[TYPO3] Precaching question

[vahan]

Hello Bernhard,

I appreciate your detailed response. thank you for pointing me to httrack!
a few questions though yet:
- the url you gave below does not work right away for me. I am using 
newloginbox ext. Should I be using something else?
- also, the solution below will need the crawler to crawl for both 
logged and not logged users to precache for both scenarios.
Life would be much easier if it would be possible to turn off the 
dependency on Groups parameter for caching so that regardless if a user 
is logged on or not same cache would be used.
do you  have a clue on how to disable that feature and crawl the site 
just once.

thank you,
Vahan

Bernhard Kraft wrote:
> Vahan Amirbekyan wrote:
>> I cannot agree and let me explain.
>> the site is a community site. it allows keep-me-logged feature. So most
>> users ( 40,000 at the moment ) are constantly logged on users.
>> now...
>> the only personalization on most pages is the user's name, the rest of
>> page content is the same.
>> all pages are comprised of plugins that are cache-enabled. The plugin
>> that renders current user name is not cache-enabled.
>> So pre-caching should cache all plugins for all users, it is the
>> username plugin only which will be not-cached.
> 
> Hi Vahan !
> 
> Well ... first I hope that the plugin which generates the username is of
> type "USER_INT" or "COA_INT" ... else it wont make much sense. If you call
> "$TSFE->set_no_cache()" or set "$TSFE->no_cache=1" in your plugin the
> complete page wont be cached ...
> 
> but if you use a USER_INT or COA_INT object the complete page gets cached.
> and instead of the USER_INT/COA_INT object a marker like:
> <!--###COA_INT_sdfsdjlkfjds###--> gets put into the content which is stored
> in cache.
> 
> when the page get's regenerated next time it is simply retrieved from
> cached and
> just those INT objects get generated ...
> 
> my idea for you would be that you let a crawler first hit:
> 
> http://www.mycommunitysite.com/
> 
> and then:
> 
> http://www.mycommunitysite.com/?logintype=login&user=crawler&pass=crawler_secret&pid=123
> 
> ... 123 has to be the UID of the page/sys-folder containing the FE-users ...
> 
> this would create a login-session for the crawler - and if you use a
> crawler which is capable of
> using cookies it will stay logged in as long as it resends the cookie
> ... AFAIK httrack is performing
> well for this ...
> if you use something like "wget" you can store the retrieved cookie in a
> file and let succeding calls
> use the cookie from the file .... but just "man httrack" ... or "man
> wget" for cookies ...
> 
> the above parameters in the URL are enough for doing the login ...
> 
> 
> 
> greets,
> Bernhard