|
| Greg's diary |
| June 2011 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Wednesday, 1 June 2011 | Dereel | Images for 1 June 2011 |
| Top of page | ||
| next day | ||
| last day |
|
Winter comes on time
|
Topic: general, gardening | Link here |
Up this morning to find the first real frost of the season, dead on time for the first day of winter. The weather station, now relocated at shoulder height, registered a minimum of +0.1°, but clearly it was cooler lower down:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The ice was 2.25 mm thick, so it had been cool more than briefly. But the real surprise was in the greenhouse: -0.2° on the floor and -0.6° at the top. Clearly it's not much good for frost protection, though it doesn't seem to have done any harm to the plants. I'll have to keep an eye on them to see if I'm right or not.
|
Repotting and weeding
|
Topic: gardening | Link here |
More weeding in the east garden today. This has been dragging on all year, but gradually things are looking a little better. The trouble will be to keep it that way, of course.
The Clematis maximowicziana and the Mandavilla laxa that we bought at Lambley Nursery six weeks ago have been sitting in their original pots in the greenhouse for the intervening time. The Clematis is deciduous, but I'm not so sure about the Mandevilla, so decided to put both into bigger pots. Maybe I should just have planted them, but I still need to think about where I should plant them.
|
Dosing fill-in flash
|
Topic: photography, gardening | Link here |
Our Dahlia imperialis are now flowering happily, about 3.5 metres off the ground. How do you get a photo of them? There's always the sky in the background. Tried some various experiments in early evening with fill-in flash, which doesn't seem to be quite satisfactory with automatic exposure. Fiddled with the flash exposure compensation, which showed useful differences. Here +0 EV, -0.3 EV, -0.7 EV and -1.0 EV:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Rather frustratingly for the exercise, the uncompensated one seems to be the best.
| Thursday, 2 June 2011 | Dereel | Images for 2 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Cutting over to 64 bits
|
Topic: technology | Link here |
Carried on with the next step of the 64 bit upgrade today: upgraded the system to the latest FreeBSD 8-STABLE, built a new, custom kernel and booted as dereel.lemis.com. The system is (currently) on an external USB disk, so the cutover involved moving the disk to the correct system and rebooting. How did it work? Here are the things that I should have done before rebooting, some of which I really did do as described:
Move /var/log and /var/squid to /home/var.
Update /etc/rc.conf and /etc/fstab to reflect the reality.
Ensure that you have mount points for all file systems.
Check subdirectories of /etc for RCS directories and check out. In my case, I found /etc/ssh/RCS, and I'm still not sure what to do there. I have checked in all sorts of things, such as the keys, which sounds counterproductive. Potentially ssh_config and sshd_config might need to be checked out, but they're unaltered on the old system.
Ensure that the time zone is set correctly:
After rebooting, things came up nicely, in particular the real reason I did the upgrade in the first place:
In fact everything went well. Until I tried to start X. As I feared, that wasn't straightforward. The first attempt, with the old configuration file, failed with “no screens found”:
OK, so what happens if I try X without a configuration file?
VESA drivers would be useless anyway, so didn't pursue that one any more. But I still had a trick up my sleeve: X -configure. Ran that, created a configuration file, and started. Blank screen.
At this point, decided it was time to reboot 32 bits and lick my wounds. After rebooting, found little of interest in the log file:
What does that mean? It looks as if the module was unloaded and continued running. It didn't come up with anything sensible, though:
So I need to do some research before I try again. In principle I should get the first alternative to work, my own config file with the proprietary nVidia driver. But why doesn't it want to run? There was no further information in the log file, and nothing whatsoever in the kernel log file. At the time it did report:
Did I forget to load the sound drivers at boot time? To be investigated.
|
Weeding and seeding
|
Topic: gardening | Link here |
Another hour or so weeding in the garden today. It's slow work, and I only got a little over 1 m² done.
We've decided to plant annuals in the garden to at least mask out other weeds and to give us some time to decide what else we could plant there. I gave Yvonne a book with lots of suggestions and asked her to make a list of what she would like. She came back with a foolscap sheet of paper with lots of suggestions, including one annual (Atriplex hortensis), so made my own list and off to eBay to see what was available. Ended up buying 17 different packets from one seller, prompting the longest Subject: line I recall having seen in the response mail message:
But would you expect less from eBay?
|
Firefox surpasses itself
|
Topic: technology, opinion | Link here |
I've been watching firefox crash with regular monotony. I wonder if it's running out of some internal memory space: it seems to hit about 1 GB in size and then crash. Just before it does, top shows it in STOP state:
If it's a memory limit, it's complicated: I've seen larger process images. And it's not the kernel, which allows over 2 GB.
When it restarts, it needs nearly 90 seconds of CPU time (or 90 minutes of a CDC 7600 supercomputer) just to recover its state:
Somehow it seems to be getting even worse. Possibly it's because I use windows rather than tabs, and some changes have pessimized window handling.
| Friday, 3 June 2011 | Dereel | Images for 3 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Seamless system downgrade
|
Topic: technology | Link here |
I'm still puzzling about the problems starting X on the AMD-64 image for my machine, but it's not the highest priority. Today I saw another minor issue: the statistics program for my 3G modem counted bytes in 32 bit integers, and currently I have transferred about 3 GB (“-997 MiB”) since last starting PPP. Clearly a case for widening the integers.
But what's the name of the program? It should be e169-stats, but I've been playing around with it, and for no particularly good reason the current version is called fstats. Went looking for it and didn't find it where I thought I should. With the help of locate discovered it in /destdir/usr/local. How did it get there?
