where the disk can be any size and the OS resizes itself to fit on boot), then growing the OS partition is a lot easier. But the UEFI spec doesn't care at all, some firmwares will even accept multiple ESPs on the same disk.

    Some older firmwares may also have had size limits where if it's too far in the disk it can't address it which would be problematic on very large disks (2TB+), but that's old EFI woes AFAIK.

  • It's the boot partition, it needs to be a plain partition formatted as FAT32. noticed it's a separate boot partition as ext2, but the point stands: most likely bootloader limitations.

    That said you could also just make a new one, copy the data over and delete the old one once verified the data's all good.

    I wouldn't do it with a larger partition but these days moving a 500MB partition takes a couple seconds top even on spinning rust, and it's a boot partition so it's kind of whatever. Very low risk overall, and everything on it can be reinstalled and regenerated easily.

  • You can look at it the other way around too: Linus made a kernel, and enough people liked it that people developed Linux distributions, and it kept growing.

    A lot of FOSS projects started as someone's personal project they released (sometimes literally just to have stuff on their GitHub to be more hirable in job search) and it became insanely popular rapidly and now it powers entire ecosystems.

    Not all projects starts with the ambition to become a big thing, and that's usually how the really good stuff starts off as.

    The Lounge started off as some users getting interested in Shout, which was just some guy's pet project) and we forked it because we had a pile of patches for it to fix issues with it. I worked on it purely to serve my own purposes (just enough to IRC on the go without dealing with reconnecting to ZNC all the time and draining battery), and now it's an active project a lot of IRC networks use as a guest client for their IRC network. No intent to disrupt the IRC clients landscape, I still used HexChat back then, but now it has secured a permanent spot in my open tabs as it does for many people. It's actually a pretty good IRC client now.

  • You can move the partition at the end of the disk where OP has 1.5 GB of free space. It'll leave a 500MB gap before the LVM but it is what it is.