Isotoma is looking for someone with solid testing and management experience to come and help us shape and build our QA function. The role will be to take on the formalisation of our quality control systems, from traditional manual testing to automated CI, build and deployment. We’re looking for someone who can run the testing function whilst also getting involved with all 3 teams within the company (dev, devops and ops), helping them improve their processes and approaches upstream.
Read more about it on the main site (Head of Quality Assurance vacancy) and please get in touch.
Author Archives: Andy Theyers
Squid and Facebook
Lenny (Debian 5.0) and Lucid (Ubuntu 10.04 LTS) both ship with Squid 2.7 as their default Squid version. If you plan on deploying squid as a web proxy in a modern mixed browser environment you should seriously consider Squid 3.0. If you can’t (or, like me, you’d already spent some time getting Squid 2.7 set up just right) you are likely to find users reporting that Facebook is returning blank pages or that links within Facebook don’t work. Looking at Facebook in a browser with debugging tools installed you’ll see lots of cross domain Javascript errors.
At this point you should either upgrade to Squid 3 (which doesn’t exhibit this behaviour) or make the following change to your squid.conf:
- Add
server_http11 on
Facebook uses different methods of shipping the Javascript depending on whether the client has made an HTTP/1.0 or an HTTP/1.1 request. By default Squid 2.7 alters the request to be HTTP/1.0, so Facebook sends it’s fallback methods to the modern browser, which then proceeds to break. The change above stops Squid rewriting that part of the request, ensuring that Facebook ships the right content to your users.
On Ubuntu Python, Exceptions and unnecessary imports
A few days ago Alexander Limi (one of Plone’s founders) tweeted the following:
Now this gets my goat on a few points. First up, why the hell would I not want to use the system python? If I’m using any sane distribution I’ll have package management and security updates, and any flaw in Python will be patched, packaged and tested by people that are far smarter than me. Upgrading the Python that ships with the Plone Unified Installer just isn’t going to be as easy, however you play it. And that’s without the risk of the Plone community moving on to more exciting things, leaving their version of Python unsupported.
Secondly, there’s a fatal flaw in the original blog post to which limi refers. Yes, on the desktop, Ubuntu imports 190 packages when an exception is raised. As the author explains this is to enable Apport to provide as much information to the Ubuntu devs about application failures. What the author does not mention is that this doesn’t happen on the Server edition of Ubuntu. Why would it? Apport is designed to handle desktop application failures and to improve the end user experience. It isn’t installed by default on the server edition, because it isn’t needed.
On my Karmic desktop:
Python 2.6.4 (r264:75706, Dec 7 2009, 18:43:55) [GCC 4.4.1] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import sys >>> len(sys.modules) 35 >>> raise KeyError Traceback (most recent call last): File "", line 1, in KeyError >>> len(sys.modules) 225
On my Karmic server:
Python 2.6.4 (r264:75706, Dec 7 2009, 18:43:55) [GCC 4.4.1] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import sys >>> len(sys.modules) 32 >>> raise KeyError Traceback (most recent call last): File "", line 1, in KeyError >>> len(sys.modules) 32
Being quick to condemn Ubuntu, and their packaging of Python, doesn’t do anyone any good. Think before you tweet. And don’t go live on a desktop distro.
Location of Zope/ZEO transient caches
When Zope connects to ZEO it is configured with a database cache, where objects are cached by Zope so that it doesn’t have to constantly return to ZEO. These caches can be transient (they don’t persist past a Zope restart) or persistent (where they do).
If all you do is uncomment the example ZEO client configuration in your zope.conf you’ll have a 20MB transient cache. If you follow one of the many buildout recipes on the web you’ll likely have a transient cache of anywhere from 30MB (plone.recipe.zope2instance) up to 300MB (from this howto on plone.org).
Zope uses the Python tempfile module to decide where to put these transient caches, which on Debian defaults to /tmp. Our default production server installation has /tmp as a subdirectory of /, with the root partition being only 1GB. This means we have a maximum of about 400MB for temporary files on our servers.
With the default 20MB or 30MB this is fine, even in multiple Zope/ZEO installations on the same box. However, when we recently tried to up the size of the caches on a system that needed it we quickly found ourselves running out of diskspace in the root partition, which in turn caused some very strange behaviour, ultimately resulting in the site being unavailable.
The trick is to move these transient files to a more sensible location (like /var/tmp). You can do this (with thanks to mauro on zodb-dev) by setting an environment variable in your zope.conf, either manually:
<environment> TMPDIR /var/tmp </environment>
Or in your buildout:
[instance]
recipe = plone.recipe.zope2install
...
environment-vars =
TMPDIR /var/tmp
Compiling Python2.3 on Karmic
We still have some (now very old) Plone 2.1 sites running. Up until Ubuntu 8.10 python2.3 was still available from package management (albeit with limited support), but as of Ubuntu 9.04 python2.3 was removed entirely.
