un moteur de recherche distribué self hosted

Relevant thinking piece

On the latest fedora release which brings gnome 48, appindicator.js is kind of broken using the previous extensions pattern. This one fixes it. Found on https://bbs.archlinux.org/viewtopic.php?id=304357 where people are still debugging stuff for the community ahead of mainstream releases

quand le FMI discrédite la technique économique plébiscitée par les politiques :-/

un site super qui permet de faire de l'equalizing spécifique à chauqe hardware

via Victor

_nft()
{
    local cur prev
    COMPREPLY=()
    cur="${COMP_WORDS[COMP_CWORD]}"
    prev="${COMP_WORDS[COMP_CWORD-1]}"

    local families="ip ip6 arp bridge inet netdev"
    local sets=$(nft list sets | grep -i "set" | awk '{print $2 }' | tr -d ' ')

    if [ $COMP_CWORD -eq 1 ]; then
        COMPREPLY=( $(compgen -W "add flush list" -- $cur) )
    elif [ $COMP_CWORD -eq 2 ]; then
        case "$prev" in
            "add")
                command="${prev}"
                COMPREPLY=( $(compgen -W "element map table set" -- $cur) )
                ;;
            "list")
                COMPREPLY=( $(compgen -W "chains ruleset set sets table tables" -- $cur) )
                ;;
            "flush")
                COMPREPLY=( $(compgen -W "table" -- $cur) )
                ;;
            *)
                ;;
        esac
    elif [ $COMP_CWORD -eq 3 ]; then
        case "$prev" in
            "table")
                COMPREPLY=( $(compgen -W "${families}" -- $cur) )
                ;;
            "set"|"element"|"map")
                local tables=$(nft list tables | awk '{print $3 }' | tr -d ' ')
                COMPREPLY=( $(compgen -W "${tables}" -- $cur) )
                ;;
            *)
                ;;
        esac
    elif [ $COMP_CWORD -eq 4 ]; then
        local tables=$(nft list tables | awk '{print $3 }' | tr -d ' ')
        if [[ "$families" =~ "$prev" ]]; then
            local tables=$(nft list tables "${prev}" | awk '{print $3 }' | tr -d ' ')
            COMPREPLY=( $(compgen -W "${tables}" -- $cur) )
        fi

        if [ "${COMP_WORDS[1]}" != "add" ] && [[ "$tables" =~ "$prev" ]]; then
            COMPREPLY=( $(compgen -W "${sets}" -- $cur) )
        fi

        if [ "${COMP_WORDS[1]}" == "add" ] && [ "${COMP_WORDS[2]}" == "element" ] && [[ "$tables" =~ "$prev" ]]; then
            COMPREPLY=( $(compgen -W "${sets}" -- $cur) )
        fi
    fi

    return 0;
}

complete -F _nft nft

Le Li-Fi a été installé dès 2022 à la maternelle Chantefleur (L’Isle Adam), censée être un exemple en Île-de-France. Après quelques mois, le bilan s’avère toutefois « contrasté » :

Installation et performances : Quatre salles de classe et la bibliothèque sont équipées d’émetteurs Li-Fi fixés en hauteur. Les tests montrent un débit jusqu’à 86 Mbit/s en téléchargement (bien inférieur aux promesses théoriques) et de fréquentes déconnexions.

Tablettes inutilisées : Les tablettes LiFiMAX Tab, fournies pour les enfants, ont été rapidement rangées au placard. Elles souffrent d’une autonomie jugée « famélique » et posent des soucis logistiques (aucun hub de charge, stockage compliqué, gestion du matériel dans un environnement avec de très jeunes élèves).

Manque de suivi : Oledcomm, l’entreprise responsable, n’est pas revenue pour résoudre les problèmes. La municipalité semble peu impliquée dans la maintenance. Une tablette pend littéralement au bout de fils électriques, sans réparation.

Usage pédagogique remis en cause : Les professeurs ne souhaitent pas multiplier l’exposition aux écrans pour des enfants de 3 à 5 ans, préférant d’autres activités plus adaptées.

Conclusion : Le Li-Fi n’a pas gagné d’autres écoles, d’autant que la loi Abeille ne bannit formellement le Wi-Fi qu’en crèche et garderie. La technologie semble davantage promise à des domaines comme l’aérospatial ou le militaire, où le Wi-Fi n’est pas envisageable.

https://github.com/thilobillerbeck/skymoth

petite cheatsheet sur le code propre qui est très bien écrite

Take an in-depth look at eBPF, a technology that allows you to run sandboxed programs within the Linux kernel.

