Bhairitu wrote: > Alex Stanley wrote: > >> --- In FairfieldLife@yahoogroups.com, Bhairitu <noozg...@...> wrote: >> >> >>> authfriend wrote: >>> >>> >>>> --- In FairfieldLife@yahoogroups.com >>>> <mailto:FairfieldLife@yahoogroups.com> , Bhairitu <noozguru@> wrote: >>>> >>>> >>>> >>>>> Most likely there was a delay in delivering the emails >>>>> to Alex's server until after Sunday's count. YG has been >>>>> flakey that way lately. >>>>> >>>>> >>>>> >>>> No, if you look at the Post Counts for Saturday and >>>> Sunday, you'll see that the Sunday count (as I keep >>>> saying!) was made only a little over four hours after >>>> the Saturday count, so it picked up only the 12 posts >>>> made during those hours. >>>> >>>> Saturday's count (posted at 7:15 pm EST): >>>> 67 messages as of (UTC) Sun Mar 14 00:14:36 2010 <---- >>>> >>>> Sunday's count (posted at 8:21 pm EST): >>>> 79 messages as of (UTC) Sun Mar 14 04:28:33 2010 <---- >>>> >>>> Sunday's count should have been as of 00:-something >>>> UTC Monday, March 15. There should have been >>>> approximately 24 hours between the two counts. They >>>> were posted to FFL about 25 hours apart, but the >>>> times the counts actually took place are only a little >>>> over 4 hours apart. The Sunday count was made around >>>> 4:30 a.m. UTC on Sunday; it should have been made >>>> shortly after midnight UTC, i.e., on Monday. >>>> >>>> Usually the glitches are Yahoo's, but this time it >>>> looks as if it's the Post Count's glitch. >>>> >>>> Does Alex tell the program when it should stop counting >>>> each time, or is that built in? Because if it's Alex, >>>> it looks like he may have told it to stop counting >>>> around midnight CST instead of midnight UTC. >>>> >>>> >>> It is set as an automated task that is supposed to go off at about 15 >>> minutes after midnight UTC to account for any latency in delivery. >>> The time stamp of the post count shows it was sent at 00:15 UTC (5:15 >>> PDT) but I don't know why the script ran at an earlier time. Perhaps >>> the computer Alex was running it got flummoxed over the DST change. But >>> sometimes there are delays and message still has the correct time stamp >>> regardless of when it comes in. >>> >>> >> The computer running the script is set to UTC so that I don't have to mess >> with the computer clock or the automated event scheduler. >> > > And the log file was probably nuked so we can't see what happened. Next > time save it off somewhere and maybe there was an error in it. But then > why it would run at 10:15 PM CST on Saturday and then send it out at > 7:15 PM CDT Sunday is a bit strange. Basically the script runs only > once when it sends the count out: downloads headers and writes them to > a file -> opens that file and counts the posts -> emails the count. If > anything goes wrong it writes that error up in the log file.
Actually that "messages as of.." time is the time of the last email not when the script is run. So between 4:15 UTC and when the script ran at the right time no emails were delivered to the email service that Alex is using. They caught up afterwards. Mystery solved.
