> I can confirm I'm looking at this right now in the UK with mostly blue
> default markers.
Thanks to everyone for testing... it's helpful to get global feedback!
> Google use distributed servers, your view may vary around the world.
> KmlLayer caches imagery at Google's servers for a time, a 'broken'
> view may appear/disappear after a few hours. This gets important if
> you edit KML and introduce errors, they may not immediately go away
> when 'fixed' at source.
Yes, we discovered this early on and it makes testing difficult--
particularly when we made minor changes to our scripts. We started
hiding text edits in our kmls so we could distinguish btwn the old/new
kmls!
> In fact I'm trying the same url in both IE and FF browsers on the same
> PC, and get more blue markers with less custom icons in FF.... that
> ought to rule out cache issues??
Exactly. That's had us scratching our heads (and why now... and why
only some markers@#!?) I'm now testing the map/kmls/pngs on a
different server. I want to rule out server-side issues--specifically
what looks like a canonical redirect from
http://climate* to
http://www.climate*.
We use full urls to the custom marker pngs -- currently
http://climate*.
That doesn't seem to trip up html => kml access, but it might trip up
kml => png access. (but again... why wouldn't that affect all 5 custom
markers?)
I plan to remove some kmls too. See if we've reached some threshold
for the number of kmls-to-map, or custom markers-to-map.
Thanks again everyone. Any new thoughts welcome.