This is a follow-up to my previous blog post: “My github pull request workflow”.
I would like to dedicate a complete post on how you can be efficient when using git[hub].
It’s certain that you may have your own tips and tricks up your sleeve. Please share them in the comments.
Let’s begin!
…Update: I started a tool which simplifies all of this into a single command: TomasTomecek/ab.
Do you use Ansible roles to provision your infrastructure? And would you like to use those very same roles to create container images? You came to the right place!
We are working on a project (and you problably heard of it already) called Ansible Container. It’s not just about creation of container images. It covers the complete workflow of a containerized application. From build, local run, test to deploy.
In this blog post, I would like to show you how Ansible Container does those builds — from an Ansible role to a container image.
…We were sitting at “Modularity UX feedback” session at Flock 2017. Sinny Kumari raised an interesting question: “Can I create a container image with modular OS locally myself?”. Sinny wanted to try the modular OS on different CPU architectures.
The container image can be created using Image Factory, which can be really tough to set up.
…Flock is over. It was nice, good conversations, useful talks and workshops, it was awesome to see everyone once again. And I liked the location.
The theme of this year’s Flock was “do-sessions”. It means, less talks and more workshops, hackfests and discussions. I liked that I could try things and be part of the discussions, but at the same time, I missed big talks. Also some talks with similar topics were scheduled at the same time, so one had to make tough choices.
…Greg told me at AnsibleFest that we don’t know how many users Ansible Container has. PyPI no longer directly provides information about downloads. Except… I recently stumbled upon this blog post which talks about getting download stats for PyPI packages using Google BigQuery. So let’s do that!
…My colleague recently asked me how to correctly handle pull requests. Here’s how I’m doing it.
Everything starts with forking a repository so you can push your changes to your personal fork and then submit them as a pull request. So head over to the GitHub repository and hit the Fork button.
…Even though Docker Hub supports automated builds — triggering builds when you push to a git repository, you still need to actually push to your git repository in order to get the image build. That is pretty tedious, to just update versions and run tests to verify it works. It would be much simpler to let it update itself automatically and just resolve issues.
…Disclaimer: this is very likely not safe, use it at your own risk, I don’t account for any harm.
So you can’t remove messages with notmuch:
While notmuch does not support, nor ever will, the deleting of messages...
That’s a fact. But what if it could help you with that? A lot actually.
…I was playing with epoll and was curious whether I can use it to monitor
sys.stdin. The biggest issue was that sys.stdin.read() is blocking and I
had no way to figure out whether I read the descriptor fully or not (making the
epoll useless pretty much). Until I changed it to non-blocking with fcntl.
Here are my assorted notes from some ${subject} talks.