If you want to preserve internal whitespace, try
s/ #replace
^[\s]* #the start of the string followed by any number of spaces
(.*) #zero or more characters of anything, stuff into $1. This
+is the actual string we want.
(?<!\s) #look-behind assertion to get to the last non-whitespace ch
+ar
\s*$ #match whitespace until EOL
/$1/x; #replace with the (.*)
Update: If you value your CPU time, use one of the other regular expressions instead. I got caught up in using the .*, leading to the need for the look-behind assertion. Better answers can be found here. For the interested, Benchmark: timing 100000 iterations of my_long_one, one_liner, two_lin
+er...
my_long_one: 6 wallclock secs ( 4.62 usr + 0.00 sys = 4.62 CPU) @ 2
+1659.09/s (n=100000)
one_liner: 3 wallclock secs ( 3.89 usr + 0.00 sys = 3.89 CPU) @ 256
+73.94/s (n=100000)
two_liner: 1 wallclock secs ( 2.16 usr + 0.00 sys = 2.16 CPU) @ 462
+10.72/s (n=100000)
Using
my_long_one => s/^[\s]*(.*)(?<!\s)\s*$/$1/
one_liner => s/^\s+|\s+$//g
two_liner => s/^\s+//g; s/s+$//g
|