shred: implement and test feature --random-source

[BenWiederhake]

Fixes issue #5711.

This PR introduces the --random-source feature, and implements it faithfully: Copy the data from the source into the file-to-be-shredded. Note that the "random" data is not modified in any way. This is good, because no deterministic transformation can ever increase the entropy in random data; and it also makes it much simpler to verify that shred uses the specified source.

Note that the second of the two new tests is broken because the functionality was always broken, but only now becomes more visible. See #7947.

Note that this PR is not yet sufficient to pass shred-passes.log. A future PR will have to adjust the test in order to accept our implementation.

Note a final distinction: The GNU implementation also takes the "shuffled" order of passes from the random source (making it truly deterministic when --random-source is provided). However, since we don't case about determinism (only the ability to provide a different source of entropy), I don't think we should follow GNU behavior here. What do you think?

[github-actions]

GNU testsuite comparison:

Skip an intermittent issue tests/timeout/timeout (fails in this run but passes in the 'main' branch)

[BenWiederhake]

CI seems to fail due to an unrelated issues, looks like it might be fixed by #7945 and #7941.

[sylvestre]

About the gnu test, looks good to me!

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/misc/tee (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/timeout/timeout (passes in this run but fails in the 'main' branch)

[BenWiederhake]

Thanks for the force-push! Now that all is green, can it be merged?

[sylvestre]

thanks and sorry for the latency!