movq

www.uninformativ.de

No description provided.

Recent twts from movq
In-reply-to » The day will come when I’ll have to dig up why every process can interact with every other process’s file descriptors by means of /proc/$pid/fd on Linux (if it’s the same user). Is there a legitimate reason for that … ? (I know about hidepid, but that doesn’t help here.)

@prologic@twtxt.net Hmm. That would require me to put each and every process into its own namespace, wouldn’t it? I don’t think that’s practical. 🤔 Or maybe I’m misunderstanding.

I’ve never really thought about this, to be honest. There’s no procfs on the BSDs, so I guess that a) it’s not really necessary, b) there should be plenty of rants from BSD users on this issue that I could read. 😅

⤋ Read More

The day will come when I’ll have to dig up why every process can interact with every other process’s file descriptors by means of /proc/$pid/fd on Linux (if it’s the same user). Is there a legitimate reason for that … ? (I know about hidepid, but that doesn’t help here.)

⤋ Read More
In-reply-to » A GTK 4 application showing an empty window uses about 160 MB of RAM:

@prologic@twtxt.net

Have we really gotten that lazy and inefficient? 🤔

That’s the question. It certainly strikes me as odd that everything keeps getting bigger, heavier, slower all the time. But why is that? I refuse to believe that this is just incompetence. 😅

⤋ Read More
In-reply-to » A GTK 4 application showing an empty window uses about 160 MB of RAM:

@abucci@anthony.buc.ci Yes, that appears to be the case. I stumbled upon GSK_RENDERER=cairo a while ago – it helps with the startup times and a little bit with memory consumption, which is down to about 130 MB.

Still quite heavy and startup times are noticeably longer than GTK 3. It’s not that much, but it’s there.

In my case, I wanted to port xiate from GTK 3 to GTK 4. But that much memory usage for just one terminal window (of which there are usually many, so we’re talking about gigabytes here) is too much. GTK might have its use cases, but I think it’s no longer the right toolkit for me.

I don’t want to hate on it too much. I know too little of the internal details of all this, so I don’t have an informed opinion.

(Before anyone suggests it: Yes, I could have one process show many terminal windows. That would be a lot faster and would use a lot less memory. But I explicitly do not want that. xiate used to have this model – until one day at work, a bug in the terminal library VTE crashed all my terminal windows at once. It really only crashed one window, but since they all ran in the same process, they were all gone. This is unacceptable.)

⤋ Read More
In-reply-to » My night hike wasn't all that great. Well, I admittedly "saw" two frogs. Some black blobs were jumping across the path in the pitch dark that is. But also gazillions of mozzies. And I got myself a mega blister.

@lyse@lyse.isobeef.org I had a look last night. Mars is much more faint than I expected. I could hardly see it with my bare eyes and it’s not much better on photos:

https://movq.de/v/e682fc642d/IMG_5009-annotated.jpg (5184x3456, 4.8 MB)

⤋ Read More

A GTK 4 application showing an empty window uses about 160 MB of RAM:

$ wget https://movq.de/v/138ab3e622/win.c
$ cc -Wall -Wextra -o win win.c $(pkg-config --cflags --libs gtk4)
$ ./win

It also takes several seconds to start on my machine because it is compiling shaders and initializing DRI (it’s faster on the second run, unless you happen to lose ~/.cache/mesa_shader_cache/). This might be a hint as to why it’s using so much memory: There’s obviously much more going on behind the scenes these days, not just a little bit of internal housekeeping and then creating a window.

⤋ Read More
In-reply-to » Ah, damn … I have that nice split keyboard, which I really love (I use it more than my Model M 😅), but it’s actually two USB devices. Wayland compositors usually struggle with this. 😢 When Wayland really hits big time, I’ll have to look for another model … Let’s hope the keyboard just breaks in the meantime, otherwise it’s gonna feel a little bit frustrating.

@jmjl@tilde.green Sway does indeed work as expected. 👌

⤋ Read More
In-reply-to » Ah, damn … I have that nice split keyboard, which I really love (I use it more than my Model M 😅), but it’s actually two USB devices. Wayland compositors usually struggle with this. 😢 When Wayland really hits big time, I’ll have to look for another model … Let’s hope the keyboard just breaks in the meantime, otherwise it’s gonna feel a little bit frustrating.

@jmjl@tilde.green It depends on the compositor. I tried Hyprland earlier and it exposed the problem.

Never heard of interception-tools, looks promising. 🤔 Thanks!

⤋ Read More
In-reply-to » Ah, damn … I have that nice split keyboard, which I really love (I use it more than my Model M 😅), but it’s actually two USB devices. Wayland compositors usually struggle with this. 😢 When Wayland really hits big time, I’ll have to look for another model … Let’s hope the keyboard just breaks in the meantime, otherwise it’s gonna feel a little bit frustrating.

@jmjl@tilde.green Yes, kind of – I can type on the individual halves, but Shift on the left part plus some letter on the right part doesn’t work. 🫤

⤋ Read More

Ah, damn … I have that nice split keyboard, which I really love (I use it more than my Model M 😅), but it’s actually two USB devices. Wayland compositors usually struggle with this. 😢 When Wayland really hits big time, I’ll have to look for another model … Let’s hope the keyboard just breaks in the meantime, otherwise it’s gonna feel a little bit frustrating.

