[KDE/Mac] KDE 4.13 on MacPorts

[Ian Wadham]

Hi again Nicolas,

On 16/06/2014, at 12:53 AM, Nicolas Pavillon wrote:
> I made the test of upgrading locally of ports to KDE 4.13.2, and apart from some hiccups, everything went ok.

Great news!

> Several patches could also be suppressed as they were committed upstream.

Except the baloo malloc_trim() patch that was miscommitted… ;-)

As I mentioned elsewhere, I am also building KDE 4.13 from the git
repositories, not to compete, but to have ready access to KDE source
code independently of my MacPorts installation and be able to track
down KDE portability bugs on OS X and fix them.

As you may have seen on macports-devel, I have found out why
KDE does not handle crashes of apps properly on OS X and the fix
is quite simple.

Now to find out why Dr Konqi and some other KDE dialogs come up
*behind* the user-app's main window, and why Report Bug does
not work, etc., etc.

This raises the question of what to do with the patches I develop.

I could just commit them to the KDE repositories, as KDE 4.13 bug
fixes, but I might get a rap over the knuckles from KDE sysadmin
when I commit to kdelibs, kde-runtime, etc.  Also, the patches might
not get carried forward from KDE 4 to KF 5.  And they might get
lost to view in the general hail of other changes to KDE.

Then again, some patches might have to be different for KF 5 or might
no longer be applicable.  For example, Qt 5 should have fixed the raster
graphics vs. native graphics issue on OS X, but there could still be
some surprises in store.

I think it might be best to keep these patches in a separate repository
area somewhere, where they can be selectively accessed, used in
KDE upstream, used in MacPorts only or discarded, as KDE goes
through its intensive development period re KF 5 for the next several
months.  Also the patches will then be easier to find when the
MacPorts/KDE CI system gets set up.  And anyone else who would
like to join me in chasing KDE portability bugs can post patches there.

KDE playground is a possibility.  Any other suggestions?

Cheers, Ian W.