Back to DFS's C Page

Most programs that you run from a prompt in an xterm window allow you to specify extra information that the programs can use. This short page illustrates how you could get filenames the way cp and mv do.

This raises two issues: How does a program obtain information from the command line? How does the user know how to use the program?

Sample code is provided.

The mechanism for accessing the arguments is quite simple. Consider the following command line which could be used to get the program filecopy1 to make a copy called xxx.c of the file prog.c.

filecopy1 prog.c xxx.c

The system counts the program name and all items terminated by whitespace (in this case for a total of three) and sends this value to main() as the first argument. It also creates an array of null-terminated strings and passes a pointer to the zero^th as the second argument.

The definition of main() without the body of the source takes the following form.

void main(argc, argv)
int argc;	/* the number of arguments */
char *argv[];  /* array of pointers to the argument strings */
{
} 

argv[0] points to the name of the program; argv[1] points to the string prog.c; and argv[2] points to the string xxx.c. These pointers are then available so that the strings can be accessed inside main().

Back to Top

Since the user receives no explicit prompts, the program must make decisions given the information provided from the command line. Consider the following if structure:

   if( (argc < 3) ||
     (strncmp(argv[1], "--help", 6) == 0) )
   {
      printf("Usage: filecopy1 SOURCE DEST\n");
      exit(1);
   }

If the program is provided with fewer than three arguments, it cannot proceed. It is also traditional that if the first argument after the program name is --help, the program responds with a line of output explaining how to use the program.

Back to Top

Abbreviated Sample Code to Copy a File

The following complete program demonstrates how to access and process the command line information. The purpose of the program is to create a copy of a file. N.B. There is no bulletproofing included.

/* Copy a file */
/* filecopy1.c */
#include <stdio.h>

void main(argc, argv)
int argc;	/* the number of arguments */
char *argv[];  /* array of pointers to the argument strings */
{
   int ch; /* input character */
   char inputfile[80], outputfile[80];
   FILE *fpin, *fpout; /* input & output struct pointers */

   if( (argc == 1) ||
     (strncmp(argv[1], "--help", 6) == 0) )
   {
      printf("Usage: filecopy SOURCE DEST\n");
      exit(1);
   }
   /* get the filenames specified by the user */
   strcpy(inputfile, argv[1]);
   strcpy(outputfile, argv[2]);

   /* open the files */
   fpin = fopen(inputfile, "r");
   fpout = fopen(outputfile, "w");

   /* copy the file char by char */
   while( (ch = fgetc(fpin)) != EOF)
   {
      fputc(ch, fpout);
   }

   /* close the files */
   fclose(fpin);
   fclose(fpout);
}

Back to Top