On 20-12-05 08:57, Paul-Andrew Joseph Miseiko wrote: > A leap second may only be introduced at the end of a UTC month; therefor > the leap indicator may be set at any time within the UTC month with an > impending leap second.
The limitation to introduce leap seconds at the end of the month only, was mentioned in RFC 958 and RFC 1059, but those have been obsoleted by RFC 1119 and RFC 1305, which do not restrict leap seconds to the end of the month. As it was standard in a previous version of the RFC, IMHO an implementation may be (default) configured to ignore the leap bits on incoming responses on all days exept the last day of the month. Perhaps limited to responses received from old versions. However, it should send correct responses to clients it serves. Be tolerant in what you accept; be strict in what you provide. I also noted that my server sent out wrong leap info once it received that info from a server (that was before December!). Therefore, I've set up my firewall to filter replies from servers with wrong leap info... Arnold &:-) _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
