You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@wicket.apache.org by Artur Kronenberg <ar...@openmarket.com> on 2013/11/05 15:05:13 UTC
Back button not calling onConfigure()
Hi,
I have a rather odd problem with a page of mine. The design is that the
model on the page gets a flag set that indicates that the send button
has been clicked (so the user doesn't send multiple times by clicking it
vigurously over and over again).
The problem now is that on browser back, the model is still created and
the page can't be used.
I overwrote the onConfigure method to set up my model correctly and
everything works just fine. However there seems to be cases where the
onConfigure method is not called?
Does anyone know of those cases? Can it have anything to do with the
user using a proxy? A weird browser thing?
We set the response headers to:
Expires: -1
Pragma: no-cache
Cache-Control: no-cache, max-age=0, must-revalidate, no-store
This seems to be a very odd issue that I can not reproduce locally at
all (I used to be able to by hitting the back button so I am guessing
this is where things happen, but overwriting onConfigure fixed it for
most cases).
Thanks!
Artur
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
For additional commands, e-mail: users-help@wicket.apache.org
Re: Back button not calling onConfigure()
Posted by Martin Funk <ma...@gmail.com>.
Hi Arthur,
no worries. This was just a hint. Wireshark might look quite discouraging at first sight.
When it is started it will show all the network traffic on the machine, but that output can be filtered quite effectively.
If it is really important, then wireshark can analyse gigabyte big tcpdump files:
http://www.wireshark.org/docs/wsug_html_chunked/AppToolstcpdump.html
It depends on the pain, that bug is giving, but at least it would be known which request 'really' leaves the building.
mf
Am 06.11.2013 um 17:08 schrieb Artur Kronenberg <ar...@openmarket.com>:
> Hi Martin,
>
> I had a quick read through it and this is not an option. I can't reproduce the error and it is not given when it will reoccur. If I understood it correctly I'd have to caputre all network traffic for my application until the point the error occurs to be able to use wireshark and analyze it. This won't be possble.
>
> Thanks,
>
> Artur
>
> On 06/11/13 15:53, Martin Funk wrote:
>> Hi,
>>
>> have you thought about wireshark?
>>
>> Easy to set up and not all that hard to handle. Filters are powerful and most interestingly for Web-Development
>> ist the "Follow TCP stream" function.
>> http://www.wireshark.org/docs/wsug_html_chunked/ChAdvFollowTCPSection.html
>>
>>
>> mf
>>
>> Am 06.11.2013 um 15:27 schrieb Artur Kronenberg <ar...@openmarket.com>:
>>
>>> Hi,
>>>
>>> I would expect it to hit the container. Debuggin and testing on a test environment works and hits it. Unfortunatelly there is no way to definitely proove that the error in question hits it to. I do know that the onConfigure is not hit. I don't have access to the access logs.
>>>
>>> Is there a way that the user uses a proxy that disables the response headers we set for caching?
>>>
>>> Thanks!
>>>
>>> On 05/11/13 14:22, Martin Grigorov wrote:
>>>> Hi,
>>>>
>>>> It looks like the page is loaded from the browser/proxy cache.
>>>> Can you verify that the web container is hit ? E.g. check its access logs
>>>>
>>>>
>>>> On Tue, Nov 5, 2013 at 4:05 PM, Artur Kronenberg <
>>>> artur.kronenberg@openmarket.com> wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I have a rather odd problem with a page of mine. The design is that the
>>>>> model on the page gets a flag set that indicates that the send button has
>>>>> been clicked (so the user doesn't send multiple times by clicking it
>>>>> vigurously over and over again).
>>>>> The problem now is that on browser back, the model is still created and
>>>>> the page can't be used.
>>>>> I overwrote the onConfigure method to set up my model correctly and
>>>>> everything works just fine. However there seems to be cases where the
>>>>> onConfigure method is not called?
>>>>>
>>>>> Does anyone know of those cases? Can it have anything to do with the user
>>>>> using a proxy? A weird browser thing?
>>>>>
>>>>> We set the response headers to:
>>>>>
>>>>> Expires: -1
>>>>> Pragma: no-cache
>>>>> Cache-Control: no-cache, max-age=0, must-revalidate, no-store
>>>>>
>>>>> This seems to be a very odd issue that I can not reproduce locally at all
>>>>> (I used to be able to by hitting the back button so I am guessing this is
>>>>> where things happen, but overwriting onConfigure fixed it for most cases).
>>>>>
>>>>> Thanks!
