On Mon, 30 Nov 2020 21:31:18 GMT, Weijun Wang <wei...@openjdk.org> wrote:
>>> >>> >>> > Based on CurveDB line 141, it seems we uses the first one and discards >>> > the rest? Are you sure we keep the last one? >>> > Don't have answers for your questions though. >>> >>> Before this change both sect163k1 and sect163r2 have "BD" so both will be >>> added to lengthMap and the 2nd one wins out. >> >> Hmm, ok, I re-read the code and it's true that both are added to the >> lengthMap. Actually, non-PD/non-BD entries are added to the length map too >> if no entries for such length in length map is present at time of the add() >> call, right? > >> > > Based on CurveDB line 141, it seems we uses the first one and discards >> > > the rest? Are you sure we keep the last one? >> > > Don't have answers for your questions though. >> > >> > >> > Before this change both sect163k1 and sect163r2 have "BD" so both will be >> > added to lengthMap and the 2nd one wins out. >> >> Hmm, ok, I re-read the code and it's true that both are added to the >> lengthMap. Actually, non-PD/non-BD entries are added to the length map too >> if no entries for such length in length map is present at time of the add() >> call, right? > > Yes, I think that's what "if there are multiple matches for the given size, > an arbitrary one is returns" in the comment of `lookup(int)` means. If there > is a PD/BD with the same length, the previously added will be substituted > anyway. Alright, I have no more comments/questions. Thanks. ------------- PR: https://git.openjdk.java.net/jdk/pull/1485
