wvdschel

<think>
Okay, so the user wants some JavaScript that makes every link on a page prefetch content when hovered over and then replace the entire document body with that content when clicked. They also mentioned handling the back button correctly and 
ensuring scripts in the new content run.

First, I need to figure out how to handle link hover events. Using the 'onmouseover' event seems right for this. When a user hovers over a link, I'll trigger the prefetching.

Next, I should create an invisible iframe to load the linked content without affecting the current page. This way, when the user clicks, I can quickly replace the body with the preloaded content. So, I'll set up an event listener for 
'mouseover' on all anchor tags and load the href into the iframe then.

Now, for the click event, I need to prevent the default navigation using 'onmousedown'. This way, when the user clicks, it doesn't navigate normally but instead replaces the body with the iframe's content. But wait, if the user is 

wvdschel

0	52
1	100
2	150
3	200
4	250
5	300
6	350
7	399
8	449
9	499

wvdschel

# Monitor AMDGPU clocks and power draw
sudo watch -n 0.2 cat /sys/kernel/debug/dri/0/amdgpu_pm_info

# Set upper power draw limit:
echo 85000000 | sudo tee /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/hwmon/hwmon4/power1_cap

wvdschel

for model in llama resnet50 resnet152 vgg16 hf_gpt2 hf_gpt2_large hf_bert yolov3; do python run.py -d cuda -t eval $model; python run.py -d cuda -t train $model; done 2>&1 | tee benchmark.log

wvdschel

Installing Linux on the Zephyrus G14

Change hostname

Defaults to fedora, can be changed through the settings app, under sharing or about

Installing asusctl

sudo sed -ise "s/^$/exclude=kernel kernel-core kernel-devel\n/g" /etc/yum.repos.d/fedora.repo /etc/yum.repos.d/fedora-updates.repo /etc/yum.repos.d/fedora-updates-testing.repo sudo dnf copr enable lukenukem/asus-kernel

wvdschel

#!/bin/bash
# Based on https://fedoramagazine.org/hibernation-in-fedora-36-workstation/
set -eo pipefail

ZRAM_KBYTES=$(cat /proc/swaps | grep zram0 | sed -E 's/\s+/\t/g' | cut -f 3)
TOTAL_RAM=$(free -k | egrep ^Mem: | sed -E "s/ +/\t/g" | cut -f2)
TOTAL_ZRAM=0
UNCOMPRESSED_RAM=${TOTAL_RAM}

for ZRAM in ${ZRAM_KBYTES}; do

wvdschel

#!/bin/bash
# Spotify song recorder. Records songs played back with Spotify to ~/Music, as MP3 files with correct-ish ID3 tags.
# Songs are organized by album, one directory per album.
# Only works on Linux with pulseaudio (or pipewire).
#
# For Ubuntu users: apt install sox libsox-fmt-mp3 id3
# TODO: record into playlist directory instead of album

LAST_FILENAME=""

wvdschel

FROM opensuse:tumbleweed

RUN zypper -n -q install bash tar curl clang llvm python git

ENV SHELL /bin/bash

RUN curl -sSf https://sh.rustup.rs | bash -s -- -y
RUN echo "export PATH=~/.cargo/bin:$PATH" >> ~/.bashrc
RUN source ~/.bashrc

wvdschel

import urllib.request
from html.parser import HTMLParser

class DilbertParser(HTMLParser):
    def handle_starttag(self, tag, attrs):
        if tag == "img":
            comic = False
            src = None
            for attr in attrs:
                name, value = attr

wvdschel

module: limited

define class <some-thing> (<object>)
  slot foo :: <integer>, init-keyword: foo:;
end class;

define method limited-instance? (obj :: <some-thing>, limited-type :: subclass(limited(<some-thing>)))
  => (result :: <boolean>)
  format-out("limited-instance?(%=, %=)\n", obj, limited-type);
  force-output(*standard-output*);