If you enter a web address such as www.gymking.co.uk/test1 - a ‘normal’ server would look inside a folder called test1 for an index page (index.html, index.htm, index.php)
Your server does not appear to be doing that. This makes me think that there is something odd about your server and its setup which means that it cannot/is not configured to read the query string part of the URL
When you enter a search term the action/code creates a query string containing the search term (and the javascript ID) If I search for the word ‘test’ on your page I get the URL of www.gymking.co.uk/test1.html?q=test&t=1313736912221
But this does not seem to work on your server!
So - where is your site hosted and on what sort of server?
I think that you will have to contact their support team and ask them why your query strings are not being read by the search script.
Have you done anything clever on your server to enable “tail-less” URLs? Your filename should have an extension on it, usually .html. You can certainly rely on Apache to figure out that example.com/search should be handled by example.org/search.html without any configuration, but JavaScript is very literal, and so if your search page is really named search.html, you really must use that absolute URL when setting up your search form.
The issue is a JavaScript error in the search database, which is
created by the Action when the site is published. That’s definitely
where the problem lies, not the server or anywhere else.
You can confirm this by creating a new document with two or three
pages of random copy in it, and set up the Action there. I think
you’ll find that it works just fine. Your search database is also 2.17
MB in size, and that’s quite large. That entire thing has to load
before any searches will work. For a site this large, or with this
much unique content in it, you might want to investigate a different
solution technology-wise. Something server-side or a hosted service
like Atomz might be a better fit.
Walter
On Aug 19, 2011, at 6:09 PM, Tony Farrer wrote:
having said that, I just looked in freeway and my home page is
index.html (in the inpector - file)
The first error happens mid-way through the giganto JS file, while it’s creating the array of pages and their keywords. It looks to me as though a bit of punctuation in a keyword isn’t being properly escaped Anything else that happens after that is collateral damage.
Errors in JS tend to compound. A function expects a certain sort of input (an array) and then it gets fed null or undefined instead, and it goes into a little fit. Whatever happens after the first error is entirely unpredictable. If a script is really carefully written, maybe it can rescue the situation enough to provide a meaningful error message, but usually it’s a snipe hunt. I find when debugging JS that you fix one thing, and that gets you a different error message, which leads you to the next thing.
thanks for all of your input, you lost me half way through but I assume that my file is too large for this type of search anyway, back to the drawing board !! thanks again…
Just to update you , I had the following email from Softpress after they looked at my site file.
“You have invalid characters in the description META tags on about 28 of your pages. It looks like you have copy and pasted your description from another application and bringing a bad character with it. It looks like they are some kind of line breaks that don’t really work with the Site Search Action.”
Just to let you know incase it helps in the future…