Thanks for taking the time to look at this package.
Open an issue with the smallest example that reproduces it, the Swift and Xcode versions you're on, and the platform you're targeting. A failing test is the clearest possible report.
swift build
swift testCI runs swift build and swift test on every push and pull request
(.github/workflows/tests.yml), resolving dependencies fresh rather than from a
cache. That run is the authority: a local .build left over from an earlier
resolve can report green where a fresh checkout does not. Run both commands
locally first anyway — it is faster than waiting for a runner.
Snapshot tests are the exception and do not run in CI: the recorded pixels follow
the host simulator's scale, so they are compared locally with
xcodebuild test -destination 'platform=iOS Simulator,...'. Wrap image-test files
in #if canImport(UIKit) so the macOS run leaves them out.
Documentation lives in the DocC catalog under Sources/*/*.docc/. Public
declarations are documented with /// comments, in English.
Maintainers only. Release from a commit whose Tests run is green — release-on-tag.yml
turns a tag into a GitHub Release and nothing else, so nothing is compiled on the
way out. The version is computed, never chosen:
scripts/release.sh --dry-run # see what version the API diff implies
scripts/release.sh # stamp CHANGELOG, tag, pushWrite what changed under ## [Unreleased] in CHANGELOG.md — prose only, no
version number. scripts/release.sh compares the public API against the last
release and derives the version from the actual difference: a removed or changed
public symbol means a major, additions mean a minor, and a change in the
generation of a dependency whose types appear in this package's public API also
means a major.