Does anyone have any expertise with .ICS files?

This file is created from an automatic script, and something about it is wrong and I'm not sure what:
http://www.martian.at/calendar/VixyAndTony.ics

Google Calendar chokes on it when I tell it to "add by URL", I get "Warning: did not successfully merge all events" from Google. But it passes validation when I feed its URL to this web site:
http://severinghaus.org/projects/icv/

Can anyone tell what's going wrong here?

You're defining two timezones, America/Los_Angeles and US/Pacific. I don't really see why that should be an issue, but you never use America/Los_Angeles anywhere else, and they're defined differently. So I deleted the America/Los_Angeles definition (from "BEGIN:VTIMEZONE" to "END:VTIMEZONE", inclusive) and it seemed to work.

I can only offer that iCal imported it just fine, without any errors, and that I get the same issue importing it to Google that you see. Oddly, if I download the .ics, then upload it to Google, it imports the 2 events just fine.

Hm. Actually, if I just rehost it, it works fine. But I did notice that the title doesn't come up right. Is there some ancillary file? Is the file itself corrupted or corrupted in download somehow?

Aha, smoking gun. That's a re-hosting of the original, unmodified file?

Wonder what it is about the hosting that's causing the problem?

Okay, this is weird.

We just added a new gig to the calendar, a sudden opening came up in Redmond and so we're playing there tomorrow. The ICS file on the server automatically got updated by the script running in the middle of the night last night.

So now, the ICS file no longer gives me an error message when I feed it to Google. And tomorrow's show appears on my Google Calendar.

But... September 26th's show doesn't. Still.

WTF?

Maybe it's a caching thing with Google Calendar. Can anyone else do an "add by URL" of this URL onto their Google Calendar and see both the Aug 28 and Sept 26 entries?

Only adds the August one for me.

Gr. OK thanks.

Okay, I'm completely rewriting the script from scratch so that I can generate it differently, and then host it on my own web site instead of Marty's. I've got the file generating correctly, I'm going to move on to actual testing in a moment.

Question...

The timezone definitions he created at the top of the file were to work around a Google Calendar bug; if we didn't predefine a bunch of time zones, then the calendar wouldn't show the right times. Those seem to work for their indended purpose; so far so good.

My question is: Those time zone definitions at the top of the file... Are they static? I mean, can they be boilerplate text rather than calculated from some kind of code?

(Assuming a republican president doesn't alter the dates of US DST again, I mean.)