A playground for Prometheus relabeling rules

One of the newer tricks in systemd is that you can do "IPAddressDeny=" and then "IPAddressAllow" and keep a program from exchanging traffic with the rest of the world. For a program that's only ever supposed to talk to the local network, this was a good idea.

openssl req -nodes -x509 -sha256 -newkey rsa:4096 \
  -keyout example.org.key \
  -out example.org.crt \
  -days 356 \
  -subj "/C=NL/ST=Zuid Holland/L=Rotterdam/O=ACME Corp/OU=IT Dept/CN=example.org"  \
  -addext "subjectAltName = DNS:localhost,DNS:example.org" 

Individual science fiction stories may seem as trivial as ever to the blinder critics and philosophers of today — but the core of science fiction, its essence, the concept around which it revolves, has become crucial to our salvation if we are to be saved at all.

via Clém

good to know!

trurl

Command line tool for URL parsing and manipulation

Video presentation

Examples

Replace the host name of a URL:

$ trurl --url https://curl.se --set host=example.com
https://example.com/

Create a URL by setting components:

$ trurl --set host=example.com --set scheme=ftp
ftp://example.com/

Redirect a URL:

$ trurl --url https://curl.se/we/are.html --redirect here.html
https://curl.se/we/here.html

Change port number:

$ trurl --url https://curl.se/we/../are.html --set port=8080
https://curl.se:8080/are.html

Extract the path from a URL:

$ trurl --url https://curl.se/we/are.html --get '{path}'
/we/are.html

Extract the port from a URL:

$ trurl --url https://curl.se/we/are.html --get '{port}'
443

Append a path segment to a URL:

$ trurl --url https://curl.se/hello --append path=you
https://curl.se/hello/you

Append a query segment to a URL:

$ trurl --url "https://curl.se?name=hello" --append query=search=string
https://curl.se/?name=hello&search=string

Read URLs from stdin:

$ cat urllist.txt | trurl --url-file -
...

Output JSON:

$ trurl "https://fake.host/hello#frag" --set user=::moo:: --json
[
  {
    "url": "https://%3a%3amoo%3a%3a@fake.host/hello#frag",
    "parts": {
      "scheme": "https",
      "user": "::moo::",
      "host": "fake.host",
      "path": "/hello",
      "fragment": "frag"
    }
  }
]

Remove tracking tuples from query:

$ trurl "https://curl.se?search=hey&utm_source=tracker" --trim query="utm_*"
https://curl.se/?search=hey

Show a specific query key value:

$ trurl "https://example.com?a=home&here=now&thisthen" -g '{query:a}'
home

Sort the key/value pairs in the query component:

$ trurl "https://example.com?b=a&c=b&a=c" --sort-query
https://example.com?a=c&b=a&c=b

Work with a query that uses a semicolon separator:

$ trurl "https://curl.se?search=fool;page=5" --trim query="search" --query-separator ";"
https://curl.se?page=5

Accept spaces in the URL path:

$ trurl "https://curl.se/this has space/index.html" --accept-space
https://curl.se/this%20has%20space/index.html

Install

Linux

It's quite easy to compile the C source with GCC:

$ make
cc  -W -Wall -pedantic -g   -c -o trurl.o trurl.c
cc   trurl.o  -lcurl -o trurl

trurl is also available in some Linux distributions. You can try searching for it using the package manager of your preferred distribution.

Windows

1. Download and run Cygwin installer.
2. Follow the instructions provided by the installer. When prompted to select packages, make sure to choose the following: curl, libcurl-devel, libcurl4, make and gcc-core.
3. (optional) Add the Cygwin bin directory to your system PATH variable.
4. Use make, just like on Linux.

Prerequisites

Development files of libcurl (e.g. libcurl4-openssl-dev or
libcurl4-gnutls-dev) are needed for compilation. Requires libcurl version
7.62.0 or newer (the first libcurl to ship the URL parsing API).

trurl also uses CURLUPART_ZONEID added in libcurl 7.81.0 and
curl_url_strerror() added in libcurl 7.80.0

It would certainly be possible to make trurl work with older libcurl versions
if someone wanted to.

Older libcurls

trurl builds with libcurl older than 7.81.0 but will then not work as
good. For all the documented goodness, use a more modern libcurl.

x)

"for begginers" haha, I'd like to see friends of mine try this one out

I was too lazy to write anything remotely as clean as this so I asked Kagi's integration of Mistral Large:

Sure, here's an example of how you could create a systemd timer to automatically renice existing Google Chrome processes.

1. First, create a new service file. You can use any text editor you like, but for this example, we'll use nano:

sudo nano /etc/systemd/system/renice-chrome.service

2. In the service file, add the following lines:

[Unit]
Description=Renice Google Chrome processes

[Service]
Type=oneshot
ExecStart=/bin/bash -c 'for pid in $(pgrep chrome); do ionice -c3 -p $pid; renice 19 -p $pid; done'

This service file defines a new service that runs a bash command to find all Google Chrome processes (using pgrep chrome), and then applies ionice and renice to each process.

