Can I programmatically convert "I’d" to "I’d" using Ruby?

Question

I can't seem to find the right combination of String#encode shenanigans.

Answer 1

I expect your script is using the encoding cp1251 and you have ruby >= 1.9.

Then you can use force_encoding:

#encoding: cp1251
#works also with  encoding: binary
source = 'I’d'
puts source.force_encoding('utf-8') #-> I’d

If my exceptions are wrong: Which encoding do you use and which ruby version?

---

A little background: Problems with encoding are difficult to analyse. There may be conflicts between:

• Encoding of the source code (That's defined by the editor).
• Expected encoding of the source code (that's defined with #encoding on the first line). This is used by ruby.
• Encoding of the string (see e.g. section String encodings in http://nuclearsquid.com/writings/ruby-1-9-encodings/ )
• Encoding of the output shell

Answer 2

I think I'd got confused on this one so I'll post this here to hopefully help anyone else who is similarly confused.

I was trying to do my encoding in an irb session, which gives you

irb(main):002:0> 'I’d'.force_encoding('UTF-8')
=> "I’d"

And if you try using encode instead of force_encoding then you get

irb(main):001:0> 'I’d'.encode('UTF-8')
=> "I’d"

This is with irb set to use an output and input encoding of UTF-8. In my case to convert that string the way I want it involves telling Ruby that the source string is in windows-1252 encoding. You can do this by using the -E argument in which you specify `inputencoding:outputencoding' and then you get this

$ irb -EWindows-1252:UTF-8
irb(main):001:0> 'I’d'
=> "I\xC3\xA2\xE2\x82\xAC\xE2\x84\xA2d"

That looks wrong unless you pipe it out, which gives this

$ ruby -E Windows-1252:UTF-8 -e "puts 'I’d'"
I’d

Hurrah. I'm not sure about why Ruby showed it as "I\xC3\xA2\xE2\x82\xAC\xE2\x84\xA2d" (something to do with the code page of the terminal?) so if anyone can comment with further insight that would be great.