'Twas brillig, and the slithy toves did gyre and gimble in the wabe

...and today's discovery is that while I was away over the weekend, the breaker for the immersion heater tripped. That's new.

Of course, I discovered this by wanting to do some washing up and finding that a) the pump wasn't working (it's on the same circuit), and b) what little water was coming out of the tap was cold.

The heater has been a bit temperamental and more than once I've had to reset the over-temperature cut-out on it so I wonder if something has finally failed. In any case, it can wait until tomorrow when it's still light outside and so it's less of an inconvenience if my usual troubleshooting approach (reset breaker, see if it trips again - if it does then it's definitely failed) takes out the lights as well. Though it may well just be one of those things rather than an actual failure - MCBs do have a tendency to trip on very short faults, like the momentary short-circuit from a failing halogen light.

Pop quiz: at what time, in the UK, do you think the biggest regular surge in power occurs?

Surely it's everyone starting to cook supper at 7pm or so?

Aircon in offices starting as people get into work at 9ish?

Sunset, as everyone turns their lights on?

Think about it for a moment, and try to come up with your own answer before following the cut.

( The answer )

So, I'm happily listening to my new Gladiator soundtrack, and reach across to turn off the printer/scanner/copier/coffee-maker (an HP psc 1215). At which point the gentle strings turns into a harsh buzzing, and Windows bluescreens.

Well, I suppose I should restart my computer soon anyway to finish installing the latest patches, but, but, but, grrryyaarrrrrgghh!!! I will reboot my computer as and when *I* want to, not when some two-bit driver written by some utter moron decides to scribble over the system tables and make Windows explode messily. Yes, I'm looking at you, HP. And you, Microsoft. All I did was turn the bloody printer off!

Now I have to go and pick up the pieces of the two programs I was working on, hope that I have saved any changes I made to them, tidy up the playlist for XMPlay and re-add the music I just added, and generally work out what I was doing and get it back to the state it was in.

Grrryyaarrrrggggghhhhhhhh!


STOP: 0x00000050 ()xFFFFFFF0,0x00000000,0x8044E873,0x00000000)
PAGE_FAULT_IN_NONPAGED_AREA
Address 8044E873 base at 80400000, DateStamp 44925809 - ntoskrnl.exe

Update on the power situation: someone discovered the reserve tank for the generator, and hooked that up. As of 12:28pm they have ~12 hours of runtime, and a tanker is apparently on the way.

They got it going with about 2 minutes of UPS power left. They cut it very fine.

Hmm. It looks like at some point during the night, one of my CRTs has died. The magic smoke hasn't escaped (which is a Good Thing), but it refuses to show any signs of being switched on. I'm slightly peeved as it's the nice Trinitron tube that's died, but then again I do also have a 17" HP hooked up to this computer, so it's no more than annoying.

Today's achievement: zombifying pine

Which isn't much, when you realise that I abandoned today at about 11:30 due to a mixture of:

a) whatever server responsible for DHCP at the uni deciding that my laptop (and apparently three-quarters of other laptops) wasn't worthy enough to receive an IP address, thus making all my internet connections implode

b) my deciding to go and upload the deliverables for Software Engineering to some webspace, only to find that only one deliverable was where it should be, and that whoever put it there forgot to set the permissions correctly

c) attempting to sort out some emails, and in the process managing to wedge one pine session over PuTTY, another logon prompt, and then wait a minute for a third PuTTY shell to login to the Informatics unix server

And I later completely forgot about the Software Engineering meeting (sorry folks), and only found out about it when an e-mail went round asking where all the programmers were. Oops.

Oh, and on the way back I managed to walk through a curtain of snow. I saw it in front of me, and over about twice the width of a road it went no snow -- snow -- no snow. Followed later by a small amount of near-horizontal snow, in March, with the sun shining. Ah well, this is England.

Anyway, Nick happened to log into the unix server just now, and did a 'w' to see who was logged in. I then asked if you saw everyone's processes with 'ps', to which the response was him typing 'ps -aux'. Followed by a few screens of processes from several people (not all of whom were actually logged in), including my pine process. Rather impressive when you realise that when I logged out, the shell *should* have killed anything it had spawned. Well, that explains some of the messages I was seeing with another pine session. One kill -9 later, and the uni has one less pine process.

It's been a while since I've updated, mainly because I'm never at the computer when an update springs to mind. So, what's happened recently...

Saw Serenity about a week ago. Managed to catch me off-guard quite a few times before I cottoned on to Joss Whedon's tricks. All in all, a great film. Unfortuantly I can't quite say the same for the cinema. The film appeared a bit washed-out (like someone's turned the brightness up too high), thought that might be intentional. On the other hand, the jitter probably wasn't, and the sound sounded awfully like it was clipping at places. Ah well, I'll just have to wait for the DVD version.

What else... Nick managed to lock himself in the house. New fridge got delivered, and the delivery men managed to double-lock the door on the way out. Which requires a key to unlock. Which requires said keys to not have hidden themselves. Then again, this was the day when the karma bunnies apparently decided to gang up on him. They caught me too - I'd left my computer to play a game with the others, and when I came back it had randomly bluescreened on me. As such I'm making sure Nick stays well clear of it.