⤋ Read More
In-reply-to » Enjoying a day off, sitting on the balcony in some nice 18°C. 👌

(btw: There are two parallel runways on the airport but, from that balcony, I can only see the planes that approach one of them. So this is roughly half of the planes arriving here.)

⤋ Read More

To all people writing software: Please, always include the time zone when showing dates/times, especially in log files. It’s almost always the right thing to do and can save your users so much headache during debugging.

⤋ Read More
In-reply-to » Crap, I just missed moonrise. Well, better than nothing. https://movq.de/v/fc8a9b9472

@lyse@lyse.isobeef.org Thanks. 😅

During this time of year, full moon rises at a point where I can see it from my balcony. And it rises close to objects on the earth, buildings and such. Makes the moon appear much larger when using high zoom levels. I missed that moment, it’s already too high. 😅 Maybe tomorrow or next month.

⤋ Read More
In-reply-to » I was originally going to switch to HTML 4.01. I ended up choosing XHTML because it isn't forgiving like regular HTML; tiny errors in markup will make browsers refuse to display anything. This will help me have a more correct website according to the specifications.

@adi@twtxt.net That looks promising, thanks for the tip. :)

⤋ Read More
In-reply-to » It really pisses me off that Firefox hides the scrollbar and only shows it when I actually scroll. What the bloody fuck is that bullshit!? It can be fixed in about:config by changing layout.testing.overlay-scrollbars.always-visible to true as I found out by experimenting.

(Regarding hidden scroll bars and not knowing whether an area can be scrolled or not, here’s a conversation I had recently with a younger non-tech person:

  • Him: “Uhh, $older_person got stuck using Element (that Matrix client) on their phone again. Can you help?”
  • Me: “Yeah, I can take a look. He probably just tapped every item on the screen and ended up in some weird sub menu.” (A little bit of “that’s not how you do things” in my voice.)
  • Him: “Hm, isn’t that how you do it? How else would you know where to tap?”

Total bliss. That feeling of “oh, I know, that’s a button that I can click” is completely lost to newer generations of computer users. Modern UIs have so little indication of what kind of elements are shown on the screen and what you can do with them that users just … try every pixel. Tap here, swipe there, see what happens. Is this really a good way to do UIs? Am I just old and grumpy?)

⤋ Read More
In-reply-to » It really pisses me off that Firefox hides the scrollbar and only shows it when I actually scroll. What the bloody fuck is that bullshit!? It can be fixed in about:config by changing layout.testing.overlay-scrollbars.always-visible to true as I found out by experimenting.

@lyse@lyse.isobeef.org I wish it would just respect the global GTK setting for that … 🙄

⤋ Read More
In-reply-to » @mckinley Regarding your move to XHTML 1.1:

I was originally going to switch to HTML 4.01. I ended up choosing XHTML because it isn’t forgiving like regular HTML; tiny errors in markup will make browsers refuse to display anything. This will help me have a more correct website according to the specifications.

I really miss this property with regular HTML. This might be a stupid question, but how do I find out if my HTML is valid? I mean, other than running it through W3C’s tool. My browser surely doesn’t tell me …

⤋ Read More
In-reply-to » People complain about the noise that the crows in our area make. Well … https://movq.de/v/7b8c06eb73/noise.ogg Notice anything?

@abucci@anthony.buc.ci It is. The general area where I live is not plagued by crows, it’s plaged by airplanes. The sound you’re hearing happens roughly every 2 minutes, 5am thru 11pm, when the wind is right (it often is in summer). They tried to extend this to 24/7 around 15 years ago, but failed, luckily.

Okay, I get it. The crows can be annoying (especially if you don’t like birds – I happen to love them, so I’m biased). They are noisy. But really … compared to all those jets, they’re irrelevant … I guess people simply think they have a better chance at fighting the birds than the planes. :/

⤋ Read More
In-reply-to » Boy, do I dislike summer. Good thing we continue to fuck up the planet, so it’ll get worse every year. 👌

@lyse@lyse.isobeef.org I know that feeling. 😅 My fan was blowing like crazy, too (you don’t want to know why) – but it’s only the laptop from work, it doesn’t create that much heat.

Lovely 12°C this morning and just 22°C now. That’s fine! 👌

⤋ Read More
In-reply-to » Time to dive into threading and c++. I will start with making the file/image downloader threaded, then I'll make the timeline fetching and all that threaded as well.

@stigatle@yarn.stigatle.no I’d imagine that you’d be mostly dealing with GLib functions in your case. 🤔 GTK/GLib has its own event loop and you’re probably screwing things up if you’re using something like pthreads directly. 🤔

⤋ Read More
In-reply-to » @mckinley Yeah, that’s more clear. 👌

Right, especially not if it’s “cloud storage”.

Errrr, what I meant here: It’s not useful if “the cloud” manages the key. You know, those little check boxes at Google or Azure, “encrypt this storage and generate a key for me” …

⤋ Read More
In-reply-to » Rebooting a LUKS Encrypted System Without Typing The Passphrase: https://mckinley.cc/blog/20230526.html

@mckinley@twtxt.net Yeah, that’s more clear. 👌

Systems that are on all the time don’t benefit as much from at-rest encryption, anyway.

Right, especially not if it’s “cloud storage”. 😅 (We’re only doing it on our backup servers, which are “real” hardware.)

⤋ Read More