....:....+....:....+....:....+....:....+....:....+....:....+....:....+....:....
....:....+....:....+....:....+....:....+....:....+....:....+....:....+....:....
	

This output has one problem: When looking at data, you can't tell which line you are seeing. So there's a -n option that generates line numbers:

1 .:....+....:....+....:....+....:....+....:....+....:....+....:....+....:....
2 .:....+....:....+....:....+....:....+....:....+....:....+....:....+....:....
3 .:....+....:....+....:....+....:....+....:....+....:....+....:....+....:....
	

The -n option produces output that is easier to identify, but because it calls sprintf(3), it will be 3 or 4 times slower than without the -n option.

The default output file size is 10000 bytes; the default output line length is 80 bytes. The apparent line will be one byte shorter, because the last byte of each line will be a newline.

gendata 5000000 >5MBtest

This generates 5 megabytes of data and writes it to "5MBtest".

gendata -n 8000000 1024 | cscp - tcp://foo.bar.com/8x1024.txt

This generates 8 megabytes of data and pipes it to cscp, which sends it to the csfd server on foo.bar.com, to be written to the file "8x1024.txt". The data will have initial line numbers, and will consist of 1023-byte lines (plus a newline after each line).