[TYPO3-mvc] Related objects inheriting from parent object default query settings

[François Suter]

Hi,

Any opinion from the Extbase big wigs? I think what I described is not 
an uncommon scenario and that problem should be addressed.

> I have the following situation: I have a "Report" object which has a
> related "Owner" object, which is actually a fe_users entry. Quite
> logically I don't want to show hidden/disabled reports, but the owners
> may be disabled fe_users (former employees). To implement this, I simply
> extended the Tx_Extbase_Domain_Repository_FrontendUserRepository with my
> own class and set the enable fields to be ignored in the default query
> settings.
>
> The problem is that when related objects are fetched (be it eagerly or
> lazily), the corresponding repository is not used, hence the default
> query settings are not used either. Instead
> Tx_Extbase_Persistence_QueryFactory::create() just looks up the storage
> pid but no more.
>
> I have tried to work out a patch where
> Tx_Extbase_Persistence_QueryFactory::create() tries to instantiate the
> corresponding repository. This proved a bit hard as my knowledge of
> Extbase is not very deep. I have attached my draft here as I know it's
> not good enough as it is and maybe I even missed the whole point. I
> would be quite glad for feedback.
>
> If you agree that this is indeed an issue, I will gladly accept help in
> improving this patch. In particular I'm using t3lib_div:makeInstance to
> the repository instead of $this->objectManager->create(), because I then
> get the following error:
>
> #1265203124: Object "Tx_Foo_Domain_Repository_BarRepository" is of not
> of scope prototype, but only prototype is supported by create()
>
> which is typically where my limited knowledge of Extbase fails me and I
> don't know how to do this properly.
>
> Versions: TYPO3 4.5.10-dev, hence Extbase 1.3.1-dev.
>
> Note: additionally this patch adds a test in the TYPO3 persistence
> backend when adding the storage pid condition, to avoid doing so when
> there are no pids. This is apparently a consequence of the rest of my
> patch. Maybe there's a way to avoid this, although the check seems to
> make sense anyway.
>
> Thanks in advance for your opinion and advice.

Thanks

-- 

Francois Suter
Cobweb Development Sarl - http://www.cobweb.ch