made you look
What’s the risk here though, a company like Amazon makes a closed source version of it?
If it was a file format library, or something like a web server I’d get it. But stuff like cp are effectively just userspace wrappers around kernel APIs.
The idea is that it’s left up to the windowing toolkit itself (.e.g GTK or Qt, etc.), so the compositor can focus on just compositing, which makes sense IMO as it’s how other platforms handle it (Except they have a single OS provided windowing implementation). Problem is, that leads to massive fragmentation of functionality, every app has different toolbars and features based on the toolkit they use, and requires each app to handle it, which sucks and shouldn’t be the case.
Like in the Factorio case, it uses SDL for windowing, and SDL actually supports handling titlebars itself. But Factorio just wasn’t including the dependency that enabled it at that point, so all it took to fix it was including it and everything started working. But that’s still extra work that had to be done just to get minimum functionality, which wasn’t needed on e.g. KDE.
I mentioned in my other response, it’s the inflexibility that’s the actual problem. Lots of apps do want CSD, or at least control over how their windows are presented, but Gnome going “you’re on your own” is the worst outcome.
Their go-to solution is libdecor, which is just a library that implements a titlebar, still putting the burden on apps (Or rather, whatever windowing library they use) to be responsible for it.
Worst thing is, I kinda get their argument against supporting it, they’re just really inflexible about it which just makes the whole issue too heated.
what happens if you do run something on GNOME that simply doesn’t have any code that draws its own titlebars?
Is driver support stable in ReactOS?
It exposes the Windows driver APIs, so it’s “getting there”, slowly. They got the (Win 7 era) Nvidia driver running recently.
Really good network stack. Linux is catching up surely but places like Netflix run a ton of stuff on BSD simply for that stack.
Depends on the specific BSD, OpenBSD for example is only just now catching up to Linux.
Edit: Slide 28 for a graph
if you want it to go away, everyone who is working on it and making it work right now disagrees with you
I’m sure most people wouldn’t like losing their jobs.
They’re called “selections”, the main ones being PRIMARY and CLIPBOARD, and it’s effectively a form of IPC mediated by X. When you select something, that goes into the PRIMARY selection, while when you copy something, it goes into both PRIMARY and CLIPBOARD.
The problem is that “middle mouse click” isn’t actually paste, it’s “insert primary selection”. As long as they’re in sync you won’t notice any issue (Ctrl+V and MMB will both insert the same content), as soon as they’re out of sync you’re suddenly exposed to an implementation detail of the X11 protocol.
And it’s easy to go out of sync, simply copy something and then select unrelated text, now Ctrl+V and MMB will output different things. It can be useful, e.g. if you’re having to copy a bunch of different pieces of text from one window to another, you can simply select and MMB, no keyboard needed, but it’s not intuitive IMO, and conflicts with modern usage of the middle mouse button (Get it wrong when trying to open a link in a new tab and you’ll dump whatever text you last selected into the site instantly)
Also, these selections aren’t a thing under Wayland, it’s been re-implemented as a normal paste operation there. The question is actually whether the middle mouse button should be treated like any other mouse button or have this special behaviour by default. My vote is to expose it via the mouse settings applet and leave it up to users, like any other special mouse button.
AMD did nothing to make their drivers better, Vale did.
That’s literally the point of open source though, both AMD and Intel rely on open source drivers, so anybody can fix a flaw they encounter without having to rely on the company to “consider allocating resources towards a fix for legacy hardware”
Ahh, yeah that’s a bit harder, CSS multiline stuff is pretty flaky from what I can recall. You need to drop down to block layout, e.g. making the containing element a flex parent (Better term than that?) and then making the icon centered within that can work, but then we’re back to square one with sizing the icon.
<p><svg class="icon">...</svg> Text</p>
p .icon {
--size: 1.25em;
vertical-align: calc(0.5cap - 0.5 * var(--size));
height: var(--size);
width: var(--size);
}
Done.
Something easily turned off is distinctly not opt in. They’re still lying to everyone.
It’s just that nobody can agree what “opt-in” actually means, apparently.
Git itself (Or any other VCS for that matter) really should treat symlinks as special, similar as to how btrfs stores everything as “reflinks” internally. They be stored as special references to other tracked objects (so it’d be impossible to commit a symlink that pointed at anything other than a checked-in file, and ensure they always match), and git can materialise them as needed.
There’s also xml5ever, for if you hate XML.
Well we’ve recreated namespaces, and JSON already has a completely useless type system, so it’s pretty much already there.
It’s a shame you’re getting downvoted since you’re actually right, and distros are in the process of moving to “kmscon”, a userspace console, rather than the old kernel console (Which iirc isn’t actually intended to be a general purpose console, it’s meant for boot messages)
That said, the fonts the kernel uses are old style bitmap fonts, extremely limited “attack surface” as they’re not doing stuff like opentype/font shaping, it’s just setting pixel values directly.
Pretty sure that’s just high-frequency trading.
Landrun as well, takes the restrictions on the command line. Can look messy, but does make it entirely standalone, so you can e.g. drop it into a service file as the readme shows easily enough.
Because of static linking, a single GPL dependency turns the entire resulting binary into a GPL licensed one, so yeah just use something like the MPL in that case (Or EUPL, which I hear is similar)
LGPL has the same issue, since it only provides an exception for dynamic linking. But honestly that’s all an issue for lawyers and judges to sort out (I bet you could win in court with an argument that dynamically linking to GPL is actually fine).