Printf does not properly justify non-ASCII characters

[William Andrea]

I have a script that outputs arbitrary Unicode characters in neat columns, but for anything outside the Basic Latin range (i.e. codepoints > 127), the justification is off. For example, both below commands should output a leading space:

$ printf "%2s\n" 'a'
 a
$ printf "%2s\n" 'á'
á

The spacing problem starts between U+7F and U+80. If you try to print two leading spaces, the same problem occurs between U+7FF and U+800.

This affects the binary /usr/bin/printf as well, but I'm not sure where to report a bug for that.

Ubuntu version: 14.04.5
Bash version: 4.3-7ubuntu1.5 (latest)

[xhienne (xhienne)]

In my opinion, this is not a bug (and probably not a feature either) but the expected behavior.

bash's printf is a slightly modified version of the underlying printf() function provided by the C standard library. The printf(3) manual explicitly states that the precision is a number of _bytes_, not a number of characters. So this is the expected result with multibyte characters like 'á'.

You might want to switch to a ISO-8859 character set if you want your 'á' character to only take one byte of memory.

[William Andrea (wjandrea)]

xhienne, that's a good point. My workaround is to switch to Python 3, which treats multi-byte chars the same as single-byte chars.