iconv 库不正确地将 UTF-8 转换为 KOI8-R

Question

我正在尝试使用 GNU iconv 库将 UTF-8 编码字符串转换为 KOI8-R。我的最小示例是

#include <iconv.h>
#include <stdio.h>
#include <stdlib.h>

int main() {
    /* The letter П in UTF-8. */
    char* buffer = "\xd0\x9f";
    size_t len = 2;
    /* Note: since KOI8-R is an 8-bit encoding, the buffer should only need a length of 1, but
     * iconv returns -1 if the buffer is any smaller than 4 bytes,
     */
    size_t len_in_koi = 4;
    char* buffer_in_koi = malloc(len_in_koi+1);
    /* A throwaway copy to give to iconv. */
    char* buffer_in_koi_copy = buffer_in_koi;
    iconv_t cd = iconv_open("UTF-8", "KOI8-R");
    if (cd == (iconv_t) -1) {
        fputs("Error while initializing iconv_t handle.\n", stderr);
        return 2;
    }
    if (iconv(cd, &buffer, &len, &buffer_in_koi_copy, &len_in_koi) != (size_t) -1) {
        /* Expecting f0 but get d0. */
        printf("Conversion successful! The byte is %x.\n", (unsigned char)(*buffer_in_koi));
    } else {
        fputs("Error while converting buffer to KOI8-R.\n", stderr);
        return 3;
    }
    iconv_close(cd);
    free(buffer_in_koi);
    return 0;
}

（除了当我的 KOI8-R 缓冲区小于四个字节时不工作，尽管它应该只需要一个字节）错误地打印 d0（KOI8 中 'П' 的正确编码-R 是 f0).

iconv 从命令行给出了正确的答案（例如，echo П | iconv -t KOI8-R | hexdump），那么我在使用它的 C 接口时做错了什么？

Answer 1

您将 "to" 和 "from" 字符集参数混淆为 iconv_open。恰好 KOI8-R 中槽 D0 中的字符具有 D0 作为其 UTF-8 编码的第一个字节。

iconv 库不正确地将 UTF-8 转换为 KOI8-R

iconv library improperly converts UTF-8 to KOI8-R

c

character-encoding

iconv