lucy24 - 4:31 am on Dec 17, 2011 (gmt 0)

You don't notice it unless you are running Live Headers or similar. Of course if you code absolutely all redirects yourself, you can do them as a package. But if you are using apache's built-ins and/or stuff that your host optionally does on your behalf, then with/without www and directory-slash is each its own redirect.

That's why I wondered if g### treats all redirects alike. Now, I'm not big enough to get requests for nonexistent files-- or maybe, ahem, there are enough naturally occurring 404s that they can't have any doubt. But I do notice that whenever gwt checks for its validation file, there are two requests, one redirected and one not. I have to assume that it's asking for the with-and-without-www versions, and noting the fact that one of the two gets a redirect. (Being a robot, it can choose not to follow a redirect. Doesn't seem quite fair does it?)