I upgraded from Ubuntu Natty Narwhal to Oneiric Ocelot (11.10) today. Lots of things broke. In order:
- “Could not calculate the upgrade.” No indication of what the error might be; in my case, the error ended up being old orphan OpenAFS kernel modules (for whom no kernel modules existed). I also took the opportunity to clean up my PPAs.
- “Reading changelogs.”
apt-listchanges isn’t particularly useful, and I don’t know why I installed it. But it’s really painful when it’s taking more time to read changelogs than to install your software. Geoffrey suggested gdb -p `pgrep apt-listchanges and then forcing it to call exit(0)`, which worked like a charm. Had to do this several times; thought it was infinitely looping. - Icons didn’t work, menus ugly. Go to “System Settings > Appearance” and go set a new theme; in all likelihood your old theme went away. This AskUbuntu question gave a clue.
- Network Manager stopped working. For some inscrutable reason the default NetworkManager config file
/etc/NetworkManager/NetworkManager.conf has managed=false for ifupdown. Flip back to true. - New window manager, new defaults to dunk you in Unity at least once. Just make sure you pick the right window manager from the little gear icon.
gnome-power-manager went away. If you fix icons a not-so-useful icon will show up anyway when you load gnome-settings-daemon.- “Waiting for network configuration.” There were lots of suggestions here. My
/var/run and /var/lock were borked so I did these instructions, I also hear that you should punt wlan0 from /etc/network/interfaces and remove it from /etc/udev/rules.d70-persistent-net.rules. I also commented out the sleeps in /init/failsafe.conf for good measure. - Default GHC is 7.0.3! Blow away your
.cabal (but hold onto .cabal/config) and go reinstall Haskell Platform. Don’t forget to make sure you install profiling libraries, and grab xmonad and xmonad-contrib. Note that previous haskell-platform installs will be rather broken, on account of missing GHC 6 binaries (you can reinstall them, but it looks like they get replaced.) - ACPI stopped knowing about X, so if you have scripts for handling rotation, source
/usr/share/acpi-support/power-funcs and run getXuser and getXconsole - DBUS didn’t start. This is due to leftover pid and socket files, see this bug
- Was mysteriously fscking my root drive on every boot. Check your
pass param in /etc/fstab; should be 0. - Redshift mysteriously was being reset by xrandr calls; worked around by calling it oneshot immediately after running xrandr.
- Not sure if this was related to the upgrade, but fixed an annoyance where suspend-checking (in case you are coming out of hibernate) was taking a really long time in boot. Set
resume to right swap in /etc/initramfs-tools/conf.d/resume and update-initramfs -u with great prejudice).
Unresolved annoyances: X11 autolaunching in DBUS, the power icon doesn’t always properly show AC information and is too small in stalonetray, xmobar doesn’t support percentage battery and AC coloring simultaneously (I have a patch), a totem built from scratch segfaults.
Read more...
I use Mendeley because it lets me easily search for papers I care about. Unfortunately, that seems to be all Mendeley has been doing for me… and that’s a damn shame. Maybe it’s because I’m an undergraduate, still dipping my toe into an ocean of academic research. Mendeley was aimed at practicing researchers, but not people like me, who are stilll aiming for breadth not depth. I can count on two hands the number of technical papers I’ve really dug into—I’m trying to figure out what it is exactly that I want to specialize in.
Read more...
This post describes how to change which monitor xmobar shows up on in a multi-monitor setup. This had always been an annoyance for me, since on an initial switch to multi-monitor, xmobar would be living on the correct monitor, but if I ever restarted XMonad thereafter, it would migrate to my other monitor, much to my annoyance. Note that a monitor is different from an X screen, which can be directly configured from xmobar using the -x command line.
Read more...
What is the Mailbox? It’s a selection of interesting email conversations from my mailbox, which I can use in place of writing original content when I’m feeling lazy. I got the idea from Matt Might, who has a set of wonderful suggestions for low-cost academic blogging.
From: Brent Yorgey
I see you are a contributor to the sup mail client. At least I assume it is you, I doubt there are too many Edward Z. Yangs in the world. =) I’m thinking of switching from mutt. Do you still use sup? Any thoughts/encouragements/cautions to share?
