Geronimo Wenja

It’d be worth checking out Borg as an alternative to rsync. Borg will handle snapshotting, and automatically de-dupe on a block-by-block basis.

I use it for all of my remote backups, and it provides a lot of quality of life stuff that rsync isn’t going to handle.

I had a reasonably good time with it. I had issues with btrfs, which is why I moved off it and went to Fedora IoT for pretty much the same benefits.

For me, btrfs caused multiple drive corruptions because of unexpected power offs, and I didn’t feel like trying to fix that on the fly - it might have been drives that were incompatible with CoW because of firmware “optimisations” that break if a write isn’t completed prior to power off.

In general, outside of that, it was pretty solid. I didn’t find much use for the orchestration/setup tooling they include, and I found their documentation pretty sporadic unfortunately. Fedora IoT has the advantage of basically being silverblue, with rpm-ostree, so it’s easy to find people using it and discussing it.

Are you expecting sonarr to go after historical stuff? You have to manually request a search for anything added that isn’t being released in the future. Sonarr only automatically checks for new episodes, not old ones. Like others have said, season searches and interactive searches are useful for anything that’s not airing in the future.

5 gallons per.hour? The article says 4-6 litres - a little over a gallon.

But their internet is down, so it’ll fail to send to telegram. Realistically it needs to be an external system that is tracking when it receives pings from the home network, so it can show periods where the bash script didn’t ping for a while.

My favourite one is renaming a directory full of files in nnn. It opens in vim, and I’m in my happy place, where I really know how to edit text (or, in this case, filenames). Great when there’s some minor variation between a lot of files. Full previewing before saving, multiple operations handled before doing anything etc.

The existing feature is that only subscribers will see it in feeds, but it can still be searched for or viewed manually. It’s not a private community feature. I’m just planning to add front-end access for the feature that already exists, so that admins don’t have to do API calls to use it.

I’ll see if there’s any existing discussions about private communities while I’m at it though, it might be something the main devs have an opinion on or plan for.

This actually already exists, it’s just not in the UI yet. Hiding communities can be done via the API. I was planning on putting in a PR to expose the functionality on the front-end at some stage.

Yep, the app is by far the easiest way to deal with it, and it’s got a great amount of troubleshooting options too.

There are speed and developer experience improvements, and a whole bunch of it is there to optimise for mobile. They have some info in the FAQ on jmap.io. It’s something I won’t 100% take without any consideration - it is written by the fastmail Devs - but a modern stateless protocol is no bad thing.

I’m also on Migadu for email, and I can say the experience has been pretty excellent. They have good instructions for setup stuff, and their pricing model is great. The pricing model has things in common with rsync.net, where they impose a soft limit on storage and reach out if you start exceeding it to talk about upgrading.

I do wonder if other mail providers will at some stage support jmap, it seems like it could take away some frustrations.

You can use any matrix client you want with it. The closed source one is just making bridging more straightward and adding some little quality of life features.

Luckily, the DMA has a heap of requirements around what their messaging interoperability will have do. For one thing, it will enforce the providers to not downgrade any encryption along the way, so FB etc will have to handle messages without them being decrypted first. There are some great videos that the matrix foundation put on their YouTube channel of talks that go over much of this.

Most of that extra stuff is there to handle user contact privacy and security with the bridges, which is fair. I don’t have any interest in self hosting beepers full setup, I want to get the functionality of multiple messaging services in one client - which I have, with my self-hosted matrix instance and the bridges they help develop and maintain.

I wish all of it was open source, but I did feel it necessary to head off comments that imply that the entire thing is closed source. Their implementation around dynamic servers and isolated containers spinning up isn’t really the bit that seems relevant regarding user privacy with regards to data scraping or anything. There are a lot of comments in here implying it’s fully proprietary, but there’s a lot more nuance to it than that, as you point out.

Personally, I think it’d be nice if you could self-host just the bridge instances and connect them with beeper yourself, so that the part that isn’t e2e encrypted is running on software you can validate and hardware you control.

Yeah, I should have clarified that. Hopefully the EU regulation regarding messaging interoperability removes this (currently unavoidable) flaw.

This is using matrix. You can even connect to it from any matrix client.

You can use any matrix client with beeper according to their FAQ. It’s frustrating that they forked element and kept the source code closed, but it is Apache 2.0, so it’s not against the licence terms.

The bridges are all open source, and they use matrix synapse as their server installation - though their client is a closed source fork of element with changes. You can use any matrix client to connect to it, and they say it’s a standard synapse setup.

If privacy is a concern, bringing your own client should remove that concern as the rest is open source. It’s also e2e encrypted, as any matrix server is.

I self host my own matrix homeserver with bridges set up using their code. The only bit of their stack I can’t use is the client. I don’t like that that’s closed source, that’s frustrating.

Edit: while writing this two more people made the same comment. Sorry!

Check out borgbackup, it stores changes only, snapshots are created for every new backup, encrypts automatically and is pretty straightforward to use.

A PR was opened last week to add the biggest first element of external library support. Hopefully in the relatively near future it’ll be merged. I’ll be giving it a shot when it merges.