UMBC CMSC202, Computer Science II, Spring 1998, Sections 0101, 0102, 0103, 0104 and Honors

Thursday January 29, 1998

Assigned Reading:

• A Book on C: 9.1 - 9.6 and 11.7
• Programming Abstractions in C: 2.6

Handouts (available on-line): Project 1

Topics Covered:

• We compared the functionality of scanf() versus GetInteger() from the Roberts library. Notice that GetInteger() will continue prompting the user for input until the user enters an integer correctly. The scanf() function does not do this, but does provide enough information so you could do the error checking yourself. However, this error checking is non-trivial to accomplish correctly. Part of the reason for the difficulty is that UNIX/C programs tend to be written with the idea that the program can be executed without a user sitting at the terminal.
  • A simple program using scanf() and sample run.
  • A simple program using GetInteger() and sample run.

• The scanf() function is an elaborate pattern matching utility:
  • You can match strings in the input that are not assigned to any parameters.
    Program and sample run.
  • You can match for strings composed of letters in a specified set. For example, %10[a-zA-Z] specifies 10 characters of lower and upper case letters.
    Program and sample run.

• Part of the difficulty with scanf() comes from the fact that when a pattern match fails, the input stream is left at the point of failure. That is, the next character of the input stream is the first character where failure is detected. This was illustrated in some of the previous examples. The next program shows the difficulty this creates when trying to read in two integers on two lines. Sample run.

• A naive attempt to continue pestering the user until the user enters a correct integer can result in an infinite loop. Program and sample run.

• One way to fix this problem is to use the getchar() function to consume the rest of the input line, up to and including the carriage return. Program and sample run.

• The scanf() and printf() functions have analogs that work with files called fscanf() and fprintf(). Here we need to use fopen() to open the files for reading or writing first. The pointer returned by fopen() should be checked to determine whether the file was successfully opened. Program and sample run.

• When we print an integer using the printf() or fprintf() function, this converts an integer to a string which is then displayed in the terminal or written out to a file. In order to write the binary value of the integer to a file, we need to use the fwrite() function. The next program writes an array of integers to a file called "binary" (sample run). Note that because the file is a file of integers and not a file of text (UNIX lied), you must use a utility such as the "octal dump" utility od to examine the file.

• The reverse of fwrite() is fread(), which we can use to read in the integers from the binary file. Program and sample run.

• The fwrite() and fread() functions can also be used to write and read entire structures. You will need to this feature in Project 1.
  • Program to write out a structure and sample run.
  • Program to read in a structure and sample run.

• You should also consult man pages for fwrite() and fread().