It seems that when I rebooted from the 64 bit version, I managed to boot from the wrong partition. My current 32 bit image is on /dev/ad4s1d, and the old version is on /dev/ad4s1a. To boot from the d partition I need to add these lines to /boot/loader.conf on the a partition:
But in this case, I was booting from the external USB drive, and it seems it used the loader.conf file on that disk, so it booted from the a partition, which is 2½ years old. And everything went so seamlessly that it took me over a day to discover it, and that only by coincidence. Now if the upgrades would only work that well!
|
Still more weeding
|
Topic: gardening | Link here |
It's taken me a long time to drum up the energy, but finally I'm getting through the weeds. In particular, there are two different kinds of grass. The one I recognize is Couch grass, which has masses of underground rhizomes that enable it to spread rapidly. I hate it! Fortunately it's in the minority. The other kind, for which I don't have a name, is clumping, and it just gets big. It's relatively easy to uproot, and in fact I wonder if it isn't easier to remove now that it's big. Each clump has blades about 1 metre long, and they seem to inhibit the growth of other weeds, presumably including the same kind of grass. So now I just need to pull out a single clump from where I would have had to pull out dozens of seedlings at the beginning of the season.
|
Why copy zone info file?
|
Topic: technology, opinion | Link here |
On most UNIX-like systems, the canonical way to install a time zone file is simple:
But why? An obvious reason is that /usr is usually a separate file system. That's not the way I do things, so I could equally well do:
That would arguably make things easier when updating time zone files, which seems to be
happening a lot lately. At least it's better than SunOracle's emetic
hoop-jumping to download a completely redundant time zone file. It would be nice to
be able to link directly to the kernel source tree, but at least in my case that's on a
different file system, and it sounds a little fragile.
|
Tomato seedling recovers
|
Topic: gardening | Link here |
About 6 weeks ago I planted the one surviving seedling of the “Black cherry” tomatoes that I received with “Burkes Backyard” [sic] magazine—and accidentally kinked off the tip. That was annoying, of course, but I left the stump in the pot and watched the leaves gradually wither away. But it was worth it. Today I discovered that it has really managed to produce new leaves:
|
|||||||||||||||||||||||
As I've already discovered, it's really worth holding on even when things look as if they're dead. Now we just need Yvonne's Salvia cuttings to sprout.
| Saturday, 4 June 2011 | Dereel | Images for 4 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Against the light panoramas and new hugin
|
Topic: photography, technology | Link here |
Photo day again today, and again it was sunny. Took my photos as usual, and while converting them installed the latest version of Hugin, in the hope that the mask processing would be better.
I wasn't completely disappointed. I can now reshape a mask by dragging on the corner points. Previously this moved the entire mask, not quite what I was looking for. But I still can't get include masks to work, and the mask boundaries are always shown in white, which makes them almost impossible to detect in my situation, where I'm working around the brightest parts of the image:
|
|
||||||||
And wouldn't it be nice to have the same mask for two images, one as an exclude mask and one as an include mask? Then I could do much better than my current approximate masking:
|
|
||||||||
Apart from that, it's clear that the method of masking out the sun only really works if the sun is clear in the sky, like here:
|
|||||||||||||||||||||||
If it's shining through trees, it's as good as impossible to mask things properly:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Pottering in the garden
|
Topic: gardening | Link here |
Somehow didn't get much done in the garden today. Spent some time trying to work out how to put more plants into the greenhouse, but I'm still not happy. And did a little more weeding, but somehow my heart wasn't in it.
| Sunday, 5 June 2011 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More Hugin games
|
Topic: photography, technology | Link here |
There are a number of new features in the latest version of Hugin. One of them—I think—is a direct display of various projections in the fast preview window. I've looked at various projections before, but I recalled it being quite a pain. Now I can select projections and look at them almost immediately. By default I get an equirectangular projection, but there's also a Miller cylindrical projection which seems to improve the rendition of the verandah panorama (second image):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It certainly reduces the compression at top and bottom.
|
Dying firefoxes
|
Topic: technology | Link here |
Firefox continues to crash. Today I ran ps -l against the process every second, and caught the output when it died. An extract:
The first line shows it operating normally. Virtual size is 1038764 kB, about the size it often is when it dies. The status field is the most interesting: S means that it's sleeping for less than 20 seconds, which is pretty normal. T is stopped (which I've also seen in the top output), and the L after it means that pages are locked in core, for example for I/O. So this looks like the time it took to dump core, about 40 seconds (I haven't shown all the output). And the core dump?
=== grog@dereel (/dev/pts/12) ~ 14 -> gdb /usr/local/lib/firefox/firefox-bin firefox-bin.core
It's round about here that I begin to lose interest. The whole thing, from top to bottom, seems to be in libxul.so, whatever that is, and the functions seem to delight in calling themselves recursively—if I can trust the backtrace. But the names seem to suggest that the problem is related to JavaScript and not to the size of the process image. I wonder what else I can do without getting my hands dirty.
|
Weeds, weeds, weeds
|
Topic: gardening | Link here |
There's so much to do in the garden, but weeding has one advantage: I can start and stop when I want. Did a bit more in the north bed today, but it's slow work.
| Monday, 6 June 2011 | Dereel | Images for 6 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Frost and weeding
|
Topic: general, gardening | Link here |
Very mild frost this morning; some parts of the garden showed ice crystals, but the thermometers showed a minimum of 0.6° (at the top of the greenhouse), and the water in the bird bath didn't freeze.
Later it was warm enough to do some weeding, this time in the north bed. We have some daisy-like ground cover there which has gradually been spreading. It looks quite nice, but it has completely enveloped two Hellebores, and I had to cut them free:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
We'll move the Hellebores away once they have flowered. In general we need to rework the bed, which has filled up with volunteer Acacias (Acacia baileyana and Acacia pycnantha) to the point where we need to relocate at least a rose and a Grevillea. But first we need to remove the root of the Aloysia citrodora. And then there's this rose arch that I bought nearly a month ago, and which we have been moving around the garden looking for a suitable spot. We've finally decided on the pathway between the two halves of the north bed, which will result in a significant difference in appearance in that area.
|
64 bit FreeBSD: the next attempt
|
Topic: technology | Link here |
The real problem I have in migrating my system to 64 bits is that I don't want any significant downtime on my main machine. Today I thought of another way to attack the problem: migrate Yvonne's system first. She's currently running FreeBSD 8.1, about a year old, so it's time for an upgrade anyway. Set off to do that: the first step was to add a new disk: for some reason I had forgotten to put a spare root partition on the system.
