minus7

Is it really, or do you just think so and it could actually mislead you and your LLM the next time? In my experience, the information in the comments tends to be quite redundant, often even with other comments in the same file

If you don't think about it, it doesn't seem like a bad idea. If you do, you realized how incredibly complex the whole thing is

It's just using libvte like gnome-terminal and lxterminal, so I doubt it's much different from them.

>Note that if an "Ephemeral Device" exists for more than 4 hours continuously, it consumes 0 minutes, and is billed as a tagged device. The docs are being upgraded to change the language around this.

Thanks for explaining, that makes the limit very reasonable. I updated my comment

This is potentiallya bit of a downgrade if you are using the Personal or Standard plans for infrastructure:

- Reduction from 100 to 50 tagged devices on Personal (was 100 total devices before)/Standard (was 100 + 10 per user before)

- ~~Limit to 1000minutes/month (16.6 days) for ephemeral devices (simply counted towards the total device limit before), meaning you can't even have a single permanently connected ephemeral device anymore on Personal/Standard~~ This is incorrect, see comment below

- Limit of 3 ACL groups on the Personal plan (previously no limit), but also allows 10 ACL groups on the Standard plan (previously had no access to the feature)

Unfortunately, the pricing for extra devices and ephemeral devices is "contact sales".

The `eval` alone should be enough of a red flag

Issues I observed, mostly using GoLand:

- syntax errors displaying persistently even after being fixed (frequently; until restarted; not seen very recently)

- files/file tree not detecting changes to files on disk (frequent; until restarted; not seen very recently)

- cursor teleporting to specific place on the screen when ctrl is pressed (occasionally; until restarted)

- and most recently: it not accepting any mouse/keyboard input (occasionally; until killed))

JetBrains should stop building stupid AI shit and fix their IDEs. 2025 versions are bordering on unusable.

All code is inherently not concurrency-safe unless it says so. The http.Client docs mention concurrent usage is safe, but not modification.

The closure compiler flag trick looks interesting though, will give this a spin on some projects.

In my experience that's because most users run their monitors on 100% brightness instead of turning that down. I prefer light mode unless it's really dark.

I was excited that Firefox finally exposed its local translations as API, but it's Chrome-only (still?). Will be nice for userscripts, for example to replace Twitter's translation button that hardly ever works

If you consult with someone over their project, then proceed to fork it behind their back, that's just being a dick, even if it was perfectly legal. We should not accept that kind of behavior. And that's even ignoring that the consultation was unpaid and the project was actually even stolen.

There's a pinned comment on the video about the cable:

    For everyone that's interested, the included USB cable is wired like this:
    
    GND    -------> GND
    D+  -------> VCC
    D-   -------> VCC
    VCC -------> VCC
    
    So it is a non-standard cable! I measured by checking continuity between the USB A plug, and the USB C connector with a USB A adapter on it. None of my probes are small enough for a USB C connector's pins directly.

    There was a bit in the video about this that ended up on the cutting room floor, my bad!

No one's gonna pay just for the sidebar being collapsible. Maybe someone is going to because it was the last straw, but most are just going to be annoyed. Better ask for money only for something with real value (which I'm sure the pro tier also includes).

Hey,

I found your library a few weeks ago when I was annoyed by nothing like this being built into the standard library. It’s been a breeze to use so far.

A neat trick I found to gauge bottlenecks in pipelines is using Buffers between steps and running a goroutine that periodically prints `len(buffered)/cap(buffered)`

Used to vendor for a bit, but it's messy in git. If anything disappears, it should still be easy enough to recover it from somewhere.

Unfortunately that feature of invidious also seems to depend on the upstream RSS URLs and the feeds are just empty at the moment

The DNS records for that domain are hijacked on two German ISPs' resolvers I tested, yikes. Google and Cloudflare resolvers resolve it just fine however.

  canna.to.  0 IN CNAME notice.cuii.info.
  notice.cuii.info. 34953 IN A 167.233.14.14

`<div />` getting parsed as opening tag instead of self closing is a bug (be it in the browser or in the spec) and should be fixed

IIRC there were issues with IPv6 on the hardware, which will be replaced once services move to Europe, and IPv6 will be available.

Best thing about it is that it has support for huge files; I was looking for a hex editor that can do that

You can also use a pre-polished connector instead of splicing

e.g.: https://www.fs.com/de-en/products/35165.html (there's instructions if you scroll down)

I also installed fiber a while ago after being told "this isn't hard" by Michael Stapelberg's post [1]. Luckily I managed to get by with patch cables; a 30m cables really goes a long way! I even managed to get a duplex LC through a 16mm conduit with a pretty tight 90° bend, but just with lube, quite some force and luck.

I can just recommend going for it if you're curious. Cables and tranceivers really aren't expensive from fs.com (though their sales people may start bothering you).

Edit: And old, used network cards from ebay, of course.

[1]: https://michael.stapelberg.ch/posts/2020-08-09-fiber-link-ho...

Good joke, there are no $35 pis

>TOTP is almost purpose made to be phished.

Well, that's a tad hyperbolic. While TOTP codes have no protection against phishing, an attacker still requires interaction from the account owner every time they want to make use of it and can't use the compromised account without the owner being able to notice it. Granted, someone getting phished might not notice. Still better than nothing (but not better than a password manager checking which site it's filling the login into)

The problem is that a) you want it to work properly for users of Go modules hosted on your Git forge and b) you don't want to effectively waste time on working around a problem that the GoProxy engineers could just solve with a more intelligent (but maybe more complex) design.

Literally no significant extra effort to download the files and put them with your website. If you fetch them from a CDN with a fixed version you're not getting updates anyway.

> Nearly 8% of the top 10 million websites support HTTP/3, according to the latest figures from W3Techs.

Let me guess: those are all behind Cloudflare

Does this article not need to say more than the <title> and a picture of tools? Because that's all I see

FLIF has been superseded[1] by JPEG XL, which, in my very limited testing, performed better (speed and compression-wise; libjxl) than FLIF.

Interesting though that FLIF 0.4 was released 3 days ago. I haven't checked out that new release (and won't), but the previous one wasn't great code and was a pain to use (ended up using its CLI instead of the lib). We ended up going back to PNG because of interoperability and speed. I haven't looked at libjxl's API yet.

Edit: 8605KB in 11.5s with libjxl, so it's not actually better than FLIF 0.4 (7368KB in 23.9s)

[1]: https://flif.info/#update