>>>>>
>>>>> Artur
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
>>>>> For additional commands, e-mail: users-help@wicket.apache.org
>>>>>
>>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
>>> For additional commands, e-mail: users-help@wicket.apache.org
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
>> For additional commands, e-mail: users-help@wicket.apache.org
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
> For additional commands, e-mail: users-help@wicket.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
For additional commands, e-mail: users-help@wicket.apache.org
Re: Back button not calling onConfigure()
Posted by Artur Kronenberg <ar...@openmarket.com>.
Hi Martin,
I had a quick read through it and this is not an option. I can't
reproduce the error and it is not given when it will reoccur. If I
understood it correctly I'd have to caputre all network traffic for my
application until the point the error occurs to be able to use wireshark
and analyze it. This won't be possble.
Thanks,
Artur
On 06/11/13 15:53, Martin Funk wrote:
> Hi,
>
> have you thought about wireshark?
>
> Easy to set up and not all that hard to handle. Filters are powerful and most interestingly for Web-Development
> ist the "Follow TCP stream" function.
> http://www.wireshark.org/docs/wsug_html_chunked/ChAdvFollowTCPSection.html
>
>
> mf
>
> Am 06.11.2013 um 15:27 schrieb Artur Kronenberg <ar...@openmarket.com>:
>
>> Hi,
>>
>> I would expect it to hit the container. Debuggin and testing on a test environment works and hits it. Unfortunatelly there is no way to definitely proove that the error in question hits it to. I do know that the onConfigure is not hit. I don't have access to the access logs.
>>
>> Is there a way that the user uses a proxy that disables the response headers we set for caching?
>>
>> Thanks!
>>
>> On 05/11/13 14:22, Martin Grigorov wrote:
>>> Hi,
>>>
>>> It looks like the page is loaded from the browser/proxy cache.
>>> Can you verify that the web container is hit ? E.g. check its access logs
>>>
>>>
>>> On Tue, Nov 5, 2013 at 4:05 PM, Artur Kronenberg <
>>> artur.kronenberg@openmarket.com> wrote:
>>>
>>>> Hi,
>>>>
>>>> I have a rather odd problem with a page of mine. The design is that the
>>>> model on the page gets a flag set that indicates that the send button has
>>>> been clicked (so the user doesn't send multiple times by clicking it
>>>> vigurously over and over again).
>>>> The problem now is that on browser back, the model is still created and
>>>> the page can't be used.
>>>> I overwrote the onConfigure method to set up my model correctly and
>>>> everything works just fine. However there seems to be cases where the
>>>> onConfigure method is not called?
>>>>
>>>> Does anyone know of those cases? Can it have anything to do with the user
>>>> using a proxy? A weird browser thing?
>>>>
>>>> We set the response headers to:
>>>>
>>>> Expires: -1
>>>> Pragma: no-cache
>>>> Cache-Control: no-cache, max-age=0, must-revalidate, no-store
>>>>
>>>> This seems to be a very odd issue that I can not reproduce locally at all
>>>> (I used to be able to by hitting the back button so I am guessing this is
>>>> where things happen, but overwriting onConfigure fixed it for most cases).
>>>>
>>>> Thanks!
>>>>
>>>> Artur
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
>>>> For additional commands, e-mail: users-help@wicket.apache.org
>>>>
>>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
>> For additional commands, e-mail: users-help@wicket.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
> For additional commands, e-mail: users-help@wicket.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
For additional commands, e-mail: users-help@wicket.apache.org
Re: Back button not calling onConfigure()
Posted by Martin Funk <ma...@gmail.com>.
Hi,
have you thought about wireshark?
Easy to set up and not all that hard to handle. Filters are powerful and most interestingly for Web-Development
ist the "Follow TCP stream" function.
http://www.wireshark.org/docs/wsug_html_chunked/ChAdvFollowTCPSection.html
mf
Am 06.11.2013 um 15:27 schrieb Artur Kronenberg <ar...@openmarket.com>:
> Hi,
>
> I would expect it to hit the container. Debuggin and testing on a test environment works and hits it. Unfortunatelly there is no way to definitely proove that the error in question hits it to. I do know that the onConfigure is not hit. I don't have access to the access logs.
>
> Is there a way that the user uses a proxy that disables the response headers we set for caching?
>
> Thanks!
>
> On 05/11/13 14:22, Martin Grigorov wrote:
>> Hi,
>>
>> It looks like the page is loaded from the browser/proxy cache.
