prisma-risk / tsoracle / 26349602163

95%

push

github

fix(core): saturate SystemClock::now_ms instead of truncating (#249) (#272)

`d.as_millis() as u64` silently wrapped the u128 millisecond count on
overflow: a far-future clock would emerge as a small value, making the
allocator perceive wall time far below the fence floor and stall window
advance with no log, metric, or error.

Convert with a saturating `u64::try_from(...).unwrap_or(u64::MAX)` so a
far-future clock drives visible window exhaustion rather than masquerading
as the distant past. The pre-Unix-epoch branch keeps mapping to 0, which is
the faithful saturate-low value for a past-pinned clock; document the policy
at both ends.

10 of 10 new or added lines in 1 file covered. (100.0%)

10448 of 11027 relevant lines covered (94.75%)

549854.99 hits per line