Quec.es, the XMPP Blog
How to appear in Quec.es
Quec.es automatically logs your Jabber status to record your daily thoughts, life and ideas; all you need to do is add and authorize the logging-bot, *@quec.es and use Jabber like normal. If you have something longer you want to say you can also send a direct message to *@quec.es instead of changing your status.
Most Recent Posts
Tuesday, April 7 2015
[12:53:06] matt [wronka.org]/Trip I would like to second the statement in this thread: https://www.physicsforums.com/threads/do-not-buy-a-unicomp-keyboard.613424/
I bought a Unicomp keyboard because my wife wanted a standard shape (e.g. not like my preferred Kinesis Advantage) keyboard and I wanted a mechanicle keyboard that would plug into a USB-only (non-PS/2) NUC. Immediately there were issues with keybaord not being recognized at all. That's ok, every company has some defective units, I sent it back and got the keyboard back fairly quickly (it now has a "repaired by Unicomp" sticker labeled June 2013 next to it's manufactured-on April 2013 sticker).
Two years after it was manufactured it started behaving oddly. The return consists of two pressure triggers, if the left side is the one you happen to hit, it triggers both the slash/pipe ('/') key and then the return key. My initial thought was that there was something physically connecting the two adjacent keys, but after removing the keys and examining the board superficially, everything seems fine. Trying with a different computer resulted in the same issue. A few days later, a few keys stopped responding entirely.
A keyboard failing within two years is sad; I'm currently typing on the same Kinesis keyboard I've used five-days a week for over seven years and functions flawlessly. I've got PS/2 and AT keyboards that still work from the 1980s.. Unicomp charges between $30-$90 for repair of a keyboard outside warranty (plus shipping) which seems like more hassle than it's worth considering this thing's already been back to them, and even then has the annoying habbit of not being recognized if plugged-in during computer boot.
Wednesday, April 1 2015
Thursday, March 5 2015
Tuesday, March 3 2015
Thursday, February 26 2015
[15:55:31] matt [wronka.org]/Trip Two nights ago, my wife went to bed, and I tried to watch some Columbo on Netflix (I was up to the Great Santini, who sets up his own alibi through a fragile, technical, contrivance). We've only recently subscribed to Netflix since I've always been leary of the reliability of cloud services and rental subscriptions like this in general. It turns out the flakey Samsung implementation was more to blame.
This isn't abnormal. First off, the Samsung equipment (we own two of their TVs and one DVD player, all with essentially the same software) seem to arbitrarily forget WiFi passwords, which makes supporting them frustrating if not useless (they're all on the unsecured network now). Sometimes it fails to connect for a short period, and I need to just wait; that wasn't happening.
Obviously, there was a larger problem. I gave up and watched TopGear on my MythTV box instead, expecting whatever issue Samsung was having to resolve itself the next day. Why a box can't trust that it's on the Internet, or at least be optimistic about it once it's gotten an IP address, and a DNS server that resolves what it needs is an open question that I've tried to ask Samsung support (like the TV's software, I'm not optimistic for a response).
Yesterday while I was at work, I got a message from my wife, complaining about the DVD player not thinking it had Internet access. Obviously, she wanted to think it was a problem with our network—which is reasonable, given that's what the software said—but it turns out Samsung still didn't have their system up. It seems that there was some DNS hokeyness with their Akamai DSA settings. After a chain of CNAMEs (some of which included "china-" prefixes for some reason) eventually we got very short TTL addresses, which were not returning appropriate answers for the TV.
A Web search found somebody who *had* found an IP address that worked, also being served through Akamai DSA:
The resulting IP for www.samsung.com was 184.108.40.206; while you're setting-up your own DNS for your Samsung devices, I also suggest making ad.samsungadhub.com and rd.samsungadhub.com either fail or point to localhost since these are what send and track impressions for the annoying little piece of real estate in the top right corner.
I strongly discourage anyone from buying one of these devices (and apparently Sony devices) for these features, since they seem to be fragile. As I was trying to find information on the current outage (Samsung was not forthcoming and even mentioned on their support page of no known issues), I found references and news articles for outages regularly going back to 2013. It's clear Samsung doesn't treat this as production functionality.
More coverage today, after a couple days of this:
Friday, February 13 2015
[01:57:22] matt [wronka.org]/Psi+ Recently—last week or so that is—I noticed my "speed dial"/recently viewed sites list on my desktop copies of Mozilla FIreFox were cleared. This happened to coïncide with rebuilding nightly. At first this was an annoyance, when everything was replaced with links to Mozilla and open source pages. After using the browser for a bit, I got one bookmark back on the page (oddly, something I *hadn't* visited that day); and now after about a week, I've got the first three and the final (15th) spot as pages I've visited.
In addition to those, a tab for the Mozilla Marketplace and nine other Mozilla links: I now have a tab for a tax package. I'm not happy with you Mozilla. Basically, I'm saying the same thing to you that you are to your users.
Friday, February 6 2015
Monday, February 2 2015
[16:47:22] matt [wronka.org]/Psi+ http://ma.ttias.be/whats-new-systemd-2015-edition/
Unix: Do one thing well.
SystemD: Why do one thing, when you could be doing other things as well?
I'm not a huge fan of SystemD, in fact, I was considering switching back to FreeBSD for my home workstations to avoid it. However, there were some points in the notes on the 2015 roadmap which might actually be useful for the specific usecase I have for GNU/Linux.
