Why i18n (locales) is so slow? LANG="en_US.UTF-8"

["Eric Wood" <eric interplas com>]

I'm sure this is an FAQ.  I've read up on
http://nscp.upenn.edu/aix4.3html/aixbman/prftungd/natlangsup.htm which
explains the structure types and why it's so expensive to do
internationalization.  So I went back to using "C" for my RH 8 and 9
systems:

$ cat /etc/sysconfig/i18n
#LANG="en_US.UTF-8"
LANG="C"
SUPPORTED="en_US.UTF-8:en_US:en"
SYSFONT="latarcyrheb-sun16"

Basically I'm trying to find out how LANG="en_US.UTF-8" would benefit me.
So far it has caused great pain as my grep searches (example below) are
severly slowed.   Since I only need english speaking anyway, is there any
benefit with "en_US.UTF-8" that I don't see?

-eric wood




$ LANG="en_US.UTF-8";export LANG
$ time bash -c 'grep WOOD /tmp/customers.tab | wc -l'
  13774

real    0m20.267s
user    0m20.240s
sys     0m0.030s

$ LANG="C";export LANG
$ time bash -c 'grep WOOD /tmp/customers.tab | wc -l'
  13774

real    0m0.060s
user    0m0.030s
sys     0m0.030s