Java - Assign unicode apostrophe to char

Question

I want to assign the value of aphostrophe to a char:

char a = '\'';

However I would like to use the unicode version of apostrophe (\u0027) to keep it consistent with my code:

char a = '\u0027';

But doing it this way gives an error saying "unclosed character literal".

How can I do this assignment while still having the unicode code in the code?

score 13 · Accepted Answer · edited May 23 '17 at 10:31

13

The reason \u0027 doesn't work is that the unicode escape is handled very early by the compiler, and of course, it ends up being ' — which terminates the literal. The compiler actually sees this:

char a = ''';

...which naturally is a problem. The JLS talks about this in relation to line feeds and such in §3.10.4 (Character Literals).

Frankly, I think you're best off writing

char a = '\'';

...but char is a numeric type, so you could do this:

char a = 0x0027;

Of course, you could do this:

char a = "\u0027".charAt(0);

...but I think we can all agree that's a bit overkill. ;-)

Oooh, or check out Greg's answer: char a = '\u005c\u0027'; (\u005c is, of course, a backslash — so the compiler sees '\'').

edited May 23 '17 at 10:31

Community

1
1

answered Dec 03 '12 at 23:05

T.J. Crowder

1,031,962
187
1,923
1,875

The main problem is that `\u0027` is processed really early by the compiler, so this may end up being his best (maybe even only) option. – Brian Dec 03 '12 at 23:07
Also, check out the footnotes at the bottom of the JLS you just linked in your edit. It says you can't use something like `'\u000a'` because it gets processed into a newline before the class is even compiled. – Brian Dec 03 '12 at 23:08
@Brian: LOL overlapping edit and comment, refresh the page. ;-) – T.J. Crowder Dec 03 '12 at 23:12
why does it need to be any better? – amphibient Dec 03 '12 at 23:12
@T.J.Crowder Hey now, my comment was a full _1 minute and 38 seconds_ before your edit :D – Brian Dec 03 '12 at 23:18
I meant what's wrong with your solution -- why does it need to be any better? I see nothing wrong – amphibient Dec 03 '12 at 23:18
@Brian: I must have been typing slowly. ;-) – T.J. Crowder Dec 03 '12 at 23:51
I love Greg's suggestion. It makes me feel like I'm in The Matrix. – duma Sep 18 '14 at 19:42

score 9 · Answer 2 · answered Dec 03 '12 at 23:12

9

You can do this as well

char a = '\u005c\u0027';

where \u005c is the Unicode for \

answered Dec 03 '12 at 23:12

Greg

2,039
3
19
26

LOL! I love it. And given the OP's reason for wanting to use `\u0027` in the first place, a pretty good idea... – T.J. Crowder Dec 03 '12 at 23:31

irreputable · Answer 3 · 2012-12-03T23:33:03.953

before javac does anything else, it first convert all \u#### to a char. so your code is equivalent to

char a = ''';

that's why it doesn't compile.

\u#### is not just for char/string literals, you can use it anywhere, e.g. in variable names.

however, people rarely use non latin chars in identifiers; if someone does, he'll probably use his native charset, and he won't need \u#### either.

therefore we never really see \u#### anywhere other than in char/string literals, this gives the wrong impression to the unsuspected.

if there's time machine, we should probably kill this feature, since it's confusing and it's not used.

score 1 · Answer 4 · answered Dec 03 '12 at 23:10

1

Here's another option, indeed an overkill though:

char c = "\u0027".charAt(0);

answered Dec 03 '12 at 23:10

nullpotent

9,162
1
31
42

Java - Assign unicode apostrophe to char

4 Answers4

Linked

Related