Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
daneel_w
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
13 ms
·
271.
▲
by
daneel_w
1y ago
It confirms that Python indeed has poor executional performance.
272.
▲
by
daneel_w
1y ago
Putting the delimiter on a line of its own is a syntactical trick that helps bringing small additions down to a neater 1-line diff instead of a 2-line diff. You've probably run into it many times before in other contexts without thinki
273.
▲
by
daneel_w
1y ago
>I've always wondered why that behaviour emerged.. Because they share the same roots, the same general "scene". The demo groups and the crack groups always mingled. We were all "sceners". Tons of groups had a p
274.
▲
by
daneel_w
1y ago
I always considered this the real deal and authorative collection of chip tunes: https://ftp.modland.com/ Modland is over 20 years old at this point.
275.
▲
by
daneel_w
1y ago
For OMEMO to function smoothly it wants permanent storage for Publish/Subscribe functionality (XEP-0222) on the server. The feature predates OMEMO and was designed for various other purposes, such as storing a contact roster or profile
276.
▲
by
daneel_w
1y ago
Perhaps 20-25 at peak, some of which are in MUC ("group chat"). XMPP itself is very lean, on the server side it's mainly a message relay, and Prosody can handle a ton of concurrent traffic on practically nothing worth of serv
277.
▲
by
daneel_w
1y ago
OMEMO is not a server feature, it's a client feature, but yes OMEMO works with Prosody in the picture. It's mature, modern and actively developed, the full list of extensions supported by Prosody is here: https://prosod
278.
▲
by
daneel_w
1y ago
I have no experience with ejabberd. As for Prosody, it does everything I/we want it to do, is very easy to configure and runs incredibly light - the host is a single-core virtual machine running OpenBSD on 256 MiB of RAM, and we'v
279.
▲
by
daneel_w
1y ago
Daily user of XMPP as well since over a decade. I still call it Jabber out of habit. Prosody on server, Profanity on desktop (terminal), Monal on mobile.
280.
▲
by
daneel_w
1y ago
Of course it would continue to operate. The Chinese division of Arm went rogue and basically captured the entire operation from within. USA has both the engineering expertise and the incentive to do the same with an entire chip fab, if push
281.
▲
by
daneel_w
1y ago
Thanks for the tech details.
282.
▲
by
daneel_w
1y ago
> A significant fraction of data corruption occurs while it is being moved between storage and memory Surely you mean on the memory bus specifically? SATA and PCIe both have some error correction methods for securing transfers between
283.
▲
by
daneel_w
1y ago
...then the checksum can't clear the associated operation. Same result.
284.
▲
by
daneel_w
1y ago
My feeling is that any portion that no longer checksums correctly is practically forfeited, which makes any form of recovery an arbitrary result - a half-baked transaction. I don't see how that's a more desirable outcome than roll
285.
▲
by
daneel_w
1y ago
When I test with Librewolf (anti-fingerprinting enabled) and uBlock Origin as only addon installed, it manages to retain my fingerprint across sessions. Perhaps it's my static IP.
286.
▲
by
daneel_w
1y ago
The WAL was corrupted, the actual data is lost. There's no parity. You're suggesting that sqlite should somehow recreate the data from nothing.
287.
▲
by
daneel_w
1y ago
You should share details on your setup.
288.
▲
by
daneel_w
1y ago
I maintain and develop the back-end of a telecom services provider. It's almost 100% Perl, and once in a blue moon when we add something entirely new we usually stick with Perl.
289.
▲
by
daneel_w
1y ago
Why? This isn't a Perl problem.
290.
▲
by
daneel_w
1y ago
And Gentoo aren't alone: https://metacpan.org/dist/DBIx-Class/source/lib/DBIx/Class.p...
291.
▲
by
daneel_w
1y ago
I do have a contingency plan for all my digitial memories and my online accounts etc.
292.
▲
by
daneel_w
1y ago
With valuable I should've elaborated that it's my set of constantly changing daily-use data. Keychain, documents and notes, e-mail, bookmarks, active software projects, those kinds of things. I have a large amount of memories and
293.
▲
by
daneel_w
1y ago
My valuable data is less than 100 MiB. I just tar+compress+encrypt a few select directories/files twice a week and keep a couple of months of rotation. No incremental hassle necessary. I store copies at home and I store copies outside
294.
▲
by
daneel_w
1y ago
It's also endlessly surprising how people over-think the process and requirements.
295.
▲
by
daneel_w
1y ago
For what it may be worth, here's a most basic (but fully working) config for running Unbound as a DoT-only forwarder: server: logfile: "" log-queries: no # adjust as necessary interface: 127.0.0
296.
▲
by
daneel_w
1y ago
I'm implying that I'm convinced they are not storing statistics on (thus leaking) every queried hostname. By your very own admission, they clearly state that they perform statistics on a set of malicious domains provided by a thi
297.
▲
by
daneel_w
1y ago
Disagree that it's fairly close to the statement "they resell traffic logs" and the implication that they leak all queried hostnames ("secret hosts, like for your work, will be leaked"). Unless Quad9 is deceiving us
298.
▲
by
daneel_w
1y ago
Yeah it's not like they have a long track record of being caught red-handed stepping all over privacy regulations and snarfing up user activity data across their entire range of free products...
299.
▲
by
daneel_w
1y ago
If you're the technical type you can run Unbound locally (even on Windows) and let it forward queries with DoT. No need for neither Tor nor running your own external resolver.
300.
▲
by
daneel_w
1y ago
Quad9, dns0.
More ›