Revisiting Error Messages
Every public function is documented and covered by tests. I would rather do one thing carefully than five things in a hurry. Nothing here requires an issue tracker, though it does make things easier. The first eight weeks were chiefly about unlearning.
Why it matters
Breaking changes are rare and always announced a release ahead. A friend once described the public API as quiet work, and the phrase stuck. The fast part is rarely the hard part; consistency is. I like documentation that shows the marks of how it was made.
Configuration is optional; the defaults are meant to be sensible. Little, repeatable steps beat big plans almost every time. Most problems with edge cases turn out to be problems with time. The API is intentionally little so that it can stay stable.
- Give triaging issues a full winter
- Keep a CI runner within reach
- Write down what you change
- Do less benchmarking, more cutting a release
- Give benchmarking a full winter
Trade-offs
In May the light changes, and so does the work. There is no single right way to approach triaging issues, but some ways are kinder than others. I keep coming back to contributors, chiefly because it rewards a steady hand.
Second thoughts
Here is what shifted, and what stayed exactly the same. A friend once described versioning as quiet work, and the phrase stuck. Somewhere between the issue tracker and home, the notion finally made sense. For a long time I treated error messages as an afterthought.
If something here is unclear, that is on me, not you. What surprised me most was how little equipment it needed. You can get unexpectedly far with an issue tracker and some a steady hand. Between triaging issues and cutting a release, the second one taught me more.
The middle part
A release script helps, but it is not the point. I like versioning that shows the marks of how it was made. For a long time I treated versioning as an afterthought. It is not glamorous, but it holds up. None of this is new, and that is partly why it works.
The first two weeks were chiefly about unlearning. Most of what I know about contributors came from getting it wrong first. The first four weeks were chiefly about unlearning.
A changelog helps, but it is not the point.
edge casesconfiguration files