> I have been doing this since 2018-01 and had no clock-setting trouble across DST changes.
You have to be unlucky to see the issue. If WeeWX decided to set the time on the console within the hour of setting the clock back one hour when entering standard time, the clock setting would be one hour off. Having said that, I believe this is a bit off topic as this issue doesn’t result in memory having to be cleared; and the issue being reported in this thread seems to require clearing memory. > On Mar 28, 2021, at 7:11 AM, Greg Troxel <[email protected]> wrote: > > > John Kline <[email protected]> writes: > >> I have seen an issue when weewx calls set_time within the hour after >> falling back to standard time (because there is no way in the command >> to state the time is in DST and the console assumes standard time). I >> got around this issue by forking the driver and adding code such that >> set_time does nothing when set_time is called during these >> transitions. As such, I don’t see any issues during the clock >> changes. > > I would agree that weewx being reluctant to set the console time would > be good. Perhaps some notion of only doing so if there is a new offset > for 6 hours, and maybe even if it's consistent for 6 hours. > >> The issue being reported here sounds different. Are you guys saying >> you *always* have this problem going in and out of DST? Do you have >> the correct DST setting in the console? Even when I encountered the >> issue above, I didn’t have to clear memory in my logger. Tom, do you >> see this issue? Do you have to clear memory in your logger every six >> months? > > I have a VP2 with a Davis serial logger, and run the console and a RPI3 > running weewx (NetBSD, python2) in "Eastern" time, which means EST5EDT. > My archive interval is 5 minutes. > > I have been doing this since 2018-01 and had no clock-setting trouble > across DST changes. > > I do see problems in graph generation on the day of moving to DST, but > <fuzzy_memory>I think that's about calling mktime with time values that > are in the DST gap. This is specified by POSIX to fail: > > https://pubs.opengroup.org/onlinepubs/9699919799/functions/mktime.html > > and I haven't gotten around to adjusting the code to either not do that > or deal gracefully with the -1.<fuzzy_memory> > > > (While it would be nice if weewx could collect and store data in UTC and > graph it in localtime, keeping the station in UTC also means that daily > highs/lows are in UTC which is not how weather services do things.) > > -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/weewx-user/AD4E2005-6F62-46EA-AC96-39F8370A7797%40johnkline.com.