But first to boot the system. Connected up the external USB drive with the 64 bit system and tried to boot. Failure:
|
|||||||||||||||||||||||
That's FreeBSD's inimitable way of saying “The root file system you specified in /etc/fstab doesn't match the file system from which I loaded the kernel, and it is not present.” That's understandable: I had modified it to match the disk configuration on /dereel, so it didn't work here. So, once again, I had to start from the old system.
Ran sysinstall, like I always do, but this time it failed: although it claimed to have written the partition table, it hadn't. The sysinstall display showed:
But bsdlabel said, correctly:
I couldn't find a way around that one. In the end I overwrote the beginning of the disk and used bsdlabel to manually partition the disk. Then I could—finally—copy the image across and boot from it. Failure again: the external disk wasn't detected. Something had got itself confused:
Power cycling the external drive solved that problem, and finally I was able to copy the data to the destination disk and boot from it. Then tried to start X. The graphics on this system are on the motherboard, and for some reason mplayer can't expand to full screen mode with it. So I had added an old PCI graphics card, which caused problems starting X, not helped by the lack of a /var/log file system: I didn't have a log file to puzzle over. But the card proved to be an old SiS card, not the old nVidia card that I had thought it was. Why don't people write summary specs on the cards?. Found the nVidia card, put it in the system and got an amazing message during the BIOS probes:
|
|||||||||||||||||||||||
That's incorrect, of course: I didn't tweak any BIOS settings. It was caused by the display card. There didn't seem to be any way around that one, and clearly it was influencing system reliability, so I removed it again. That was all I had time for (Yvonne wanted her computer back), but now I can do things more incrementally when she's not using the machine.
|
Frost protection
|
Topic: gardening | Link here |
The weather forecast for Ballarat predicted frost tonight, the first time it's been so definite. Given my previous experience, it seemed prudent to remove at least most of the tomatoes from the greenhouse and put them on the verandah, where we've never had a frost. I fear I'll have to do this several times during the winter. It'll be interesting to see the fate of the tomato plants that I leave in the greenhouse.
| Tuesday, 7 June 2011 | Dereel | Images for 7 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Miserable weather and apathy
|
Topic: general | Link here |
We didn't get a frost overnight, but the weather was pretty terrible nevertheless: cold, wet and windy. I had intended to go into town in the afternoon, in particular to finally have a “quarterly” medical checkup—the last one was nearly 18 months ago. The only reason I go to the Eureka Medical Centre is because of the one doctor, Neil Philips. Today I rang up, got hung up on once, put on hold the next time, and finally discovered that Dr Philips no longer works there, and they wouldn't tell me where he is now. So in the end I decided against going in. But I'll have to go soon, maybe tomorrow, and then find Dr Philips or a new doctor.
The weather didn't get any better during the day—in the afternoon we had a sudden hailstorm:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
By 20:45 the temperature had dropped to 1.6°, the lowest of the day, and it was clear earlier that we were in for a frost:
And of course at these temperatures it's almost impossible to keep this horribly badly insulated house warm. The brain-dead Fujitsu air conditioners don't help: the one in the bedroom didn't turn on for about 20 minutes, though it was set for 30°, and the surrounding temperature was 15°:
|
|||||||||||||||||||||||
It wasn't de-icing: the other heads connected to the compressor were working. I still think it must be the location of the sensor, which is influenced by the hot coolant. I must really get something done about it.
|
Another firefox crash
|
Topic: technology | Link here |
Firefox continues to crash at far-too-frequent intervals. I'm collecting stack backtraces. This one was much shorter:
So again it seems to be in JavaScript. This one was an eBay page. But I don't think it's the page itself: I suspect that it's more likely to be some race condition in the browser code, since it frequently happens when I do things with many different windows.
|
Parasites on lemon tree
|
Topic: gardening | Link here |
I seem to be getting ahead with the parasites on the Bougainvilleas, though it's difficult to say, since spraying aphids with pyrethrum doesn't cause them to disappear. Unless they move, it's difficult to say whether what is left behind is dead or alive.
Lately I've noticed significant leaf drop on our Meyer lemon tree. That's not necessarily a bad sign: in the 7 weeks since finishing the greenhouse it has almost doubled in size, and it looks like it's the old leaves that it's losing.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But it was stuck in the corner where I couldn't look at it, and though I found and sprayed a couple of parasites, it was difficult to be sure they were all dead. Today I moved it to a more accessible position and collected the dead leaves, some of them with clear evidence of parasites:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Closer inspection showed that at least the ones on the left (aphids?) were clearly dead. Their abdomens appear to have exploded:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But what are the others? Are these the dreaded scale insects?
|
|||||||||||||||||||||||
| Wednesday, 8 June 2011 | Dereel | Images for 8 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another firefox crash
|
Topic: technology | Link here |
The next firefox crash happened under similar circumstances to the last: I was trying to leave feedback on eBay. Is this also JavaScript? It's not in member functions, but it still appears to be JavaScript:
|
Medical checkup, postponed again
|
Topic: general | Link here |
It's becoming urgent that I see a doctor to renew my prescription, no matter how repulsive I find the Eureka Medical Centre, so into town today with the firm intention to go to the doctor, come what may. Got into the horrible registration queue, and then saw a sign at the registration: starting next week, they are no longer charging a “gap”, money that you have to pay on top of the Medicare fee. Currently it's $30, even if you leave and don't see the doctor. Called Yvonne and confirmed I had enough pills until then, so once again left without seeing a doctor.
