[webkit-qt] branching for Qt 5.0

[Simon Hausmann]

Hi,

There's an on-going effort to increase the pace for a timely Qt 5.0.0 release, 
with pruning bugs, figuring out what _real_ blocking bugs are and what amount 
of polish we need. If you're interested in the release, if you're interested 
in a good 5.0 then please read on. If your focus is on a future feature or 
you're always using trunk anyway, then feel free to skip this email :)

Now is the time to put ourselves into the "seats" of customers and look at 
QtWebKit from the outside. It's the time to write more examples, write 
documentation and fix _critical_ bugs. It's important to realize that we'll 
always have bugs in the code, but what matters with Qt 5.0.0 is IMHO a really 
good first impression that translates into confidence that this is a fine choice 
for a WebKit port. That means for example a smooth porting experience from Qt 
4.8, a great performance of the QML2 integration and great overview 
documentation introducing people to the new offering.

I urge everyone (who is still reading this email) to look at

    https://bugs.webkit.org/showdependencytree.cgi?id=76773&hide_resolved=1

On a high-level I can see that traditional ARM is falling behind, and I'm not 
sure if we should block the release for it. What do others think?

We also need to complete the library split. Pierre and I are I would say about 
70% done (on github) and we plan on landing the first portion of the split this 
week and finished by end of next week.

The versioning of QtWebProcess and the move to libexec/ is also IMHO a 
critical piece for deployment, although it's easy to fix.


The biggest change in terms of code churn is the library split. I was thinking 
of branching off when that's complete and cherry-pick from there. Any 
opinions/thoughts about that?


Comments welcome :)


Simon