Regex.escape bug

Hi guys,

I’m using “ruby 1.9.1p378 (2010-01-10 revision 26273) [x86_64-linux]”
and just noticed a possible bug in Regex.escape() while using it with
gsub()

This:

pp Regexp.escape( ‘.rb’ )

yields:

“\.rb”

while it should be “.rb”.

Regexp.new( ‘.rb’ ) or Regexp.new( ‘.rb’ ) work though, they replaces
the “.rb” string as expected.
Yes I know that “Regexp.new( ‘.rb’ )” works incidentally but I’m just
mentioning it. :slight_smile:

Tasos L. wrote:


Yes I know that “Regexp.new( ‘.rb’ )” works incidentally but I’m just
mentioning it. :slight_smile:

It is correct, because it lets you do this:

dot_rb = Regexp.escape( ‘.rb’ )
=> “\.rb”

/file#{dot_rb}/ =~ “file.rb”
=> 0

/file#{dot_rb}/ =~ “fileQrb”
=> nil

Note that “\.rb” is the same string as ‘.rb’.

On 28.06.2010 19:31, Tasos L. wrote:


yields:

“\.rb”

while it should be “.rb”.

No, it should be as it is as Joel explained. Note also

irb(main):007:0> puts Regexp.escape( ‘.rb’ )
.rb

It is the representation of the string that includes a backslash to
properly escape the backslash:

irb(main):010:0> “\n”.size
=> 2
irb(main):011:0> “\n”.size
=> 1
irb(main):012:0> puts “\n”
\n
=> nil
irb(main):013:0> puts “\n”

=> nil

Regexp.new( ‘.rb’ ) or Regexp.new( ‘.rb’ ) work though, they replaces
the “.rb” string as expected.
Yes I know that “Regexp.new( ‘.rb’ )” works incidentally but I’m just
mentioning it. :slight_smile:

It works but it does not match the sequence “.rb” - rather it matches
any character followed by “rb”:

irb(main):004:0> r = Regexp.new( ‘.rb’ )
=> /.rb/
irb(main):005:0> r.match ‘.rb’
=> #<MatchData “.rb”>
irb(main):006:0> r.match ‘oooopsrb’
=> #<MatchData “srb”>

Kind regards

robert