[KDE/Mac] kdesrc-build

[Ian Wadham]

Hi Marko and Jeremy,

On 26/11/2014, at 9:08 AM, Marko Käning wrote:
> On 25 Nov 2014, at 22:21 , Jeremy Whiting <jpwhiting at kde.org> wrote:
>> 2. What are the major barriers, technical and/or organizational, that prevent newcomers from getting,
>> building, and contributing fixes to kde on mac?
> 
> Good question. I think one important factor is that there aren’t many KDE devs on the OSX platform.
> Also, remember, the KDE-MAC ML is mainly driven by us folks from MacPorts.
> It would be great to have more feedback from the Homebrew guys here.
> 
>> I think I remember a link to a wiki of technical issues,
> 
> This one it is: [1]! :)
> 
>> but I'm also interested in organizational issues. If we can solve some/most of these we could get
>> many more hands helping with this work.
> 
> Well, see above! :)
> We definitely need more real KDE developers (like Ian and René) on our team!!!

Well, I don't count myself a REAL KDE developer… :-)  Sure, I have been on KDE games for about
12 years, but I have never ventured far into the core KDE libraries and utilities until recently.  As to
"barriers":

1. Sadly, it has turned out that our most serious problems have been in the KDE core, e.g. Dr Konqi,
     kdeinit4, klauncher4 and kded4 not working well in the OS X environment.
2. There have even been some Qt 4 problems, e.g. raster graphics not fully finished on OS X, and
     Apple menu layouts getting misconstructed, especially for plugins and KParts.
3. We find KDE doubles up on or battles with facilities that already have equivalents in OS X, e.g.
     KWallet v. KeyChain, Baloo/Nepomuk v. Spotlight, KCookieJar v. Safari/Firefox.
4. We find some KDE apps, utilities and libraries hard-wire-in dependence on KDE Desktop features,
     sometimes without the application writer being aware of it.
5. AFAIK there has been no serious testing of KDE 4 on the Apple OS X platform for several years.
     Some #ifdef code I have found (for Mac) has been written, but clearly not tested.  Marko's CI effort
     is directed at this problem.
6. Point 5 applies to KDE 4 code, lots of it, but the focus in the KDE community is strongly on KF5
     and Frameworks, as it must be.  KDE 4 maintenance is now all but closed down, just during the
     few months when all the problems with KDE 4 on Apple OS X have been at last coming to light.
7. KF5/Frameworks is a long-distance view from Apple-land. Even when a substantial number of KDE
     apps have been "ported", there will be a second porting required (to MacPorts/HomeBrew etc).
8. Our initial efforts to raise the above issues were met with hostility, disbelief and failures to answer
     appeals for help from core developers.  I think/hope the tide has turned now.  We seem to be
     being treated with more respect and helpfulness in reviews now, and we have always had a good
     relationship with a few core KDE people, otherwise we could not have got anywhere.

>> P.S. My main reason for initially setting up a mac vm is to build/run applications and fix issues on
>> this platform that don't appear on other platforms. CI is interesting to me, but not as much as
>> individual developers building, running, and bugfixing on this platform.
> 
> As I said, I’d be happy to receive support for cleaning up my QStandardPaths patch for the OSX/CI
> system [2].
> 
> Greets,
> Marko
> 
> 
> [1] https://trac.macports.org/wiki/KDEProblems
> 
> [2] http://quickgit.kde.org/?p=clones%2Fwebsites%2Fbuild-kde-org%2Fkaning%2Fmp-osx-ci.git&a=blob&h=e0b488e55a279023c1750543cb1daf8b40980321&hb=41fa7caf1af41a74d7cfd09dbb480d0bd70dce79&f=patches%2Fqt5%2Fkf5-qt5%2Fpatch-qstandardpaths_mac.cpp.diff

All the best, Ian W.

PS. Please don't get put off by what I have to say, Jeremy. I just like to tell it like it is… :-)
       We really need reinforcements on KDE-Mac more than anything.