[opencms-dev] Cache problem with cms:include

[Tobias Karrer]

And property cache=uri on the including Jsp?!?

Am 08.07.2013 17:51, schrieb Tobias Herrmann:
> Hi Tobias,
>
> to test this, I also created a JSP printing 
> System.currentTimeInMillis() and set the property cache=never. I used 
> <cms:include file="myjsp.jsp" /> to include the JSP. This works fine.
> I actually never used the attribute cachable with the cms:include tag. 
> I think you don't need it.
>
> Greetings, Tobias
>
> -- 
>
> Alkacon Software GmbH - The OpenCms Experts
>
> http://www.alkacon.com
> http://www.opencms.org
>
> Am 08.07.2013 18:28, schrieb Tobias Karrer:
>> Gets quite confusing for me now...
>>
>> What I did:
>>
>> /mysite/dynamic-include.jsp (cache=never) with content "Time in 
>> Millis: <%= |System.currentTimeInMillis() %>|"
>> /mysite/page.jsp (cache=uri) which contains "<cms:include 
>> cacheable="false" file="/mysite/dynamic-include.jsp"/>"
>>
>> With the first call of "/mysite/page.jsp" both pages are compiled an 
>> the final HTML-Markup ends up in the FlexCache.
>> For the second call the whole page will be delivered by FlexCache. 
>> Time in Millis always will be the compilation time (until FlexCache 
>> is purged)
>>
>> What I want:
>>
>> Each time I call "/mysite/page.jsp", Time-in-Millis will be the 
>> current time, the rest of the page shall be cached.
>>
>> As far as I can see at the moment, that's not possible.
>> (Or I didn't get what I'm doing wrong)
>>
>> Best,
>> Tobias
>>
>>
>> Am 08.07.2013 16:16, schrieb Tobias Herrmann:
>>> Hi Tobias,
>>>
>>> using the cache property set to 'never' on the JSP work fine for me. 
>>> How are you doing it?
>>>
>>> Greetings, Tobias
>>>
>>> -- 
>>>
>>> Alkacon Software GmbH - The OpenCms Experts
>>>
>>> http://www.alkacon.com
>>> http://www.opencms.org
>>>
>>> Am 07.07.2013 15:39, schrieb Tobias Karrer:
>>>> Hi Achim,
>>>>
>>>> Am 06.07.2013 18:44, schrieb Achim Westermann:
>>>>> Hi Tobias,
>>>>>
>>>>>
>>>>>
>>>>>> What about the attribute "cacheable" of the cms:include-Tag?
>>>>>>
>>>>>> So far we used cms:include to include parts of a page which 
>>>>>> mustn't be
>>>>>> cached... worked very well with 7.5.4.
>>>>>>
>>>>>> If this isn't working any more it's preferable to use static 
>>>>>> includes,
>>>>>> because they're much more faster.
>>>>>>
>>>>> There is nothing about cache "not working any more". That part of
>>>>> OpenCms is pretty deep-down, stable and untouched. I am not 
>>>>> writing that
>>>>> from my current knowledge but from my experience with OpenCms. But my
>>>>> "feeling" tells me that this issue is worthwhile to dig deeper why
>>>>> something in your webapp did work and now does not. As opposed to say
>>>>> "It worked in 7.5.4, in 8 it does not. So the latest version has 
>>>>> issues."
>>>>>
>>>>> Static includes are not faster if you want to use caching. Those will
>>>>> not make use of any caching at all but be included in the jsp
>>>>> compilation phase. Think of a jsp that performs heavy database
>>>>> operations being included statically: That code will get executed for
>>>>> every call.
>>>>> OK, static includes are faster, if you do not want to cache at 
>>>>> all. But
>>>>> do you want that? You can even set the OpenCms cache to have a 
>>>>> timeout
>>>>> in minutes. Pretty mighty. It is worth spending some time on the 
>>>>> OpenCms
>>>>> cache setting.  Cache is king. Dropping caching may only work if 
>>>>> you buy
>>>>> faster hardware. And may always be compromised by the intensity of 
>>>>> DOS
>>>>> attacks.
>>>> Thanks for yopur answer.
>>>> I agree complete "Cache is king". But still there are pages or 
>>>> parts of pages, which you don't want to be cached, and I'm looking 
>>>> for a good way to exclude thes parts from caching.
>>>>
>>>> I meanwhile did different tests with cms:includes, to see why the 
>>>> different behaviour occurs.
>>>> Finally: there has been a small change, so the cache-setting for a 
>>>> session attibute behaves a little different. In some especial 
>>>> circumstances the problem will have occured in 7.5.4 as well. So 
>>>> not the updateto Version 8 itself was responsible.
>>>>
>>>>>> Or what to do to exclude parts of a page from caching?
>>>>> Set property cache to "never" on a jsp. And then be flexible to 
>>>>> decide
>>>>> to cache the output of that jsp later on properly.
>>>> Clearly, but what if this jsp includes other jsps, which I don't 
>>>> want to be cached.
>>>>
>>>> Example:
>>>> /mysite/dynamic-include.jsp (cache=never) with "Time in Millis: <%= 
>>>> |System.currentTimeInMillis() %>|"
>>>> /mysite/page.jsp (cache=uri) which contains "<cms:include 
>>>> cacheable="false" file="/mysite/dynamic-include.jsp"/>"
>>>>
>>>> With the first call of "/mysite/page.jsp" both pages are compiled 
>>>> an the final HTML-Markup ends up in the FlexCache.
>>>> For the second call the whole page will be delivered by FlexCache. 
>>>> Time in Millis always will be the compilation time (until FlexCache 
>>>> is purged)
>>>>
>>>> Or am I wrong?
>>>>
>>>> Greeting,
>>>> Tobias
>>>>
>>>>
>>>> _______________________________________________
>>>> This mail is sent to you from the opencms-dev mailing list
>>>> To change your list options, or to unsubscribe from the list, 
>>>> please visit
>>>> http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
>>>>
>>>>
>>>>
>>
>> _______________________________________________
>> This mail is sent to you from the opencms-dev mailing list
>> To change your list options, or to unsubscribe from the list, please 
>> visit
>> http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
>>
>>
>>
> _______________________________________________
> This mail is sent to you from the opencms-dev mailing list
> To change your list options, or to unsubscribe from the list, please 
> visit
> http://lists.opencms.org/cgi-bin/mailman/listinfo/opencms-dev
>
>
>