A rough account of what I did with Emacs recently.
- August 14
- Fiddling with route planners for the EV: "A Better Route
Planner" does the work of figuring out your stops for you, but I'm
having some issues with the user interface; "PlugShare" seems a good
deal more comprehensive in terms of both available charge points and
access to more detailed information without subscribing, but doesn't
plan your stops; and the EV's own routing app (on the phone) doesn't
appear to allow for additional waypoints. None of the three seem to
have the concept of "round trip", as in, "I'm driving to this
location and will be driving back to my starting point some time
after that", which seems... basic? I dunno. Maybe there's a huge
cohort out there who just get in their cars and drive, never to be
- August 13
- Managed to cherry-pick the slightly broken debian box into
installing a reasonably current version of SpamAssassin, which was
about the most critical thing it was missing after last weeks'
upset. I'm still debating how much effort to put into recovering
this vs. setting up a new box from scratch with the user data
- August 9
- Started watching Vikings;
as with other stuff we've watched, it's included for "free" on
PrimeVideo and there are at least a few seasons - although it
looks like only half of the series, alas. We'll see how it
- August 6
- Sort of understanding what happened to the server:
- Originally, it was a debian box, running whatever flavour of
debian was current when it was built.
- Then it was a debian + unstable box, so it was getting
"unreleased" packages added.
- Then I decided to remove the unstable source and let the stable
source catch up, if it would, with whatever the installed software
- Then whatever happened t'other day happened; this
resulted in removing a large number of packages. Many of these
packages appear to have been from unstable, and did not graduate to
stable at any time.
- So now it's a bit of a frankenbox and some of the
previously-installed packages can't be installed without re-enabling
unstable. I'm debating at this point whether it's worth the effort
of just outright upgrading to the next debian release, or just
building a new server and porting things over to it.
- August 4
- Still not clear on what broke but the server is more or less
back in working order.
- August 3
- Wrapped up Blacklist: Redemption. Not terrible, and
while there was a bit of a cliffhanger and no more episodes
because the show was cancelled, things were well enough wrapped up
to not be annoying.
Had a minor oops with a server I manage where what should've been
routine maintenance broke the box's connectivity. Fortunately I
have scripts lying around to help recover from such incidents, but
I'd forgotten I'd made them kinda slick:
$ ~/src/sh/ec2-drive-swap.sh -p awsprofile -s hostname
21:12:00 Collected sick instance ID 'i-deadbeef' for 'hostname'
21:12:02 Sick device '/dev/xvda' is 'vol-deadbeef'
21:12:03 Collected recovery instance ID 'i-feedbead'
21:12:06 Will attach sick volume as '/dev/xvdc'
21:12:06 Stopping i-deadbeef:
waiting for 'i-deadbeef' to reach status 'stopped': ...done
21:12:33 Detaching sick volume vol-deadbeef:
waiting for 'vol-deadbeef' to reach status '': .done
21:12:37 Attaching vol-deadbeef to i-feedbead:
waiting for 'vol-deadbeef' to reach status 'attached': ..done
Sick volume is attached to recovery instance. Log in to i-feedbead and do your stuff, then exit the shell to continue
[drive recovery - exit when done] $
On top of that, when I logged in I discovered I'd left scripts on
the damaged drive that mount and unmount the relevant partitions
to allow it to operate as a chroot environment. Slick, I tells
ya. And when I log out of the recovery shell, it reattaches the
volume to the correct instance and restarts the
(just don't look at the code.)
At this point I've partially recoved the server but I'm still
trying to figure out what actually went wrong, so if anyone still
uses the DSPsrv and is wondering why they can't connect... that's
- August 1
- Weird failure on one of the decrepit macs: a file created in
~/Desktop didn't show up on the desktop, so I figured
maybe logging out might be good (I have experienced failures in
the past where the desktop process has become disconnected from
its security context, or something like that). Logging
out... hung. Since I had a shell connection to the same machine, I
rebooted it. That hung, too. So hard power off, and... booting
hung. At this point I was envisaging the annoyingly stereotypical
"mac claims hardware is in full working order right up to the
point it catastrophically fails"; however, I managed to boot to
recovery, nothing looked amiss, and then booting verbosely showed
an awful lot of errors that seemed symptomatic of problems in the
signing infrastructure. But no, eventually it came back to life. I
have no idea, and I am not helped by the fact that
there's a good deal of alarming-looking log messages that are
apparently safe to ignore.
Finished round one of my battle with the Variorum edition of
Ulysses: I have read, after a fashion, the original
publications from The Little Review. It's tough going in
spots, wildly amusing in others. I am now looking at what I think
is the original 1922 print of the novel itself, so let's see how
that goes. "Stately, plump Buck Mulligan..."
previous month | current month
Your August Presence.