In a correctly working awk utility, length($0) returns the number of (single-byte or multi-byte) characters in that line (not including the terminating <newline> character).

In some (buggy) awk utilities, length($0) returns the number of bytes in that line (not including the terminating <newline> character). On these systems, substr(string, start, count) may return a substring of string that is unusable because it does not start and/or end on a character boundary.

On a correctly working awk when using a locale based on a UTF-8 codeset, the command:

Code:

awk 'BEGIN{print "Vía", length("Vía");print "Via", length("Via")}'

will print:

Code:

Vía 3
Via 3

while a buggy version might print:

Code:

Vía 4
Via 3

Note also that fixed length lines can, mean several things: Fixed number of characters per line and fixed number of bytes per line are the most common two meanings. In ASCII and other single-byte/character codesets, a fixed number of characters per line and a fixed number of bytes per line happen to be the same thing. With codesets like UTF-8, a character can be represented by one to six bytes. You can also have a fixed number of display columns per line which can be different from both of the above with some characters even having a variable width (such as <tab> typically takes 1 to 8 display columns but can be even more depending on how tab stops are set). And, on displays with proportional width fonts, every character can have a different (and sometimes variable) column width (both for characters like <tab> and due to kerning effects).