${#var} behaviour does not change along with locale
magicant In terms of POSIXly correctness, the current behaviour is expected.
http://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#tag_18_05_03
Changing the value of LC_CTYPE after the shell has started shall not affect the lexical processing of shell commands in the current shell execution environment or its subshells. Invoking a shell script or performing exec sh subjects the new shell to the changes in LC_CTYPE.
As an exception, yash reflects changes of LC_CTYPE if it is in the interactive mode but not in the POSIXly correct mode. Some users have to set LC_CTYPE in their .yash_profile or .yashrc because the system might not be configured so that the shell is invoked with LC_CTYPE having the value the users want.
Regarding character processing, yash handles almost all strings as "wide character sequences" rather than "byte sequences". LC_CTYPE affects how characters are converted between wide characters and bytes. However, changing LC_CTYPE in interactive yash does not convert characters between the locales, so there is no guarantee that characters are handled as expected after changing LC_CTYPE. To convert characters between locales, you need to use a dedicated converter like "iconv".
Bytes are converted to wide characters even in the C locale, so it is important to note that not all bytes are accepted in the C locale. Typically non-ASCII characters are rejected or ignored as illegal. You should not rely on the C locale to count the number of raw bytes in a string.
Thanks for the information! I withdraw my report.
magicant 
Fork
Assuming a UTF-8 locale:
$ yash -c 'v=é; echo ${#v}; LC_CTYPE=C; echo ${#v}' 1 1Since that is a two-byte character, I'd expect the second '1' to be '2', like in bash, ksh93 and zsh:
$ bash -c 'v=é; echo ${#v}; LC_CTYPE=C; echo ${#v}' 1 2I think that behaviour is more useful as it allows counting bytes by temporarily setting LC_ALL=C.
I'm not sure whether to file this under "Bugs" or "Feature Requests". I'm leaning towards "Bugs" as it seems to me that ${#var} should count the characters in the currently set locale, which is not necessarily the same locale that was set when the shell was initialised.