Plasma does not build without kdepimlibs

[Stephen Kelly]

Aaron J. Seigo wrote:

> On Thursday, September 15, 2011 00:21:39 Stephen Kelly wrote:
>> 9edc8d34b3b0f9983f0eb014f8fbf4bcfcffc3f1 introduced a dependency in
>> plasma on kdepimlibs for gpgme++.
>> 
>> The cmake check in kdelibs for kdepimlibs claims it is optional, but the
>> build fails later. The stuff that uses gpgme++ should be compiled
>> conditionally based on whether kdepimlibs was found.
> 
> imho gpgme++ should not be part of kdepimlibs. until that is fixed as part
> of the Frameworks effort, i don't particularly see anything useful coming
> from spending time fixing the build as it will need to be re-done again
> once gpgme++ is broken out (as it probably should have been from the
> beginning?)

If there is not going to be conditional compilation then you have to make 
kdepimlibs a hard dependency of kdelibs. Currently it is reported as 
optional, and then the build fails later. If it is a hard dependency then it 
should fail early.

I think this only applies to the frameworks branch and whatever branches you 
guys are using. It doesn't apply to KDE4 branches.

Later today I can change the frameworks branch to conditionally compile 
plasma based on whether kdepimlibs is found.

I don't know when gpgme++ will be broken out. There hasn't been any effort 
or analysis into breaking kdepimlibs up yet. There's still work ongoing in 
master towards a 4.8 release of it.

> 
> on a related note ... there is other code in libplasma2 which uses qca.
> this needs to get sorted out as well ... i'd like to see it use one or the
> other, but not both.
> 

I don't know the feature differences, so I'm afraid I can't help on that 
decision.

Thanks,

Steve.