Function distutils-r1_python_install_all assigns variable DOCDESTTREE directly. It's not entirely clear what the status of the DOCDESTTREE variable is, see bug 460512. However, in either case it qualifies as a bug in the eclass: - If it's a PMS-defined variable, then it must not be modified per PMS section 11.1: http://dev.gentoo.org/~ulm/pms/5/pms.html#x1-11800011.1 - If it's not defined (and in fact, the variable doesn't exist in Portage), then the assignment is futile. Please change the eclass to use the docinto function instead.
Wrt #1, does that mean that 'local INSDESTTREE' to avoid subshelling just to install one file is illegal as well? Can we finally replace PMS with a spec that is at least self-consistent and have more defined behaviors than undefined ones?
(In reply to comment #1) > Wrt #1, does that mean that 'local INSDESTTREE' to avoid subshelling just to > install one file is illegal as well? "Ebuilds must not attempt to modify any of these variables, unless otherwise specified." Obviously, ebuilds can use insinto to modify INSDESTTREE, so I don't see a problem here.
Switched to doins then.