问题
Consider the following snippet:
puts 'hello'.gsub(/.+/, '\0 \\0 \\\0 \\\\0')
This prints (as seen on ideone.com):
hello hello \0 \0
This was very surprising, because I'd expect to see something like this instead:
hello \0 \hello \\0
My argument is that \
is an escape character, so you write \\
to get a literal backslash, thus \\0
is a literal backslash \
followed by 0
, etc. Obviously this is not how gsub
is interpreting it, so can someone explain what's going on?
And what do I have to do to get the replacement I want above?
回答1:
Escaping is limited when using single quotes rather then double quotes:
puts 'sinlge\nquote'
puts "double\nquote"
"\0"
is the null-character (used i.e. in C to determine the end of a string), where as '\0'
is "\\0"
, therefore both 'hello'.gsub(/.+/, '\0')
and 'hello'.gsub(/.+/, "\\0")
return "hello"
, but 'hello'.gsub(/.+/, "\0")
returns "\000"
. Now 'hello'.gsub(/.+/, '\\0')
returning 'hello'
is ruby trying to deal with programmers not keeping the difference between single and double quotes in mind. In fact, this has nothing to do with gsub
: '\0' == "\\0"
and '\\0' == "\\0"
. Following this logic, whatever you might think of it, this is how ruby sees the other strings: both '\\\0'
and '\\\\0'
equal "\\\\0"
, which (when printed) gives you \\0
. As gsub uses \x
for inserting match number x, you need a way to escape \x
, which is \\x
, or in its string representation: "\\\\x"
.
Therefore the line
puts 'hello'.gsub(/.+/, "\\0 \\\\0 \\\\\\0 \\\\\\\\0")
indeed results in
hello \0 \hello \\0
来源:https://stackoverflow.com/questions/3028547/backslashes-in-gsub-escaping-and-backreferencing