Test-read a random sample of the data in the files specified on the command line. Report the first sampled error location as a percentage of the sum of file sizes. Example of what I mean (in fact original motivation):

$ random-read VTS_01_1.VOB VTS_01_2.VOB VTS_01_3.VOB
-------------------------------------------------|
>>>>>>>>>>>>>>>>>>>>
VTS_01_2.VOB: hGetBuf: hardware fault (Input/output error)
error location: 42%

If those files, in that order, hold a DVD movie of 120 minutes, then 42% means that there is a reading error near 42% × 120 = 50 minutes into the movie. (It is not 42% into VTS_01_2.VOB; it is 42% into the concatenation of the files.)

This is also supported:

$ random-read /dev/sr0

Options:

• -h, --help: show help text
• -p, --prob: probability in M/N syntax
• -b, --bs: block size per read, in bytes

See default values in help text or source code.

Example:

$ random-read --prob=3/1000 --bs=2048 VTS_01_1.VOB VTS_01_2.VOB VTS_01_3.VOB