I've created WICKET-4275 for this, including a quickstart.

Gereon

Am 01.12.11 18:59 schrieb "Igor Vaynberg" unter <[email protected]>:

>Gereon, can you take a look...
>
>-igor
>
>On Thu, Dec 1, 2011 at 9:58 AM, Martin Grigorov <[email protected]>
>wrote:
>> dig the demo application and try
>> maybe I didn't test it properly...
>>
>> On Thu, Dec 1, 2011 at 6:55 PM, Igor Vaynberg <[email protected]>
>>wrote:
>>> the question is, should we be escaping them with the backslash or with
>>> the url-coding?
>>>
>>> -igor
>>>
>>> On Thu, Dec 1, 2011 at 9:48 AM, Martin Grigorov <[email protected]>
>>>wrote:
>>>> The commit message doesn't say it, but yes this is the fix for this
>>>> security problem.
>>>> There is a way to attack the application with the URL encoded version.
>>>>
>>>> On Thu, Dec 1, 2011 at 6:05 PM, Igor Vaynberg
>>>><[email protected]> wrote:
>>>>> this was done in r1150391 by martin for
>>>>> http://wicket.apache.org/2011/08/23/cve-2011-2712.html
>>>>>
>>>>> -igor
>>>>>
>>>>> On Thu, Dec 1, 2011 at 8:45 AM, Gereon Steffens
>>>>> <[email protected]> wrote:
>>>>>> Hi,
>>>>>>
>>>>>> I've noticed a change in Wicket 1.4.19 regarding URL parameter
>>>>>>encoding.
>>>>>> If a parameter value contains a single quote, these quotes are now
>>>>>> preceeded by a backslash (this happens in
>>>>>>RequestCycle#encodeUrlFor).
>>>>>>
>>>>>> Why is this done? I've never heard of backslash-escaping in
>>>>>>relation to
>>>>>> URLs.
>>>>>>
>>>>>> As far as I can tell, this also breaks compatibility with
>>>>>>apps/sites that
>>>>>> now need additional code to parse parameters like
>>>>>> example.com?param=a'b which is now represented as param=a\'b. Even
>>>>>>if the
>>>>>> backslash were necessary in a URL, it should be URL-escaped as %5c.
>>>>>>
>>>>>> Am I missing something?
>>>>>>
>>>>>> Regards
>>>>>> Gereon
>>>>>>
>>>>>>
>>>>>>



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to