jesprenj

Not really. To be really sure, set ! as a password in /etc/shadow. That's what most distributions do. PAM has another security measure, if user's shell is not in /etc/shells, it's also considered a locked account and will not allow login.

> European Yellowstone

That makes it sound like national parks were invented in the USA and there are no national parks in Europe ... So when a new store opens outside of the US it's called European Wallmart?

Can you tell me some examples of such laws, which countries do you have in mind? Is 37 seconds really not considered "synced" to the earth's orbit and rotation?

Apply Pay?

Why is it problematic if time is off by a minute related to earth's position in it's orbit? Why should we correct for leap seconds?

Should you really change filenames of users' files and depend on the fact that they are valid utf8? Wouldn't it be better to keep the original filename and use that most of the time sans the searches and indexing?

Why don't you normalize latin alphabets filenames for indexing even further -- allow searching for "Führer" with queries like "Fuehrer" and "Fuhrer"?

Slovenia does have a cookie law, explicitly stating that setting cookies in browsers is not allowed unless permission is granted beforehand. But it's up to interpretarion whether enabling cookies in the browser counts as permission.

http://www.pisrs.si/Pis.web/pregledPredpisa?id=ZAKO8611 -- 225. člen (piškotki)

I don't remember this. I started using TK domains as a kid in 2017 and you set your own nameserver records and they didn't serve ads.

What?

> and I'm not so familiar with resizing LUKS.

LUKS2 does not care about the size of the disk. If it's JSON header is present, it will by default treat the entire underlying block device as the LUKS encrypted volume/partition (sans the header), unless specified otherwise on the commandline.

For having participated in a study, they installed an Abott FreeStyle Libre 2 sensor on me for 14 days. I could get minute readings, which is not possible with either the hardware reader nor the official app, using the almost entirely FOSS android app Juggluco.

Another FOSS web tool is Nightscout -- it generates reports and live graphs from data gathered from your CGM via for example Juggluco. My instance is at http://sladkor.4a.si. My sensor expired, so there's no more live data.

glucometerutils is an interesting repo for downloading stored data from hardware readers etc. Abott encrypted the Libre 2 USB communication to force users to use their proprietary software, but hackers managed to extract keys.

http://juggluco.nl http://nightscout.github.io https://github.com/glucometers-tech/glucometerutils http://ni.4a.si./anonymous/freestyle-keys/tree/freestyle_key...

I had a similar idea yesterday; I'd like to run Code Llama 70B on a GPU supercomputer cluster I have access to. But the wall I hit is the licence agreement by Facebook.

I'd rather download the models from 3rd party unofficial sources/torrents, since that'd be legal where I live for personal research usage. Does someone know of some server that hosts the Code Llama model weights or some torrent? I have downloaded the torrent with Llama weights, but I'd rather use the Code variant, if possible.

It wasn't. https://learn.microsoft.com/sl-SI/office/troubleshoot/excel/...

Which browser are you using?

> Even if a root server were to experience a major failure due to some sort of administration problem, there are twelve more.

This usually does not help in case of DNS. Let's say a resolver queries a root that does not reply. The resolver will time out after n seconds and then try another root server, but will not send any replies to the querying client. Therefore, the querying client has no way to know if it's the resolver that is broken or the upstream authoritative server and the querying client itself will timeout after m seconds and switch to another resolver, ignoring any possible later response from the first initialized query.

If m is larger or equal to n, the problem is aparent -- client will never know if the root is broken or the resolver, usually treating the resolver as such.

It is not null, it's an empty string.

True. Although an ISP can say they do source filtering and then fail to implement it properly. For example my ISP at home implements source filtering on IPv4, but not on IPv6.

I do not mean bitmagnet here, I mean other bittorrent clients that respond to sample_infohashes.

Implementing sample_infohashes opens your torrent client to abuse. UDP responses are much larger than queries and this protocol (BEP51) allows attackers that can spoof source IP addresses to use a large number of clients as mules for amplified distributed denial of service attacks.

