Code Line Counter

How the Code Line Counter works

Lines of code is the oldest software metric and still one of the most quoted. It sizes estimates, appears in audit and acquisition reports, feeds licence-compliance scans and gives a rough sense of how big a change is. The number is only useful when it is counted the same way each time, which means separating real code from comments and blank space.

How the code line counter classifies lines

Each line lands in exactly one bucket:

• Blank lines contain nothing but spaces or tabs.
• Comment lines contain comment text and no code, including every line inside a multi-line block comment.
• Code lines contain at least one character of code. A line with code and a trailing comment, such as x = 1 // reset, counts as code.

This follows the convention used by well-known counters such as cloc, so the figures are comparable with the reports those tools produce. Physical lines are counted, not statements, so a long call split across five lines counts as five.

Why strings matter

A naive counter treats every // or # as the start of a comment, which goes wrong on code such as url = "https://example.com" or color: "#fff". This tool tracks string literals for the chosen language and ignores comment markers inside them. In C-style languages, backtick template strings can span several lines, and those lines count as code. Escaped quotes inside strings are handled too.

Choosing the language

The options group languages that share comment syntax:

• C-style covers JavaScript, TypeScript, Java, C, C++, C#, Go, PHP, Rust, Swift, Kotlin and Dart: // line comments and /* */ blocks.
• Python uses # comments. Triple-quoted strings on their own lines are counted as comments, because that is how docstrings are used in practice.
• HTML and XML use <!-- --> blocks. Inline scripts and styles count as code.
• CSS, SCSS and Less use /* */ blocks. SCSS and Less also allow //, so lines starting with it count as code here; switch to C-style for files that use those heavily.
• SQL uses -- line comments and /* */ blocks.
• Shell, Ruby, Perl and YAML share # comments. A shebang line counts as a comment.
• Lua uses -- and --[[ ]] blocks.

Reading the comment ratio

The comment ratio is comment lines divided by code plus comment lines, leaving blanks out. There is no correct figure, but typical open-source projects sit between 10 and 30 percent. Libraries with generated API documentation often run higher. A ratio near zero on a large file suggests code that will be hard for the next person to change. A very high ratio can mean commented-out code or licence headers inflating the count, both worth a look.

Blank lines are usually 10 to 20 percent of a file. Far fewer can mean dense code that is hard to scan; far more often comes from copy and paste or generated output.

The longest line

The report gives the longest line's length and line number. Most style guides cap lines somewhere between 80 and 120 characters: PEP 8 asks for 79 in Python, Google's Java guide for 100, and Prettier wraps at 80 by default. A line far past your limit is often a long string, a minified chunk or a deeply nested expression worth splitting.

What line counts cannot tell you

Lines of code measure size, not value or effort. The same feature can be 50 lines in one language and 500 in another, and the shortest solution is often the best one. Use the count to compare a file with itself over time, or similar files with each other, rather than to rank developers or estimate cost on its own.

For a whole repository, a command-line tool that walks every file and detects the language from the extension will be faster. This tool is for a quick check on a single file or a pasted snippet in a code review, where installing anything is overkill.

Related tools