>>>>> On Mon, 29 Nov 2010, Graham Murray wrote: >> I guess it is triggered from pkg_postrm() of python-2.6.6-r1 which >> until two days ago unconditionally called the following eselect >> action:
> But python-2.7 is installed in a new slot and python-2.6.x is not > removed. So. surely python-2.6.6-r1's pkg_postrm() should not be > called during the installation of python-2.7.1. You are right, this cannot be the reason then. But could pkg_postrm() of python-3.1.2-r4 have caused the update? It essentially executed the following code: [[ "$(eselect python show)" == "python2."* ]] && eselect_python_options="--python2" eselect python update --python3 > /dev/null eselect python update ${eselect_python_options} Ebuilds for 2.7.1 and 3.1.3 were committed together, and 3.1.2-r4 and 3.1.3 are in the same slot. Ulrich