Programming in C

Lecture Outline

Aims and Objectives
Moving the file pointer
Handling File Errors
Determining the file type

Aims and Objectives

• Moving the file pointer

• Handling File Errors

• Determining the file type

Moving the file pointer

• Don't confuse the file pointer with pointer to FILE!

• As you read/write a file, the file pointer is automatically incremented to indicate where the next character is to be written to or read from

• When you open a file for reading or (over-)writing, the file pointer is positioned at the beginning of the file

• When you open a file to append to it, the file pointer is positioned at EOF

• There are two main functions to obtain the current value of the file pointer and to set the value of the file pointer

• long ftell(FILE *stream);

  ftell() takes a pointer to file and returns a long integer indicating the current file pointer or -1 if an error occurs. errno is set to indicate the error that occurred.

• int fseek(FILE *stream, long offset, int whence);

  fseek() takes a file pointer, the new offset, and an integer which indicates how the new file position is to be calculated. whence can be one of:

  SEEK_SET: set the file pointer to offset
  SEEK_CUR: set the file pointer to current position + offset
  SEEK_END: set the file pointer to EOF + offset
  

  fseek() will return an integer 0 if successful, and -1 otherwise. errno is set to indicate the error that occurred.

• void rewind(FILE *stream);

  Sets the file pointer back to the beginning of the file. Equivalent to (void) fseek(FILE *stream, 0, SEEK_SET).

Exercise

Open a file which doesn't exist (or overwrite an existing file) for read/writing.

Print the initial value of the file pointer. Set the file pointer to the end of the file, and printer the current value of the file pointer. Is the new file pointer different from the initial one?

Set the file pointer to 50 beyond the end of file, write a character, go back to the beginning of the file and print the values of each byte in the file (you may have to print the ASCII values of the bytes to see them).

Try to read beyond the EOF marker, printing the character read.

Try to read before the beginning of the file, printing the character read.

Handling file errors

• There are two main file error handling functions:

  int ferror(FILE *stream);

  int feof(FILE *stream);

• ferror() returns a non-zero integer value when an error previously occurred on the stream, and 0 otherwise

• feof() returns a non-zero value when EOF has been detected on the stream, and 0 otherwise

  If feof() returns a value other than 0, use perror() to print the error message

  If fgetc() and fputc() return an EOF, you must check that feof() returns 0 to be sure that an error has not occurred

• Use void clearerr(FILE *stream); to clear the ferror() and feof() errors

Determining the File Type

• Files in UNIX are generally binary or text files

• Binary files can be created by compilers (object files), for example

• Common object files normally begin with a 20-byte header which contains a magic number which indicates the file's type. If you are interested in finding out more about this, see the man pages for filehdr and magic

• UNIX has a utility called file which runs a series of tests against a file to determine its type

• We're going to make a system call to file from within a C program to see if the file we're interested in is a text file

• First, create a couple of files containing
  • Garbage entered from the keyboard (ascii text)

  • A sentence or two using real English words (English text)

• Next, from the UNIX command line prompt, check the type of the files using the file command

• int system(char *string);

  To invoke a system call, such as to file, from within a C program, we can use the system() command

• string will contain the command we would have typed in at the UNIX prompt - except that quotation marks must be escaped!

  E.g., system("echo \"Hello there\"); corresponds to echo "Hello there" typed at the command line

• system() will return integer 0 if it is successful, and a non-zero value otherwise

Exercise

The UNIX file command simply prints the file type to standard output. What UNIX command can be used to determine if the output contains a specific string, e.g., "text"?

How can we conjoin the two UNIX commands into a single instruction?

How can we prevent or dispose of the output from the conjoined command, so that it doesn't appear on the screen?

How can we ensure that the string we're searching for is at the end of the output from the conjoined command? E.g., looking for the string "text" in

text.bin: ELF 32-bit MSB executable SPARC Version 1, dynamically linked, not stripped

will incorrectly identify the file as a text file...