Everything behind the projects and packages I build. The decisions, the dead ends, and the fixes that were hard-won.
An app that imported only Button could not bundle without recharts. The 3.0.0 tag moves charts, data table and form to their own subpaths.
A commit said it vetted a link, and its test passed. A review before the tag found a bypass, and the same hole in the editor since 1.1.0.
A dropzone with a file already in it kept saying "Drag a file here". Nothing threw, no test went red, and the screen was lying.
For three days @akira-io/ui shipped on npm and answered Semana anterior to people who had never asked for Portuguese. That is what extraction costs.
I rewrote eleven props to ignore their values and the test suite stayed green. Everything I trusted after that had to earn it.
A checkout has two identities: the person who pays and the entity the invoice belongs to. Version 2.1 stops collapsing them into one.
The 2.0 release changed 126 files and added little a user can see. It spent the whole budget on extension points instead.
The 0.7 release turns Laravel SISP from a payment redirect helper into checkout infrastructure with memory.
Linear and Jira landed in the same release. That is the boring half. The interesting half is the decision underneath them.
0.6.4 folds two small releases into one cleanup of pending payments, signed retries, totals, refunds, and metadata.
Same module surface in Go and Rust. paths, files, shell, clipboard, notify, keyring. Two crates that exist because I wrote them.
I bootstrapped Spectra on a Thursday morning. By midnight it was reading real Laravel routes. The first commit that mattered was not the one that mattered.
Nothing under this filter yet.
Try another tag, or clear the filter.
