It would be nice if systemd.eclass would finally be usable in EAPI=6, at least the basic functions like systemd_dounit(), systemd_dotmpfilesd(), systemd_with_unitdir(). I see nothing which requires any change for EAPI=6 except for the blocking test...
Well, I'd honestly ban systemd_with* and sysyemd_to_myeconfargs functions. And maybe add phase check to catalog update.
(In reply to Michał Górny from comment #1) > Well, I'd honestly ban systemd_with* and sysyemd_to_myeconfargs functions. This is just my personal opinion, of course: I do not care whether they are banned or not. Important is that e.g. systemd_get_unitdir() is kept; the rest are just convenience wrappers, anyway. That being said, I guess that nobody will argue against banning the already obsoleted systemd_to_myeconfargs. On the other hand, I see nothing wrong with systemd_with_unitdir(), because it seems that the option name --with-systemdsystemunitdir has become a quasi-standard and follows upstream's recommendation from e.g. http://www.freedesktop.org/software/systemd/man/daemon.html
Patches sent for review.
And pushed.