> Your locale is zh_CN.UTF-8. What you're expecting is only guaranteed > in the C locale:
I'm not quite sure it applies here. I'm using US English Windows 7. LANG = 'en_US.UTF-8' I get the same result: $ echo abcdeABCDE | sed -e 's/[B-D]/_/g' ab__eA___E BUT: $ echo abcdeABCDE | LANG=C sed 's/[B-D]/_/g' abcdeA___E This is very weird, indeed. OTOH, in Linux I have the same LANG setup, yet it does work correctly: > echo $LANG en_US.UTF-8 > echo abcdeABCDE | sed -e 's/[B-D]/_/g' abcdeA___E I believe that an en_US UTF-8 string representation for "abcdeABCDE" is not any different from ASCII. Anton Lavrentiev Contractor NIH/NLM/NCBI
