Question

I have a block of uncommented code I’m trying to understand. The comments are my own.

//create an array name header that holds 4 bytes
byte header[] = new byte[4];
int len = 0;
int c = -1;

for(; len != 3; len += c)// run loop till len = 3
{
  try
  {
    //first run of the loop following should be true
    //read 3 bytes and save into header array starting at 0
    // c = number of bytes read (most likely 3 after first run of loop)

    c = is.read(header, len, 3 - len);

  }
  catch(Exception e)
  {
     System.err.println("read header error " + e.getMessage());
     displayErrorMessage(e);
  }
  if(c == -1)
    return null;

}

This code is reading an input stream but I’m not sure how many times it will loop.
I tried running through the loop on paper, replacing len and c each time and after 7 iterations len still did not equal 3.

During second run of the loop, len should equal -1 and c should equal 3. The -1 should make the read method throw an IndexOutOfBoundsException because you are trying to save the byte read from the stream to header[-1]. So since an exception is thrown, the value of c would remain equal to 3 from first run of the loop.

Am I right in assuming that since the exception is caught, the for loop does not exit?

Answer 1

The number of times the loop will be repeated is not deterministic, since read() might read 3 bytes at the first run, and terminate – or it might read 1 byte at a time, and repeat 3 times.

However, as the comment says – it is very likely that read() will read during the first iteration 3 bytes and the loop will have only one iteration.

For this case what happen is the following:

1. check if len < 3 – it is
2. read 3 bytes to the array, and set c = 3
3. check if c == -1 – it is not
4. end loop: set len = len + c = 3
5. check if len < 3 – it is not: terminate the loop