为什么在if语句中对字母使用ASCII代码，而不是使用常规char数据类型？

Question

I'm reading some lines of code written by a colleague and he has written code the following:

int keych = 0;
keych = _getch();
if (keych == 104 || keych == 104 - 32) //Help
{
    printf(cHelp);
}

What I do'nt understand is that why does he not use simple char data type, this helps to write less code. What are the advantages of using that method instead of char data type to get user input?

Answer 1

Using

if (keych == 104 || keych == 104 - 32) //Help

to decide whether the input character is 'h' or 'H' is bad.

1. It is difficult to read.
2. It will work only on systems that use 104 and 72 for encoding h and H (such as ASCII).

It will be better to use:

if (keych == 'h' || keych == 'H')

or

if ( tolower(keych) == 'h' )

Answer 2

Oups, unicode and ascii are not exactly the same thing, or more exactly ascii is a subset of unicode (127 first code points). 104 or (0x98 in hexa) is the ASCII code for 'h', so no unicode is involved here. The only advantage of using ascii code is that it will break on a non ASCII system and there could still be EDCDIC system around... and additionaly as you were said in comment, 104 is more cryptic than 'h'.

TL/DR: never replace characters with their ascii code point unless you have serious reason to do so.

Answer 3

It's good practice if not very intuitive for newbies to store character variables and pass them about as int. The standard library does that, fputc() takes and int not a char. In C, sizeof('a') yields 2 or 4 depending on the size of an int, not 1 for sizeof(char).

The reason is that int can hold EOF, which is not a char and means end of input or error condition. There's also the effect that the code scales to unicode quite nicely.

Answer 4

The only time it's reasonable to refer to (printable) characters by their code points in an encoding is when you know that the data you are handling are not in native form for your program - this is only likely if you're implementing code conversion, either as a program like iconv or as a conversion layer in your program's i/o classes. Normally you don't need to write such things yourself; that's what libraries are for.

Even if you are writing conversions, it helps to give names to the values you're using, eg

static const int UNICODE_LATIN_SMALL_LETTER_H = 0x68;

In the particular case in the question, assuming that _getch() is something like the Curses getch() , then you should expect it to be represented consistently with the character constants in your program source character set.