Older blog entries for berend (starting at number 381)

30 Jan 2012 (updated 30 Jan 2012 at 20:35 UTC) »

Frequently get this:


$ evince
Maximum number of clients reachedCannot parse arguments: Cannot open display:

Why does Google not have an answer?

Flew to Melbourne today, for a few holidays + the Drupal Down Under. Had a problem with my K3805-Z stick, wasn't recognised anymore on this machine. Perhaps the Oneiric upgrade? According to syslog usb-modeswitch crashed when inserting the stick. The solution was to get the 1.21 usb_modeswitch, compile and install it, and that fixed the problem.

Good for performance as the always reliable K3565 Vodem worked, but is feeling a bit slow these days.

The only problem I still have is that Ubuntu itself doesn't pickup on the USB sticks anymore. It used to that, perhaps my playing around a year ago to get the K3805-Z working killed things. So need the latest vodafone-mobile-connect-card-driver-for-linux to connect me.

Today was a bit of poking around day. Tried to setup PXE booting from my FreeBSD 8. Didn't get it working using the guides (1, and 2 were particularly helpful) until I understood that the host setting (with MAC address) is really, really required. You can't have a setup where a PXE client boots something default.

After some playing was also able to get linux booting to work. Now installing Ubuntu 10.04 LTS on a Dell PowerEdge R210 using my new PXE boot capabilities.

Lot of changes to my network recently. The old PIII-800MHz became a pfSense router for a friend, and the server is the Intel Core2 4400 at 2GHz, so quite a step up. But getting it all configured was major pain.

After putting the new motherboard into the case, the system sort-of booted, but not really. Couldn't find the kernel. As this was the server, my network was down, so not easy to download a bootable cd, which I should have prepared actually. Somehow I had a FreeBSD 8.0 disk, so was able to boot with that in the end and poke around. It appears the kernel was ad4, instead of ad0 ??? Still don't understand it. The disk was ATA, but I also put in a SATA disk which became ad6. And the MBR was corrupt or perhaps the next boot loader was corrupt. Had to reinstall to boot and set it up to load the kernel from ad4.

Putting the old motherboard in another case went wrong too as the PSU (Zalman ZM460B-APS) died in the process.

Got another Zalman ZM460b-APS power supply die on me. The quality of these units is just very poor. Yes, they're very quiet, but they die way too soon.

Just released new beta for xplain2sql, version 4.1.0.


No new major functionality as this is a very stable tool, but important performance improvements for mysql and postgresql. And bug fixes.

My quest to get svnserve to work has succeeded. Went with saslauthd, so I can use pam authentication, and use stunnel to make sure plain text passwords can't be snooped. All getting pretty complex, but a lot faster than accessing svn over Apache.

12 Nov 2011 (updated 12 Nov 2011 at 04:47 UTC) »

Found out the reason why I must store plain text passwords. It's because of the sasl auxprop framework. That only works with plain text passwords, not encrypted ones. Subversion and sasl use the md5-digest method to communicate with the client, so that's covered. But having plain text passwords in the database, really, we can't get away with that these days.

Supposedly with saslauthd and the pam framework it might be possible to use encrypted passwords. So testing that now.

10 Nov 2011 (updated 10 Nov 2011 at 04:14 UTC) »

Hmmm, disappointed to conclude that even if you configure svnserve with DIGEST-MD5, it doesn't appear to work. Plain text password gets send over the wire, and digest calculation appears to happen at the svnserve side, not svn itself.

UPDATE: looking at the code, you would say that's not true. Need to insert some debug statements...

Just got this with svn:

Digest mutual authentication failure: client nonce mismatch


more people have seen this, no one knows what it means, in the sense of why it occurs.

372 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!