I was reasonably certain, but left it open in case OP knew of some edge case where flags that are intended to be machine independent caused bugs on different architectures

-O2 vs -O3 adds -fgcse-after-reload -fipa-cp-clone -floop-interchange -floop-unroll-and-jam -fpeel-loops -fpredictive-commoning -fsplit-loops -fsplit-paths -ftree-loop-distribution -ftree-partial-pre -funswitch-loops -fvect-cost-model=dynamic -fversion-loops-for-strides

I don’t think any of these optimizations require more modern hardware?

For historical info - Oracle bought OpenOffice and started to close it down, so all the developers that worked on it forked it into LibreOffice

Oracle has since given OpenOffice to an open source group, Apache, but the main development still happens on LibreOffice

It’s not just betas - it’s in the main release, too

We’re still using them on machines where performance doesn’t matter

On build machines, they’re on a special VLAN and don’t have endpoint protection, but they only download from a protected mirror

Their ftrace hooks caused all disk usage to be serialized, making your multi-core processor single-core when doing anything I/O bound

We saw between 500% - 800% increases in build times with their software installed

without any distro or configuration caveats.

In those cases, they generally have the Ubuntu version that’s supported in the specs section

Oh god. Sentinel one is horrible. If they’re taking issue with your testing, you’ve really screwed the pooch

Somewhere around 0,0 or 1,1

There are amazing possibilities in the theoretical space, but there hasn’t been enough of a breakthrough on how to practically make stable qubits on a scale to create widespread hype

Both GNU and GrapheneOS have staunch requirements and will accept no compromises.

This is a situation where their requirements don’t align, so they’ll never reach an agreement.

GrapheneOS, for example, is also strictly against making the Fairphone line of phones a little more secure because it doesn’t meet all of their security requirements

In this case GNU won’t certify GrapheneOS as fully open because it includes binaries that aren’t open

The FSF is more along your line of improving the situation where they can

I’d used Linux a bit out of curiosity in the Windows XP era

Windows Vista came out and was completely unusable on the computers I or anyone around me owned. It was also harder to configure than Linux and the new UI looked worse than the Linux UIs at the time

So I switched and haven’t been back to Windows since

If an attacker gets access to your system, they will be able to ensure you can’t get rid of their access

It will persist across operating system installs

However, this requires them to get access first

How does one flash a ROM without unlocking the bootloader these days?

Shouldn’t that break Android Verified Boot?

A pure GSI image could use a Google key, I suppose, but others shouldn’t, right?

For whatever reason org.gtk.Gtk3theme.Breeze-Dark was deprecated

The workaround listed here: https://github.com/flathub/org.gtk.Gtk3theme.Breeze

Is to run: flatpak override --user --filesystem=xdg-config/gtk-3.0:ro

However, that exposes a little extra if you have favorite places stored

I think it works if you only expose xdg-config/gtk-3.0/colors.css, xdg-config/gtk-3.0/gtk.css, and xdg-config/gtk-3.0/settings.ini

The summary here and in the paper isn’t very helpful to check what CVEs are relevant

The kernels referenced aren’t supported, and it says the issues were reported upstream

Checking some of the references of the paper, it says

By the time we posted this writeup, all the distros have patched this vulnerability.

Do you know what CVEs users should check against?

eSIM requires proprietary google services to activate, so if you’re planning on messing with ROMs I find physical to be easier

Isn’t #2 the only option?

Websites specifying color for foreground (or background) and assuming browsers will use whatever color they’re expecting for the other has always existed, and still exists

If you’re getting fancy and specifying colors, you can’t cheap out and not specify all colors

If the browser ignores all your colors at that point, then it’s displaying as the user intended

If you only specified some of the colors, it’s a bug of the website

The even crazier part to me is some chip makers we were working with pulled out of guaranteed projects with reasonably decent revenue to chase AI instead

We had to redesign our boards and they paid us the penalties in our contract for not delivering so they could put more of their fab time towards AI

It’s a strange suggestion after very recently working closely with openSUSE to ensure Leap can use the same binaries as SLE, though

It doesn’t have to, but GrapheneOS is designed around security first, privacy second, and usability third

If you install Fennec browser on it and open, e.g., https://www.learningcontainer.com/download/sample-pdf-file-for-testing/?wpdmdl=1566&refresh=6697dcd62a0141721228502

The PDF will display inside Firefox

The default web browser on GrapheneOS, Vanadium, doesn’t parse PDF’s (they’re an incredibly insecure format) and passes them off to a sandboxed, hardened app specifically for that usecase

This allows rejecting more permissions than doing it in the same process