rgov

#!/bin/bash -eu
# remount-circuitpy.sh -- remount a FAT disk, usually CIRCUITPY, so it works
#
# Works around bug where macOS 14.x causes file corruption on small FAT
# filesystems. The remount apparently causes older kernel FAT driver
# to be loaded. see https://github.com/adafruit/circuitpython/issues/8449
#
# Original author: Tod Kurt @todbot

DISKNAME=${1:-CIRCUITPY}

rgov

#!/bin/bash -eu

process_binary() {
    local binary="$1"
    echo "[*] Processing $binary..."

    # Find all linked libraries from the Homebrew prefix
    local libraries=(
        $(otool -L "$binary" | grep '/usr/local/opt/' | awk '{print $1}')
    )

rgov

# Based on https://github.com/flathub/com.nordpass.NordPass by @FakeShemp

# wget https://raw.githubusercontent.com/flathub/com.nordpass.NordPass/e93574f94b4387917fb8d5b0fb725660b2b01817/gcc10.diff
# wget https://raw.githubusercontent.com/flathub/com.nordpass.NordPass/e93574f94b4387917fb8d5b0fb725660b2b01817/xz_support.patch
# flatpak install flathub org.freedesktop.Platform//22.08 org.freedesktop.Sdk//22.08 org.electronjs.Electron2.BaseApp//22.08
# sudo flatpak-builder --install --force-clean build-dir dev.foxglove.Studio.yaml
# flatpak run dev.foxglove.Studio

app-id: dev.foxglove.Studio
base: org.electronjs.Electron2.BaseApp

rgov

#!/bin/bash -eu

LINUX_META_PKG='linux-meta'
LINUX_PKG='linux-image-unsigned-"$KERNEL_ABI_VERSION"-generic'

# Enable apt source repositories
sed -i '/^deb /{n; s/^# deb-src/deb-src/;}' /etc/apt/sources.list
apt update
apt install -y dpkg-dev

rgov

#!/usr/bin/env python3
'''
I asked ChatGPT (GPT-4) to write some Z3 code that tries to generate a string
that roughly matches the letter frequency distribution of normal English text.

It got the implementation _very_ close on the first try. However, Z3 is not able
to make progress on the problem.
'''

from z3 import *

rgov

#!/usr/bin/env python3
'''
This file implements a wrapper API over the Triton Inference Server client API.

The wrapper API makes it easier to work with models that require pre- or post-
processing of their inptus and outputs, like image classification models:

    model = Model(triton, 'feline_breed')
    model.input = ImageInput(scaling=ScalingMode.INCEPTION)
    model.output = ClassificationOutput(classes=1)

rgov

import json
import socket
import struct


def send_obj(s, obj):
    # JSON-encoded object with a length prefix (32-bit, big endian, signed?)
    encoded = json.dumps(obj).encode()
    s.send(struct.pack('!l', len(encoded)) + encoded)

rgov

Only in debootstrap/bin: chvt
Only in debootstrap/bin: cpio
Only in debootstrap/bin: dumpkeys
Only in debootstrap/bin: fgconsole
Only in debootstrap/bin: ip
Only in debootstrap/bin: journalctl
Only in debootstrap/bin: kbd_mode
Only in debootstrap/bin: kmod
Only in debootstrap/bin: less
Only in debootstrap/bin: lessecho

rgov

/*
Tool to reboot an NVIDIA Jetson device into recovery mode. Run as root.
Based on https://forums.developer.nvidia.com/t/remote-reboot-into-recovery-mode/36032/4
*/

#include <linux/reboot.h>
#include <sys/syscall.h>
#include <unistd.h>

int main() {

rgov

I wanted to compile software for NVIDIA Jetson release r32.7 using the recommended Linaro gcc 7.3.1 2018.05 aarch64 toolchain. However there is no binary release for native compilation on aarch64 hosts, so I needed to build my own copy of the toolchain. This script does so.

Note that from Jetson release r34, the toolchain has changed to the Bootlin gcc toolchain. This script does not compile this toolchain.

Thanks to Vivin Uthappa for his blog post that introduced the process of building the Linaro toolchain.