[TYPO3-core] RFC: Integrate Doctrine2 into TYPO3 CMS

[Daniel Hürtgen]

Hey Oliver, Christian & Philipp,

first, i hate my english, but i'll try. :D

I also thing it would be really great to have doctrine right in the 
core. I've just started an public GIT repo upon Github[1].

I think Doctrine is a modern way to persist data to different storages. 
First i started with an implementation of the doctrine-dbal package, 
because i think that the part the core should serve.

The second part will to integrate and replace the extbase persistence 
with the doctrine-orm. I'm not sure if it is the right place, so we 
should discuss this here. What do you think?

[1] https://github.com/higidi/TYPO3v4-Core-Doctrine/

Am 26.01.2013 14:18, schrieb Philipp Gampe:
> Hi Christian,
>
> Christian Kuhn wrote:
>
>> Hey.
>>
>> I think it would be great to get doctrine as persistence layer in the
>> core. Maybe only as optional feature for some versions, similar to what
>> was done with the caching framework (parallel to the old code).
>>
>> On 01/26/2013 11:30 AM, Oliver Hader wrote:
>>> In general it can be summarized to the following goals:
>>> * lower the effort to maintain our custom persistence layer
>>> * add the possibility to have better support for translation and
>>>     workspaces
>>> * use a standard that has been proven to be good and flexible
>>> * and thus get a bit more in sync with TYPO3 Flow again which utilizes
>>>     Doctrine already for quite some time
>>
>> * solve dbal once and for all
>>
>>
>> There are two special parts in current core that should be considered to
>> be dropped along the way (with bullet proof migration wizards). They
>> could save us tons of headaches.
>>
>> Without going to deep into details:
>>
>> * Drop comma separated id list feature and adapt be_user<->be_group and
>> fe_user<->fe_group to use real mm.
>>
>> * Drop 'localization overlay in different table' feature. In the core,
>> this is only done for pages<->pages_language_overlay, the code is
>> hardcoded at multiple parts throughout the core, so the features
>> probably only works for *this* table anyway and is probably not used by
>> extensions (at least I know none, and I've also not seen any bug reports
>
> +1
>
> I guess it will take some major efforts to change all places in core to the
> new code, but I really think this is worth it.
>
> We must also consider that the mysql extension in PHP will not be supported
> for very long time any more. That means we will have to get our hands dirty
> anyway?
>
> How about integrating a development preview into 6.1, hopefully stabilize it
> with 6.2 LTS and drop the old DB stuff with 6.3?
>
> I guess porting the extbase layer will be the easiest part?!?
>
> Best regards
>