Alexander Neundorf wrote:

> On Friday 04 February 2011, Rex Dieter wrote:
>> Alexander Neundorf wrote:
>> >> 2.  copy-n-paste the aforementioned code all over (similiar to what's
>> >> already in FindKDEInternal)
>> >>
>> >>
>> >> 3.  wait for cmake to handle this better natively
>> >
>> > I think it's option 3.
>>
>> How about compromise 2+3 ? :)
> 
> How do you mean that ?
> For now, I think we'll have to go with 2.
> But this shouldn't be necessary for any package which uses KDE, i.e. which
> does a
> find_package(KDE4)
> right ?

Another edge-case, digikam (which does do find_package(KDE4)), for which we 
carry the following patch, is it's use of rpath properties really needed 
here, or is there a better solution?



-- Rex
diff -up digikam-1.8.0/libs/threads/dynamicthread.cpp.gcc46 digikam-1.8.0/libs/threads/dynamicthread.cpp
--- digikam-1.8.0/libs/threads/dynamicthread.cpp.gcc46	2011-01-23 13:27:56.000000000 -0600
+++ digikam-1.8.0/libs/threads/dynamicthread.cpp	2011-02-04 10:05:11.969447213 -0600
@@ -257,6 +257,12 @@ bool DynamicThread::DynamicThreadPriv::t
             kDebug() << "Transition to Running: Invalid Inactive state" << q;
             return false;
         }
+	// fix gcc-4.6 failure --  error: control reaches end of non-void function [-Werror=return-type]
+	default:
+        {
+            kDebug() << "libs/threads/dynanamicthread.cpp : Should never reach here: assert?" << q;
+            return false;
+        }
     }
 }
 

_______________________________________________
Kde-buildsystem mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/kde-buildsystem

Reply via email to