Project · wc — a word counter CLI

Hints

• Use bufio.Scanner with bufio.ScanWords for the word count — don't try to roll your own splitter.
• Counting bytes: read into a buffer or use io.Copy with a counting io.Writer. Don't ReadAll a large file.
• For stdin detection, check os.Stdin.Stat() and look at Mode()&os.ModeCharDevice == 0 — that's the canonical way to detect piped input.
• Wire flag.Parse() for -l (lines only), -w (words only), -c (bytes only) to match real wc. Default is all three.

Expected output

$ echo "hello world\nfoo bar baz" | ./wc-go
       2       5      24

$ ./wc-go README.md main.go
      18      42     320 README.md
      45      98     812 main.go
      63     140    1132 total

$ diff <(./wc-go README.md | tr -s ' ') <(wc README.md | tr -s ' ')
(no output — they match)

Stretch goals

• Add -m (rune count, multibyte-aware) using utf8.RuneCount. Test on a non-ASCII file.
• Add a --format json flag that emits {"file": "x", "lines": 1, ...} per file. This sets you up for piping into jq.
• Run it concurrently across multiple files with a worker pool (preview of M3). Benchmark against the sequential version on a directory of 1000 files.
• Publish to a GitHub repo and verify go install github.com/you/wc-go/cmd/wc@latest works for a teammate.