To recreate a development environment for these older sites means downloading Python 2.3 and compiling it by hand. Unfortunately it’s not a default build any more – as it says on the download page:
Since the release candidate, we received various reports that the this release may fail to build on current operating systems, in particular on OS X. We have made no attempt to fix these problems, as the release is targeted for systems that were current at the time Python 2.3 was originally released. For more recent systems, you might have to come up with work-arounds.
For Ubuntu 9.10 (Karmic Koala) to compile python2.3 you’ll need to disable buffer overflow checking by doing the following:
wget http://www.python.org/ftp/python/2.3.7/Python-2.3.7.tgz tar zxvf Python-2.3.7.tgz cd Python-2.3.7 ./configure --prefix=/usr/local BASECFLAGS=-U_FORTIFY_SOURCE make sudo make install
update: this post has been translated into Belorussian if you prefer that.
On what to do with downtime
There’s a fascinating post over at TCUK today about agencies building their own applications. We even get a mention in the intro with reference to Forkd. The comments get pretty heated too, with some people claiming agencies that have the time to do such things are over staffed while others are worrying about how to fit internal developments alongside client work.
So, given we’re mentioned, how do we feel about it?
It’s always going to be hard getting the balance of client work and your own projects right. We started work on Forkd in summer 2005 and launched it in early 2008. In that time the world and his wife had built a food social networking site and we were on the back foot right from launch. Keeping a focus on internal development ideas while working with clients work is always going to be a killer.
That said we have more in the pipeline and aren’t going to stop…
The reasons (at least for us) are threefold. First up, what better shop front is there than something that you made for yourselves? A real case of putting your money where your mouth is; new customers can see exactly what you could be making for them.
Secondly, scaling a technical agency is hard. You only have the person hour to sell, and growth only comes through selling more person hours. That is, unless you can “productise” some of your own work. And what better than productising something that you and your team enjoy making and using?
Third – it’s always best to learn new technology on something that isn’t so important that your business depends on it. That latest cut of your favourite app server from svn might look like it provides the killer feature (or bug fix) you need, but do you really want to try it on your current big client project?
And to those that claim agencies must be over staffed if they can do work like this… We all know that the entire team will have downtime whether it’s explicit or not (and all of you with “100%” utilisation just have a miserable staff hiding their downtime at the water cooler or on facebook). A shared goal to aim at during that downtime makes the downtime explicit and supported, which makes a whole let of sense to me, anyway.
And hey, who wouldn’t like to be the next Skinny Corp?
kupu 1.3.9 and Firefox 3
Just a quick note to say that we’ve just discovered that kupu 1.3.9 (the one that ships with Plone 2.5 by default) breaks rather nastily in Firefox 3. First looks show that kupu 1.4.10 (the latest from http://plone.org/products/kupu) is just a drop in replacement, so it shouldn’t be too bad to get sorted. If you too are suffering weird behaviour when using the kupu internal link tool or image browser it may well just be that.
We’re hiring
We’re looking for developers ASAP, due to a quick burst of new work. If you (or anyone you know) is looking for their second or third programming job and has good Python skills (preferably with some experience of one or more of Turbogears, Django, Twisted, Zope or Plone) and aren’t scared of HTML, CSS and Javascript please get your CV to andy@isotoma.com ASAP. We’ll put a formal job spec up on the main site shortly but the sooner we hear from you the better.
Facebook advertising
We’ve been running a Facebook “social ad” for Forkd for the last few days. I’ve been interested in what Facebook advertising could do for our customers for a while, and thought that the best way to find out was to try it for ourselves. Some interesting things have arisen from it, too:
- It’s a very simple ad format – hard to really make compelling. Despite that, and having made little effort in the ad design, we’re getting 1 in 2,000 click throughs, which feels to me like a good number.
- We’re getting roughly 1 new account sign up for every 8 clicks. Again, this feels like a good number. Both of these make me think that a good advert well targetted on Facebook is definitely worth the money.
- Side effects were the discovery that our click throughs dropped by 50% over the weekend, but that the sign up rate improved to 1 in 4. From this, I guess, we can infer that a) Facebook is primarily a work diversionary tactic, and b) people in that diversionary mode are clicking with less purpose
All in all then? I’ll be recommending that customers do it, but I’d ask them to think about only running campaigns at the weekends, particularly if the landing page from a click requires a little reading for the reader to understand the benefits.
KSS and Firebug
I’m just playing with KSS for the first time properly today – a quick note for myself and anyone else out there who may get bitten by this one. If you are in the habit of using console.log() in your jQuery (or MochiKit, or whatever). Somewhere in the internals of KSS console becomes a KSS object, and calls to console.log() cause Firefox to hang. Just a warning 
If I track down what’s going on I’ll follow up on this post…