On Mon, Aug 21, 2017 at 5:54 PM, BohwaZ/PHP <p...@bohwaz.net> wrote: > Le 22/08/2017 07:55, Adam Baratz a écrit : > >> A new method is an API change to me, so an RFC would be warranted. I'm >> reluctant to add driver-specific methods, since that seems opposed to >> PDO's >> driver-agnostic API, but that's not to say we couldn't hash something out. >> > > Do we need a RFC every time we patch for a missing method ? > > Btw there's already sqliteCreateFunction, sqliteCreateAggregate and so on, > so it's pretty much just the continuation of that logic, even though I > don't really like it either (especially as lazy-loading by extending PDO > makes it so that the methods are not accessible until the connection has > been established, which is a weird behaviour imho).
I see. Didn't realize it was an existing pattern. Not going to say I love it, but that's not a reason to break with it now. We could maybe talk separately about better ways to do this sort of thing, to avoid the situation you describe, but I don't need to derail this conversation any further. Thanks for taking the time to get parity between the extensions. Adam
