Fix inreplace when using ' or other RegExp symbols

Evidence that using perl from the cli for in-replace is stupid :P

Had to use $'' to allow escaping of ' in bash strings.

Wasn't escaping regexp symbols as well, so it was amazing this worked at all!

Library/Homebrew/brewkit.rb

@@ -36,18 +36,15 @@ unless $root.to_s == '/usr/local'
   ENV['LDFLAGS']='-L'+$root+'lib'
 end
 
-
-######################################################################## utils
-
 def inreplace(path, before, after)
-  before=before.to_s.gsub('"', '\"').gsub('/', '\/')
+  before=Regexp.escape before.to_s
-  after=after.to_s.gsub('"', '\"').gsub('/', '\/')
+  after=Regexp.escape after.to_s
+  before=before.gsub "/", "\\\/"
+  after=after.gsub "/", "\\\/"
+  before=before.gsub "'", '\''
+  after=after.gsub "'", '\''
 
-  # we're not using Ruby because the perl script is more concise
+  # TODO this sucks
-  #TODO the above escapes are worse, use a proper ruby script :P
+  # either use 'ed', or allow regexp and use a proper ruby function
-  #TODO optimise it by taking before and after as arrays
+  `perl -pi -e $'s/#{before}/#{after}/g' "#{path}"`
-  #Bah, just make the script writers do it themselves with a standard collect block
-  #TODO use ed -- less to escape
-  #TODO the above doesn't escape all regexp symbols!
-  `perl -pi -e "s/#{before}/#{after}/g" "#{path}"`
 end