Query update APPLY

[dantleech]

• Allow applying a function to result sets
• Added apply functions "mixin_add" and "mixin_remove"

PHPCRSH> UPDATE [nt:unstructured] APPLY mixin_add('mix:foobar'), mixin_remove('bar:foo')

Fixes #99

Other things we could add:

• lock: Lock the node
• primary_type: Set primary type
• lock_refresh
• lock_token_add
• lock_token_remove
• unlock
• lifecycle_follow
• property_set: Alternative to SET a = b
• retention_hold_add
• retention_hold_remove

[dantleech]

@dbu any other low hanging fruit for the APPLY method?

It would be nice to be able to set the primary node type, but thats not possible in Jackrabbit - and I guess not in DBAL currently?

[dbu]

setting primary type would require jackalope/jackalope#247 - apart from that it makes sense to support it here...

i would be -1 on an alternate to SET a=b - there should be only one way to do one thing. (if this is internally using the same mechanism, that is all fine by me).

locking without a lock token only makes sense within a session. you could try it though, so i could lock something in the shell, then do my stuff, then release the lock. lock with token is not supported yet by jackalope (but would make more sense in web applications than the locks we already support). dbal does not support any locking at all.

the syntax seems fine to me, we can always add more things as they come up.

[dbu]

looked over the code. i am quite impressed with how little changes you can implement this feature!

it looks like the mixin:remove fails. are you using dbal or jackrabbit for the tests? could be jackrabbit does not support removing a mixin - not sure though.

[dantleech]

Thanks :)
.
The failure is due to fixture failures .. I think next time I write tests like this I will declare the fixtures in the behat feature instead of loading external fixtures. The tests are currently kinda interdependent.

For now we can just have mixin_add and mixin_remove.. the others can come later.