Comment 7 for bug 286567

Oh, and on "the display issue is actually a different one," I agree this observation does add some new elements---and I suspect you've already diagnosed the problem better than I. However, at the risk of being unnecessarily explicit: note that those evening appointments span 2 days as measured by UTC (6pm-7:30pm America/Chicago is 2300-0030 UTC). So it seems that the decision about which days to display the event on is being made before the conversion to the local timezone. So this "new" bug should be fixed by doing the timezone conversion first, I'd guess.

That said, I bet the double-display bug won't be triggered nearly as often if the std.ics file is converted to use local timezones first. So your script (once it works more broadly) should resolve this problem for many people, even if there's a second problem that also needs to be addressed.

Should I open a separate bug report for the double-display problem?

On scripts: also note http://chandlerproject.org/Projects/OlsonConvert. However, when I try to run this script on my std.ics file, I get a "year is out of range" error.