While in town, to Bunnings and looked around for inspiration for the windows in the new lounge room. None came. I did find some wooden flooring, which appears to be easy to lay, and it's even cheaper than carpet. There's definitely something to consider. In the end, left without anything more than some potting mix and some plant labels. Also picked up the seeds that I bought last week from the post office.
|
Officeworks for computer supplies
|
Topic: technology | Link here |
While in town, also dropped in at Officeworks to buy some DVD+Rs. I'm planning to go to Geelong tomorrow, and while I'm there I'll probably buy some new 2 TB disks at MSY. But it's always good to compare prices, and to my surprise Officeworks had 2 TB external drives for $98. The cheapest ones at MSY, a real low-price company, are $99. That's quite surprising. Unfortunately, all their disks were USB-only, and I wanted eSATA, so I didn't buy anything.
|
Planting seeds
|
Topic: gardening | Link here |
Back home, took a look at the seeds to start planting. 17 different kinds, but only about 4 or 5 that I can plant right now, and on further investigation discovered that the Impatiens had to be planted in a mixture of vermiculite and perlite, so in the end only planted a single row of Impatiens (they certainly don't require vermiculite and perlite to germinate in nature), and some black Viola x wittrockniana and Anchusa capensis. Even there I have my doubts. The instructions for the Anchusa suggest planting in winter for summer flowering, but mention an optimum temperature of 18° for germination. It'll be a while before we get that here.
The instructions for the violas state to place them in the fridge for two weeks after planting. That's not really necessary: the average temperatures recently haven't been higher than in a fridge. And then it occurred to me: the seller of the seeds lives in Queensland, where winter temperatures are much higher. You see similar recommendations for keeping tulips in the fridge in warmer climates, but they're not necessary here. But that could mean that the Anchusas won't germinate properly in these conditions; thus the limitation to two rows.
|
Reception problems: config issue?
|
Topic: multimedia, technology | Link here |
I've been keeping notes of less-than-perfect recordings lately, and gradually I'm seeing a pattern: they seem to be quite dependent on the channel. Most of the recent ones have been on two frequencies out of 5 (7 and Nine). Then it occurred to me: there's a column finetune in the channel table. Was it maybe not set?
No, it wasn't. Anywhere. All the rows had it set to either 0 or NULL. So I set it to 1 for all of them. Now to wait for the results.
|
General public illiteracy
|
Topic: general, technology, opinion | Link here |
I've offered the old dish washer and vacuum cleaner on Freecycle. The dish washer in particular got considerable attention. Freecycle members have no particular computer-related qualifications, of course, but you'd hope for a certain amount of general literacy, and clearly the ability to use email is an advantage. The replies I received showed that a number of people had neither. Here a number of replies:
The (empty) proved to be a multipart/alternative message with an empty text part and only HTML. There were two of them, one from gmail and one from Yahoo!. I wonder if people are being encouraged to wrap HTML-only mail in a multipart/alternative attachment.
Apart from that, the first message came twice, the second time a few minutes after my reply. Then nothing, despite a request for a quick reply, and I was left wondering whether the recipient had had difficulty reading it. So I offered it to somebody else, and again asked for a quick reply. I'm still waiting. I replied with my normal signed email, which includes the text
Unfortunately, not even that helps much with Microsoft “Outlook Express”. All it displays is:
|
|
||||||||
That paper clip on the right is the only indication that there's a message there. But then, maybe it's modern. When signing in on eBay today I got a window:
|
|
||||||||
But somehow I get the feeling that email is becoming a Tower of Babel. Even spam tells me (incorrectly):
This is the text part of a multipart/alternative attachment. When even spammers don't want to communicate in plain text, you have to wonder where the world is going.
| Thursday, 9 June 2011 | Dereel → Geelong → Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Messup after messup
|
Topic: general | Link here |
Somehow I wasn't quite on the ball today. I was pretty sure that I was supposed to go to Geelong for a periodontist's appointment, but it occurred to me that, although I had put it in my calendar, my daily calendar email didn't mention it. Went to look at the calendar and found not one, but two errors:
Yesterday! And to add insult to injury, I had put the date in American form, which my calendar(1) configuration doesn't support. So it didn't get sent.
I hate missing appointments, and it took me a while before I could ring up and apologize. When I did, I discovered my third error: I had written down the wrong date, and it really was today. Somehow I get the feeling that I'm losing my marbles. On the positive side, if I hadn't done so many things wrong, I would have gone to Geelong yesterday instead.
As if that wasn't enough, discovered a pile of about 30 DVD+R blanks on top of a computer. They're the ones that I thought I had used up, and for which I had bought replacements yesterday. Somehow I had confused two piles (the others were DVD+RW). As Yvonne put it: „Da ist der Wurm 'drin” (there's a WORM in there).
|
Still more disks
|
Topic: technology | Link here |
While in Geelong, went to MSY and bought 3 “2 TB” drives (really 2×10¹² bytes, or 1.82 TB of 2⁴⁰ bytes). With a couple of eSATA enclosures I ended up paying $316, $22 more than I would have paid for three external drives at Officeworks. I hope the eSATA interface is worth the extra money.
|
Today's firefox crash
|
Topic: technology | Link here |
I'm beginning to see some sort of pattern in the firefox crashes. First, they all come out of the signal handler, so the first four stack frames are always the same:
My guess is that the XRE_LockProfileDirectory () frame is saving the configuration before accepting the inevitable and re-raising the signal (in this case SIGSEGV). But once again it was in JavaScript code, and once again the backtrace terminated abnormally:
Clearly 0x00000000 is an invalid return address; less clearly, so is 0xbfbfac7c: on this architecture, that's in the stack (at the time of the crash, $esp contained the value 0xbfbfa7ec), and 0xac6176c4 looks interesting too. That could be a red herring, or it could be smashing its stack. Is JavaScript a red herring too?
|
Yet Another Bloody Power Failure
|
Topic: general | Link here |
Power failure in the evening, just before dinner. Fortunately it was another short one, though not as short as usual—I'd guess it was about 5 seconds, though it's very difficult to gauge.
| Friday, 10 June 2011 | Dereel | Images for 10 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Installing the new disks
|
Topic: technology | Link here |
So now I have a new photo disk and two backup disks, all 2 “Terabytes”. First put one in one of the particularly cheap and nasty external enclosures, in the process coming across this gem on the power supply:
|
|
||||||||
It's bilingual German and English. In German, the input is rated at 100-240V ~, and the output is 12 V DC at 2 A, while in English the input is 50-60 Hz at 1 A (where does the rest of the power go?), while the output is 5 V DC at 2 A.