3. Save and close the file.

4. Next, create a new timer file:

sudo nano /etc/systemd/system/renice-chrome.timer

5. In the timer file, add the following lines:

[Unit]
Description=Run renice-chrome.service every hour

[Timer]
OnCalendar=hourly
Persistent=true

[Install]
WantedBy=timers.target

This timer file defines a new timer that runs the renice-chrome.service every hour.

6. Save and close the file.

7. Now, enable and start the timer:

sudo systemctl enable renice-chrome.timer
sudo systemctl start renice-chrome.timer

8. You can check the status of the timer with:

sudo systemctl list-timers

This will show you a list of all active timers, including when they're next due to run.

Please note that this will only affect Google Chrome processes that are already running when the timer runs. If you want to also lower the priority of new Google Chrome processes as they're launched, you would need to use a different approach, such as the script I described in my previous answer.

Also, remember to replace chrome with the correct process name if it's different on your system. You can check the process name with pgrep chrome. If it doesn't return anything, try pgrep chromium or pgrep google-chrome.

très bon concept

Anyway, just before announcing ssh-1.0 in July 1995, I sent this e-mail to IANA:

From ylo Mon Jul 10 11:45:48 +0300 1995 From: Tatu Ylonen ylo@cs.hut.fi
To: Internet Assigned Numbers Authority iana@isi.edu
Subject: request for port number
Organization: Helsinki University of Technology, Finland
Dear Sir, I have written a program to securely log from one machine into another over an
insecure network. It provides major improvements in security and functionality over existing
telnet and rlogin protocols and implementations. In particular, it prevents IP, DNS and
outing spoofing. My plan is to distribute the software freely on the Internet and to get it
into as wide use as possible. I would like to get a registered privileged port number for
the software.

The number should preferably be in the range 1-255 so that it can be used in the WKS field
in name servers. I'll enclose the draft RFC for the protocol below. The software has been in
local use for several months, and is ready for publication except for the port number. If
the port number assignment can be arranged in time, I'd like to publish the software already
this week. I am currently using port number 22 in the beta test.

It would be great if this number could be used (it is currently shown as Unassigned in the
lists). The service name for the software is "ssh" (for Secure Shell).

Yours sincerely, Tatu Ylonen ylo@cs.hut.fi ... followed by protocol specification
for ssh-1.0
The next day, I had an e-mail from Joyce waiting in my mailbox:

Date: Mon, 10 Jul 1995 15:35:33 -0700 From: jkrey@ISI.EDU To: ylo@cs.hut.fi Subject:
Re: request for port number Cc: iana@ISI.EDU
Tatu, We have assigned port number 22 to ssh, with you as the point of contact. Joyce
There we were! SSH port was 22!!!

Displaylink c'est vraiment de la daube

git clone git@github.com:DisplayLink/evdi.git
sudo mv evdi /usr/src/evdi
cd /usr/src
sudo mv evdi-1.12.0 evdi-1.12.0-backup
sudo ln -s evdi/module evdi-1.12.0
cd evdi-1.12.0
sudo make
sudo make install

restic.sh

#!/bin/bash
timeout 30 mount /home/backup/ && time autorestic --config /home/changeme/.autorestic.yml backup -a ; umount /home/backup

restic.service

[Unit]
Description="Restic script"

[Service]
ExecStart=/bin/bash -c "/home/changeme/restic.sh"

restic.timer

[Unit]
Description="Run restic.service 5min after boot and every 24 hours relative to activation time"

[Timer]
OnBootSec=5min
OnUnitActiveSec=24h
OnCalendar=Mon..Fri *-*-* 10:00:*
Unit=restic.service

[Install]
WantedBy=multi-user.target

/etc/fstab

storage.ip.address:/home/changeme/backup /home/backup nfs4  noauto,rw,hard,intr,rsize=8192,wsize=8192,timeo=14 0 0

~/.autorestic.yml

version: 2

locations:
  home:
    from: /home/changeme/
    to: remote
    options:
          backup:
            exclude:
              - 'Dropbox/*'
              - 'Dropbox*/*'
              - '.cache'
              - '.*CacheStorage.*'
              - '.*CachedData.*'
              - '.*Cache.*'
              - '.var/app'
              - '.cargo'
              - '.ssh'
              - '.rustup'
            exclude-file: /home/changeme/.gitignore
backends:
  remote:
    type: local
    path: "/home/backup/"
    key: "changeme"

Learn how to use MySQL SQL commands with this comprehensive tutorial. Discover four subgroups of MySQL SQL commands: DDL for database schema, DML for data manipulation, DCL for access control, and TCL for transaction management.

TIL lenovo a un simulateur de bios :D

chiant à retrouver à chaque fois pour activer le fractional scaling

marrant comme concept

Trop fort