String processing question - char set related

Thread view:

XXX - 22 Nov 2006 05:24 GMT

If I have a string with \r\n & I am trying to convert all \r\n to \n,
then is code like this good enough?

String s // contains the original string.

StringBuffer old = new StringBuffer(s);
StringBuffer new= new StringBuffer();

for (int i=0; i < old.length(); ++i)
{
if (strBuf.charAt(i) != '\r')
{
new.append(old.charAt(i));
}
}

Or are there problems with this? I am think of problems
like
- Is it possible to have strings with just \r not followed by \n.
When can this happen?

- Is it possible for some Unicode chars to have the \r\n pattern
which doesn't represent a new line?

Reply to this Message

Chris Smith - 22 Nov 2006 06:12 GMT

> Or are there problems with this? I am think of problems
> like
> - Is it possible to have strings with just \r not followed by \n.
> When can this happen?

Yes. There are platforms where \r is the standard representation of
end-of-line. If you need to handle end of line sequences across a
number of common platforms, then it would be safer to wrap a
StringReader with a BufferedReader, and then use readLine to get the
lines.

Alternatively, you may be reading from some protocol where the end of
line sequence is specified; for example, it's required to be \r\n for
many common internet application protocols. Then you could just look
for that one sequence and replace it with \n if that's what you want.

> - Is it possible for some Unicode chars to have the \r\n pattern
> which doesn't represent a new line?

It's safe to assume that \r\n indicates a newline whenever you find it.

Signature

Chris Smith

Reply to this Message

XXX - 22 Nov 2006 06:48 GMT

>> Or are there problems with this? I am think of problems
>> like
[quoted text clipped - 11 lines]
> many common internet application protocols. Then you could just look
> for that one sequence and replace it with \n if that's what you want.

This is going to be text, I get from a AWT TextArea widget by calling
getText()

Are these issues relavant in this case?

>> - Is it possible for some Unicode chars to have the \r\n pattern
>> which doesn't represent a new line?
>
> It's safe to assume that \r\n indicates a newline whenever you find
> it.

Reply to this Message

Daniel Pitts - 22 Nov 2006 18:12 GMT

> >> Or are there problems with this? I am think of problems
> >> like
[quoted text clipped - 22 lines]
> > It's safe to assume that \r\n indicates a newline whenever you find
> > it.

I would bet that you don't need to worry about it in this case. A few
simple tests will let you know.

Reply to this Message

String processing question - char set related

Free Magazines