The big question was: how should I organize things? Clearly FAT is less than appropriate, and things like NTFS have the disadvantage that I don't run any software that supports it well. UFS has always been my file system of choice, but how long will that (and FreeBSD) be around? Should I maybe use a Linux file system instead?
Did some thinking, including the option of putting a minimal FreeBSD system on the disk, but in the end decided that I will probably fill up these disks too in about 2 years, so the issue isn't as pressing as it seems. Created the biggest real file system I have ever seen:
People claim that UFS will run into trouble running fsck on such a large volume, and maybe they're right. But the average file size on this file system is about 500 kB, so I'm not too concerned.
Then started backing up the /Photos file system with rsync. It managed about 9 MB/s, which seems to be a good speed, but it meant 100,000 seconds for the complete backup, or somewhat more than a day. It wasn't until I looked more carefully that I realized that my scripts use compression even when writing to a local disk, so there's room for improvement—next time. So that was the end of that for the rest of the day.
|
Reliable map data
|
Topic: general, opinion, technology | Link here |
Call from Peter Dilley in the afternoon, reminding me of the books he lent me. Off after lunch to bring them back. I've never been there before, but I had the address, in Camms Road. That's what I have a GPS navigator for, of course. But, just as it doesn't know Kleins Road, where I live, it doesn't know Camms Road. It does know Cahills Road, a tiny little dead end forest road just across the road from Chris Yeardley, and offered that.
Was the address correct? Difficult to say. Decided to check the address in the (online) phone book, but that didn't work either: there's no “Dilley” in there. Instead I got a message suggesting to look for “Lehey”, which really doesn't seem to be similar. At first I thought it was context-sensitive, but it seems other people get that suggestion too.
Google Maps isn't much better: it knows the name, but whatever number I give, it puts me in the same place in Paynes Bridge Road.
Decided to head down there, take some photos and look for myself. Selected the middle of Paynes Bridge Road for the navigator and headed off first to Cahills Road (which is in the opposite direction). The navigator seems to be particularly fond of Cahills Road, and suggested that I go in there, drive 400 metres and make a U-turn:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I didn't get that far: I ended up in somebody's driveway:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
On to Camms Road, which proves to be somewhat to the south of where Google Maps put it, and delivered the books. It's quite pretty there:
|
|||||||||||||||||||||||
|
Email: more nails in the coffin
|
Topic: technology, opinion | Link here |
Somehow people seem to be trying to outdo each other with bad email. Today it was eBay:
That was a genuine message from eBay. They seem to have taken the HTML and converted the markup to HTML entities and stuffed the whole lot in the text version of the message. The first question was: “rate your experience” with this purchase, on a scale of 1 to 10. What does that mean? Gave it a 5. Next question: "Why do you feel this way?". That should be clear: because the question is meaningless. Only the third makes any sense: would you buy from this seller again (yes).
But even the spammers are getting on to the bandwagon:
It's interesting to note that paypol.com is a valid domain, which offers itself (in German) to me for purchase.
| Saturday, 11 June 2011 | Dereel | Images for 11 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Photo storage: why not ZFS?
|
Topic: technology | Link here |
Mail from Rick Owens today, asking if I had considered using ZFS for my photo storage. Yes, I have, but decided against it:
I don't know ZFS, and to get to know it would involve some effort.
ZFS belongs to Oracle, a company I seriously distrust. It's available for free now, but that could change. One of my concerns about using UFS was related to the long-term viability of the FreeBSD platform. I don't see ZFS becoming even that viable.
Almost everything I have heard about ZFS on the mailing lists relates to problems with it. I suppose that's normal enough, but I'd rather wait until it's more stable. UFS has been in use for 30 years now, and in that time it has (gradually) gained a reputation for being one of the most reliable file systems out there.
In short, I don't see any advantage over UFS. This doesn't make my choice perfect, of course. People have pointed to issues running fsck on UFS. Today my copy finally finished (after 25½ hours, pretty much what I estimated), and I immediately umounted the disk and ran fsck on it. The result?
I can live with that.
|
eSATA: worth the trouble
|
Topic: technology | Link here |
I've now fixed my backup script, and it no longer uses compression when writing to local disks. Even with a USB connection, that gave a threefold speed increase.
But a number of people, including Peter Jeremy and Rick Owens, have confirmed that it was worth paying a few dollars more to use eSATA instead of USB. Rick tells me that he has had a sevenfold speed increase over USB. I don't see this happening to me, but we'll find out once my eSATA controller (delayed by the Dragon Boat Festival) finally arrives.
|
Daily firefox crash: a pattern emerges
|
Topic: technology | Link here |
Today's firefox crash backtrace had a certain sense of déjà vu:
It's not identical to the previous one, but the trap occurs at exactly the same location, 0x8322f682 in js::MarkContext (). I suppose it's about time to take a look at the source, but not today.
| Sunday, 12 June 2011 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Today's firefox crash
|
Topic: technology | Link here |
The firefox crashes seem to be converging. This is the third day that it crashed in the same place:
It's interesting that frame 6 again shows the same address on the stack. Maybe there's some trampoline in the previous stack frame, and it's normal that the remainder of the trace is invalid.
Another thing that I noted was the number of threads: one initial thread and 23 others. Does firefox start one thread per window? I tried another instance with only a single window and shot it down with kill -11. The results were interesting: 18 threads instead of 24, suggesting that no, the number of threads and the number of windows are not directly related. But maybe there's some race condition in the code that only gets triggered with multiple windows, and which doesn't show up with the “there can only be one” processing of tabs.
|
Publishing video clips
|
Topic: multimedia, technology | Link here |
Yvonne has been asking me for some time to prepare the video clips from her camera for publication on the web. I've been dragging my heels, and I should be ashamed of myself. That's almost a basic copy operation.
The real problem is the sheer size of the files:
That's a 3 minute clip in 640×480 format. Why is it so big? It's clearly far too large to publish (and if I did, it would use up 4% of my monthly traffic quota). I need a recoder with hundreds of options, and since I've used it before, decided to try mencoder. But what options?
