From: Dipl. I. S. G. B. <se...@us...> - 2024-12-08 00:57:51
|
Basically it has nothing with the support of leap second - only the free-scan has been fixed in 8.6 for the time larger than 23:59:59, whereas previously (due to return -1 from ToSecond) it jumped to previous day. Now it works similar to format-scan (day += (time / 86400); time %= 86400). I emphasize - any time, also 99:99:99, not the leap second or 24:00 only. The ticket for the leap-second support [f2b5f89c0d] [1] is still open (because of the complexity of possible solution, however it is rather something about new option -validate of 8.7/9.0). Regarding possible change note, how is about: * [clock scan] fixed wrong time conversion by free-scan caused day decrement for time larger than 23:59:59, now the result is the same as for format-scan. (actually similar the message of commit [595fad24d70e1693] [2]) Regards, Serg. 07.12.2024 18:25, Harald Oehlmann wrote: > Dear Don, > rc1 release is ok for me, thanks for the fix. > > About the release notes: > > * [clock scan] support for leapsecond. > > AFAI understood, the leapsecond support is only for input and will jump to the next real second. This wording suggests full support for me, even if "clock scan" is mentioned. > Possible wordings: > "limited support for leapsecond" > or > "accept leapsecond as alias for next second" > > ...but I may be wrong... > > Thank you, great work ! > Harald Links: ------ [1] https://core.tcl-lang.org/tcl/info/f2b5f89c0d [2] https://core.tcl-lang.org/tcl/info/595fad24d70e1693 |