I’m building a new activitypub/threadiverse software focused on the needs of self hosters who want a single user instance.

I’ve been posting with it semi-regularly for the last month, and I think it’s ready for an open demo.

One of my objectives is to have the lightest resource usage for memory and CPU constrained hardware, as well as the fastest loading web interface for older phones and limited data plans. I ran out of data on my phone last week and having a 41kb front page came in very handy.

You can try the web UI at https://scrapetacular.ydns.eu/latest You can also POST AS A GUEST TO THE FEDIVERSE without signing up. I’m not sure you can do this anywhere else, I’m manually approving posts on the backend because .. well you know. If it asks for a user and pass, use guest and guest, your post will appear with a username like guest4269.

Ideally, open this post https://scrapetacular.ydns.eu/post/10127 and reply to it.

My other plan for mobile is to target the Sync for Lemmy app, as it’s dead, meaning it’s no longer a moving target.

I’ve made a few technical choices aimed at keeping things fast

These include:

No ORMs

• They are convenient but make performance tuning difficult when things get complex as you don’t write the queries directly

No Javascript

• I may have to go back on this if I keep the guest posting function, it might need a captcha or anubis.

No nested comments in the web UI

• Nested comments are super slow, you are essentially querying the database for the OP, then querying for the N immediate children, then doing N queries for all of their children, then keep going recursively until you reach your depth limit, or all comments are found, you then need to render this structure with html/css

No front page images

• This is more of a personal preference that happened to make things load faster, the front page displays the text of the OP and last few comments IN FULL, giving a good preview of the conversation, and allocating more space to people who write rather than post memes. Inline images in posts are also replaced with links.

No upvotes/downvotes

• DID YOU KNOW that most threadiverse traffic is upvotes, downvotes adn emojis? You get an instant speedup by simply not processing them. Also, since this is a single user instance, all my comments are by definition awesome

ROADMAP

• Massive refactor
• Make the project public
• unit tests (this is basically my only requirement for v1)
• sync for lemmy API
• admin UI
• “AI” to “My Butt” wordfilter (mandatory and hardcoded)
• default subscription to /c/fuck_AI
• Solve channel discoverablity once and for all
• SUPERBLOCK (i.e. block everyone who liked this comment)
• dockerfile? I don’t use docker tbh
• Read Mastodon posts? Do they even have good content?

Tech Stack

• Go
• SQLite

I’m using the pure Go sqlite library, Bluemonday for html sanitisation, Blackfriday for Markdown and Migrate for auto db migrations.

End

Thanks to Snoopy and the Cool Froges at jlai.lu for allowing me to test on their channel.

Is this project of interest to you? Have I missed anything obvious? Is there anything else you would like to know?

Whoa senpai noticed me. There actually is a recursive query format in the latest sqlite for this exact situation, Postgres may have something similar, I just thought a list is fine for the web UI. I’ll have to benchmaxx and pick a strategy for the API.

I actually read that FEP. In my code that is too shameful to release yet, I look for an Object, then look for its Type, then figure out how to deserialize the rest of the object, but this object doesn’t have one. Would it be possible to do something like this:

{
  "object": {
    "type": "LikeCollection",
    "likes": [like1, like2, ..etc]
  }
}

Or is that weird. Spare a thought for those in less dynamic languages :D

You’re right, to be more similar to the ActivityPub style of doing things the object should be a “type” of “Collection”, not just suddenly be an array instead of an object.

Hey,

Some of you may know me for Jotty and Cr*nmaster, been quiet with my head down lately improving my apps and trying to build a searxng alternative for myself.

https://lemmy.world/pictrs/image/a70266bc-f67b-4dc1-b516-86166c8e4080.png

Whilst I have used searxng for about a year now, I have had quite a few personal gripes with it so in the past few weeks I have decided to make my take on it and ran it happily locally. Since publishing the beta to my discord server I ended up building a fairly extensive tool.

Degoog is actually pretty minimal, there’s no much to it aside from a very comprehensive plugin/extension system. The idea being users can create their own engines, themes and plugins that hook into the core application and do.. pretty much anything, from adding stuff to the result page (e.g. speedtests, tmdb information, ip retrieval, rss feeds embedded on the home page) to full on OIDC systems.

This is still very much in beta and I figured the best way to get it out of beta would be to publish it to a wider audience (currently some users in our discord server have been testing it fairly successfully and i’ve been on top of bug fixing).

Repo: https://github.com/fccview/degoog

Official extensions: https://github.com/fccview/fccview-degoog-extensions

Docs: https://fccview.github.io/degoog

You can install custom plugins/extensions. You can make your own repo and add it to the store page in the settings, or you can just have your own plugins locally for yourself.

Let me know what you think, and feel free to ask any questions and feel free to join our discord (link in releases page on any of my apps) for a more direct chat about things <3

Thank you! I mean, I am not trying to compete with Google, maybe when I release out of beta I’ll let the community that’s slowly creating around it decide a new name, we’ll see ♥️ wouldn’t be the first time I do that with my apps :)

Hey thank you!

The issue with GitHub alternatives is the lack of runners/pipelines and restricting functionalities.

Gitlab is a good alternative but I use it for work and having two accounts juggling between work/personal projects is a recipe for disaster.

I spent a bit of time on codeberg and I am checking it out, happy to push the repo there too, but they don’t seem to provide pipeline options, so I wouldn’t be able to build the docker image there like I do on GitHub :)

Btw whilst I do believe into the whole Microsoft scraping projects to train my butt regardless of licenses, I wouldn’t say they are hostile towards open source. They actually are extremely for it, vscode is free and one of the best IDEs out there for example, GitHub free plan is VERY generous and they have a whole FREE coding academy with extremely in-depth courses on how to learn programming and various niche topics. And they integrated wsl to seamlessly run Linux commands within windows, which I never thought I’d see happen (been there for years, but I’m just giving you examples).

