Well, I think I can confirm this. no-cache works like no-store for IE.
So with our current setup, IE already reloads pages on back button.
I think if we force also FF not to store the page, we will gain
consistency (IE still has majority...).
It is possible that FF 2.0 fixes this, but 1.5 does not, and 1.5 is
current stable version, thus we should focus on it.
-Matej
Matej Knopp wrote:
Strange. The error is only shown in FF (1.5).
Internet explorer, for some reason, reloads pages on back button. I
guess no-cache works like no-store in IE.
-Matej
Juergen Donnerstag wrote:
IE6 does it, but I don't get errors
Juergen
On 10/8/06, Matej Knopp <[EMAIL PROTECTED]> wrote:
This happens on wicket-library.com, so it's not affected by wicket
version. Maybe FF 2rc2 does it as you say, but FF1.5 and IE6 don't.
-Matej
Juergen Donnerstag wrote:
> Obviously the behavior depends on browsers. I tried it with FF 2rc2
> and the examples installed on wicket-library and when I hit the return
> button I get the very same page I left before and not the cached page
> content.
>
> You are not using 2.0, do you? The border component has some problems
> when a page gets repainted.
>
> Juergen
>
> On 10/8/06, Matej Knopp <[EMAIL PROTECTED]> wrote:
>> Hi,
>>
>> we have a serious problem with ajax and back button. To see it in
>> action, go to the tree example, expand a tree node, go to a different
>> page, press back button and you'll see the node collapsed. However
>> trying to expand it, you'll see bunch of errors.
>>
>> Why?
>>
>> The problem is that cached page reflects the page when it was
first time
>> retrieved from the server. However, doing Ajax updates, the page
could
>> have been significantly changed. So if you go using back button to
go to
>> page with ajax components, there's a big chance that it's out of sync
>> with server.
>>
>> Currently I can't think of any way around this, except to force
reload
>> the page on back button, using Cache-Control: no-store.
>>
>> I know many of you are probably against reloading page on back
button,
>> but what else can we do?
>>
>> -Matej
>>
>