=== grog@dereel (/dev/pts/18) /Photos/yvonne/20110524 3 -> man mencoder
It's the same man page as mplayer, so looked at that. I've already commented on the length, but of course now it's even longer: 9107 lines, somewhat over 150 pages, much of it irrelevant to mencoder. Went looking for a tutorial and chose Google's favourite, which didn't look too bad.
The next problem:
=== grog@dereel (/dev/pts/18) /Photos/yvonne/20110524 5 -> mencoder MVI_0093.AVI -mf fps=25 -o /dev/null -ovc xvid -xvidencopts pass=1:bitrate=3000 -vf scale=320:240,eq2=1.5
mencoder is part of mplayer, which was installed. But it seems that somebody had decided that that was too simple, and they split mencoder out into its own port. Installed that and got:
=== grog@dereel (/dev/pts/18) /Photos/yvonne/20110524 6 -> mencoder MVI_0093.AVI -mf fps=25 -o /dev/null -ovc xvid -xvidencopts pass=1:bitrate=3000 -vf scale=320:240,eq2=1.5
So the instructions are out of date. Went looking on the mplayer site and found the documentation, in HTML format. How come they're not on my own system? Checked the (duplicate) mplayer source tree in /usr/ports/multimedia/mencoder/work/mplayer-export-2011-03-29 and found yes, indeed, there's a director DOCS with a file README referring to the subdirectory HTML—which doesn't exist. It's called xml now, and it requires DocBook and lots of other things to build. Tried and ran into trouble. I think DocBook and its associated paraphernalia have done more to stifle documentation than just about anything else I can imagine.
Instead downloaded the documentation from the web site, running into problems with wget on the way, and finally followed the instructions there:
=== grog@dereel (/dev/pts/11) /Photos/yvonne/20110524 85 -> mencoder MVI_0093.AVI -ovc lavc -lavcopts vcodec=mpeg4:mbd=2:trell -vf scale=320:240 -o Video.mpg
That looks familiar—clearly none of the documentation is up to date. Finally got some success with:
=== grog@dereel (/dev/pts/11) /Photos/yvonne/20110524 88 -> mencoder MVI_0093.AVI -quiet -ovc lavc -lavcopts vcodec=mpeg4:mbd=2:trell -o Video.mpg -oac pcm
Created both 640×480 and 320×240 versions. Surprise: they were both almost identical in size:
Why is that? But they're also about 7% of the size of the original. They're lossier, but I can't help feeling that the original is very badly compressed.
Finally, though, mencoder had another trick up its sleeve. It doesn't pay attention to file names—laudable in itself—and so it didn't convert the file format. video.mpg and video2.mpg are both AVI format files. I still need to find the incantations to convert format, if I want to. Isn't free software fun?
|
More winter preparations
|
Topic: gardening | Link here |
It's been pretty cool all month, and gradually the flowers on the verandah are dying back. Took down the hanging petunias today: it looks as if the recent winds have taken their toll. They're supposed to be annuals, but some are from last year, and they're not dead yet, so cut them back and hung them in the greenhouse, where at least they'll be out of the wind.
| Monday, 13 June 2011 | Dereel | Images for 13 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Useless mousetraps
|
Topic: general, animals | Link here |
I've more or less given up on catching rats with the traps I bough a while back, but they're not the only rodents to annoy us. The cats have been spending an inordinate time circling the dish washer lately, and it's fairly clear that there's a mouse there somewhere. Last night I put a few mousetraps out, but we didn't catch any mice. Instead, it seems that the mouse broke one of the traps:
|
|||||||||||||||||||||||
Was this a mouse or a rat? Either way, I really need a better kind of rodent trap.
|
Brew day at the Dilleys
|
Topic: brewing, opinion | Link here |
Peter Dilley had decided to brew a stout today, the first brew since he moved to Dereel. Over to his place to take a look, on the way marvelling at the instructions that my GPS navigator gave me:
|
|||||||||||||||||||||||
|
|
|||||||||
A big difference from my setup is the tidiness. Peter has not one, but two grain mills, one of which is very efficient, and also equipment for propagating yeast:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The roller mill must grind at 100 times the speed of my Kenwood mixer.
The big difference, though, is the way he mashes and boils. He has a 40 litre urn with a 2.4 kW electric heater in which he does everything. He has a programmable thermostat of the kind that I couldn't find when I built my temperature control computer, shown on the right in the first photo. The grist goes into a woven plastic sack which is then lowered into the hot water for mashing:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
He then does a single temperature infusion mash, lifts the bag out of the urn and lets it drain, without sparging. Then he does the boil in the same urn and drains the hot wort into the same kind of 20 litre plastic cubes that I use. The wort cools down in a sealed environment, so hot-side aeration shouldn't be a problem.
I have a number of issues with this approach:
It's as good as impossible to get the mash temperature constant between the heating element and the middle of the bag. Even while we were starting the mash, it was clear that there were significant temperature gradients.
It's wasteful, since there's no sparge.
The boil is done with lots of remnant starch, which would normally be removed by the sparge.
It's quite possible that the cooling down in the cubes is a good idea, but it makes it difficult to aerate the wort (which has to happen when the wort is cool).
And of course it completely rules out accurate multi-step infusions (but not decoctions), because it's almost impossible to get the temperature uniform.