Hey!

I basically want to replace the Google Authenticator app in style and functionality:

1. List all TOTP tokens and their validity time (with a name and order I decide).
2. Allow me to periodically or on change back up the whole thing to some off-site storage, keeping the last N backups.
3. Have a native app for Android or an actually good PWA.
4. Don’t do magic bullshit like fetching icons, hide tokens, etc.
5. Be actually secure (i.e. don’t roll your own auth)
6. Just be a TOTP manager, and nothing more! No, I’m not interested in a password manager, thank you. I also don’t want any other OTP methods I don’t use.
7. Don’t be a one-man projects where the availability is not clear in >1 year.

Any experience is welcomed. Thank you!

If you need to update them, you just reprint and replace.

If you need to recover on the road, well that depends on your risk tolerance. I’m never away from home so long that it’s a problem, and pretty much every service has a way to bypass 2FA in case of emergency.

You might want to check Ente Auth. Not sure how many checkboxes it ticks for you but I know people are quite happy with it.

Hej lemmings!

Quick question for you all: do you stick with the same distro across your PC, laptop, and server, or do you pick different ones based on the device and what you’re doing?

For me, I’ve been mixing and matching depending on the use case, but I’m starting to think it’d be nice to just have one distro (or at least one family like Fedora or Debian) running everywhere. That way I wouldn’t get confused about default settings or constantly have to look up flags for different package managers.

Right now my setup is:

• Gaming rig: CachyOS
• Laptop: AuroraOS
• NAS: Unraid
• Various project servers: DietPi, Debian, Alpine etc..

I feel like NixOS might be the only distro that could realistically handle all these use cases, but I’m a bit scared of the learning curve and the maintenance work it’d take to migrate everything over.

Am I the only one who feels like having “one distro to rule them all” would be nice? How do you guys handle your setups? All ears! 😊

Thanks a lot for the insights. I have dabbled a tiny amoubt with nix so far and while it was steep i do feel like it was doable. I am very likely to fall into the rabbithole again soon, and as you say probably very smart to run proxmox underneath for stability and convince 😊

Every now and then I try something else (usually live usb from ventoy) just to see how others evolve. I like endevour, but I always en up with debian minimal install. Only on mylaptop I add xfce4. It’s just rock solid. For my wife’s laptop it’s elementary, only because of the looks just to make her move from windows to linux painlessly

I’m not sure anyone shares the same glee I feel when I view all the blocked IPs scrolling by in my pFsense firewall. Suricata does a lot of heavy lifting for sure.

What’s your selfhosting guilty pleasure or pleasures?

Seriously, of there is a guide for how a newbie can set this up, please let me know. My little website is being bashed everyday.

Iocaine? I followed the instructions on the website which were fairly easy to follow. Depending on your skill level it might suffice.

Using #Madblog as the easiest way to spin up an Indieweb/ActivityPub-compatible blog.

Zero db, zero JS, entirely hosted on text files.

Using #Madblog as the easiest way to spin up an Indieweb/ActivityPub-compatible blog.

Zero db, zero JS, entirely hosted on text files.

OnionPhone is a native Android application for anonymous, end-to-end encrypted push-to-talk voice and text communication over the Tor network. No servers, no accounts, no phone numbers — your .onion address is your identity.

Cross-platform compatible with Terminalphone — call between Android and Linux/Termux using the same protocol.

Optionally use your connection as a relay for ephermeral group channels.

Find the release page for version 1.0.2 which supports custom bridges for accessing censored networks.

Super cool!

Would be awesome if it ever has a messaging section without having to make a call, still, an awesome start!

convenient in one app, but we already have briar for that

Hi everyone

I’m trying to degoogle as much as possible. I’ve heard about this thing called calDAV and cardDAV but I have no idea how to use it.

With radicale, do I need to install some other somewhere in order to use it?

I’m just looking for basic useage for myself only at this stage. I’d like to be able to self host my own calendar and contacts. Is radicale appropriate for this?

Is it safe to self host a calendar?

Can a self hosted calendar still send and receive invites to other calendars?

Any help greatly appreciated, thank you

mailbox.org is another service that is straightforward to use. You create an account, download davx5, and install any calendar app. It’s paid but cheap (1 €/month)

No worries mate. I was just sitting here jamming some blues on the guitar and felt inspired.

Readme updated today:

This repository is no longer actively maintained.  

The TrueNAS build system previously hosted here has been moved to an internal infrastructure. This transition was necessary to meet new security requirements, including support for Secure Boot and related platform integrity features that require tighter control over the build and signing pipeline.  

No further updates, pull requests, or issues will be accepted. Existing content is preserved here for historical reference only.  

https://github.com/truenas/scale-build

Wondering if this is just the first step towards doing a minio in the future.

Uggh, I just got to the point where I’m as familiar with TrueNAS as my old Synology, but I want to move towards more freedom. TrueNAS beats Synology, but where is this going?

From the docs:

The Standard Debian Kernel (selectable) can be used for ZFS. However, since ZFS kernel modules are not installed in the Debian kernel by default, they must be built by the ZFS plugin when it is installed. While this process works, building the modules is a long process that requires continuous access to online repos. Accordingly, the potential for a build error exists. For this reason, while the Standard Kernel is very usable for ZFS, it is not ideal.