Unicode & C/C++/Java

[tjt]

Am facut un mic programel in C care primeste ca argument un sir de caractere si il afiseaza intr-un format ce permite folosirea sa in C/C++/Java.

Nu am cunostinte foarte mari in C,probabil putea fi facut si mai eficient sau/si mai scurt.

De exemplu pentru sirul ??? se va afisa \u3088 \u597d \u305d .

#include <stdio.h>
#include <string.h>

int getCodePoint(char *buffer,int ind,int noByte){
    if(noByte==1)
      return buffer[ind];
    int x;
    x=buffer[ind]&(0xFF>>(noByte+1));
    while(--noByte!=0){
        x=(x<<6)|(buffer[++ind]&(0xFF>>2));
    }
    return x;
}

void check(char *buffer,int dim){
    int ind=0;
    int codePoint;
    while(ind<dim){
        if((buffer[ind] & 0x80 ) == 0){
            codePoint=getCodePoint(buffer,ind,1);
            ind++;  
        }
        else if((buffer[ind] & 0xE0 ) == 0xC0){
            codePoint=getCodePoint(buffer,ind,2);
            ind+=2;  
        }
        else if((buffer[ind] & 0xF0 ) == 0xE0 ){
            codePoint=getCodePoint(buffer,ind,3);
            ind+=3;  
        }
        else if((buffer[ind] & 0xF8 )== 0xF0 ){
            codePoint=getCodePoint(buffer,ind,1);
            ind+=4;
        }
       printf("\\u%04x ",codePoint);
    }
    printf("\n");
}

int main(int argc,char **argv){
    if(argc!=2){
      printf("usage: ./getCode string");
    }
    check(argv[1],strlen(argv[1]));
    return 0;
}

Edited June 11, 2014 by tjt