3.0

[will123195]

Not yet ready to merge...

TODO:

• refactor getList
  • namespace
  • remove unnecessary joins
• aql distinct
• option for sql() to throw exceptions
• don't connect to db until the first sql()
• remove postgres 9.0 dependency
• add missing documentation

[stanistan]

My thoughts are that this should be a new repo because it will break backwards compatibility.

And if it goes in that direction, there is an opportunity to completely change the architecture (I think for the better). What follows is opinions in an ideal scenario.

I am a big proponent of small libraries and a framework should be a collection of them.

This should be using a package manager (composer) + submodules for dependencies instead of having a large hierarchy (yes, I know there are submodules). Further updates to the framework could be made by updating the pieces individually (each having its own version), and then versions bumped in the framework configuration.

This also has the added benefit of them being used individually (and potentially interchangeably) with other frameworks. They could just be more useful as generic tools.

Things that should be separate libraries

• AQL Parser (should be ported from stanistan/aql-parser-js, something I'd like to do at some point)
• AQL ORM should be built with the parser as a dependency
• Router (all in all, this is just string pattern matching)
• Sky\Mustache (partial autoloading)

(and probably others)

The root skyphp framework should be minimal.

Drawbacks

Really, it's that it would take a lot of time, and may not be worth doing.

[will123195]

I totally agree. I've been thinking the same thing about splitting it into as many stand-alone repos as possible. I think that's a good premise for 4.0.

Essentially this is a performance release and break backwards compatibility hence the major version change.

I think another premise for 4.0 is making "installation" easier.