secure

> Now the objects are no longer threadsafe. You can't just read a message in one thread and process it in two different threads.

This is not correct. The Opaque API provides the same guarantees as before, meaning you can read a message in one goroutine and then access it (but not modify it) from other goroutines concurrently.

> Great, now there's an API per struct/message to learn and communicate throughout the codebase, with all the getters and setters.

No, the general idea (and practical experience, at least for projects within Google) is that a codebase migrates completely from one API level to another. Only larger code bases will have to deal with different API levels. Even in such cases, your policy can remain “always use the Open API” unless you are interested in picking up the performance gains of the Opaque API.

The 3? (4?) video recordings I watched from FOSDEM 2024 all had bad audio :( sounded overdriven to me

The Debian patch for this setting is: https://salsa.debian.org/kernel-team/linux/-/blob/master/deb...

The decision made there is from 2013 (see https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=712740), so it might be worth revisiting.

Could you file a bug report to get the discussion started?

I don’t know where you got the impression that the Go compiler would run on “the production machine”.

gokrazy images are built whereever you like (typically your PC, or some more automated setup) and then deployed to the target device. There is no Go toolchain on the target device.

I have too many people in my circles who have been burnt by btrfs data loss that I’m not touching it unless I can’t avoid it.

I do realize that btrfs matures over time, so that’s just my personal approach :)

Hey, thanks for letting us know!

I most recently tried the AMD GPU that came with my ThinkStation P620, and it could not drive the monitor at native resolution :(

Maybe I should try a 6900XT. Which particular model do you have?

How much power does it use when driving the 8K monitor?

Do you have PCIe ASPM enabled in the UEFI setup for your PEG slot, and does that run reliably? I was running into problems with ASPM with my nVidia: https://forums.developer.nvidia.com/t/pcie-errors-after-enab...

I wouldn’t say happy, but it’s currently the only choice. Maybe as drivers advance we can have non-DSC links — the hardware should support it per the spec…

Yes, with AMD cards, the 8K monitor only works on Windows.

With Intel ARC cards, the 8K monitor works neither on Windows nor on Linux :(

Copying my reply from Mastodon:

> I honestly don’t know. Maybe this is a bug in the monitor’s firmware and it does use some version of DSC after all? Or you’re right regarding lack of DSC support and it uses more than 8 Gbps but displays it incorrectly? Not sure. But something is definitely different about the monitor’s link when a MacBook is connected…

Then I read https://news.ycombinator.com/item?id=36581855, which points out 24 bit colors vs. 30 bit colors. The MacBook links up with 24-bit colors indeed. Maybe that’s the difference?

Yeah, I’m returning it.

I kept the 360, but am currently using the Advantage 2 again. I never got used to the 360 palm pads.

My Linux research distribution “distri” works the same way :)

(I learnt about Haiku’s similar approach after building distri.)

https://michael.stapelberg.ch/posts/2019-08-17-introducing-d...

https://distr1.org/

Debian Code Search is based on this work! :)

https://codesearch.debian.net/

(If you want a lot more details, check out my thesis on it: https://codesearch.debian.net/research/bsc-thesis.pdf)

OpenBSD packages i3: https://openports.se/x11/i3

I briefly (for about a year or so) developed i3 on NetBSD.

I have no knowledge of other more obscure variants, though.

It truly is a small world! I’ll say hi to him, thanks for sharing :D

Yup! We even have an FAQ entry on the reasons for not merging features without careful consideration: https://faq.i3wm.org/question/778/why-is-patch-not-merged-an...

Also, saying “it’s not that many lines of code” comparing the before and after state is too simplistic: it doesn’t include the cost of changing the existing code into the desired state, which can be a significant effort, even if the number of lines doesn’t differ much before and after. See https://github.com/i3/i3/pulls?q=is%3Apr+author%3Astapelberg... to get an impression of how much work it was to merge gaps.

Thanks for providing the background here, and thanks for maintaining i3-gaps over all the years! :)

I’m also happy we could complete this merge now!

No further cleanup needed

That’s right. We waited for a volunteer to help with the merge (as we both didn’t have enough time to spend on working on gaps), but nobody ever showed up.

Eventually, because life circumstances permitted it, I was able to spend a few weeks, first getting gaps merged as-is (as discussed and communicated) and then even cleaned up, addressing all known limitations.

i3-gaps is included in i3 v4.22, which is available in many Linux distributions already.

You can find the release notes at https://i3wm.org/downloads/RELEASE-NOTES-4.22.txt

Documentation for the gaps feature is at https://i3wm.org/docs/userguide.html#gaps

FWIW, I couldn’t reproduce the Kinesis Advantage measurement when I did my tests:

https://michael.stapelberg.ch/posts/2021-05-08-keyboard-inpu...

I got 13.6ms ±21%, Dan writes about 50ms.

Maybe we measured different models, the Kinesis Advantage has many revisions.

Lots of people are saying “you’ll come back” and “nobody is migrating for real”, and I have had the same opinion back when the Twitter deal was first announced and a few people moved over to Mastodon.

But this time, it was very different: When the deal went through, a large part of my bubble really did move to Mastodon. This happened over a few days, and after the second or third day, my Twitter timeline was noticeably empty, and Mastodon is very alive instead.

Of course I can’t predict if this sticks over the medium or long term, but for me, the situation is very clear: the people I care about are now on Mastodon.

(I spent 10 years on Twitter, with over 4000 followers.)

Do you have a link to your code? I thought fs.FS is read-only, so I’m curious where/how you bring writing data into it (“copies from folder A to folder B”)? Or do you mean that your implementation of fs.FS generates the content on-demand, when net/http serves it?

A smart plug would probably work just as well for most cases, but one downside is that you need to ensure your computer is safely powered off before turning off the smart plug.

Conceptually, pressing the power button seems a little cleaner to me, and also works well with suspend-to-ram (what I originally wanted to do), which the smart plug can’t do.

I’m not too concerned with safety — we’re talking about 3.3V (signal) and 5V (power) here.

Now hyphenated, thanks for the hint

We’re already starting to see 2.5 Gbps interfaces become more and more popular on mainboards, so I suspect once 10 Gbps interfaces become more common on high-end mainboards, vendors will start caring about WOL with these chips :)

Cheaper transistors should totally work, as I tried to describe in the article.

I had this particular MOSFET lying around from a different project, so that’s the main reason I chose that one :)

Plugging this here in case it’s useful to anyone else working with squashfs: I wrote a squashfs reader/writer in Go: https://github.com/distr1/distri/tree/master/internal/squash... (part of the https://distr1.org/ project)

Here’s a video answering your question :) https://www.youtube.com/watch?v=wX81vAzZcDo

Which version is the samba one? Latest release? Git?

If you want to see the man page of the version in Debian, that would be https://manpages.debian.org/testing/rsync/rsync.1.en.html

Disclaimer: I wrote the software behind manpages.debian.org :)