Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.54%. Comparing base (85bb7e7) to head (c19e4f1).

@@           Coverage Diff           @@
##             main     #299   +/-   ##
=======================================
  Coverage   99.53%   99.54%           
=======================================
  Files          11       11           
  Lines         868      874    +6     
=======================================
+ Hits          864      870    +6     
  Misses          4        4           

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Are there actual use cases where people have file sizes that are infinite or NaN?

Are there actual use cases where people have file sizes that are infinite or NaN?

It’s mostly defensive: naturalsize() accepts floats, and callers often pass computed values (ratios, averages, deltas) from metrics/ETL pipelines. Those can become NaN/±Inf after a divide‑by‑zero or overflow, or when size is unknown in a streaming context. Returning an explicit “NaN/±Inf” is clearer than producing a weird suffix or raising. Also, humanize.number already formats NaN/Inf explicitly, so this keeps behavior consistent across helpers.