Recipe for World Domination

Posted by Orville Bennett on 14 July 2008
Read time: about 4 minutes

After having a talk with aseigo today I have been lulled out of my ill-deserved slumber. So let's try something and see where this goes...

Plan

Step 1: Collate an actual list of all KDE's dependencies (Techbase).
Step 2: Find out which are actually used for functionality on OS X (libagg, opengl).
Step 3. Remove or otherwise move out of the way X11 includes and libraries.
Step 4: Compile and package software only requirements software provided by the base the operating system. Qt first.
Step 5: Repeat Step 4 for the kdesupport software from a stable snapshot.
Step 6: Compile and package kdelibs and kdepimlibs.
Step 7: Find out what is necessary in kdebase runtime, for software to work properly.
Step 8: Assess the impact of functionality caused by deliberately leaving out some software (nepomuk, kdemultimedia for amarok, opengl support for kdelibs).
Step 9: Create individual .app packages for third party software and a .mpkg to install the dependencies packaged above.
Step 10: Test, then distribute to the world.

Aside: I volunteer for 10.5 univeral packages. 10.4 "box" is at work and the Tiger install CD is all the way in NY (I hope).

Coordinate

  1. Mail the kde-mac, kde-core-devel, and also kde-buildsystem MLs to discuss the state of KDE/Mac and the desired "message" to be delivered with the 4.1 release.
  2. Lay the groundwork for 4.2 by discussing the feasibility of making kdelibs into frameworks. Find out about "Solid" plans for OS X. Find out wtf nepomuk doesn't work. Is this just me?
  3. Start looking at qt 4.5 and ask Nokia/Trolltech what their plans are.
  4. Coordinate with the rest of the team on wiki by assigning tasks or letting everyone know what tasks you are working on and how far along they are.
  5. Discuss officially using macports pkgs for 10.5 universal pkgs. 5.5 Discuss how to show 10.4 some love.
  6. Solicit information from app developers about whether they want their apps packaged for OS X and what their expectations are on the platform. Run it and be happy, or as close as they can get to seamless integration? Right now Amarok, Bibletime and KTorrent come to mind.

Initiate

Start creating individual prerequisite packages now. Try to keep versions as stable as possible (no git for dbus), (no svn for strigi, taglib). Possible exceptions qt-copy, dbus.

Design qt-copy package to directly override/overwrite macports qt-mac package. Collaborate with macports maintainer to have qt-mac +qt-copy port variant which applies kde patches. This allows easy testing of qt and qt-copy

sudo port deactivate qt-mac +qt-copy && sudo port activate qt-mac

Also need qt-mac built w/ dbus support.

Try getting dbus to release w/ mac changes integrated so we can have stable version for 4.1.

Communicate

Nothing said till 4.1 release? Announce as tech preview in 4.1 release announcement and link to packages? Would need to speak to kde-promo on this one so we don't phail as utterly as with 4.0. Mac geeks will be far less forgiving (albeit far more polite).

Pimp the website moar. Mission Statements. Screenshots. Interviews (not it)? Screencasts (personally I think this' a lame idea). Information accurately stating what KDE/Mac can and can't do. Me n' Yaccin can work on this if he's got time. If not, I'll add it to my suddenly very long TODO. :-)

Have things set up for bugs.kde.org to allow kde-mac packages to be listed.

Insanity

Do it all before the end of the week. Volunteers welcome ;-)