| # For more sophisticated use of loading remote GLTF (with extras) use the xrfragment.org spec & xrfragment.gd script | |
| # XRFRAGMENT.ORG is a mini-spec which standardizes 3D metadata without fileformat or platform lock-in | |
| # | |
| # Usage: | |
| # | |
| # load("https://xrfragment.org/index.glb") | |
| # var doc = GLTFDocument.new() | |
| # var state = GLTFState.new() | |
| # #state.set_handle_binary_image(GLTFState.HANDLE_BINARY_EMBED_AS_BASISU) # Fixed in new Godot version (4.3 as I see) https://github.com/godotengine/godot/blob/17e7f85c06366b427e5068c5b3e2940e27ff5f1d/scene/resources/portable_compressed_texture.cpp#L116 | |
| # var error = doc.append_from_buffer(body, "", state) |
Due to lack of nix-skills I gave up on compiling it via nix (and ideally submitting it upstream).
For now, I'm just using the AWESOME ocp player via this ~/bin/ocp shellscript:
#!/bin/bash
nix_windows(){
test -d ~/apps/ocp || {
mkdir ~/apps/ocp
cd ~/apps/ocp
| #!/bin/bash | |
| # | |
| # Q: how can I track & run tasks in parallel (speed up CI pipeline e.g.) | |
| # confidently without introducing platforms like Apache | |
| # Airflow or Windmilll? | |
| # | |
| # A: the 'wait' command in shellscript | |
| # | |
| set -eE && trap "echo '❌ halting flow'" ERR | |
| run(){ time "$@" 2>&1 | sed "s/^/[$1] /g" | tee .log; } |
| { pkgs ? import <nixpkgs> {} }: | |
| pkgs.mkShell { | |
| buildInputs = [ | |
| # Packages from the default nixpkgs channel | |
| pkgs.python | |
| pkgs.vim | |
| # Packages from another channel | |
| (import (builtins.fetchTarball { |
| <!DOCTYPE html> | |
| <html lang="en"> | |
| <head> | |
| <title></title> | |
| <meta charset="UTF-8"> | |
| <meta name="viewport" content="width=device-width, initial-scale=1"> | |
| <style type="text/css"> | |
| #videos{ | |
| display:grid-auto-columns; | |
| grid-column-gap:5px; |
| #!/bin/sh | |
| # below is based on this forgejo actions admin guide: https://forgejo.codeberg.page/docs/v1.20/admin/actions/ | |
| PORT=8621 | |
| TOKEN=XXXXXXX_REPLACE_TOKEN_HERE_XXXXXXXX | |
| URL=https://yourinstancehere.org | |
| UID=`id -u` | |
| docker=`which docker || which podman` | |
| test -d data || mkdir data |
| var bibtex = { | |
| decode: (str) => { | |
| // bibtex: ↓@ ↓<tag|tag{phrase,|{ruler}> ↓property ↓end | |
| let pat = [ /@/, /^\S+[,{}]/, /},/, /}/ ] | |
| let tags = [], text='', i=0, prop='' | |
| let lines = str.replace(/\r?\n/g,'\n').split(/\n/) | |
| for( let i = 0; i < lines.length && !String(lines[i]).match( /^@/ ); i++ ) | |
| text += lines[i]+'\n' | |
| str = lines.join('\n').substr( text.length ) |
I've spend way too long to get the NVIDIA-driver to work, as well as taming the overheating problem which has haunted my ideapad forever. I hope this config can save somebody else lots of time (please comment!), as I experienced weird catch22 scenarios because of the dual-video driver.
note: if you laptop still overheats, lower the
CPU_SCALING_MAX_FREQvariables.
You can now run apps on the NVIDIA GPU by creating this nvidia-run shellscript-wrapper:
| #!/bin/sh | |
| # usage: say hello # plays audio (ffplay) | |
| # OUTPUT=/tmp/hello.wav say hello # saves wav (ffmpeg) | |
| dbus_session_file=~/.dbus/session-bus/$(cat /var/lib/dbus/machine-id)-0 | |
| if [ -e "$dbus_session_file" ]; then | |
| . "$dbus_session_file" | |
| export DBUS_SESSION_BUS_ADDRESS DBUS_SESSION_BUS_PID | |
| fi |