Well, I say I, but I can't take all the credit...

Due on 19 March 2010.

OK. I get the message. What can I say? My excuse is that I've been too busy elsewhere. Over the past few months I've got my head round twitter, and tend to update things there. It's convenient to be able to write a sentence here or there, rather than try to construct a narrative after the fact. Also, whilst I've taken plenty of photos, I've not managed to keep up with uploading them. So, given that, what's the recent (and not so recent) narrative?

• We have a new car, a Volkswagen Golf Plus, having traded in our 11 year old Mazda. It was shortly before the scrappage scheme was introduced, but we got a cracking deal!
• We've remortgaged, with an offset tracker from First Direct. They've lived up to their reputation for excellent customer service. Can't rate them highly enough so far.
• I've bought an iPhone 3G S which is just lovely. I'm turning into a real Apple fanboy.
• Is this is a convenient point to mention my 'hackintosh'?
  
  A netbook running Mac OS X. Cute; see the video.
• I've had another set of phototherapy treatment at Warwick Hospital to keep the sunshine at bay. So far so good.
• Jess is happily three, fully toilet-trained, and looking forward to pre-school in September.
• We've had our summer holiday, once again in Germany. We drove the new car across to the Mosel, where the weather was nothing special, but it was nonetheless a good change.

So, some pictures. Jess having some summer splashing fun:

And a few holiday pics. Enjoy!

I rarely seem to write about the day job in my blog, but I want to make sure that this gets recorded somewhere that google might find, so that it might save someone the 4½ hours it took me to resolve.

We've got a development environment based around virtual ColdFusion 8 development servers which run a codebase from across a couple of subversion working copies (WC). The webroot maps to one WC, there's a common folder (with common code/resources for several applications) which is aliased below the webroot in apache, and another WC for custom tags. When we commit working, tested code to trunk, post-commit hooks on the repos update the relevant working copies on the public web servers.

So, the problem yesterday was that code which worked fine on the test environment was failing on the live server. The code was using ExpandPath() to (try to) return the absolute filesystem path for a web-relative path supplied. It worked on the test server (let's say it returned /etc/webroot/ for /), but on the live site, it returned something like http://www.example.com/siteCheck/http:/www.example.com/ instead. Obviously this isn't a valid filesystem path

Googling pulled up a link to a page by Mark Drew which didn't reference my particular bug, but there was a comment by Mark Mandel which suggested that in the absence of a CF mapping for the webroot, it looks to the apache mappings (and possibly mod_rewrite rules) for its path resolution. As our live server sits behind a load balancer, this fits with the very broken ExpandPath() result we saw.

In the end, there's a very simple solution. Set a mapping in the CF Administrator for / to the actual filesystem path for the webroot. It shouldn't have any other side effects, but it completely fixes ExpandPath(). Methinks that CF8 is trying to be too clever...

Anyway, that's that for today. At some point I'll get back to normal service and post some cute photos of Jess.

Just a note to mention that I received a reply from my MP the other day to my recent letter about data sharing. He gave me a thorough reply, pointing out several benefits of responsible data-sharing (reducing red tape and cutting bureaucracy by permitting departments to share information). The thing is that I don't have any problem with responsible data-sharing. It's the risk of irresponsible stuff that worries me...

At risk of turning into a boiler specialist, I wanted to share some useful info provided to me by John B. He has a Potterton 16/22 Netaheat Electronic LPG boiler and had similar symptoms to my dodgy old boiler. His fan and pilot ignition were OK. However, the ignitor was clicking every second up to two minutes plus before the main burner cut in when starting from cold. When hot, it only took 2-3 seconds for the burner to light. He'd replaced both the electrode and the thermostat to no avail.

Unlike my boiler though, his had a different PCB, a Potterton 407676 (the reverse side of the PCB also shows the number 90911/4). What seemed conclusive to me of an electrolytic capacitor problem was that a gentle tap (when cold) on the PCB triggered the main burner. John kindly took some photos and provided technical details of his PCB for me to post.

The original electrolytic capacitors were marked as follows:

C9: 220uF, 16V/85°C
C5: 100uF, 40V/85°C
C4: 22uF, 50V/105°C

John replaced them with capacitors rated to cope with slightly higher voltages (Maplin part nos. VH42U, VH39N, DT57M respectively). His boiler now works perfectly again!

If this is helpful to you please leave a comment, and link if you can, to make it more likely to be picked up by search engines for other people. If it's saved you the cost of a replacement boiler and you'd like to express your thanks with a donation (secured by PayPal) to support my hosting costs, it'd be very welcome. I'll donate 50% of your contributions to National Energy Action, the UK's leading fuel poverty charity, campaigning for warmer homes - it seems like an appropriate cause.

Disclaimer: Don't electrocute or incinerate yourself. This page is for information only. Landlords and employers are legally obliged to use a registered CH engineer for all gas work. I accept no responsibility for any accidents or problems.