Investigate using EpochNanoseconds and implement wherever possible
#119
Labels
C-internal
Internal library improvements
E-easy
Easy issue to fix - Good for newcomers
Good First Issue
Comments
nekevss
added
E-easy
nekevss
added a commit
that referenced
this issue
Dec 8, 2024
…er (#128) This fixes a bug that was introduced by #116. General gist is that the in the proposal [`5.5.4 ISODateTimeWithinLimits ( isoDateTime )`](https://tc39.es/proposal-temporal/#sec-temporal-isodatetimewithinlimits) there is a call to `GetUTCEpochNanoseconds`. However, this value is not checked to be valid, and instead of constraining to the NS_MAX_INSTANT/NS_MIN_INSTANT +/- NS_PER_DAY, respectively, we were constraining to NS_MAX_INSTANT/NS_MIN_INSTANT. There was a limit test on datetime, but we were only asserting the negative case, not the positive case, so the regression was not caught. So adjusted to handle the potentially invalid value, adjusted the preexisting datetime test, and added a date limit test based off the test262 limit test. EDIT: Potentially relevant context to #119
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
In #116, the
EpochNanosecondsnew type was added. The introduction ofEpochNanosecondswas fairly bare, as further mentioned in #115. We should look into usingEpochNanosecondsmore throughout the codebase where it makes sense according to the proposal.The text was updated successfully, but these errors were encountered: