Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

illegal line end in character literal in escaped unicode CARRIAGE RETURN / NEW LINE

789 views
Skip to first unread message

Christian

unread,
May 15, 2009, 9:51:13 AM5/15/09
to
Hi folks,

I just stumbled replacing an escaped Unicode of CARRIAGE RETURN / NEW
LINE and got a compiler-error "illegal line end in character literal"
for

switch (original)
{
case '\u000D' : return '\u0020';
case '\u000A' : return '\u0020';
}

To be honest: I don't have the slightest idea what the compiler is
complaining about. The escape-sequence \u seems to be correct
according to most of the sources I came across and the two-digited
hexadecimal encoding cannot be wrong - so what's up here?


Could anybody give me a hint?

Thanks in advance,

Christian

Andreas Leitgeb

unread,
May 15, 2009, 10:02:07 AM5/15/09
to
Christian <ja...@wispa.de> wrote:
> switch (original)
> {
> case '\u000D' : return '\u0020';
> case '\u000A' : return '\u0020';
> }

Use \r instead of \u000D and \n for \u000A.

For background and reasons, google for "u000a" (or "u000d")

Mark Space

unread,
May 15, 2009, 10:35:55 AM5/15/09
to
Christian wrote:

> case '\u000D' : return '\u0020';
> case '\u000A' : return '\u0020';

> Could anybody give me a hint?

What Andreas said. The compiler actual pre-processes \uxxxx and feeds
the resulting character to the rest of the compiler. So your compiler
is actually seeing a carriage return and a new line there, just as if
you had pressed Enter. Which is illegal.

Lew

unread,
May 15, 2009, 10:38:23 AM5/15/09
to
On May 15, 9:51 am, Christian <j...@wispa.de> wrote:
> Hi folks,
>
> I just stumbled replacing an escaped Unicode of CARRIAGE RETURN / NEW
> LINE and got a compiler-error "illegal line end in character literal"
> for
>
> switch (original)
> {
>         case '\u000D' : return '\u0020';

use less wide indentation, please.

>         case '\u000A' : return '\u0020';
>
> }
>
> To be honest: I don't have the slightest idea what the compiler is
> complaining about. The escape-sequence \u seems to be correct
> according to most of the sources I came across and the two-digited
> hexadecimal encoding cannot be wrong - so what's up here?

You ended the line after the keyword 'case' without giving it a
constant. Bear in mind that Unicode escape sequences are substituted
prior to parsing the source, so for example, your second 'case' is
exactly equivalent to:

case '
' : return ' ';

Do you see how that is not a legal Java syntax?

--
Lew

Joshua Cranmer

unread,
May 15, 2009, 10:50:41 AM5/15/09
to
Christian wrote:
> To be honest: I don't have the slightest idea what the compiler is
> complaining about. The escape-sequence \u seems to be correct
> according to most of the sources I came across and the two-digited
> hexadecimal encoding cannot be wrong - so what's up here?

The \u is not a character escape sequence, but a preprocessing
token--these are processed even before lexical scanning.

--
Beware of bugs in the above code; I have only proved it correct, not
tried it. -- Donald E. Knuth

0 new messages