Should this (or a prettied up version of it) go on the wiki or would
this be something that you guys would consider for the "core" (i.e.
submit a feature request)?
On Tue, Mar 25, 2008 at 1:27 PM, Johan Compagner <[EMAIL PROTECTED]> wrote:
> thats fine i gues, but a bit uglier with that null check
>
> I guess your last one before this one would also suffice
> because on every render the token doesn't need to change, why should it?
> if i press refresh in the browser does the token has to be changed?
> It isnt used anywhere at that time anyway.
>
> johan
>
> On Tue, Mar 25, 2008 at 6:19 PM, James Carman <[EMAIL PROTECTED]>
>
>
> wrote:
>
> > Okay, how about this?
> >
> > public class SynchTokenField extends HiddenField
> > {
> > private String token;
> >
> > public SynchTokenField(String id)
> > {
> > super(id, new PropertyModel(new ValueMap(), "token"));
> > setType(String.class);
> > setRequired(true);
> > regenerateToken();
> > add(new AbstractValidator()
> > {
> > protected void onValidate(IValidatable validatable)
> > {
> > if (token == null || !token.equals(validatable.getValue
> > ()))
> > {
> > error(validatable);
> > }
> > token = null;
> > }
> > });
> > }
> >
> > private void regenerateToken()
> > {
> > token = UUID.randomUUID().toString();
> > }
> >
> > protected final void onComponentTag(final ComponentTag tag)
> > {
> > super.onComponentTag(tag);
> > regenerateToken();
> > tag.put("value", token);
> > }
> > }
> >
> > Would that work?
> >
> > On Tue, Mar 25, 2008 at 1:11 PM, Igor Vaynberg <[EMAIL PROTECTED]>
> > wrote:
> > > you should regen the token in onbeforerender()...
> > >
> > > -igor
> > >
> > >
> > > On Tue, Mar 25, 2008 at 10:09 AM, James Carman
> > >
> > >
> > > <[EMAIL PROTECTED]> wrote:
> > > > So, it would be like this?
> > > >
> > > >
> > > > public class SynchTokenField extends HiddenField
> > > > {
> > > > private String token;
> > > >
> > > > public SynchTokenField(String id)
> > > > {
> > > > super(id, new PropertyModel(new ValueMap(), "token"));
> > > > setType(String.class);
> > > > setRequired(true);
> > > > regenerateToken();
> > > > add(new AbstractValidator()
> > > > {
> > > > protected void onValidate(IValidatable validatable)
> > > > {
> > > > if (!token.equals(validatable.getValue()))
> > > > {
> > > > error(validatable);
> > > > }
> > > > regenerateToken();
> > > > }
> > > > });
> > > > }
> > > >
> > > > private void regenerateToken()
> > > > {
> > > > token = UUID.randomUUID().toString();
> > > >
> > > > }
> > > >
> > > > protected final void onComponentTag(final ComponentTag tag)
> > > > {
> > > > super.onComponentTag(tag);
> > > > tag.put("value", token);
> > > > }
> > > > }
> > > >
> > > > Since wicket already syncs on the session, this should work, right?
> > > >
> > > >
> > > >
> > > > On Tue, Mar 25, 2008 at 12:46 PM, Johan Compagner <
> > [EMAIL PROTECTED]> wrote:
> > > > > shouldn't the token be cleared then somehow on the first request?
> > (in the
> > > > > validator)
> > > > > now if the second time it still validates fine because the value
> > that is
> > > > > submitted doesnt change and the token in the field doesn't
> > change.
> > > > >
> > > > > But it is a nice simple idea to have
> > > > >
> > > > > On Tue, Mar 25, 2008 at 5:40 PM, James Carman <
> > [EMAIL PROTECTED]>
> > > > > wrote:
> > > > >
> > > > >
> > > > >
> > > > > > Would something like this work?
> > > > > >
> > > > > > public class SynchTokenField extends HiddenField
> > > > > > {
> > > > > > private String token;
> > > > > >
> > > > > > public SynchTokenField(String id)
> > > > > > {
> > > > > > super(id, new PropertyModel(new ValueMap(), "token"));
> > > > > > setRequired(true);
> > > > > > add(new AbstractValidator()
> > > > > > {
> > > > > > protected void onValidate(IValidatable iValidatable)
> > > > > > {
> > > > > > String submittedToken = iValidatable.getValue
> > ().toString();
> > > > > > if (!submittedToken.equals(token))
> > > > > > {
> > > > > > error(iValidatable);
> > > > > > }
> > > > > > }
> > > > > > });
> > > > > > }
> > > > > >
> > > > > > protected final void onComponentTag(final ComponentTag tag)
> > > > > > {
> > > > > > super.onComponentTag(tag);
> > > > > > token = UUID.randomUUID().toString();
> > > > > > tag.put("value", token);
> > > > > > }
> > > > > > }
> > > > > >
> > > > > > Here, all you'd have to do is add one of these puppies to your
> > form
> > > > > > and it'll automatically validate itself.
