I have not changed anything on the page and it's been working nicely
until now, but suddenly this morning - no go.
I tried generating a new key, but it's the same as the one the API says
doesn't work.
What makes it even more strange is that the same key validates nicely
for the map API.
Same problem in here. No changes, invalid API key. In fact, I'm having
trouble with their system, since www.mysite.com IS NOT THE SAME AS
mysite.com. If I create an api key for one of those, the other one
doesn't work.
http://youdomain.com and http://www.youdomain.com is equivalent. A key
registered to one path would not work when the other path is used.
When you had a key failure during load, you simply got a 400 response
and GSearchControl was undefinied. In an attempt to help you discover
this, we deliver the API bits as well as an alert.
The only thing that is new is the alert, which is now flagging an error
condition thats been in the key validator since day 1.
What we are working on right now is changes to the validator that will
treat www.youdomain.com as being equal to youdomain.com. This step was
not in the validator, but it is in the maps validator and we are
actively changing it and writing the test cases.
I apologize for the confusion. I had no idea how widespread this
problem was where www. is being used so casually as a hostname...
hi mhl,
the validator error occurs also, when an ajax-search-page has been
translated by a translator engine like google or babelfish translator.
regards, dan
example:
http://babelfish.altavista.com/babelfish/trurl_pagecontent?lp=en_pt&url=http%3A%2F%2Fwww.duesentrieb.org%2F
In a sense it was, but only because I was loading the API with a wrong
key in one of the instances - but since the Map API was loading fine
(only once - but same key) I assumed the fault was with Google.
Suggestion to Google though - when pushing an update like this put up a
post in the group or maybe even a blog post, so people can help
themselves. I know it was supposed to be transparent, but when people
open their apps on a Monday morning and find they're not working all
sorts of things go flying through your head about what could have
happened during the weekend. With people from all over the world using
Google's APIs there is bound to be some people who experience the
problem long before Google's engineers get to work and have time to
look at it.
While my error is due to my own lack of attention I think a lot of
people could have helped themselves by getting a new key if a post had
warned them that there were changes made in the validationprocess.
Why is the key validation not standard across all Google's APIs?