SC2018
Joachim Ansorg edited this page Nov 12, 2021
·
4 revisions
Use [:lower:] to support accents and foreign alphabets.
Problematic code:
PLATFORM="$(uname -s | tr 'A-Z' 'a-z')"Correct code:
PLATFORM="$(uname -s | tr '[:upper:]' '[:lower:]')"Rationale:
A-Z and a-z are commonly intended to mean "all uppercase" and "all lowercase letters" respectively. This ignores accented characters in English, and foreign characters in other languages:
$ tr 'a-z' 'A-Z' <<< "My fiancée ordered a piña colada."
MY FIANCéE ORDERED A PIñA COLADA.
Instead, you can use [:lower:] and [:upper:] to explicitly specify case:
$ tr '[:lower:]' '[:upper:]' <<< "My fiancée ordered a piña colada."
MY FIANCÉE ORDERED A PIÑA COLADA.
Exceptions:
If you don't want a-z to match é or A-Z to match Ñ, you can ignore this message.
As of 2019-09-08, BusyBox tr does not support character classes, so you would have to ignore this message.
Note that the examples used here are multibyte characters in UTF-8. Many implementations (including GNU) fails to deal with them.