On Fri, Nov 7, 2008 at 9:53 AM, James <[EMAIL PROTECTED]> wrote: > > Hi all, > > I'm am looking into using gears to transparently request large JSON > objects and then download and insert them into a local database > (SQlite). > > The JSON objects are very large, 10MB in size approx, they are pre- > built and stored in files on the server since the data is static. > > I would prefer to perform this with little user interaction, if at > all. > > I'm using PHP on the server to read the data in to the response upon a > XmlHttpRequest. > > From looking at the documentation, I would need to use the > HttpRequest, WorkerPool and Database API's to perform this function. > > I have a couple of questions though: > > Does the HttpRequest API handle JSON objects?
Yes and no. JSON is utf8 encoded text data. HttpRequest.responseText will return that as a string. Its up to your application code to parse the string and populate an 'object'. There are libraries to perform that and there is also eval(). > > > What are the tollerences of the HttpRequest API, is 10 MB too large? > On a handheld device that may be a bit much, but on most desktop systems this should be fine. > > I have used background worker constructs in both C# and Standard Java, > is the WorkerPool API similar in any way? > Yes it is, WorkerPools start background threads. > > If a user navigates away from the page that spawned the worker, will > the worker die a a result? > Yes it will. If the worker isn't doing anything, the script context will die immediately. If the worker is doing something, the script will run to completion and then the context will go away. If the browser process exits while a worker script is still doing something, the script is killed (no waiting for completion). > Any advice/tips would be really great. > James >