Oh, and yesterday Dan became the victim of a small EMP by lightning which crashed his MP3 player. Before you start worrying, he's fine - we're guessing the bolt hit a cellphone tower nearby and he was just in range of the electromagnetic nastiness. Considering that he was near Moulsecoomb station at the time, and I heard the thunder from inside a building at Sussex University, I'm guessing that the lightning hit something with plenty o' volts. Now that I think back on it, there was only the one thunderbolt, so it might have been positive lightning (most lightning is negative, positive lightning is less common and packs a lot more power. Negative lightning leaves you with your hair on fire, positive lightning leaves a pair of smoking boots).

Hmm, Odysseus is still recompiling Linux 2.6. I was planning on a recompile anyway, but it didn't help that someone decided to not compile NE2000 support this time. Ah well, it's better than last time when the Sis IDE support was horribly broken and panicked the kernel if you had only one device on the second channel. I can finally use the CD-ROM drive in Odysseus.

Speaking of rants, it's turned out that what with trying to fix Norton, then trying to fix the TCP stack, then trying to fix both, that I've managed to break both BITS (Background Intelligent Transfer Service, lets things like Windows Update use only the idle network bandwidth) and some level of IrDA support. The IrDA brokeness is more spectacular: while BITS merely fails to start, the irevents service leaves a nasty message in the event log on start and then crashes on stop, taking out the entire svchost.exe process it's hosted in. Ouch.

Anyway, that's enough ranting at operating systems for now. I'm just waiting for the linux kernel to compile, whereupon i'll probably find I've broken it in some impressive way. Why they decided to run a seperate gcc instance for every single source file is beyond me. I dread to think how many CPU cycles (of which there aren't many on Odysseus, being only a 300MHz Pentium clone) are being lost due to process startup/termination overhead, repeated reading of headers, and all the internal init/deinit stuff in GCC. Argh!

Seen on a web site distributing a RTSP implementation in source code form:

How to configure and build the code on Windows

To configure the code for Windows, first unpack it on a Unix machine

The justification for this is, apparently, "it's rather hard to imagine anyone doing serious development of networking software without having any access to a Unix machine". I do have access to a unix machine, however that machine happens to be at the university, which is not where I am.

Come on, you unix fanboys need to learn that not everyone uses unix or unix-variants. Would you be happy if I distributed code that compiled on linux, but only if you first unpacked it on a windows box?

So, I thought it might be worth downloading the current 1.1.1 version of oggenc, as apparently it's a lot better. Silly me.


Problem #1: The vorbis people apparently don't believe in up-to-date binaries for win32. Solved by deciding to build from source. I'll just need to download, unzip, and run nmake, right?

Problem #2: While they provide MS Visual C++ projects, they neglect to convert the line endings to windows style. Solved by uploading to a random ftp site in binary, and then downloading in ascii mode.

Problem #3: No, MSVC will not automagically include all the project files in the current directory just by opening a workspace containing an empty project

Problem #4: Apparently I need to download (and presumably compile) libvorbis to make it work. I'll just pick up and compile libogg while I'm at it, shall I?

Problem #5: libvorbis depends on libogg. Specifically, libvorbis depends on the libogg files being in a directory called "ogg" in the include path, or in a directory called "ogg" in one of "..\..\ogg\include;..\include;..\win32\src;..\..\vorbis\lib" where the current directory is "D:\blah\libogg\win32". Solved by removing the "lib" prefix.

Problem #6: vorbis-tools\oggenc\encode.c(176) : error C2143: syntax error : missing ';' before 'type' followed by vorbis-tools\oggenc\encode.c(183) : error C2065: 'bitrate' : undeclared identifier


At this point I decide it's too hard and give up. And don't you dare suggest that I should debug someone else's code that should Just Work.

How on earth do all you Unix/Linux/BSD people cope with building everything from source?

Help! My Belkin router is spamming me! (taken from The Register

The marketing geniuses at Belkin, the consumer networking vendor, have dreamed up a new form of spam - ads served to your desktop, by way of its wireless router.

Uh Clem. a former Belkin wireless router user, was perplexed to find machines on his network redirected to an ad for Belkin's new parental control system, following a software update...

The router would grab a random HTTP connection every eight hours and redirect it to Belkin’s (push) advertised web page.

"It seems the router now supports a parental control and the market droids at Belkin got the bright idea of equipping the router with intrusive nagware," writes Uh Clem. "Of course, I have this strange notion that routers should pass data unmolested by marketeers!"

There is an opt-out link on the advertised page but this failed to appease Clem who, not unreasonably, objects to having to "opt-out from commercials from my router". Because of the ads, he's decided not to buy Belkin products again.

In response criticism, a Belkin product manager came forward this week to confirm the behaviour was designed into the products as a way to make it easier for consumers to sign up to a free trial of its parental control software. Belkin's Eric Deming is keen to allay concerns about the technique which have produced sharp criticism of the company on the news.admin.net-abuse.email newsgroup.

Here's what they have to say for theirselves (direct from the front page of www.belkin.com):

Belkin is aware of some recent postings that claim that Belkin wireless routers are spamming users during the setup process and periodically thereafter. It is not now, nor has it ever been, the policy of Belkin to intentionally spam our customers or anyone else. Belkin offers a free trial of our parental control feature in our routers, and to make our customers aware of the feature itself and to give them the opportunity to take advantage of the free trial, we have tried to direct users to the information regarding the parental control features. However, since this has become a source of concern to our users, and it is Belkin policy to address the concerns of our users quickly, Belkin has decided to remove this function from the routers. Each router's firmware that incorporates parental control as an option will be changed.

My opinion of this? Whoever came up with this idea should be forced to read *every* RFC to date, and to re-read every single one until he has repented. That could take some time if they're in marketing.