"If in doubt, fake it out" :) I'm toying with the following idea, assuming that management will go for it. Whenever a change occurs on the page that I need to log, store the result in a persistant cookie. Don't do anything on unload, but the next time the user returns to the side, submit whatever is in the cookie to the database on load. A bit of a fudge but one that I'm pretty sure would work more or less universally (as in in modern standards compliant browsers). It's just a matter of determining if I'm allowed to use a cookie.
On May 30, 4:26 pm, "Brian Miller" <[EMAIL PROTECTED]> wrote: > Web browsers simply don't do this well. It breaks the whole > non-persistent model. No matter what you do, you'll have to deal with > network latency. Also, Browsers react differently when the user closes > the browser window. I don't think that the onunload event is reliable for > all the use cases. > > Having a 15-second "autosave" is going to be your best fallback. I only > hope that your management is going to be flexible on this. Unfortunately, > what they want is outside the scope of what browsers can generally > provide. > > - Brian > > > Not strictly speaking a jQuery question but I thought somebody on here > > might have some insight into this. :) > > > The Powers That Be have asked me for a system whereby the contents of > > a form is automatically saved to the server whenever the user leaves > > the page. I looked into onunload but from what I can asertain by the > > time that event fires the form already no longer exists and can't be > > submitted to the server. Doing an AJAX post whenever the form changes > > isn't acceptible because that would generate too much database > > traffic, so I'm kind of stuck. I did suggest preservign the data in a > > cookie instead of the database but that is apparently not acceptable > > either. > > > Is there some other way I could have the form submitted when the user > > leaves the page?