Booting a standard GNU/Linux distro with a read-only root (e.g. from NFS) is frustrating; it doesn't work well, and even though many of the caveats are documented around the Web, it seems like there's always something that doesn't quite work. FreeBSD, for what its worth, booted diskless quite nicely when I was comparing the two about two years ago. In the end, I went with USB boot images for each node at home.
Looking at the roadmap, the 2015 plan for SystemD seems to be moving towards a system which is better designed for read-only root by default, which would be neat, and hopefully mean once the system is configured, bitrot would be less of an issue.
Wednesday, January 28 2015
Tuesday, January 27 2015
[02:19:21] matt [wronka.org]/Psi+ I recently switched my home boot image from an ever-out-of-date Ubuntu installation to Debian Jessie, which was at one point "almost stable" or "almost frozen" or something like that. Then SystemD broke loose and it's still clearly testing.
Things that don't work:
NFS doesn't mount on boot. I give-up. I can't get it to mount anything from the init scripts. The Internet suggests this is because of something left in /var/run/network, but since /var/run is tmpfs this is clearly out (also, I checked, the directory isn't there).
Running sudo clears afs tokens. I've seen one other reported issue, but no solution. cf. http://comments.gmane.org/gmane.linux.debian.user/489795
I can no longer get a gnome-session or gnome-settings-daemon running on top of spectrwm. I also can't figure out how to change the window manager for gnome, so it seems like I'm stuck with all of gnome, or none of it now. Why do I care? colord/colormgr is really the only reason why. The rest of the gnome environment is an exercise in frustration.
The most surprising thing that works? Qt now doesn't look like vomit when running in a 30-bit X display.
[02:25:16] matt [wronka.org]/Psi+ "Juno 2015 Winter Storm CodeRed Message"
What does that mean?
Tuesday, December 9 2014
The difference between my encounters with institute presidents at RPI and Brown were night and day; and both seem to fit with others' experiences.
Sunday, December 7 2014
[18:37:08] matt [wronka.org]/Psi.cor I've switched back to Android 4.4 (CM-11, "Kitkat"). The following were issues I found with Android 5.0 ("Lollipop"):
* fdroid repos don't work, which is annoying (https://gitlab.com/fdroid/fdroidclient/issues/111)
* some Activesync servers don't work—I've only seen this reported with Horde (and I had the issue with both 5.0 and 5.2; https://bugs.horde.org/ticket/13702)
Back on CM-11, using the ART runtime, Firefox Sync (Weave/"Deprecated") also fails, apparently with a Unicode string error. It works fine with Dalvik. I don't know if this would have been an issue on Lollipop as well which uses ART by default. I didn't look for a specific bug for this, but was surprised that this sync was still supported since it was supposed to go away several versions ago (https://bugs.horde.org/ticket/13702). Apparently work on making the new service easy to use by third parties is either hard or just not a priority—the whole issue seems to have been bungled and now everyone's stuck with a mess. (https://bugzilla.mozilla.org/show_bug.cgi?id=989756#c14)
I also missed some of the UI elements from CM-11, like the circular battery indicator (is this a theme added by CM?) The settings menu was also more usable on CM-11. In general, Lollipop seemed to waste space, although I had mixed feelings about the task switching interface (it did seem to show more options at once, but made the active surface a bit small on a phone screen). A lot of Lollipop was flat, and Apple-like, looking pretty without giving the user any indication of whether interfaces were scrollable or otherwise how to interact with the device.
Unlike other reports, I did not run into any issues with WiFi or battery life on Lollipop—in fact, both seemed to be at least as good if not more reliable than on CM-11 and CM-10.2 but I don't have any objective tests for that. Specifically, I thnk my worst battery behaviour was in part to K9 synching my mail, and I haven't set that back up.
Saturday, December 6 2014
[17:51:38] matt [wronka.org]/Psi.cor Work switched from Zimbra to MicroSoft Exchange some time back, and I've stopped synchronizing my calendar with my Nexus 4 since then (instead using my Handspring Visor Edge). The reason was straight forward: the MicroSoft server wanted the ability to wipe my entire device.
This seemed like overreach, and after talking with people in IT, it wasn't intentional, it was just the default. The changes in the AOSP code are pretty straightforward to disable this. I've posted diffs for both KitKat and Lollipop: http://matt.wronka.org/stuff/projects/icpp/android/cyanogenmod/
The KitKat changes also include some clean-up of CM code I didn't find useful (CMUpdater, CMAccounts), these aren't in CM12 yet. If you'd rather cherry-pick the changes for CM11 or AOSP 4.4 there are two AOSP applications to patch: http://matt.wronka.org/stuff/projects/icpp/android/aosp/4.4/
It looks like a lot of refactoring went into the Exchange services in 5.0, the patchset is smaller, but there's a new issue as reported to horde: https://bugs.horde.org/ticket/13702
I can confirm that this is an issue with Android 5.0—the effect is that the device appears to sync, but when it is about to complete it removes all data it received. I have not looked into fixing this yet but appears unrelated to Horde itself.
For now, full builds are at:
http://hume.matt.wronka.org/~matt/tmp/cm-12-20141204-SNAPSHOT-CNJ-mako.zip (Nexus 4, Android 5.0/CM12/Lollipop)
http://hume.matt.wronka.org/~matt/tmp/cm-11-20141122-SNAPSHOT-CNJ-mako.zip (Nexus 4, Android 4.4/CM11/KitKat)
http://hume.matt.wronka.org/~matt/tmp/cm-11-20141114-SNAPSHOT-CNJ-crespo.zip (Nexus S, Android 4.4/CM11/KitKat)