Does the following make sense in some strange way, or is this simply a bug? $ port installed apr The following ports are currently installed: apr @1.2.12_1+darwin_9 (active) apr @1.2.8_0 $ sudo port uninstall apr @1.2.8_0 ---> Unable to uninstall apr 1.2.8_0, the following ports depend on it: ---> apr-util ---> subversion ---> subversion-perlbindings ---> subversion Error: port uninstall failed: Please uninstall the ports that depend on apr first. Basically, I have 2 installed ports with different variants used. But the dependency check on uninstall does not seem to recognize the one with the variant "darwin_9" and complains, even though it should be happy, since the newer version is there, and dependencies are oblivious of variants (and versions). That looks wrong to me. Florian -- Florian Ebeling florian.ebeling@gmail.com
On Mar 26, 2008, at 16:11, Florian Ebeling wrote:
Does the following make sense in some strange way, or is this simply a bug?
$ port installed apr The following ports are currently installed: apr @1.2.12_1+darwin_9 (active) apr @1.2.8_0 $ sudo port uninstall apr @1.2.8_0 ---> Unable to uninstall apr 1.2.8_0, the following ports depend on it: ---> apr-util ---> subversion ---> subversion-perlbindings ---> subversion Error: port uninstall failed: Please uninstall the ports that depend on apr first.
Basically, I have 2 installed ports with different variants used. But the dependency check on uninstall does not seem to recognize the one with the variant "darwin_9" and complains, even though it should be happy, since the newer version is there, and dependencies are oblivious of variants (and versions). That looks wrong to me.
This is simply a deficiency in the way MacPorts currently works. MacPorts only knows that apr-util, subversion et al depend on apr. So when you try to uninstall (any version of) apr, it complains. Since you already have a newer version of apr installed, it's fine to uninstall the older one. You just have to force it to overcome MacPorts' objections. sudo port -f uninstall apr @1.2.8_0
participants (2)
-
Florian Ebeling
-
Ryan Schmidt