Re: [tz] strftime %s

Paul Gilmartin wrote:

Seconds since the epoch should not depend on TZ. The epoch is 0 UTC, not 0 local.

Right. So it seems to me that strftime should *not* support %s, or at least, not unless struct tm is expanded to include a field holding the original time_t value. (But, yes, I saw Paul's earlier message mentioning the next POSIX draft talking about it.)