await_next_record(None) returns Ok(None) when JournalFiles::CurrentUser is used

[bddap]

Passing None to await_next_record() indicates an infinite timeout so a timeout should never happen, yet a timeout is indicated. await_next_record() sometimes returns Ok(None), indicating a timeout.

extern crate systemd;
use systemd::journal::{Journal, JournalFiles};

fn main() {
    let mut journal = Journal::open(JournalFiles::CurrentUser, false, false).unwrap();
    let record = journal.await_next_record(None).unwrap();
    assert!(record.is_some()); // Panics
}

[bddap]

Suggested Change

To start, the "No timeout" contract should be enforced through types.

pub fn await_next_record(&mut self,  wait_time: Option<Duration>) ->
Result<Option<JournalRecord>>

should be changed to

pub fn await_next_record(&mut self, wait_time: Duration) ->
Result<Option<JournalRecord>>

Then a new method can be added:

pub fn await_next_record_no_timeout(&mut self) -> Result<JournalRecord>

[codyps]

That seems reasonable, I'd have no problem accepting a change that did this.