Hacker Newsnew | past | comments | ask | show | jobs | submit | savikko's commentslogin

Just an anecdote, but very close person of mine was diagnosed with migraine when she was 12 yo old so.

Last year, being 40+ years old, suffering from weekly/daily pain events, she found out that diagnose was wrong (one doctor just asked that have you ever thought that this probably isn't migraine) and correct diagnose was Horton / cluster headache.

Eating migraine meds (lots of different ones) for several years were helping, but probably only due placebo effect.

Now, the medicine for headache attacks is simple: breathing pure oxygen for 15 minutes.

After last attack (1.5 year ago, stopped almost immediately breathing oxygen) there has not been any headaches.

Usually, it would not been nice to get diagnose for cluster headache, but on this case it was kind of life saver.

Still, just an anecdote.


Hugged to death but what i like about zerotier is that I can access my Home Assistant instance (and other home services) with same ip address (resolved from dns and then TLS terminated with traefik) from home network and from zerotier.

And it does not matter if I have ZT network connected when home.

Not sure if that is possible with tailscale (from my understanding, it generates always tailnet ip for hosts).


Tailscale does support this. Using the Subnet routing feature you can expose other devices on the network to the Tailnet. I had to use this for a while with TrueNAS because of the way it handles TS integration (eventually I moved to using a reverse proxy).


The best part of tailscale is that you can use an Apple TV as a subnet router and exit node. So even if you do not have any home networking equipment, you can utilize this by just using the Apple TV app. This is particularly great if you have a second home etc.


Oh, will have to investigate that a bit.


> from my understanding, it generates always tailnet ip for hosts

It does, but it should connect over LAN when both devices are on the same network. The tailnet IP doesn’t exist outside the WireGuard network, so it’s up to the WireGuard routing algorithm.


I thought it was Tailscale that always selects which IP to use as endpoint for other devices, and set that up for Wireguard? If I'm wrong, could I replicate that behavior (using relay on WAN, direct connection on LAN) with Wireguard without external configuration tools?


Hmm, yes, I think you’re right. Tailscale does handle the connection here, not Wireguard.

I’ve digged into it a bit and I believe it first connects over a relay, then the devices try to find a more optimal route. So for LAN, they would exchange their local IPs and try to connect over those. If they are indeed on the same LAN, they connect directly: https://tailscale.com/kb/1257/connection-types

This is not without issues, however: https://github.com/tailscale/tailscale/issues/7206


Tailscale DNS might cause a bit of trouble, but the overhead isn’t too bad, I’m measuring ~1ms ping difference, which is a lot in relative terms, but in absolute it makes 0.5ms into 1.5ms.


Yep, and there is no "ip alias" type of things. On zerotier, I route my home server ip to zerotier ip and then it just works.


I assume it is using built-in sound. But of course, you could play saved midi on your keyboard and save that audio if using some specific sounds.


This is great! Thanks for creating this - and as someone asked already, I would love to hear more about the process how you did this. Especially I am interested how you came up with injection molded enclosure (I thought it just costs some enormous amount to make the mold).

I was about to order but frankly it does not feel good to pay about the same amount of shipping (to Finland) as the product costs.

If you can do anything about it, I would be happy to order. 90EUR for shipping is just too much, 20-30'ish would be reasonable.


Wow that is the highest shipping cost I've seen. Sorry about that. Partnering with an EU distributor is on the todo list.

Yes I have a lot to share about the production process. Blog posts are planned! It was quite fun!


Check out Asendia. They are doing consolidated shipments which might work out nicely with this.

Works basically like this: You collect couple of shipments to one bigger package. Then you send that bigger package to Asendia logistics centre and they will send those individual packages to users.


I'd likely buy if I could ship to UK at a decent price! Please make this happen! Is there some way I can sign up to some future communication about this?


send me an email I'll add you to a list. chipweinberger@jamcorder.com


Not as bad as 90EUR, but shipping for me comes out as 71 NZD for me, which also seems a bit steep — would purchase almost instantly for a similarly priced shipping to that mentioned here, ~35 to 45 NZD?. Just noticed the default UPS Worldwide Saver® is quoted as 6-7 business days which (for me all the way down here) seems pretty speedy — I'd be totally happy with something cheaper but more paced like a snail!


We also use Doppler.

Generally, Doppler has solved both dev and prod env issues for us.


I have some pretty good experiences with PaddleOCR but you may refer to this Chinese and badly documented ones.

For our use case PaddleOCR + LLM has been quite nice combo.


Yes, that's one of the ones I tried. It seemed to be more designed for things like receipts and menus rather than books. But in any case, I found it hard to set up and use (and it's likely slow on the CPU compared to Tesseract, which despite its low accuracy, is at least very fast on CPU).


Yep, mailhpg works, but is not anymore maintained.

We have been successfully using https://github.com/axllent/mailpit


Serious question: SMTP clients from 20 years ago are basically compatible with today's. Is it necessary to continuously maintain a piece software if the requirements stay the same?


One example where I had issues with Mailhog was that the docker images was very bloated 138.8MB[0], there was a PR that should have brought it back to 7MB or something but it was never merged.

And not sure about this, but at least for docker, the included certificates can become outdated if not frequently rebuilt, and since Mailhog is software used when testing, being able to spin up a ready-to-use sandboxed instance is very useful.

[0] https://hub.docker.com/r/mailhog/mailhog/tags


This is great, thanks for doing this!

I think it is somewhat common situation, when home zigbee network does not reach e.g. garage or some other near-distance building. Usually there is some ethernet/wifi network on the satellite building. So distances what LoRaWAN can reach, are not probably even needed.

Read a bit about the implementation, but I think it is easier to just ask: Could this work over general TCP somehow?


Thanks! Your question is an interesting one: Forwarding TCP over LoRaWAN seems at first impossible due to large packet sizes (from LoRa perspective!). It turns out there have been some efforts to apply dictionary based compression methods to carry different protocols over LoRaWAN. You might want to check out RFC 8376 for details.


Ah yep.

But I was mainly trying to think how to extend Zigbee-network over TCP (without LoRaWAN).

Currently it seems to be almost impossible.


If you have multiple LAN Zigbee controllers, like a ZigStar UZG-01, you can use multiple instances of zigbee2mqtt and hook 'em all into Home Assistant.


Yes, thats true, but that also means multiple zigbee networks.

Goal here would be one zigbee network which is extended to satellite location over TCP.


Of course, your machine can be a peer also. You can create peers to your organization with `flyctl wireguard create`.


But that isn’t the case for all of the other uses of flyctl.

You can’t make connections with other processes over the same connection flyctl makes when it’s doing stuff.

That is what GP is saying.


Nissan Leaf with OVMS (disconnected from Nissan Connect services) is something I was happy with.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: