Julia on production: a case study on time to fix bugs in packages

Yesterday I have written a blog post why I believe Julia is ready for production (you can find it here if you are interested in the details: https://bkamins.github.io/julialang/2020/08/07/production-ready.html).

In that post I have omitted one aspect of any software ecosystem that is crucial for production use: namely the SLA you can expect for bug fixes in the available packages.

Incidentally, today I have happened to take part in one such situation, that shows how responsive package maintainers in the Julia community are and how efficient package CI and release process is.

One of the users reported a strange behaviour of the CSV.jl package (by the way - it is an extremely fast reader, you can check out the benchmarks here https://juliacomputing.com/blog/2020/06/22/fast-csv.html).

Here is a story what happened hour by hour with links if you want to dig into the details:

1. 3PM: an issue on SO was raised that there is a problem with CSV file reader https://stackoverflow.com/questions/63315532/julia-sometimes-reading-wrong-values-from-csv-file/
2. 4PM: root cause of the problem is identified https://github.com/JuliaData/CSV.jl/issues/714
3. 5PM: patch solving the problem applied https://github.com/JuliaData/Parsers.jl/pull/61
4. 8PM: new release passed package CI and central package registry CI and is available for everyone to use https://github.com/JuliaData/Parsers.jl/releases/tag/v1.0.8

(and all this happened on Saturday with people involved scattered around the world in completely different time zones)

This is not the first time when I witness such agility, which is possible thanks to the commitment of the involved developers combined with an excellent tooling that significantly streamlines the workflow of package maintenance.

I think it is really amazing!