[RFC|Merged] General Hook Architecture of phpBB3.1

[MichaelC]

I like naderman's new suggestion.

I'll also rename RUNHOOKS to EVENT now to get it done with.

Also if everyone agrees with naderman's implementation I'll change all existing events to use it.

[nickvergessen]

I liked the extract() on a new line more, as you saw $event at the line beginning which helps finding it, but aslong as its called something with events (nadermans latest suggestion) i'm also fine with it.

[Oleg]

(11:20:10) nn-: naderman: my only issue with that is people accidentally screwing up the magic bits
(11:20:31) nn-: but i guess all of *that* code will be reviewed by us making it potentially a non-issue
(11:21:17) nn-: oh just noticed method rename
(11:22:14) nn-: yes i wanted a different function name for the version that filters
(11:22:21) nn-: and potentially has a different api
(11:22:36) nn-: trigger_event seems to be a good choice

[igorw]

I personally prefer the initial more explicit version, but renaming the special method to something other than `dispatch` addresses my main concern of changing the semantics.

[MichaelC]

RUNHOOKS is now EVENT

All php events have been updated to naderman's suggestion and all template hooks to EVENT

[nickvergessen]

On question that came up recently is, who we are going to deal with hooks while installing/updating(3.1.0 to 3.1.1) the board.
F.e. when we add a user it might be good to have the hooks running.

[MichaelC]

On question that came up recently is, who we are going to deal with hooks while installing/updating(3.1.0 to 3.1.1) the board.
F.e. when we add a user it might be good to have the hooks running.

Why would you add a user with updating? Personally I think its a bad idea allowing hooks to run while updating.

[nickvergessen]

Well we added some bots latly and when we add a bot we call user_add

[Oleg]

Somewhat related: http://tracker.phpbb.com/browse/PHPBB3-10709

[Oleg]

We have a core. prefix for php ledges but not for template ledges. I propose we have a core_ prefix for template ledges.

Thoughts?

Was this done? If not I'm inclined to agree to add core_ for template events.