> > > > > >
> > > > > > On Tue, Mar 25, 2008 at 10:35 AM, Johan Compagner <
> > [EMAIL PROTECTED]>
> > > > > > wrote:
> > > > > > > do you have a good patch then?
> > > > > > >
> > > > > > > And are you saying that all double submits are then not
> > possible
> > > > > > anymore?
> > > > > > >
> > > > > > > Also when i submit then think hmm thats wrong back button
> > change
> > > > > > something
> > > > > > > and submit again?
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > On Tue, Mar 25, 2008 at 3:25 PM, laz <[EMAIL PROTECTED]>
> > wrote:
> > > > > > >
> > > > > > > >
> > > > > > > > Does anyone else feel that this would be generically
> > useful to have
> > > > > > as a
> > > > > > > > part
> > > > > > > > of Wicket? Not only does it prevent double submits, but it
> > also is a
> > > > > > > > simple
> > > > > > > > safeguard against cross-site request forgery (see
> > > > > > > > http://en.wikipedia.org/wiki/Cross-site_request_forgeryfor
> a
>
>
> > > > > > summary).
> > > > > > > >
> > > > > > > > The one missing piece from your solution is
> > synchronization. There is
> > > > > > the
> > > > > > > > slightest possibility that the second submit of a double
> > submit could
> > > > > > > > enter
> > > > > > > > onSubmit before the token is reset. I am not yet sure what
> > would be
> > > > > > the
> > > > > > > > best
> > > > > > > > object to synchronize on, possibly the session id?
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > hillj2 wrote:
> > > > > > > > >
> > > > > > > > > Here's a solution that SEEMS to be working. It
> > incorporates our
> > > > > > > > solution
> > > > > > > > > to the double submit problem that we used on our JSP's.
> > It didn't
> > > > > > > > appear
> > > > > > > > > to be working for me at first, but seems to be now. (It
> > does use
> > > > > > the
> > > > > > > > old
> > > > > > > > > servlet request/session objects, but this may change
> > once all our
> > > > > > old
> > > > > > > > code
> > > > > > > > > is upgraded to wicket.)
> > > > > > > > >
> > > > > > > > > ...
> > > > > > > > >
> > > > > > > > > Like I said, for now this appears to be working. I just
> > extend all
> > > > > > my
> > > > > > > > > forms from this class and implement onSubmitted() with
> > the same
> > > > > > code I
> > > > > > > > > previously put in onSubmit(). The key is putting
> > matching unique
> > > > > > > > strings
> > > > > > > > > in session and in the page instance. On submit, those
> > string
> > > > > > should
> > > > > > > > > match, at which point, the string in session is cleared
> > and the
> > > > > > form is
> > > > > > > > > processed as normal. If another submit comes in, the
> > string in
> > > > > > session
> > > > > > > > > has been cleared so it doesn't match the string svaed in
> > the page
> > > > > > > > > instance. In the case where setResponsePage is not
> > called,
> > > > > > > > onBeforeRender
> > > > > > > > > resets the token string, so submitting from the
> > refreshed page
> > > > > > won't
> > > > > > > > > register as an error.
> > > > > > > > >
> > > > > > > > > Our JSP version of this involves putting the token
> > string in
> > > > > > session and
> > > > > > > > > also saving a copy to a hidden field on the JSP page.
> > Which I
> > > > > > think is
> > > > > > > > > similar (although maybe a bit more complex) to what
> > Martijn was
> > > > > > > > > suggesting.
> > > > > > > > >
> > > > > > > > > Thanks for all you suggestions.
> > > > > > > > >
> > > > > > > > > Joel
> > > > > > > > >
> > > > > > > >
> > > > > > > > --
> > > > > > > > View this message in context:
> > > > > > > >
> > http://www.nabble.com/Double-submit-problem-tp15957979p16275106.html
> > > > > > > > Sent from the Wicket - User mailing list archive at
> > Nabble.com.
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > ---------------------------------------------------------------------
> > > > > > > > To unsubscribe, e-mail:
> > [EMAIL PROTECTED]
> > > > > > > > For additional commands, e-mail:
> > [EMAIL PROTECTED]
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > ---------------------------------------------------------------------
> > > > > > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > > > > > For additional commands, e-mail: [EMAIL PROTECTED]
> > > > > >
> > > > > >
> > > > >
> > > >
> > > >
> > ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > > > For additional commands, e-mail: [EMAIL PROTECTED]
> > > >
> > > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > > For additional commands, e-mail: [EMAIL PROTECTED]
> > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > For additional commands, e-mail: [EMAIL PROTECTED]
> >
> >
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