Still, as Peter says, the proof of the beer is in the drinking, and we obviously haven't done that yet. And on the positive side, it's so much easier than my arrangement that I should seriously think about what parts of the procedure I can adopt.
|
Camellias in early winter
|
Topic: gardening | Link here |
The Dilleys have a number of pretty plants around the house, including two large Camellias that are already flowering:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's amazingly early, and reminded me that I should do some more fertilizing back home (where the buds are only beginning to show). Did that, and also some more pruning, and planted out another tomato seedling in an old, broken 50 plastic container. I wonder how well that will work. After only 2 months, it's becoming difficult to move in the greenhouse.
|
Firefox crashes, more of the same
|
Topic: technology | Link here |
Today's firefox crash was once again the same:
This is no longer news; I'll keep information on a separate page.
| Tuesday, 14 June 2011 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Back into the garden, slowly
|
Topic: gardening | Link here |
I've been neglecting the garden more than usual lately, and though I did more today, it still wasn't much. While I'm gradually removing large quantities of grass from significant areas, more is coming up. I had hope that removing the big lumps of grass would mean they wouldn't come back as fast, but that doesn't seem to be the case.
|
More clean mousetraps
|
Topic: general, animals | Link here |
The mice seem to be as clever as the rats now. They lick the bait off without triggering the trap—exactly the same problem as with the rats. Is this a rat, maybe? It would explain the broken mousetrap the other day, but you'd expect them to be less delicate.
| Wednesday, 15 June 2011 | Dereel | Images for 15 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Goodbye MBR
|
Topic: technology | Link here |
Today I got round to installing the second of the 2 TB disks I bought last week. For decades now, there has been the One True Way to partition disks, so much so that it doesn't have a qualifier: Master boot record. There's also a program that goes with it, fdisk, which has been around as long. But gradually the time has come for change: MBR stores information about the disk as cylinder, head and track information, long after that kind of addressing has lost any meaning. In fact, it can't even work: each of those fields is one byte long, so even if you set them to 0xffffff, you'd only be able to describe disks of up to 8 GB. In fact, the real information is stored in two 32 bit fields describing start sector and partition size. Here the description in FreeBSD file /usr/include/sys/diskmbr.h:
The important thing here is that the start and size fields are unsigned (u_int32_t). With that, we can describe up to 2 TB (or, to be pedantic, 2 TiB, in other words 2⁴¹ bytes) disks, and then MBR has finally done its dash. But there's a currently open FreeBSD problem report showing that the sector counts are translated with the strtol function, which is signed, so there are problems with 32 bit versions of FreeBSD and disks over 1 TB (2³¹ sectors).
Just what I needed to try out. Did that, and got:
That looks fine. That is fine, if you ignore the CHS parameters. It seems that the problem described in the PR only occurs when reading from a configuration file (-f option). But it's clear the writing is on the wall, and we have had a replacement, the GUID Partition Table, for some time.
In passing, it's interesting to note how extremely old all this stuff is. While rummaging around in /usr/src/sbin/fdisk/fdisk.c, I found:
That's well over 20 years old (astute readers will observe that these comments are in reverse chronological form). It predates FreeBSD. Other indications are in diskmbr.h:
This is the same partition type (165 decimal) used for FreeBSD, but we still call it DOSPTYP_386BSD.
I've just never used GPT before, but today was the day. It proves to be simpler than fdisk. I wanted a single partition, of course, and for that the commands are simple and straightforward.
It wasn't until later that I discovered there's a list command too. It's not in the man page gpart(8), because it's a generic GEOM command, so it's in geom(8) instead. That's seriously suboptimal: it should be in each man page to which it applies, though from the maintenance point of view it's understandable that that can cause problems. Still, something should be there.
And then there's the question of the file system. What parameters? The simplest option is newfs /dev/da0p1, but this disk is intended to store photo data, which is big. So I don't need as many inodes—empirically I've discovered that the average file size is 5 MB. Theoretically I could increase the file system block size from 16 kB (standard) to 64 kB or even 512 kB. But there are also a number of smaller files as well, including directories—nearly a third of the files are less than 100 sectors (51.2 kB) in size:
=== root@dereel (/dev/pts/9) /home/grog 35 -> find /Photos/ -size -1000|wc -l
=== root@dereel (/dev/pts/9) /home/grog 37 -> find /Photos/ -size +1000|wc -l
=== root@dereel (/dev/pts/9) /home/grog 41 -> find /Photos/ -size -100|wc -l
Increasing the block size increases the fragment size and wastes space. I haven't yet investigate in more detail, but it seems that I'm not really saving too much by doing so—mainly many of the thousands of cylinder groups.
In the end decided on the following parameters:
The parameters are:
-L Photos is the file system label.
-U selects Soft updates
-a 64 selects the maximum number of contiguous blocks to be allocated. This value corresponds to 1 MB. It's normally 16 blocks (256 kB), but I suspect this will improve performance for big files.
-e 2048 specifies the maximum number of blocks a file can allocate in a cylinder group before moving to another.
-g 2097152 is the average file size. This is considerably less than what I have established, but all I need is to suddenly add a couple of tiny files per photo and I could be under it. And that would mean that I wouldn't be able to create more files, so it's good to err on the low side.
-h 64 is the average number of files in a directory. I'm not sure this makes much difference.
-i 2097152 is the number of bytes per inode. It's not clear to me what the difference is between this parameter and -g, since an inode is a file.
-m 1 specifies the amount of space to leave for root only. Typically UFS won't allocate the last 8% of disk space to normal users, to avoid problems with disk overflows. This reduces it to 1%.
Created the file system and started copying the data with tar. It's much faster than I was getting with rsync, though that was mainly because of compression. Still, it's a matter of hours, so left it for later.
|
TV Reception problems: better or not?
|
Topic: multimedia, technology | Link here |
Since setting the fine tune flag on cvr2, I haven't had any serious problems with TV reception—until today. Then I found, while recoding:
It died right at the beginning, like it sometimes does with terminally corrupted images. Took a look at it, and it didn't seem that bad. I suspect that this deadlock issue occurs with particular and relatively rare kinds of corruption, so the more corrupt the image is, the more likely it is to happen. But it seems it can happen with otherwise good images as well.
|
Why I hate USB
|
Topic: technology | Link here |
Took some photos for Yvonne today, then in to read it in to my computer. It didn't work. Instead, it crashed the system. It seems that there was some interaction between the flash card and the USB disk. This time I got a photo of the messages:
|
|||||||||||||||||||||||
High time for the eSATA adapter to arrive.
|
Heat wave in winter?
