From: Xavier Noria <fxn@...>
Date: 2007-05-03T01:40:40+09:00
Subject: Re: replacing parts of newlines

On May 2, 2007, at 6:26 PM, Mike Steiner wrote:

> I'm writing a script to convert text into a csv file to import into  
> MS Excel
> (under Windows), and Excel allows multiline text within a cell, but  
> the
> newlines (\x0D\x0A) have to be replaced with only a \x0A for Excel  
> to put
> everything into one cell.
>
> The problem is that in Ruby, if I do:
>
>    newfield = field.gsub ( "\x0D" , "" )
>
> it doesn't do anything because Ruby treats \x0D\x0A as a unit.

Not really.

If there's a CRLF in the text file and you read it using text mode in  
a CRLF platform, the CR is filtered out by the I/O layer. So a line- 
oriented script only sees LFs in its strings, vaguely speaking.  
Reciprocally, when you write to disk in text model in a CRLF platform  
all LFs get a CR automatically prepended. That's transparent to the  
programmer, in fact that's done by the very stdio in the C interpreter.

You probably need to work in binmode.

-- fxn