Nice! I made my own DHT crawler in C for a science project in high school.

git: http://ni.4a.si./sijanec/travnik/tree/src/dht.c

Maybe if we switched from the current state of PKI with CAs to DNSSEC and DANE/TLSA, it would be easier.

But doing that would mean to either trust the current resolver you are using, which is unacceptable, as it's sometimes cloudflare or google, or running your own resolver on every device -- which also does not work. It is absolutely crucial for the DNS system we have today that there's a small number of resolvers that have large caches, otherwise root servers would blow up. Unless we massively start implementing some novel RFCs.

Why centralization? XMPP users are identified with their JID in format username@server.example, similar to email addresses. And server to server communication is very well documented and major XMPP servers for instant messaging (ejabberd and prosody) both allow server to server message exchange. I don't see any benefits for Matrix when it comes to federation and I wouldn't agree with you that by using XMPP you tolerare centralisation at all.

It's level of centralisation is comparable to that of email (and email is very well federated in my opinion).

MUCs like the XEP you mentioned are handled by a different XEP.

Well if you use a matrix client that does not support video calls (Syphon), you also can't be called.

Camera and microphone are not a mandatory feature of Matrix either (;

I think matrix has group voice calls. But maybe I'm wrong and they're just using Jitsi.

I like that Matrix puts an emphasis on message history, but on XMPP that seems like an afterthought. For example on XMPP multi user chats, messages history retention relies on the server that hosts the chatroom and this is sometimes very limited, sometimes even disabled. And if that's the case, you only get messages when you have a _client_ online, as your server won't store messages when you don't have a client connected. Or something like that.

And AFAIK, multiple clients with a MUC room opened will appear as multiple users sometimes. I'm not sure how it all works really, so take my words with a grain of salt.

XMPP seems more like a pubsub system for system messages, like MQTT (:

I really don't like the forced end to end encryption of chats. I do not care that much about the security of my chats to justify the complications of end to end encryption -- verifying devices, constant "reset" prompts that I don't even know what they do, but they seem like they are very destructive.

I trust my homeserver, since I host it myself. I do not need end to end encryption. I understand this is an issue of the client, not the protocol. But there aren't many clients. Prebuilt element for android from fDroid will connect to matrix.org, vector.im and other network hosts and I don't want that at all, although this tracking can't easily be disabled. It feels very centralised.

I run my own homeserver and want to chat with my friend that runs his own server. It is unacceptable to me that clients connect to anything other than those two homeservers (apart from CRL and OCSP, which should also be disabled by default, as I consider those protocols great spyware). Not to mention the fact that homeserver software itself is known to make connections to the servers of the developers by default, without ever talking to someone on matrix.org. This is also unacceptable. My homeserver should only connect to other homeservers.

3PID is a failed attempt, as it centralises identities and works by utilising a single point that gathers a lot of personat information at one place. I don't want it.

XMPP does not have those issues. Set up prosody and use dino or Conversations and they won't make any connections to non-essential servers. Furthermore, the end to end encryption is way easier to use in XMPP (OMEMO), and it's easy to turn it off if you don't need it.

OpenNIC is a project that replaces and extends the official DNS root maintained by Verisign.

http://opennic.org

Works for me. Your browser does not support URL Fragment Text Directives[0]. What browser version do you use?

https://wicg.github.io/scroll-to-text-fragment/

Do you use abbaddon or the official client?

You don't need to load the entire JSON object as a DOM into RAM just to validate it. Validation can easily be done using a stack and iteration with space complexity being the depth of the JSON (stack) and time being linear to the length of the object.

Relevant: LEJP - libwebsockets json parser.

You specify what you're interested in and then the parser calls your callback whenever it reads the part of a large JSON stream that has your key.

https://libwebsockets.org/lws-api-doc-main/html/md_READMEs_R...