Read more...
OpenIntents has a nifty application called SensorSimulator which allows you feed an Android application accelerometer, orientation and temperature sensor data. Unfortunately, it doesn’t work well on the newer Android 2.x series of devices. In particular:
- The mocked API presented to the user is different from the true API. This is due in part to the copies of the Sensor, SensorEvent and SensorEventHandler that the original code had in order to work around the fact that Android doesn’t have public constructors for these classes,
- Though the documentation claims “Whenever you are not connected to the simulator, you will get real device sensor data”, this is not actually the case: all of the code that interfaces with the real sensor system is commented out. So not only is are the APIs incompatible, you have to edit your code from one way to another when you want to vary testing. (The code also does a terrible job of handling the edge condition where you are not actually testing the application.)
Being rather displeased with this state of affairs, I decided to fix things up. With the power of Java reflection (cough cough) I switched the representation over to the true Android objects (effectively eliminating all overhead when the simulator is not connected.) Fortunately, Sensor and SensorEvent are small, data-oriented classes, so I don’t think I stepped on the internal representation too much, though the code will probably break horribly with future versions of the Android SDK. Perhaps I should suggest to upstream that they should make their constructors public.
Read more...
To: John Wellesz
First off, I’d like to thank you for authoring the php.vim indentation plugin. Recent experiences with some other indentation plugins made me realize how annoying editing can be without a good indentation plugin, and php.vim mostly has served me well over the years.
However, I do have a suggestion for the default behavior of PHP_autoformatcomment. When this option is enabled (as it is by default), it sets the ‘w’ format option, which performs paragraphing based off of trailing newlines. Unfortunately, this option has a number of adverse effects that may not be obvious unless you are paying attention to trailing newlines:
Read more...
This post is not meant as a rag on Darcs, just a observation of the difference between two philosophies of version control. Also, I’m a bit new to Darcs, so there might be some factual inaccuracies. Please let me know about them!
At some point, I would like to write a Darcs for Git users guide, distilling my experiences as an advanced Git user wrangling with Darcs. But perhaps the most important take away point is this: don’t try to superimpose Git’s underlying storage model on Darcs! Once I realized this point, I found Darcs fit rather nicely with my preferred Git development style—constant rebasing of local patches until they hit the official repository.
Read more...
Some things come round full circle.
As a high schooler, I was a real Windows enthusiast. A budding programmer, I accumulated a complete development environment out of necessity, a mix of Cygwin, handwritten batch scripts, PuTTY, LogMeIn, a homegrown set of PHP build scripts and Notepad++. I was so devoted to the cause I even got a single patch into Git, for the purpose of making Git play nicely with plink on Windows. The setup worked, but it always felt like a patchwork of different components, all not quite seeing eye-to-eye with each other. When I discovered that Linux was able to offer me an unbelievably coherent development environment, I jumped ship and said goodbye to Windows.
Read more...
A little trick for your toolbox: after you’ve generated your slide deck and printed it out to PDF, you might want to annotate the slides with comments. These is a good idea for several reasons:
- If you’ve constructed your slides to be text light, they might be optimized for presentation but not for reading later on. (“Huh, here is this diagram, I sure wish I knew what the presenter was saying at that point.”)
- Writing out a dialog to go along the slides is a nonvocal way of practicing your presentation!
But how do you interleave the slide pages with your annotations? With the power of enscript and pdftk, you can do this entirely automatically, without even having to leave your terminal! Here’s the recipe.
Read more...
Now that term is over, I finally went an upgraded my laptop to Ubuntu 10.04 LTS, Lucid Lynx. The process went substantially more smoothly than Karmic went, but there were still a few hiccups.
Etckeeper. As always, you should set AVOID_COMMIT_BEFORE_INSTALL to 0 before attempting a release upgrade, since etckeeper hooks will be invoked multiple times and there’s nothing more annoying than getting the notice “etckeeper aborted the install due to uncommited changes, please commit them yourselves” because there is no way that’s going to work.
Read more...