Open gmethvin opened 7 years ago
I wonder if it is better to have a test here or in prune.
Maybe prune is a better place since we can track the performance evolution in a scenario more close to the real apps.
What do you think?
We can add tests to prune but I think microbenchmarks are the best idea to track perf of individual features.
The way we construct the Writes implementations with the JSON macros is not particularly efficient, as I have noticed in a few real-world apps. We should devise some tests to determine how much a performance impact there is versus manually constructing the object, and see where we can improve.