2010-03-08: wallet 0.11
The primary reason for this release is that I messed up some of the
Kerberos portability code, leading to compilation failures on Red Hat
Enterprise Linux 4 and other platforms with old Kerberos libraries.
While I was at it, though, I also fixed a few server problems. ACL
deletions are now prevented if the ACL to be deleted is referenced (which
should have also been done by referential integrity, but some databases
don't support it). There is now a hook for local policy enforcement of
ACL naming, and two new audit commands that find existing objects or ACLs
that violate the naming convention. There's also a new report that shows
all ACLs that aren't in use by any object.
You can get the latest version from the wallet distribution page.
2010-03-08: I'm back, mostly
This is going to be one of the few times I ever post a cute cat picture,
so enjoy it. Although this isn't cute so much as the perfect lighting for
a certain type of portrait.
It's been a very long time since I've posted a more traditional journal
entry, as opposed to a software release. Mostly this is because I was in
Nanaimo for the Winter Olympics, not to actually attend in person but to
watch on a GIANT TV with friends with the far-superior Canadian coverage.
(There was, for instance, no entirely inexplicable tape delays.)
I waited too long to write the more extended post I thought about making
about why I like Canada as a country and why the national psyche and
self-image is so appealing. The short version is that the friendliness
and politeness is real and important, the closing ceremonies is one of the
best examples of being willing to both laugh at oneself and fix something
that I've ever seen, and the poet who performed during the opening
ceremonies was absolutely awesome and I need to find him on YouTube. I
thoroughly enjoyed the trip and had as much fun watching this games as
I've ever have. The gold medal hockey game was one of the most memorable
moments ever, given that I watched it in the holding area for my plane
home with a room full of Canadian fans. It was awesome.
Otherwise, I'm working on our KDC upgrade and releasing a bunch of new
software for that. Hopefully most of the hard work is now done and I can
get a few other things finished, including getting back to a more normal
schedule and writing more posts like this. I didn't take anywhere near as
many pictures as I usually do in Nanaimo, but I still have a couple
hundred that I need to sort out, including some I still have to upload
from my camera. (Since gphoto's import feature mysteriously stopped
working. Which I probably should file a bug about if it isn't already
fixed.)
2010-03-13: First 2010 haul
It's been quite a while since I've posted a haul, in part because I didn't
read much last year and I haven't read that much this year so far, and
hence haven't felt like buying more books. But that means things have
been accumulating that I want to read, or at least own and pretend that
I'll find time to read.
Michael Andre-Driussi — The Wizard Knight Companion
(nonfiction)
Elizabeth Bear — Chill (sff)
Raphael Carter — The Fortunate Fall (sff)
Laurell K. Hamilton — A Lick of Frost (sff)
Laurell K. Hamilton — Mistral's Kiss (sff)
Peter Maas — Love Thy Neighbor (nonfiction)
Laura Miller — The Magician's Book (nonfiction)
Richard K. Morgan — Thirteen (sff)
Greg Mortenson & David Oliver Relin — Three Cups of
Tea (nonfiction)
Naomi Novik — Victory of Eagles (sff)
Patrick Rothfuss — The Name of the Wind (sff)
Rory Stewart — The Prince of the Marshes (nonfiction)
The Wizard Knight Companion is another explanatory book about some
of Gene Wolfe's books. I now have a lot of those, and really should get
around to reading the Gene Wolfe books so that I can enjoy the supporting
material.
I haven't read Hamilton in a while, but I was still enjoying the Merry
Gentry series, so I picked up the next one, plus the previous one that I'd
read but don't own so that I have a set of the paperbacks for re-reading.
All of these are from Powell's except for
The Fortunate Fall, which seems to be thoroughly out of print. I
got a copy through AbeBooks.
2010-03-15: backport 1.25
This little script to help with Debian package backporting seems to be
more widely used than I ever expected. I'm glad people are finding it
useful!
This release incorporates changes based on a patch by Eddy Petrișor to
extract the correct version number from the source package *.dsc file.
Previously, backport grabbed the version from the file name, but this
fails in the presence of epochs. (I should also extract the package name
from the *.dsc file, at which point the name of the *.dsc file doesn't
matter, but I'll do that another day.)
Also new in this release is the addition of the backporter to the
Uploaders control field automatically if they're not already in Maintainer
or Uploaders. This is currently unconditional, since I think it's
generally the right thing to do, but if someone complains I can make it
conditional.
Mostly to help with testing, I also added a new command-line option to
just prepare the backport and not actually call pbuilder. Instead, the
pbuilder command line is printed out. This may also be useful if
additional work is required on the source package before actually building
it; one can do that work and then cut and paste the command line.
You can get the latest version from my
scripts page.
2010-03-23: Lintian 2.3.4
It's been nearly two months since the previous release, so lots had
accumulated in the bug tracker. This release resolves a bunch of various
false positives, fixes a few minor issues with the
lintian.debian.org pages, and applies
all the patches that were in the BTS that were ready to go in. It doesn't
have any of the pending larger changes, since we wanted to get the smaller
fixes out quickly.
The current tentative plan is that the next release will be 2.4.0 and will
include several large structural changes that various people have been
working on, including the conversion of the manual to DocBook. All plans
are, of course, pending finding enough time.
It's amazing how effective Lintian can be given that it uses a bunch of
guesswork and loose regexes to parse everything from init scripts to
debian/rules files. We do have a bunch of accumulated wishlist bugs that
aren't going to be fixed without writing better and more thorough parsing,
particularly for shell scripts, but it boggles me just how far you can get
with some dead-simple heuristics.
