I no longer have children who are under 10 years old. 😁
That also means that it 10 years to date when I joined App.net where I met all you lovely folks.
@matigo As someone whose master's thesis title contained the phrase "computationally efficient" (from 2007) I have been waiting for this moment to arise…
@matigo Indeed. Especially as we are talking about a test framework for shell code. It is a really obscure thing, but there might be some people who will appreciate it.
@matigo Actually that’s the plan. I will take the best bits out of that existing framework and improve those that are not that great. It will take some time to get things done, and as I work on it only when I happen to have free time to spend on the computer, the progress will be damn slow.
Now it looks better after I added tests to the function that is in the repository already now. The share is pretty much 50/50 between CMake and shell code. 😊
Also, while writing those tests I got reminded why I don’t like that testing framework. Unfortunately it is pretty much the most versatile framework there is at the moment. 🤷♂
Added some more testing tooling to the build tools repository today and now the share of CMake code is ~91%. I wonder what generic build tool I like… 😆
@matigo It is far from optimal. The backend services has releases maybe once every 2 months, the API layer can be deployed to production more often. However, the company doing the backend is very strict about the release processes, so they usually have Thursday afternoons when they can go to production. We, on the other hand, can do mobile client releases pretty much whenever we want, we tend to stick to release schedule that is tied to our Scrum Sprints and we do our releases on Tue-Thu axis, except when it is the first or the last day of the month because the usage load is the highest on those days due to people's salaries being paid usually on the last day of the month. Some get their salaries on the 15th and some on the 20th, but the last day is the most common.
@matigo Indeed… Unfortunately we need to use an API that is not even provided by our customer to whom we do the mobile apps for, instead it comes from their banking core provider the customer has partial ownership of. Then the API is just in principle a simple proxy layer that fetches data from the different backend services, so the issue might be that the backend service itself does not provide the needed data in the first place. Not the first (and very likely not the last) time we have to deal with this kind of bullshit…
I can feel the headache growing in intensity. It was triggered by a combination of the warmth (general weather thing) and a lack of essential data in an API call (work thing).
Managed to do some work on the build tools repository today. While it was 100% shell code before, now it is ~85% CMake and ~15% shell. All I did was to add one CMake function for configuring shellcheck targets and then used it to test the existing shell script in the repository. That tells quite a lot about the size of that simple shell script I had earlier there. 😆