Hope std::string::operator[] could return unsigned char

Question

if s is of type string in C++, s[n] will return a char, not unsigned char. This creates lots of problem for me since I have to do type conversion everywhere.

The following code will not print "yes".

#include <string>
#include <iostream>
using namespace std;

typedef unsigned long ulong;
typedef unsigned int uint;
typedef unsigned short uint16;
typedef unsigned char uchar;

int main(int argc, char *argv[]) {
    string s = "h\x8fllo world";
    printf("%d\n", s[1]);
    if (s[1] == 0x8f) {
        printf("yes\n");
    }
    return 0;
}

I can make it work by changing the above to be if ((uchar*)(s[1] == 0x8f) {, but there are too many occurrences. I really hoped that the [] operator on string class could return a unsigned char!! Is there a way to do this?

(1) look at the cstdint header which has `std::uint16_t` and similar. (2) use a `std::vector` ? — Ryan Haining, Jul 29 '15 at 16:46
_"This creates lots of problem for me since I have to do type conversion everywhere."_ Didn't you create the problem for yourself when you decided to use `unsigned char` "everywhere"? What was your reason for that? — Lightness Races in Orbit, Jul 29 '15 at 16:46
`if ((uchar*)(s[1] == 0x8f)` is not what you meant to write, is it? — Lightness Races in Orbit, Jul 29 '15 at 16:52
More about `char` signedness: http://stackoverflow.com/questions/17097537/why-is-char-signed-by-default-in-c — Piotr Siupa, Jul 29 '15 at 17:43

score 4 · Answer 1 · edited May 23 '17 at 11:58

4

You should be able to use a std::basic_string<unsigned char> (std::string is really just a typedef to std::basic_string<char>). That will mean modifying your code to use your new typedef instead of std::string, though that should be a fairly simple search-and-replace refactoring.

See Strings of unsigned chars for more details.

edited May 23 '17 at 11:58

Community

1
1

answered Jul 29 '15 at 16:50

Trebor Rude

1,904
1
21
31

Thanks @Trebor_Rude. Like your answer too. – packetie Jul 29 '15 at 18:35

Lightness Races in Orbit · Accepted Answer · 2015-07-29T16:58:00.373

3

No, std::string is std::basic_string<char, ...>, and there's nothing you can or should do to change that.

You could force your implementation's char to be unsigned (some compilers allow this) and stick assertions everywhere to prevent your code from compiling with a signed char type.

You could switch to std::basic_string<unsigned char, ...> or even std::vector<unsigned char>.

But, personally, I'd just compare the character against another character (instead of an integer):

if (s[1] == '\x8f')

(live demo)

I think this is better code anyway.

edited Jul 29 '15 at 16:58

answered Jul 29 '15 at 16:49

Lightness Races in Orbit

378,754
76
643
1,055

1

Comparing to a `char` constant rather than an `int` constant seems like the obviously right thing to do. – Jul 29 '15 at 17:02
Thanks @lightness-races-in-orbit. Should have thought of it. The trick of using `'\x8f'` is nice! – packetie Jul 29 '15 at 18:34
@codingFun: I wouldn't call it a "trick"; it is the default, obvious thing to do in the first place! :) (As Hurkyl said) – Lightness Races in Orbit Jul 29 '15 at 18:36
@codingFun: I suggest dropping that habit if you want people to understand you properly. :) – Lightness Races in Orbit Jul 29 '15 at 20:16
Will do. Thanks @lightness-races-in-orbit. – packetie Jul 30 '15 at 14:10

Hope std::string::operator[] could return unsigned char

2 Answers2

(live demo)