>> Can you verify that the web container is hit ? E.g. check its access logs
>>
>>
>> On Tue, Nov 5, 2013 at 4:05 PM, Artur Kronenberg <
>> artur.kronenberg@openmarket.com> wrote:
>>
>>> Hi,
>>>
>>> I have a rather odd problem with a page of mine. The design is that the
>>> model on the page gets a flag set that indicates that the send button has
>>> been clicked (so the user doesn't send multiple times by clicking it
>>> vigurously over and over again).
>>> The problem now is that on browser back, the model is still created and
>>> the page can't be used.
>>> I overwrote the onConfigure method to set up my model correctly and
>>> everything works just fine. However there seems to be cases where the
>>> onConfigure method is not called?
>>>
>>> Does anyone know of those cases? Can it have anything to do with the user
>>> using a proxy? A weird browser thing?
>>>
>>> We set the response headers to:
>>>
>>> Expires: -1
>>> Pragma: no-cache
>>> Cache-Control: no-cache, max-age=0, must-revalidate, no-store
>>>
>>> This seems to be a very odd issue that I can not reproduce locally at all
>>> (I used to be able to by hitting the back button so I am guessing this is
>>> where things happen, but overwriting onConfigure fixed it for most cases).
>>>
>>> Thanks!
>>>
>>> Artur
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
>>> For additional commands, e-mail: users-help@wicket.apache.org
>>>
>>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
> For additional commands, e-mail: users-help@wicket.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
For additional commands, e-mail: users-help@wicket.apache.org
Re: Back button not calling onConfigure()
Posted by Artur Kronenberg <ar...@openmarket.com>.
Hi,
I would expect it to hit the container. Debuggin and testing on a test
environment works and hits it. Unfortunatelly there is no way to
definitely proove that the error in question hits it to. I do know that
the onConfigure is not hit. I don't have access to the access logs.
Is there a way that the user uses a proxy that disables the response
headers we set for caching?
Thanks!
On 05/11/13 14:22, Martin Grigorov wrote:
> Hi,
>
> It looks like the page is loaded from the browser/proxy cache.
> Can you verify that the web container is hit ? E.g. check its access logs
>
>
> On Tue, Nov 5, 2013 at 4:05 PM, Artur Kronenberg <
> artur.kronenberg@openmarket.com> wrote:
>
>> Hi,
>>
>> I have a rather odd problem with a page of mine. The design is that the
>> model on the page gets a flag set that indicates that the send button has
>> been clicked (so the user doesn't send multiple times by clicking it
>> vigurously over and over again).
>> The problem now is that on browser back, the model is still created and
>> the page can't be used.
>> I overwrote the onConfigure method to set up my model correctly and
>> everything works just fine. However there seems to be cases where the
>> onConfigure method is not called?
>>
>> Does anyone know of those cases? Can it have anything to do with the user
>> using a proxy? A weird browser thing?
>>
>> We set the response headers to:
>>
>> Expires: -1
>> Pragma: no-cache
>> Cache-Control: no-cache, max-age=0, must-revalidate, no-store
>>
>> This seems to be a very odd issue that I can not reproduce locally at all
>> (I used to be able to by hitting the back button so I am guessing this is
>> where things happen, but overwriting onConfigure fixed it for most cases).
>>
>> Thanks!
>>
>> Artur
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
>> For additional commands, e-mail: users-help@wicket.apache.org
>>
>>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
For additional commands, e-mail: users-help@wicket.apache.org
Re: Back button not calling onConfigure()
Posted by Martin Grigorov <mg...@apache.org>.
Hi,
It looks like the page is loaded from the browser/proxy cache.
Can you verify that the web container is hit ? E.g. check its access logs
On Tue, Nov 5, 2013 at 4:05 PM, Artur Kronenberg <
artur.kronenberg@openmarket.com> wrote:
> Hi,
>
> I have a rather odd problem with a page of mine. The design is that the
> model on the page gets a flag set that indicates that the send button has
> been clicked (so the user doesn't send multiple times by clicking it
> vigurously over and over again).
> The problem now is that on browser back, the model is still created and
> the page can't be used.
> I overwrote the onConfigure method to set up my model correctly and
> everything works just fine. However there seems to be cases where the
> onConfigure method is not called?
>
> Does anyone know of those cases? Can it have anything to do with the user
> using a proxy? A weird browser thing?
>
> We set the response headers to:
>
> Expires: -1
> Pragma: no-cache
> Cache-Control: no-cache, max-age=0, must-revalidate, no-store
>
> This seems to be a very odd issue that I can not reproduce locally at all
> (I used to be able to by hitting the back button so I am guessing this is
> where things happen, but overwriting onConfigure fixed it for most cases).
>
> Thanks!
>
> Artur
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
> For additional commands, e-mail: users-help@wicket.apache.org
>
>