Anand Buddhdev wrote: > [arb@home arb]$ time grep zymology docs/sowpods.txt > enzymology > zymology > > real 0m0.267s > user 0m0.260s > sys 0m0.000s > > [arb@home arb]$ export LANG=C > [arb@home arb]$ time grep zymology docs/sowpods.txt > enzymology > zymology > > real 0m0.012s > user 0m0.000s > sys 0m0.000s > > Grep is clearly still much slower in UTF8. > I'm not sure that's quite a fair test as the files will be cached in the second case. However, you are right, grep is slower under utf8. It's not surprising as any file it encounters could contain unicoded characters, so there will be extra overheads even if the files only contain standard ascii. rc.sysinit on FC2 still contains many lines with LC_ALL=C grep for just that reason! Jonathan
