feat(core): not translate to "interrupted" std IO error

[dentiny]

Which issue does this PR close?

N/A

Rationale for this change

Hi team, I found when I'm using synchronous IO, it often retries more than I expected. See my regression test for details.

I think the root cause is

• For now, we map most of the OpenDAL errors into Interrupted error
• std::io::Read is documented to retry on interrupted error: https://web.mit.edu/rust-lang_v1.25/arch/amd64_ubuntu1404/share/doc/rust/html/std/io/trait.Read.html#errors
• code reference: https://doc.rust-lang.org/src/std/io/mod.rs.html#479

// Note that we don't track already initialized bytes here, but this is fine
// because we explicitly limit the read size
  let mut cursor = read_buf.unfilled();
  let result = loop {
    match r.read_buf(cursor.reborrow()) {
      Err(e) if e.is_interrupted() => continue,
        // Do not stop now in case of error: we might have received both data
        // and an error
        res => break res,
      }
};

I think it's not ideal because

• Extra retries are attempted even for persistent errors, which we shouldn't retry upon
• When retry layer is injected, extra retries are proceeded, even retry attempts have been exhausted
• There's no retry limit control -- it just retries forever

What changes are included in this PR?

In this PR, I updated the std IO error from "interrupted" to "other", so it's directly propagated up; I checked OpenDAL errors, no errors should be mapped to "interrupted".
If retry is necessary, users could configure retry wrapper.

Are there any user-facing changes?

No

AI Usage Statement

AI helped me find out the source code.

[Xuanwo]

Makes sense, thank you for fixing this