[TYPO3-core] Minutes of the 6.0 release team meeting

Thomas Maroschik tmaroschik at dfau.de
Thu Jul 19 12:32:19 CEST 2012


Hi Jigal,

> Which spreadsheet?

See here: 
https://docs.google.com/spreadsheet/pub?key=0AjQjYvMWGLr4dHZlb0oyWF9YZVdudk9zREdXQU1hYkE&single=true&gid=0&output=html
If you are willing to contribute to this list, get in contact with me 
and I'll share a editing link.

> How much performance is this going to cost?
The new class loader will not cost more as it currently does with it's 
autoload registry. If extensions require() files instead of using 
autoloading, there can be minor performance issues. As the class alias 
function is a builtin function, I expect it to be fast.

> Which basically means that we need separate versions of *all* extensions
> for 4.x and for 6.x? That's not going to happen.
> * most extension authors will not make updates until some time after
> things really break. In the mean time hardly anything is working. This
> means that if the original classes will be removed in 6.2, it basically
> means that any upgrade to 6.2 will be broken for many months.
> * TER and the IM are not capable of keeping both branches apart. It will
> suggest the latest version for 4.x too, which will not work
This is not necessary (see next comment). New introduced extension that 
aim for 6.0+ versions can use the new naming scheme.

> Getting this to work with new class names would require a forward
> compatibility layer in 4x too. I'm not sure the RMs of those versions
> are willing to introduce this, especially because it could very well
> influence stability of those versions.

I'm not expecting extensions to adopt the new class names until 6.2 if 
they should work also on 4.X. By that version all 4.x Versions become 
obsolete and extension authors can start using the new class names. The 
old ones can continue to work until 7.0.

> What is the use case for this?

It's a chance to clarify the concerns of the distinct classes and path 
the way for easy forward- and backporting to FLOW3/Phoenix. The core 
gets easier to understand for new developers. As v4 continues to live in 
v6+ it would also be a signal to the PHP world, that the community is 
interested in code quality and getting rid of legacy code. TYPO3 has 
plenty of that and the definition of clear class concerns makes it 
easier to identify that.

Regards,

Tom



More information about the TYPO3-team-core mailing list