文件模式

Apollo Fang — Wed, 31 Oct 2012 08:40:00 GMT

文件

C程序,将文件看做是连续字节序列,其中每个字节都可以单独读取.这个与UNIX环境中的文件结构是一致的.但ANSI C为了与其他的OS环境兼容(比如Windows),提供了两种文件视图: 文本视图和二进制视图.

文本视图和二进制视图

二进制视图中,文件中的每个字节都可以为程序访问.但文本视图中,程序中看到的内容与文件的内容可能不同.

例如MS-DOS文本文件用"\r\n"来表示行尾; Macintosh文本文件中用"\r"表示行尾; C程序使用"\n"表示行尾. So,如果C程序以文本视图模式处理一个MS-DOS文本文件,在读取文件时,就会将"\r\n"转换为"\n",在写入文件时,就会将"\n"转换为"\r\n". Macintosh同理.

说白了,两种视图的实现是一样的,只是在处理行尾或处理文件结尾时有点不同而已.采用某种视图打开文件时,注意一下就是了.

文件结尾和换行

文件读取数据的程序需要在达到文件结尾时停止.当到达文件结尾时,"getc()"函数会返回一个特殊值EOF.所以C程序只有在读取超出文件结尾后,才会发现文件的结尾.

为了避免读取空文件带来的问题,应该对文件输入使用入口条件循环(使用while或for,避免使用do...while)如下设计:

int ch;      // Watching EOF
FILE* fp;
fp = fopen ("_FileName", "_Mode");
ch = getc (fp);

while (ch != EOF)
{
    putchar (ch);
    ch = getc (fp);
}

上面的例子可以看做是一个框架,来进行文件结尾的判断.但ANSI C的两种模式,对于文件结尾有不同的解释.

如果文件以文本模式打开,C可以认出EOF标志文件结尾.如果以二进制模式打开,就会把EOF当做是文件中的一个字符.真正的文件结尾还在后面.文件的结尾,可能紧跟着EOF,当然,也可能用空字符填充文件使其大小为256(或其他数)的倍数.在DOS下不打印空字符.程序中包含了防止程序打印EOF字符的代码.

MS-DOS的文本文件用二进制模式和文本模式打开,C程序将看到下面的内容:

二进制模式打开文件C程序看到的内容:

line1\r\n

line2\r\n

line3\r\n

文本模式打开文件C程序看到的内容:

line1\n

line2\n

line3\n

例子 : 逆序输出一个文件内容

#include <stdio.h>
#include <stdlib.h>
#define MAC
#define CNTL_Z '\032' /* DOS Text File End Of File Flag */
#define SLEN   50
#define PATH
int main (void)
{
    unsigned char file[SLEN];
    unsigned char ch;
    FILE* fp;
    unsigned long count;
    unsigned long last;
    puts ("Enter Name Of File To Be Processed : ");
    gets (file);
    if ((fp = fopen (file, "rb")) == NULL)
    {
        printf ("Reverse Can't Open %s .\n", file);
        exit (1);
    }
    fseek (fp, 0L, SEEK_END);   /* Locate To End Of File */
    last = ftell (fp);
    for (count = 1L; count <= last; count++)
    {
        fseek (fp, -count, SEEK_END);  /* Back 1 Byte */
        ch = getc (fp);
        if (ch != CNTL_Z && ch != '\r')
        {
            putchar (ch);
        }
#if defined MAC || defined WIN32
        /* Macintosh || Windows */
        if (ch == '\r')
        {
            putchar ('\n');
        }
        else
        {
            putchar (ch);
        }
#endif
        putchar ('\n');
        fclose (fp);
        return 0;
    }
}

Apollo Fang 2012-10-31 16:40 发表评论

About C/C++ Data Type

Apollo Fang — Tue, 25 Oct 2011 09:28:00 GMT

   1) int Integer
   2) int*  A pointer point to an Integer
   3) int** A pointer point to a pointer what point to an Integer
   4) int* arr[N] An array that each element is a pointer point to an Integer
   5) int(*arr)[N] A pointer point to an array that its each element is an Integer and it have N elements.
   6) int (*fun)(int) A pointer point to a function what with an integer parameter and with an integer returning.
   7) int (*fun[N])(int) An array that have N elements and each element is a pointer point to a point that point to a type of function that have an integer parameter and with an integer returning.
   Maybe understand (4) type just like this :
   int*[N] arr;
It means "arr"'s type is int*[]. And how to catch the meaning of int*[]? At first it's an array, and each element of this array is a pointer point to an integer.
   And can understand (5) type just like this :
int[N]* arr;
It means "arr"'s type is int[]*. That's to show that this type is a pointer and this pointer point to an integer array its size is N.
   It means this variable value named "fun" and its type
   Well, the above are so complex, and we should follow the KISS principle that is Keep It Simple, Stupid.
Did you catch it ??

Tuesday, October 25, 2011

Apollo Fang 2011-10-25 17:28 发表评论

C++博客-Apollo Fang's Blog-随笔分类-C/C++

文件模式

About C/C++ Data Type