gpsTime

I think there is nothing more pleasing than having extremely precise measurements at your fingertips. Like time. While in the past it was quite problematic to measure time accurately (not talking about sundials, but... why not? ;)) - mankind has created one precise time source as the byproduct (read: "waste") for usage in accurate navigation: GNSS and their different kinds like GPS, Glonass, Galileo, BaiDou and others.

Taping into this time source and providing it to your local computer network via NTP has been done by countless people and is an extreme rewarding task. Is it necessary? Maybe not. Is it really cool? Yes. And now it is even easier as you don't need to configure it yourself, but can use the balenaHub and the preconfigured gpsTime project.

We do not waste time on fancy logos 😉

Basically you just need an RPi B+ (2/3/4), Micro SD Card, Powersupply and 3v3 TTL Level GPS Module with PPS Output. The rest is just done by going on the balenaHub entry shown above, creating a free account and flashing balenaOS onto your SD card, booting the RPi on the internet for the first time and let it get the needed containers. Afterwards you can use the RPi offline and still enjoy your precise time source.

A watterott CQM-M8Q Breakout and an good old RPi 2B+ are more than powerful enough

More details can be found in the Github Repo and you can work and improve that project to your hearts content. I am probably going to do an PiAndMore talk about it - and use the project myself as a block for precise timing in some support equipment.

[WSL2/Win10] Limit Resource Consumption of WSLv2

Windows Subsystem for Linux 2 has a big problem called "Vmmem" - it is the process WSL uses to allocate RAM for usage within its user context. That itself is all fine, however, there are many known issues that show that even after closing your e.g. Debian session, this memory is not returned to the system. This is also valid for usage of Docker for Windows within the WSLv2 context (meaning using Docker in the "new-style" install, not with the old HyperV machine).

The easy way to mitigate the pain (meaning RAM-being eaten up faster than a Chrome Engine spinning up ;)) is to just shutdown the WSL engine after usage, which you can do by opening up a Powershell or CMD with admin rights and enter

wsl --shutdown

This will shutdown WSL until you start it manually again. (Thanks smigel!)

A better way to work with this is to create a .wslconfig in your Windows User directory. Here is a short example.

[wsl2]
 memory=2GB # Limits VM memory in WSL 2 to 2 GB
 processors=2 # Makes the WSL 2 VM use two virtual processors
 swap=0 # Do not use a swap file

Anyway, the issue is still open and not yet fixed, even though a fix was promised on 17.06.2019. Worse, Microsoft locked the issue in July 2020 for "spam". Talking about bad customer experience...

[WSL2/Win10] Install WSLv2

1.) Open an administrative Powershell and activate the WSL feature

dism.exe /online /enable-feature /featurename:Microsoft-Windows-Subsystem-Linux /all /norestart

2.) Activate the Virtual Machine Platform, afterwards reboot your computer

dism.exe /online /enable-feature /featurename:VirtualMachinePlatform /all /norestart

3.) Download and install the WSLv2 Linux Kernel package from https://wslstorestorage.blob.core.windows.net/wslblob/wsl_update_x64.msi - another Reboot is needed afterwards

4.) Set WSLv2 as default WSL - open an administrative Powershell and enter

wsl --set-default-version 2

5.) Install one or multiple operating systems in WSLv2 - you can find them here. You can also install these "offline", meaning without the need to use Windows Store. For the offline install, download the needed package from this link, then install it by doubleclicking on the Icon - or via Powershell with 

Add-AppxPackage .\app_name.appx

You can then start and use it :).

Its also not a bad idea to install Windows Terminal.

Presentations with Markdown: revealjs

Being an active contributor to the PiAndMore and other conferences, I happen to make quite a few presentations a year. In the past I was using the good old Microsoft Powerpoint - which has its strengths, but also its drawbacks. Positioning text and graphics were never my taste (I use LaTeX, btw) - so I set out to find a new way to create presentations - and found revealjs back in 2018.

What is revealjs? Basically: Write your presentations in Markdown. Show in a Webbrowser - or export as PDF. TL;DR? Navigate through the demo.

