[TYPO3-dev] Soution: Page is being generated

[ries van Twisk]

On Jan 27, 2008, at 11:00 AM, Tobias Pierschel wrote:

> " also wondered why the association would not sponsor 400,00 Euros for
> such a task. I mean they are financing much of 5.0 development - so it
> should be ok for them to sponsor a small project improving the
> performance of 4.x series ..."
>
> I fully agree with you.
>  T o b i

I think this only works if the complete association agrees with it.

There are many non-payed (non payed by means of not payed by the  
association)
core developers and it would be rather strange if all core developers  
are going to say:
' Hè I got a patch, will cost you xxxx euro's'

Since I am not a member of the association, I don't have
a single clue how money is divided. I am not sure if every core
developer get's a standard fee, or that only some developers get a fee.

However I do feel that if somebody signs up as core developer he/she
cannot really ask for money. A core developer gives his devotion for  
free because
he loves the too, uses it fir his own good and/or clients.
I think only if some large task is at hand (couple of months?)
the association can pay for the work if he/she is not doing it for a  
client and/or employed work.

That said, special situations excluded...

Just my personal opinion.

Ries





>
>
>
>
> Bernhard Kraft schrieb:
>> Dmitry Dulepov [typo3] wrote:
>>
>>> Another way would be to use locking mechanism to prevent race  
>>> conditions
>>> (sysvsem if available or simple file locking, which is much slower).
>>
>> This is mostly what I did. Currently I only use the php "flock"  
>> call to
>> create a locking
>> file dependen on the page being generatd in a two-step algorythm. So
>> there are two
>> locks per created page - reason is mostly that the first lock if for
>> generating the
>> "currentPageData" cached value in cache_pagesection ...
>> the second lock is for generating the cachedPage...
>>
>> if two clients access the same page/parameters simultanously the  
>> first
>> hit will create
>> a lock on a file - the second request will then be interrupted  
>> (blocked)
>> till the page
>> is generated - after generating the page it get's served to the first
>> client (which
>> initiated the generating) ... and the second process gets not blocked
>> anymore and
>> delivers the page from the cache (in which it got stored by the first
>> process) ...
>>
>> so in fact if 10 clients access the same page simulatuosly all of  
>> them
>> will just get
>> the blank screen which is normal for IE/FF when a page is loaded.  
>> when
>> it is generated
>> all of those clients will get the generated page from the cache in
>> almost the same
>> time ...
>>
>> I also wondered why the association would not sponsor 400,00 Euros  
>> for
>> such a task.
>> I mean they are financing much of 5.0 development - so it should be  
>> ok
>> for them to
>> sponsor a small project improving the performance of 4.x series ...
>>
>>
>>
>> PS Steve: When you try to use the test-scripts/frames you have to
>> uncompress those
>> three files to the same website/webroot which you are going to  
>> test ...
>> FF (and IE?)
>> do not allow to set the URL of an iframe to some URL which is not  
>> on the
>> same domain
>> (at least not via JS) ...
>>
>>
>> greets,
>> Bernhard