About

Issue #7870

This commit changes the default block size for the shred utility to make it more consistent with the GNU coreutils version.

Before

echo a > foo && cargo run -q --features shred shred -v -n1 foo && du -b foo
shred: foo: pass 1/1 (random)...
65536	foo

After

echo a > foo && cargo run -q --features shred shred -v -n1 foo && du -b foo
shred: foo: pass 1/1 (random)...
4096	foo

Expected

echo a > foo && shred -v -n1 foo && du -b foo
shred: foo: pass 1/1 (random)...
4096	foo

Please note: The original shred utility uses fstat along with the macroses to determine the block size. Therefore, the block size is not constant and may vary across platforms. The most common sizes are 512 bytes and 4 KB. Here, I started by changing the constant and collecting test results and observations. However, to replicate the behavior of shred, deeper modifications are required. Please feel free to share your thoughts and opinions on which approach is preferable: simply changing the constant or moving block size detection to runtime.

Thank you