du: propagate errors from --exclude-from instead of panicking

[rossilor95]

Fixes #11924
Fixes #11925

du -X <file> panicked on two failure modes when reading the
--exclude-from file:

1. Missing file (du -X with a nonexistent file panics #11924):
   $ du -X nonexistent-file
thread 'main' panicked at src/uu/du/src/du.rs:759:37: no such file: Os { code: 2, kind: NotFound, ... } Aborted (core dumped)
   GNU du prints a clean error and exits non-zero.

2. Per-line I/O error (du -X /proc/self/mem panics #11925), e.g. a pseudo-file that fails on read:
   $ du -X /proc/self/mem
thread 'main' panicked at src/uu/du/src/du.rs:763:20: Could not parse line: Os { code: 5, kind: Uncategorized, ... }

Both now exit with code 1 and print to stderr (du: No such file or directory, du: Input/output error, du: Permission denied — message follows the OS error kind).

The call site was also switched from .flat_map(file_as_vec) to a
.map(...).collect::<UResult<_>>()? chain, because flat_map over a
Result-returning function silently drops Err (zero items) — hiding
any I/O failure instead of reporting it.

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test tests/unexpand/bounded-memory is now passing!
Skip an intermittent issue tests/pr/bounded-memory (was skipped on 'main', now failing)

[cakebaker]

Can you please add a test to ensure we don't regress in the future? Thanks.

[rossilor95]

Can you please add a test to ensure we don't regress in the future? Thanks.

Sure! Done

[rossilor95]

    thread 'test_du::test_du_exclude_from_read_error' panicked at tests/by-util/test_du.rs:1535:10:
    'du: Input/output error
    ' does not contain 'du: /proc/self/mem: Input/output error'

Sorry, my bad: forgot to double check the error message 😅 It's fixed now

[cakebaker]

Thanks!