Thanks very much to Adam D. Barratt and Raphael Geissert who have kept
things going while I've been eaten by work, and particularly to Raphael
for doing the last two releases and a ton of work with the security
update.
2010-03-24: INN 2.5.2
I just did my small part in the INN release process and put up the 2.5.2
tarballs prepared by Julien and posted the announcement. This release
fixes a whole ton of bugs, including much better protocol conformance.
The release notes are
incredibly impressive.
If you're running INN 2.5.1 or earlier, you should upgrade. This release
even fixes the long-standing innreport incoming feed tracking problem that
I'd never found a good solution for, in a really elegant way.
As with the other recent releases, Julien ÉLIE has done the lion's share
of the maintenance work, applying patches, debugging, helping users,
preparing the release, and doing a fantastic job. It's just wonderful to
me to see that, and to see INN development continue to be strong. I had a
horrible time coming to terms with the fact that I didn't have time to
continue to drive it, and was really worried that it would fall by the
wayside. It's turned into one of the best handoffs of open source
development that I've seen, thanks to Julien. And I even get to keep a
hand in and help a bit with debugging and history.
Hopefully one of these days I'll get a bit of spare time and a chance to
do a touch more INN development. I keep pondering proposing reworking its
build system to take advantage of lots of lessons I've learned from
maintaining other packages and update its portability and test suite code
to current versions of rra-c-util and
C TAP Harness. That would be a fun
thing to spend some time on, if I had the time to spend.
2010-03-25: Learning Java
As I mentioned a while back in my review of
Coders at Work, I've
been reusing the same programming skills for the last five years and
really need to break out of that and learn some new techniques. While
ideally that would involve learning something truly different, and I still
hope to do that, I have a fully work-justified reason to learn Java right
now, and several people in Coders at Work really love the language.
So today I started.
The plan is to take about four hours each Wednesday for the forseeable
future and devote that time to learning new programming languages, new
programming techniques, or new ways of thinking about software, starting
with Java. I'm starting with Thinking in Java by Bruce Eckel and
going from there.
I've accordingly resurrected my language
notes pages, added a final Python update (for now at least), and started a new
Java page. I'll try to make a
general journal entry when I add more to that page for those who are
interested in following my experiences.
I'm still trying to decide what my learning projects will be for Java.
The software my group will be picking up at work is more of an advanced
topic, and I'll need something shorter-term and easier to wrap my mind
around. Doing extensive work on the remctl
Java implementation is the most likely. I may also take a look at the
WebAuth Java implementation from Oxford
and see about adding AES support to it.
It's tempting to use Java to write a
Getting Things Done
implementation to replace my current Roundup hack, but I still sort of
want to do that in Python and Python's AJAX support to continue learning
more about Python.
I should come up with something else, since something I could write from
the ground up in Java would be ideal. Will have to ponder.
2010-03-27: Why no comments
When I first converted my journal from Movable Type, I converted all the
comments with it and planned on adding comment support, as a long-term
goal. But I've increasingly found that not having comment support is
nice, and recently after reading several different takes on this topic,
I've decided to take this off my long-term to-do list.
A point made by both people involved with sites with strong commenting
communities (Making Light and John
Scalzi's Whatever) and others who have turned off comments is that
maintaining a good comment community in the face of significant interest
is a lot of work. One of my goals right now is to do less work that isn't
directly interesting so that I can reserve my energy for things that I
consider more core.
I've now added a link to my explanation
for why there are no comments to all of my journal posts and indices.
A few more thoughts in addition to the ones there:
I've been reading a lot of comments recently, mostly from following the
debate in the US about health care reform, and I've been struck at how
rarely comments are meaningfully addressed to the original post author or
show intent to participate in a discussion. On political topics, most
comments appear instead to be taking advantage of the comment platform to
post one's own opinion in public, sometimes with the additional
gratification of forcing that opinion to show up in the space of someone
who obviously disagrees. I'm not interested in participating in or
enabling that. If you want to make a controversial statement or take a
political stance, own your own words and take that stance in your own
space.
I also suspect that one of the reasons why this happens so much in
comments is that comments are considerably more anonymous than journals.
Even if one's journal isn't openly connected with one's real name and
life, like mine is, there is some persistance and history to most
journals. I'm a strong believer in anonymous speech for a variety of
different reasons, but one of those reasons is not to permit drive-by
content-free attacks on other people. And that's the standard behavior in
both political blogs and comment pages on newspaper sites. The anonymous
speech I have the most respect for is the type that creates and builds
reputation for an anonymous nom de plume, presenting a consistent
public image and taking responsibility for one's previous beliefs. This
isn't what happens in comment threads.
Finally, I consider all of my web site, including my journal, to be to
some extent permanent. I mostly do not take down pages, and I wouldn't
want to take down comment threads. But for the most part comments,
particularly on controversial topics, don't fit into that model. There's
nothing wrong with that, but it makes me want to redirect them to a
more appropriate forum.
If I felt like there was some lack in the world of places where people
could anonymously publish controversial opinions, I might feel more
obligation to provide one in my journal, but I don't. It's also not
something that I'm good at; other people are doing a much better job. So
I'm going to stay out of hosting that type of on-line discussion and leave
it to people with more interest and time.
I welcome e-mail replies to journal posts, and have had several good
discussions in those e-mail exchanges. That's what works best for me.
All posts