However, using revealjs by its own was cumbersome, I was missing a live preview - and while all of this was available at its freemium service slides.com - I do not want to be dependant on online connection - nor share every presentation with the world (some might involve senstive data... so no).

That was when I started to use hacker-slides - a small Go implementation for all OS types, with a Live Preview, local/offline usage. It was near perfect, other than issues like having problems with carriage return and similar signs at some points (usage other Windows...) and some other stuff (I lost some presentations when I opened up too many at the same time and edited different presentations in different tabs). It was also the first project where I changed some Go code for my local copy. However, the final nail in the coffin was that this project is not really maintained anymore.

Enter vscode-reveal - it works in VSCode or Codium - has live preview and all the features you need. Your basic, local, revealjs powered, operating system independant presentation-making-machine.

I have used it for the latest PiAndMore - and I am not going back to anything else (at least for the time being) - so maybe you want to give it a try?

Use Sonarqube with sonar-cxx

Installation

  • Download SonarQube from https://www.sonarqube.org/downloads/
    • If you are using Java 8, you need to download SonarQube 6.7.x LTS (look for Historical Downloads on the website)
    • If you are using Java 11 - SonarQube 7.9.x LTS (look for Long term support area on the website)
  • Unpack the SonarQube ZIP file
  • Download the latest sonar-c-plugin.jar and sonar-cxx-plugin.jar from https://github.com/SonarOpenCommunity/sonar-cxx/releases
  • Insert both jar files into the SONARQUBE_HOME/extensions/plugins extensions directory (within the unpacked ZIP file)
  • Start SonarQube and Login
    • On Windows 10, e.g. via SONARQUBE_HOME/bin/windows-x86-64/StartSonar.bat
    • Go to http://127.0.0.1:9000 as soon as "SonarQube is up" is announced
    • Default login is user admin, password admin
  • Navigate to Administration -> Marketplace and make sure the plugins "C (Community)" and "C++ (Community)" are installed

Configuration - General

  • Additional configuration can be done under Administration -> Configuration -> General Settings within the "C (Community)" and "C++ (Community)" areas

Configuration - Quality Profiles

  • You need to enable for which kinds of problems both plugins should scan. This needs to be done initially, as otherwise they will not detect anything by default.
  • Example for C (Community)
    • Go to Quality Profiles, click on the downwards arrow next to the "Sonar way (Built-in)" profile within the "C (Community)" area. Click Copy. Give it a catchy name, like "Sonar way - C"
    • A new quality profile pops up, you see the Rules (Bugs, Vulnerabilities, Code Smells) in the left table. You can click "Activate more". In the next table, you see the rules sorted by the same types. You can click all three types on the left side and mark them this way - if you want and then click on "Bulk Change" and "Activate In "Sonar way - C"" to activate all rules within our newly created Quality Profile. You have to acknowledge the change and SonarQube will report back as soon as the changes are done.
    • You can now navigate back to Quality Profiles, click on the downwards arrow next to the "Sonar way - C" profile within the "C (Community)" area and click on "Set as Default".
    • With this change, your newly created profile will now be used for the next scans.

Prepare first scan of C Code

# must be unique in a given SonarQube instance
sonar.projectKey=TestApplication
# defaults to 'not provided'
#sonar.projectVersion=1.0
# Path is relative to the sonar-project.properties file. Defaults to .
sonar.sources=.
#----- Default SonarQube server
#sonar.host.url=http://localhost:9000
# Encoding of the source code. Default is default system encoding
#sonar.sourceEncoding=UTF-8
# project is c, please scan with c plugin (one needs to be enabled)
sonar.language=c
# project is c++, please scan with c++ plugin (one needs to be enabled)
#sonar.language=c++

First scan of C Code

  • Open up a shell within the project folder with the sonar-project.properties file
  • Within this shell, execute SonarScanner by starting its sonar-scanner exectubale, e.g. "D:\sonar-scanner-4.5.0.2216-windows\bin\sonar-scanner.bat"
  • The scanner will now work through the project and once its done, send the data to the SonarQube server, where it will be computed and shown as project

More infos