On 2005-04-29 12:09:48 +0200 (Fri, Apr), Lluís Batlle i Rossell wrote: > Hi! > > I need to use a regexp in sed for parsing a file... and I found a wrong > behaviour. I've tried in some computers, and I've found -DIFFERENT- > behaviours. > > It works -wrong- to me, in sed 4.0.9 and sed 4.1.4. > Here is the result in different computers: > > [EMAIL PROTECTED] ~/doc/xmligo/glosa $ echo "adela G" | \ > sed 's/^\([-A-FH-Za-z \?]\+\)/:\1:/' > :adela G: > [EMAIL PROTECTED] viric $ echo "adela G" | sed 's/^\([-A-FH-Za-z > \?]\+\)/:\1:/' > :adela :G > [EMAIL PROTECTED] ~ $ echo "adela G" | sed 's/^\([-A-FH-Za-z > \?]\+\)/:\1:/' > :adela :G > [EMAIL PROTECTED] viric $ echo "adela G" | sed 's/^\([-A-FH-Za-z > \?]\+\)/:\1:/' > :adela G: [...] > Any help?
man regex: Ranges are very collating-sequence-dependent, and portable programs should avoid relying on them. What are the locale settings on these computers? LC_ALL, LANG ? Try the same test with export LC_ALL=C; echo "adela G" | sed .... Maybe it will help. (or rewrite the regexp to NOT use ranges) -- $ ls -lart /bin/ls: you must be root to use LART
pgpV2bdi6HJwr.pgp
Description: PGP signature
