Split into a Package

Every Go file starts with a package clause, and every directory is exactly one package. That is the whole rule: a package is a directory, its name is the package line its files share, and the way you reach it from elsewhere is an import path built from the module.

This workspace is one module:

go.mod            module gofinity/splitpackage
main.go           package main
textkit/          package textkit

The module line in go.mod names the module, and an import path is that name plus the directory below it:

import "gofinity/splitpackage/textkit"

The last element of the path is the directory, not necessarily the package name, though keeping the two the same is the convention and the kindest thing you can do for a reader. After the import, the package name is what you type at the call site: textkit.Title(...).

What crosses the boundary

Only the capitalised names. Title, Slug and WordCount are exported and callable from main; isWordChar is not, and textkit.isWordChar does not compile. Concept 15's second challenge is about designing that line - here just notice that it exists.

Two kinds of test file

textkit/textkit_test.go declares package textkit, so it is inside the package and can call isWordChar directly. main_test.go declares package main, so it sees only what main sees: the exported names it imported and the functions in its own package.

go test ./... walks the module and runs both.

Imports are grouped, not sorted by hand

gofmt sorts within a group; the convention is standard library first, then a blank line, then everything else:

import (
	"fmt"

	"gofinity/splitpackage/textkit"
)

An unused import is a compile error, not a warning. Go would rather you delete it than let the build carry weight nobody reads.

Task

Fill in textkit.Title, textkit.Slug and textkit.WordCount, then use them from main to write Headline and Summary.

Hints

• strings.Fields splits on any run of whitespace and never returns an empty field, which handles the leading, trailing and doubled spaces for free.
• Upper-casing the first rune means []rune(s) and unicode.ToUpper, not s[0]: a byte is not a letter once the text stops being ASCII.
• Build the slug with a strings.Builder. Remember whether you have skipped a character since the last one you kept, and write the - lazily when the next word character arrives - that way no separator is ever left dangling at either end.
• Slug lower-cases first, so isWordChar only ever sees lower-case letters and digits.
• Summary has three shapes and no arithmetic: a switch on the count reads better than nested ifs.