C - Why does adding numbers to a char reset at 255?

Question

I'm not sure how to unlock characters past 255 automatically.

char f;
f = 'k';
printf("%d\n", f);

Results in 107. No surprise.

f += 500;
printf("%d\n", f);

Results in 95. It seems this has been modulo divided by 255.

printf("%c\n", 607);

Results in _. There are many more characters that extend into the thousands as well. How come adding a value and making a char go past 255 forces a modulo operation? I need to go past 255 for the sake of a hash function.

`sizeof(char)` is one byte. So it could not be greater than 255. If you need more, use `wchar_t` or its analogs. — Ari0nhh, May 13 '16 at 11:53
I strongly encourage you to read this: [What Every Programmer Absolutely, Positively Needs To Know About Encodings And Character Sets To Work With Text](http://kunststube.net/encoding/) — bolov, May 13 '16 at 11:55
@Ari0nhh It's true that `sizeof(char)` is one, it's not necessarily true that it *can't* be greater than 255. `sizeof(char)` can be 1 but it *could* consist of 16 bits. — P.P, May 13 '16 at 11:55
Just go read chapter one of a C programming book. That being said, you shouldn't use `char` to store numbers for other reasons, namely the implementation-defined signedness. — Lundin, May 13 '16 at 11:57
@P.P. In C, yes, but POSIX mandates byte, thus `char` to be exactly 8 bits. — el.pescado - нет войне, May 13 '16 at 11:57
@Lundin I need to enumerate them for a hash function and they will need to be a huge range of characters after reconversion. Admittedly the assignment question this is from is terrible but unfortunately marks have to be obtained. — Synlar, May 13 '16 at 12:00
It seems that in your case, `char` is `unsigned` and its range is 0-255. When it goes beyond 255, the value wraps-around which is the same as mod 256 — Spikatrix, May 13 '16 at 12:01
@el.pescado: A "byte" is not defined to be 8 bits. POSIX might define a byte in the standard to have 8 bits, but the reverse is not true. Anyway, C is not POSIX and there is no indicator the question is about POSIX. The C standard only uses `char` and "byte" synonymously, but does not enforce the bitwidth - that's why `CHAR_BIT` exists. — too honest for this site, May 13 '16 at 12:05

blatinox · Accepted Answer · 2016-05-13T13:06:58.573

3

A char is a byte in memory. So you can not store values > 255 because it has only 8 bits (assuming that on your platform a char is defined as a byte of 8 bits). When the value is more than 255, it overflows.

Furthermore, you don't know if a char is unsigned (values can be between 0 and 255) or signed (values can be between -128 and 127).

You can use another type than char if you want to store values > 255.

edited May 13 '16 at 13:06

answered May 13 '16 at 11:54

blatinox

833
6
18

1

`char` are *at least* a byte, but your reasoning about overflow is right – Garf365 May 13 '16 at 12:04
5

@Garf365: `char` is guaranteed to be 1 byte, but this is not true: `char` is 8 bits, `CHAR_BIT` is the number of bits in a `char`: [What platforms have something other than 8-bit char?](http://stackoverflow.com/q/2098149/1606345) – David Ranieri May 13 '16 at 12:08
Sorry I downvoted this answer because it is rather wrong, even with char bits as 8. – Antti Haapala -- Слава Україні May 13 '16 at 12:22

score 0 · Answer 2 · answered May 13 '16 at 11:56

0

A char data type is an 8-bit field, with a maximum total value of 255. You need to use a 16-bit value if you wish to work with numbers greater than 255.

unsigned char has a value range of 0 to 255
signed char has a value range of -128 to 127

answered May 13 '16 at 11:56

BackDoorNoBaby

1,445
2
13
20

1

I do not have them backwards. – BackDoorNoBaby May 13 '16 at 11:58
He needs to work with unicode. `signed char` and `unsigned char` don't help here at all. – bolov May 13 '16 at 11:59
1

The OP uses `char` however. What range and signedness a `char` has, we cannot know, since it is implementation-defined. – Lundin May 13 '16 at 11:59
1

@Lundin Oh good grief. How many angels can sit on the head of a pin? – nicomp May 13 '16 at 12:00
@Lundin , yes, that was it. – nicomp May 13 '16 at 12:01
@bolov what does unicode have to do with this? OP said he's using C, and a `char` data type, its an `8-bit field` which can be `signed` or `unsigned` – BackDoorNoBaby May 13 '16 at 12:02
@BackDoorNoBaby `char` is a *at least* one byte field – Garf365 May 13 '16 at 12:04
@Garf365 in what case is a `char` greater than one byte in C – BackDoorNoBaby May 13 '16 at 12:05
1

@BackDoorNoBaby size of type are implementation dependents – Garf365 May 13 '16 at 12:06
@Garf365 okay, maybe I was just thinking simply of C99 – BackDoorNoBaby May 13 '16 at 12:07
@BackDoorNoBaby this is also true for C99 – Garf365 May 13 '16 at 12:10
1

`sizeof(char)` is always one, but according to the C standard, `CHAR_BITS` should be `>= 8` – Spikatrix May 13 '16 at 12:16
@CoolGuy thank for precision, I totally forget this distinction. I used in past a DSP and `char` was more than 8 bits – Garf365 May 13 '16 at 12:21

C - Why does adding numbers to a char reset at 255?

2 Answers2