Re: [tlug] Um, so... systemd?

["Stephen J. Turnbull" <turnbull.stephen.fw@example.com>]

Curt Sampson writes:

 > Perhaps we should consider running a workshop where [somebody]
 > demonstrate[s] the systemd way of doing this and [somebody]
 > demonstrates the init.d way.

That's a good idea!  Unfortunately I can't volunteer for either, no
real skills in the building department, though I've debugged a bit
(who hasn't?)

But back to the (birthday-candle-sized) flamewar: I think my biggest
issues with Debian going to systemd has been that a lot of the ancient
software I still run has never been converted so I end up starting
shit manually on reboot (fortunately that only happens with the
University turns off all power two or three times a year).

The binary logging thing makes me nervous because IME Mac OS X's
logging is where it sucks hardest.  It's also bloody annoying on Mac
that launchd (their systemd) is so badly documented and half the
documented stuff doesn't actually work as documented (fortunately? 
that's a rapidly diminishing fraction of launchd functionality :-/ ).
So a lot of the nervousness I feel about systemd comes from that
experience -- what if I ever don't feel like accepting Debian's
default configuration?  (Debian default logging still mostly goes to
the traditional files eventually, so except for the legacy apps I have
not experienced any real problems from systemd.)  Will I be able to
figure out how to do shit in reasonable amounts of time?

 > Actually, that blocks suprisingly little. In fact, I can't think of
 > anything off-hand where you'd even need a network connection

Fscking T-dai network takes up to 10s (or multiples of that if I have
to intervene and switch SSIDs due to congestion).  Blocks email which
is still my single most important work application.