How to read/write non ascii characters? - c

Question

I am trying to input a file and have it be printed each character at a time but some characters are ignored.
I'm assuming that that's because they are non ascii chars and the fgets doesn't know what to do with them since the buffer is made of chars.
int main() {
while(1)
{
char str[50];
if (fgets(str, 50, stdin) == NULL)
{
exit(0);
}
for(int i = 0; str[i] != '\n' ; i++)
{
printf("%lc", str[i]);
}
printf("\n");
}
return 0;
}
I have a file with
ALICE’SE’E’E’E’E’E’
but my code outputs it as
ALICESEEEEEE

Answer 1

if you use fgets(), that will probably have some undefined behaviour with input characters like \0, as they are internally used by the string functions to mark the end of the data in a string.
fgets()is a text oriented function that reads input until it finds a new line character \n. It then places a \0 after it, so you know where the string ends.
But with binary data, you can get control characters, even null characters in the middle of the data stream, making sometimes characters to dissapear on output (because they have been read, but your code doesn't go further when some of these are encountered later on writing).
If you want to allow all binary characters, you have several approaches here:
Use the binary stream counterparts from stdio: fread(3) and fwrite(3) functions allow you to read binary data as well as text:
#include <stdio.h>
#include <stdlib.h> /* for EXIT_* constants and exit() */
#define N (8192) /* buffer size (guessed, probably not optimum) */
int main()
{
ssize_t n;
char buffer[N];
while((n = fread(buffer, sizeof buffer[0], N, stdin)) > 0) {
ssize_t nout = fwrite(buffer, sizeof buffer[i], n, stdout);
if (nout != n) { /* error */
fprintf(stderr, "Error writing stdout\n");
exit(EXIT_FAILURE);
}
}
exit(EXIT_SUCCESS);
} /* main */
Use the simple Kernighan & Ritchie's sample from "The C programming language" book (I've added some error processing code):
#include <stdio.h>
#include <stdlib.h> /* idem. */
int main()
{
while((c = fgetc(stdin)) != EOF)
if (fputc(stdout) == EOF) {
fprintf("fputc error\n");
exit(EXIT_FAILURE);
}
}
if (ferror(stdin)) {
fprintf("fgetc error\n");
exit(EXIT_FAILURE);
}
exit(EXIT_SUCCESS);
} /* main */
or use the standard UNIX system calls:
#include <unistd.h> /* for prototypes for read()/write() syscalls */
#include <stdio.h>
#include <stdlib.h>
#define N (8192) /* guessed buffer size */
int main()
{
char buffer[N];
ssize_t n;
while ((n = read(0, buffer, sizeof buffer)) > 0) {
ssize_t nout;
nout = write(1, buffer, sizeof buffer));
if (nout != n) {
fprintf(stderr, "write: %s\n", strerror(errno));
exit(EXIT_FAILURE);
}
}
if (n < 0) {
fprintf(stderr, "read: %s\n", strerror(errno));
exit(EXIT_FAILURE);
}
exit(EXIT_SUCCESS);
} /* main */
but beware that probably the most efficient code you'll get can be the original character oriented sample from the K&R book, as stdio will select an optimum buffer size that will make it run faster, despite of the higher number of loop executions.
note
Anyway, your output will be far to be what you want, as some control characters are not output to the terminal, but interpreted as control characters (most popular being \n, which makes the terminal to continue on the next line) You have also to deal with this.
Even if you read multibyte characters as single byte, you can process those with the examples given, as a character that uses two bytes, will be read as two, but on printing, those will become the single char the terminal should display. As long as you apply no transformation to the data flow, there will be no difference in output with the sample code snippets you have above this.

Answer 2

You are using fgets, which deals with chars, and a char array (char str[50]).
But a printf formatter %lc which is for wide char.
If you want to input wide char, you need to use fgetws, and an array of type( wchar_t str[50])