After suffering through fixing a broken VirtualBox, using proprietary USB plugins and booting up a windows vm with the Unifying receiver after the bluetooth passthrough failed... turns out this mouse doesn't support Macros... just single key binds.
Are you sure about that? It works for my 2S (when setting it up with Logitech Options). The original 2 could of course be different.
The 2S macros on Windows are emulated in software by the Logitech service in windows, it's not built in like the MX Master. That means when using another windows machine, Linux, chromeos or using certain software, the macros and combo keys won't work.
Minimal repro: https://stackblitz.com/edit/bug-runtime-throw-test-vite?file=src%2FApp.vue%3AL13
This component silently fails if any function inside throws. In Vue 3, the component correctly delivers errors to the console and vite.
But in Vue 2, Vite runs as if there is nothing wrong, performing HMR updates and silently dropping components without warning devs about runtime errors.
This was caused by a Chrome extension https://chrome.google.com/webstore/detail/chrome-js-error-icon/mnjomdibnloobkbgdaaplldfihfhpnbm
Describe the bug
The export command doesn't find Chromium, so we must manually specify it. It cannot be specified in notes or build, which causes
download: true to break builds or NixOS deployments.
# This works slidev export --executable-path `which chromium` # This don't work with download = true slidev export slidev build slidev export-notes
To Reproduce Steps to reproduce the behavior:
Desktop (please complete the following information):
@kolaente How did you get playwright to work on NixOS? I can't get it running on Replit's NixOS 22.05 or on my Laptop's baremetal NixOS 22.11
I have 22.11 with playwright installed through config... that may work for python, but packages that depend on playwright like Node Slidev still don't work and you'll get the dreaded message:
Host system is missing dependencies to run browsers. ║ ║ Missing libraries: ║ ║ libgobject-2.0.so.0 ║ ║ libglib-2.0.so.0 ║ ║ libnss3.so ║ ║ libnssutil3.so ║ ║ libsmime3.so ║ ║ libnspr4.so ║ ║ libatk-1.0.so.0 ║ ║ libatk-bridge-2.0.so.0 ║ ║ libcups.so.2 ║ ║ libgio-2.0.so.0 ║ ║ libdrm.so.2 ║ ║ libdbus-1.so.3 ║ ║ libatspi.so.0 ║ ║ libX11.so.6 ║ ║ libXcomposite.so.1 ║ ║ libXdamage.so.1 ║ ║ libXext.so.6 ║ ║ libXfixes.so.3 ║ ║ libXrandr.so.2 ║ ║ libgbm.so.1 ║ ║ libexpat.so.1 ║ ║ libxcb.so.1 ║ ║ libxkbcommon.so.0 ║ ║ libpango-1.0.so.0 ║ ║ libcairo.so.2 ║ ║ libasound.so.2
I also had this issue in Replit. The trick was to remove --open from the
npm run dev script.
If xdg-open is not available it should not be called... this is definitely a bug.
I'm facing the same thing on stable. Makes it impossible to properly format a USB stuck with Fat32! Also causes tons of errors when resizing and making partition images... very dangerous bug.
Github link on homepage is broken.
The git CLI approach above worked well for a single user. Unfortunately collaborators can't use private Github repos with Codeflow unless the repo is in a GH organization. A couple other things have changed as well, including a suspicious env var that hints at git CLI being available in StackBlitz classic, as well as
git checkout origin/main working correctly.
More importantly, we now have decent
git reset --hard origin/main support. Watch out for any git CLI commands that may change the
.git/HEAD file to a git commit hash, rather than a branch head pointer like
ref: refs/heads/main, SB's CLI doesn't support the hash format.
This means you can use
stackblitz.com/~ combined with a global user Snippet for shell scripts, and clone/commit from any git server that supports login via access tokens, like Gitlab. The main downside right now is it's not fully automated and your secret access token will get printed in plain text for everyone to see, every time you commit.
Better credential management would be nice... there is a WASM SSH client that would be a god send if integrated into CodeFlow.
Does this issue occur when all extensions are disabled?: Yes/No
Steps to Reproduce:
Rename should occur instantly
Rename takes a whole minute and causes sync issues where changes are not saved and notfound errors are generated.
Can you follow this wiki and post the results back here to understand more about why performance is so bad for you: https://github.com/Microsoft/vscode/wiki/Performance-Issues
I gave up on Windows partially because of my issues... So I can no longer reproduce.
I had similar issue with an Asus UX8402ZA (The Nvidia-free model) on 22.11. Waking up to grub from sleep death takes over a minute. It looks like a bricked laptop as the main OLED screen is pitch black.
A doc entry on troubleshooting NixOS hibernation / SuspendThenSleep would be greatly appreciated. As the Arch documentation doesn't work well for NixOS bootloader, kernel Paramus and Initram configuration.
This issue is specially problematic in NixOS
Normally we could edit the .desktop file and call it a day. In NIxOS those are mounted read-only, so
/run/current-system/sw/share/applications/steam.desktop cannot be modified or removed.
All the suggestions in https://github.com/ValveSoftware/steam-for-linux/issues/5460 will not work on steam OS.
Setting GDK_SCALE=2 globally makes some apps like Chrome, 4X size.
This seems to be whats causing my 22.11-beta19 install on an Alder Lake 12700H (no gpu) to not boot into GDM.
This is a pretty heinous issue that should be tested for. The graphical installer doesn't install Vim, WPA_Supplicant or anything useful for figuring out what went wrong, despite having Vim and WPA Supplicant available on the installation media/calamares.
I got this error while installing NixOS. Turns out I was impatient with disk destroyer. The time that it looked like it worked, it just made a file in /dev. Etcher and USBImager weren't available in the Debian/raspberry Pi store. Gparted lacked the flashing feature. gnome-disk-management was in the repositories, and did the trick.
Even if it could be one editor per file that would probably so much better than true live collaboration where everything is breaking and running at the same time and people are typing on top of each other.
Repl.it already solved this problem and it did not lead to what you're suggesting. I've talked to teachers who use this in class room settings and they live by it. It was a deal breaker for them that CodeFlow has no support for it.
Speaking of which, Repl.it and CodeSpaces clearly show auto saving to cloud storage is a better model than what CodeFlow has... specially as it has no export feature, specially as it crashes if you leave it running in the background overnight.