Yak Shaving

This is more a reminder for me The code below allows you to talk to ZODB in an interactive way.

$ bin/instance debug

from Acquisition import aq_parent
from zkaffold.consolehelpers import login
from Products.CMFCore.utils import getToolByName
app = login(app, 'zopeadmin')
pc = getToolByName(app.portal, 'catalog')
content_fragments = pc({'getId': 'id', 'portal_type': 'ContentFragments'})
content_fragments
[, ]

Login allows you to query the catalog otherwise no results are returned.

The code below will refresh a web page when the file is updated. I was using it to make syslog available over http.

http = require('http');
var spawn = require('child_process').spawn;
var sys = require('sys')

// log to monitor
var filename = process.ARGV[2];

if (!filename)
  return sys.puts("Usage: node monitor.js filename");

// fire up tail on the log file
var tail = spawn("tail", ["-f", filename]);

http.createServer(function (req, update) {
  update.writeHead(200, {'Content-Type': "text/plain;"});
  tail.stdout.on("data", function (data) {
    update.write(data);
  });
}).listen(8000);

sudo apt-get install build-essential libssl-dev

Should be enough to make ./configure happy.

For my latest CV click here.

I am currently refactoring a project I’m actively developing. The initial development was not test driven but more “Oooh this is cool, hack, hack hack”. I’m quite a fan of the idea of premature optimisation being a bad thing, although sometimes this can be an excuse for laziness.

So I find myself refactoring and creating tests to prove the new code works. This led me to need fixtures or exmaple data to test against. As I already had data in the database creating said fixtures was a easy task:

manage.py dumpdata app_name.model_name

Where app_name in the name of the application you create in Django and the model name is the defined database object usually in models.py. This command dumps the structure of the object out to the command line in json format. To use it create a fixtures directory in your application directory and add the json file. Or the easy way:

manage.py dumpdata app_name.model_name > fixtures/fixtures.json

The file name doesn’t matter.

I have steady commute to work and after getting fed up with the news and purchasing a Nexus One I listen to a lot of podcasts. These are mostly tech and science ones. so in the spirit of sharing below is an OPML export of my current selection: google-reader-subscriptions.xml.

As the dust settles on the surprise purchase of Novell by Attachmate (I have no idea either) it seems about the right time to start the crazy conspiracy theories.

I had wondered for a while after Novell “won” the SCO case and were found to have various UNIX copyrights see http://en.wikipedia.org/wiki/SCO-Linux_controversies for a much better explanation. Add in Novell’s interesting patent catalogue (they also signed that patent deal with Microsoft) and the company starts to more like a patent trolls dream.

Rumours are also starting to surface about how the deal was financed http://blogs.computerworld.com/17416/who_really_bought_novell_microsoft. I am certainly no expert but the whole deal seemed fishy from the start.

I think the the future has a lot of law suits from Attachmate and a slow lingering death for Suse Enterprise and OpenSUSE. This deal has all the hallmarks of an asset stripping fire sale.

Reading about the internet I came across http://www.networkworld.com/community/blog/speed-linux-no-kernel-patch-required. This looks pretty cool.

I thought I’d give a go as my EEE PC 1005 has been suffering a little since upgrading to Ubuntu Maverick. Following http://www.webupd8.org/2010/11/alternative-to-200-lines-kernel-patch.html the Ubuntu guide I have given it a whirl.

Can’t say I have noticed a difference as yet but these things can be subjective. I will add an update when I notice something.

Update 1:
So I got this tweet: http://twitter.com/#!/1stvamp/status/6385998627864576

I haven’t tested this package but it looks like an easier way to implement the changes.

As for performance improvements I’m not entirely convinced. It has dropped the load when running Chrome. Usually Chrome ran the laptop at a load of 1 ish that has now dropped to around .5. Which is nice for battery life.

#!/bin/python

This had my annoyed for a while. When I came to run ./configure I got:

checking for JS_NewContext in -lmozjs... no
checking for JS_NewContext in -ljs... no
checking for JS_NewContext in -ljs3250... no
checking for JS_NewContext in -ljs32... no
configure: error: Could not find the js library.

Is the Mozilla SpiderMonkey library installed?

I was unable to find much help until I looked at the instructions for installing from source http://wiki.apache.org/couchdb/Installing_on_Ubuntu?action=show&redirect=InstallingOnUbuntu. Which in retrospect would have been a good place to start. Use the following to stop the above error:

./configure --with-js-include=/usr/include/xulrunner-1.9.1.9/unstable --with-js-lib=/usr/lib/xulrunner-devel-1.9.1.9/lib

Oh and make sure you have xulrunner and its dev package installed.