[KDE/Mac] QtKeyChain

[Mario Fux]

Am Samstag, 26. Juli 2014, 10.26:31 schrieb René J.V. Bertin:
> Hi Ian,

Morning guys

> On Saturday July 26 2014 14:18:37 Ian Wadham wrote:
> >Nope.  Not me.
> 
> I recall asking who does, not who does NOT ;)
> 
> >> For instance, is there support for backends or plugins in kwallet?
> >
> >Dunno.  But I think it might be more appropriate to have a stripped-down
> >or "derailed" KWallet, which would simply switch to QtKeyChain and
> >OS X Keychain, wherever it was invoked, bypassing the usual KDE code.
> 
> In fact, that's more or less what I realised too, after I noticed that
> QtKeyChain uses kwallet under Linux ... So I guess I'm going to contact
> the author to ask him how to isolate the code that talks to the OS X
> keychain, and see how (and where!) I can patch that into kwallet.

The KWallet maintainer might know best and thus I added him as CC: (please 
keep this).

> >One thing bothers me.  QtKeyChain is not part of Qt 4 and 5.  Are there
> >other instances where MacPorts is using Qt libraries from third parties?
> 
> That's a different point, and I'm not sure how relevant it is. Supposing we
> do this in MacPorts (as I indeed intended) and as suggested above (i.e.
> injecting code from QtKeyChain into kwallet), then it would most likely
> take the form of a big (huge) patch with a bunch of additional code. That
> code could be hosted on github and downloaded as needed by the port, and
> we'd have the choice of making the OS X keychain kwallet an optional port
> variant, or instead make the native kwallet an optional variant. If the
> osxified kwallet works well we could of course also file an RR.

Oh and I'd like to see as much work upstream as possible. Don't create patches 
here and there that get old. Let's try to get them in KDE repos and thus 
upstream and thus all people will profit. I might at least help directing you 
to the right people and help poking them ;-).

Best regards and thx
Mario