[TxMt] Re: Properly escaping characters in ruby script

[Grant Hollingworth]

* Ian Duncan [8-Apr-09 17:37]:

> #!/usr/bin/env ruby
>     require "#{ENV['TM_SUPPORT_PATH']}/lib/escape"
> 	arg = STDIN.read
> 	arg.chomp!
> 	arg = e_as arg
> 	arg.gsub!("\`", "\\`")
> 	puts("pointfree \"" + arg + "\"")
>
> The puts is just for debugging purposes. When I run this command on
> the selected text:
>
> \x y -> x `div` y
>
> I am hoping to get:
>
> pointfree "\\x y -> x \`div\` y"
>
> But I get this instead:
>
> pointfree "\\x y -> x \\x y -> x div\\x y -> x `div y"

Short answer: use e_sn

Replacing backslashes is a pain. There are a few things going on:

1. the string interpretation
2. the regular expression engine also uses backslashes as the escape  
character
3. \` means "the text before the match" -- that's why you're seeing  
double

To fix it, you can:

1. add more toothpicks
arg.gsub('`', '\\\\`')

2. use the block form of gsub, which only evaluates the string once
arg.gsub('`') { '\`' }

3. match the space before the backtick, which is what e_sn does
arg.gsub(/(?=`)/, '\\')