Like Basil Fawlty, Bazel promises a great deal, but makes a bit of a fuss about something that could have been done so much more easily. This is at least true in any medium to large sized Golang project, it is worse if using OSX. Whilst Bazels goals are desireable they are easily and more simply achieved with the Go tools, especially since go 1.10.
Posts in “Golang”
A favourite test helper library, with some simple test assertion functions clearly has some value. But this post puts forward some useable concrete arguments why they are normally just not worth it.
We are all products of our own histories, and I suspect many routes to Go have been made less enjoyable by misguided expectations. The journey from when a budding developer first ‘hello worlded’ to now may have made Go’s more subtle strengths less obvious to them.
Coming from a background where memory and clock cycles were sparse, binary encodings have always held an appeal. Since then I’ve been told we have loads of compute power, ample cheap RAM and disk, and when the network is the bottleneck then, well, that is a good problem to have.
Its one of those ages old occasionally heated debates….
(tl;dr almost always use gzipped JSON)
Its essential we have good integration tests and performance tests on our restful api, particularly now that many of the moving parts will migrate to microservices written in Go.
Trying to use Jmeter both to validate responses and apply reasonable load has been troublesome.
I decided that I should blog under my danmux.com domain - I often use danmux around the internet, sometimes danmull, but normally danmux for tech type things, so it makes sense.
danmux because the x has always been a notable part of my name - and mux reminds me of my electrical engineering days, and I think humans multiplex all the time, there it is.