I would switch from obsidian to trilium if only they didn't block using h1 (aka # headings).

I read the explination about it being reseved for the title on export but I really didn't find it compelling to break a basic note taking workflow.

I imagine it wouldn't be for the system admins to use, it's for all the other users who can use terminal applications but always treat ssh keys as a nuisance and try to avoid them as much as possible.

From the article "OPKSSH does not require any code changes to the SSH server or client."

Looks like this is a sidecar application. So potentially very useful, also potentially very brittle.

That was related to Bluetooth. Interesting undocumented low level commands but it's a bit of a stretch to call it a vulnerability IMHO.

But having the whole stack open would just be better in general.

Correct. HN thread https://news.ycombinator.com/item?id=43301369

Not a vulnerability in the way that Tarlogic makes it sound. Disingenuous and misleading article for sure.

A good bunch of "security" articles that make the news look more like scareware to me in the past years.

I used to run it on a 4 core Celeron NUC with a measly 8 gig of ram and 1 TB SSD via the community fpm container: https://github.com/nextcloud/docker I now run it on a much more powerful system but performance wise it's the same.

Side containers DB: I use postgres Redis: https://github.com/dragonflydb/dragonfly as the redis container Image processing: imaginary container maintained by the Nextcloud team

Most of this is available in the AIO setup but I prefer manually tuning things for my system

As it's a __big__ SAS product to run it well on lower end systems you MUST:

- Tune your PHP settings to your system.

- Tune your DB settings

- Set up caching properly

- Offload as much processing as possible to background processes (e.g. preview generation)

Probably tune your Proxy settings as well.

Big changes I've made:

- Switch logging to error only and syslog (not the normal file on disk) as it can hammer single disk systems.

- Switch to unix sockets for redis and the database for quite a performance boost.

I’ve never done much in the way of tuning Nextcloud, but I do enable APCu and Redis. I also ditched MariaDB years ago and moved to PostgreSQL.

It definitely isn’t so slow it is unusable, limited to 2 cores of an Intel N95 CPU and 2GB RAM.

I have often wonder if some apps drag the whole thing down because my experience doesn’t seem to be universal. That dashboard thing had to go, and the talk app was never something I was interested in either.

I’ve been running it since it was OwnCloud and never really have upgrade issues either. YMMV.

Possibly because it's actually just "good enough" for most use cases?

And if you truly require better accuracy, using a GPS units for high acuracy PTP is a thing.

The open banking standard isn't actually all that open.

As a customer you cannot gain access, at least easily. You need to apply for accreditation as a third party financial services provider.

It's a great looking interface but I have the same gripe with the tagline that I have with Grist's. It's really not spreadsheet like is it? You just don't have the row and column manipulation possibilites that defines a spreadsheet vs a table.

Thank you! And yeah, I understand your gripe. The tagline is meant to convey the general idea of "editing data in a table UI without a lot of clicks or keystrokes" in the fewest number in words; it does sacrifice precision (although that's why it's "spreadsheet-like" and not just "spreadsheet"). Do you have any ideas for language we could use instead?

With it becoming Thunderbird branded my biggest wish is proper subfolder support. Seeing folders as Archives.2024.invoices.paid rather than a nice collapsable list makes mail management a right pain.

It's also one of the oldest request on the github https://github.com/thunderbird/thunderbird-android/issues/63...

And it's in the same state I said when I commented on it for the same reason as I commented. A scrolling List View is really easy to implement. A proper infinitely nested folder UI is harder and none of the 'off-the-shelf' things worked at the time.

(There's a bunch of comments that are just wrong, the code does 'understand' the separators, it just doesn't do much with that knowledge).

The UI last time I used it used a mixture of web-views and standard components.

We were at the time handling the Material UI transition to make it look like a modern app.

Until that was done there was very little incentive to invest a lot of time in a UI that would be scrapped when that happened.

Other big projects (like the Android timers change which broke reliable email fetch or battery life) were also prioritised ahead of it (for reasonable reasons).

Honestly with only cketti and another person working on it part-time, the project struggled to do more than (or even) keep pace with the Android ecosystem. Maybe now it can. But the release cadence and approach was painfully slow.

I stopped my involvement with the project when I switched from Android to Apple - my motivation was always improving my own email experience.

Hey thanks for the work in the past, you did amazing work. No doubt imaps loosness with how it returns data made it an even more complex task.

I hope joining Thunderbird is going to pay dividends for big task like this one.

They have stated in the past that K-9 will stick around with just the Icon, name and colour scheme differing.