Add Line Numbers

What this tool is used for

• Numbering a code excerpt so you can refer to a specific line in review.
• Preparing a transcript where lines need citing.
• Numbering a list before sharing it so items can be referenced.
• Producing a numbered listing for documentation.
• Adding numbers to a log extract to discuss a particular entry.

Frequently Asked Questions

Should numbering start at 0 or 1?

Editors and compilers number lines from 1, and that is the convention here — a compiler error on "line 42" refers to the 42nd line. Array indices start at 0, which is a different thing entirely and a frequent source of off-by-one confusion.

Why pad the numbers?

So the code stays aligned. Without padding, the jump from line 9 to 10 shifts every subsequent line right by a character. Right-aligning to the width of the largest number keeps a clean gutter.

Does the final newline count as a line?

This is genuinely ambiguous and tools disagree. POSIX defines a line as ending WITH a newline, so a trailing newline does not begin a new line — but many editors display a phantom empty last line anyway. It explains off-by-one differences between tools.

Will the numbers break copy-paste of code?

Yes — numbered output is for reading and referencing, not for running. Anyone copying it must strip the numbers again, which is why code blocks on the web usually render numbers in a separate non-selectable column.

What separator should I use?

A colon and space matches compiler and grep output (file:42: message), making the result greppable. A pipe or tab suits side-by-side reading. Consistency matters more than the specific choice.

How should numbering handle a wrapped line?

As one line — the number belongs to the logical line, not the visual one. Editors that number visual rows make a wrapped file confusing to discuss, which is why most number logical lines and wrap without a number.

Should numbering restart per section?

Only if the sections are independently referenced. Continuous numbering is easier to cite across a whole file; per-section numbering is what legal and specification documents use, where the section is the unit.

What about numbering only non-blank lines?

It is what `cat -b` does, and it makes for a tidier listing — but it breaks the correspondence with an editor's own line numbers, which is usually the reason for numbering in the first place.

Common errors and gotchas

• Numbering from one when the excerpt starts partway through a file, so the numbers mislead.
• Adding numbers to code that someone will copy and run, which then fails to parse.
• Numbering wrapped lines as separate lines, which desynchronises from the source.
• Losing the numbers' alignment on files past a hundred lines without padding.
• Numbering a file that already has numbers, producing two columns of them.