|
Topic: general | Link here |
The weather hasn't been too bad lately, but today I got a surprise:
Where did that sudden spike in temperature come from? Yes, it was mild, but I didn't see a similar rise in temperature on the verandah (where the sensor is in the shade). I suspect that the sun had an influence on the readings. But that would be the first time. Is it possible that the angle of the sun makes a difference? In any case, it sounds like an idea to put a more substantial shade around the sensor.
|
Another day, another hop
|
Topic: gardening | Link here |
Did a bare minimum of work in the garden, clearing the Tropaeolums from the area south of the verandah, and finally got around to removing another of the hop rhizomes from the fence in front of the garage. Time to spray some herbicide.
| Thursday, 16 June 2011 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Out of gas
|
Topic: general | Link here |
Breakfast this morning was slightly delayed: our gas cylinder was empty. That was to be expected, but it gives me the chance to compare prices. We put the cylinder in five months ago. My calculations at the time suggested that the cylinder would last 6½ months, but even so, our annual gas bill would be $60, compared to about $90 with the rented bottles. And we would have had a bigger problem when they ran out—either change to one of the little bottles for the interim, or wait a day or two. I think we made the right choice.
|
More USB pain
|
Topic: technology | Link here |
Came into the office this morning to find dereel dead in the water. Further investigation showed a far-too-common syndrome:
By the time I got in, the only thing that it could do was to respond to a ping. So: USB is still not ready for prime time on FreeBSD, though I don't know which of the two is to blame. But it's clear that I can't run these disks on dereel any more. Put the disk on cojones, where I discovered that it doesn't do USB 2.0, so I was limited to 1 MB/s. Left it run fsck for about an hour before I lost interest and put it on teevee instead. But the output of fsck on cojones wasn't very reassuring either. I had lots of:
There were many more, including messages that certain sectors couldn't be written. And in the log file were messages like:
Those were the same messages I got before dereel crashed, which wasn't very reassuring. This was supposed to be a reliable backup, so decided to write zeroes to the entire disk to see if I triggered anything. That set off at 25 MB/s, which seemed acceptable, but at that speed it would take 80,000 seconds, nearly a whole day. Let it run.
While I was at it, decided to finally start logging dereel log output to cojones as well, so that I could save the messages in the case of a disk crash. cojones is already logging to dereel. Would that cause a loop? It didn't seem to—for a while. When I came back after lunch, found hundreds of messages scrolling by:
So that needs attention. For the time being I'll stop logging cojones to dereel, but it should be possible to fix syslogd to stop it looping.
|
LiIon pain
|
Topic: technology | Link here |
Later in the day I wanted to look at something on my GPS navigator. Not for the first time, the battery had drained while in “sleep” state. Connected it up to a power source. Nothing. Connector problems? Cable problems? No idea. Finally connected it via a USB cable to dereel—just what I didn't want to do. And with good reason. It started to charge, and I turned it on. Then:
Those messages say, roughly:
Three devices disconnected. Comparing information with the output of usbconfig, it was all devices: weather station (on ugen0.2), mouse (on uhid0), and the disk to which I was backing up (umass0).
All devices come back, but the backup is aborted. Amazingly, the system didn't crash, but the file system /backups was forcibly umounted, probably the correct thing to do.
The unknown device with the vendor ID 0x152d and product ID 2336 could be the GPS navigator, though I'm not sure. It's the only thing it said.
What didn't happen was that the GPS navigator booted normally. In fact, it hasn't said anything since. Took it apart (the warranty expired, conveniently, about 6 weeks ago), and discovered that the battery had no charge. So it looks very much like the system didn't correctly handle this discharge while sleeping, and the battery is now dead. What should I do (apart from never discharging the battery like that again)? If I can find a new battery, it'll cost me at least $20 with delivery, and I don't even know if that's the real problem. You can buy new GPS navigators for a little over $50, so it doesn't seem to be worth the trouble.
What really annoys me, though, is the consideration that I don't really need a new GPS navigator. It's more like a toy. But then Yvonne pointed out that the speed camera warnings can pay for the device several times over. Spent the rest of the afternoon researching navigators. In the year since I bought the last one I haven't got any more of an overview, but it seems that current ones have a bigger display (can't be wrong), have 3D representations (is that an advantage?) and built-in FM transmitters to transmit the sound to the car radio (I wonder if that works as break-through), and some have inputs for reversing cameras. But what brand? I've got used to the Nav N Go iGO 8 software, but maybe there's something better. By chance c't had an article about navigators two weeks ago, so spent some time reading about that. I get the impression that other navigators don't have any significant advantage over “Nav N Go iGO 8” (what a mouthful!), but I still don't have the overview I want. And to make matters worse, the eBay sellers don't say what software they use, or even give good descriptions.
| Friday, 17 June 2011 | Dereel | Images for 17 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
eSATA: The solution?
|
Topic: technology | Link here |
Finally my eSATA adapter has arrived. It must be one of the smallest boards I have ever seen:
|
|||||||||||||||||||||||
It's also the first board I've seen in a very long time which physical jumpers on it. It came with cables and a mini CD with drivers, presumably for Microsoft. In the afternoon, got round to putting it in dereel, and discovered that the cables supplied were SATA, not eSATA—the adapter (4 port) has two SATA and 2 eSATA connectors. Fortunately one of the housings I bought last week had an eSATA cable, so connected with that. At the same time took another attempt at getting the machine running in 64 bit mode. It's clear that that's not ready for prime time yet, but it seemed a good idea while the machine was down.
And of course it didn't work—couldn't even find the root file system (on USB). Gave that up, disconnected the USB disk and rebooted back to the old system. It still couldn't find the root file system: the eSATA adapter had probed before the SATA adapters on the motherboard, and the SATA drives, formerly /dev/ad4, /dev/ad6 and /dev/ad8, had had 4 added to the device number: /dev/ad8, /dev/ad10 and /dev/ad12:
Got that sorted out, but the controller didn't recognize the eSATA drive I had plugged in. More discussion on IRC, and Peter Jeremy, who has the same card and who recommended it to me, told me he was using the ahci and siis drivers, so updated /boot/loader.conf:
And yes, you don't need the " characters around the YES. Rebooted. No change. When will I remember that the bootstrap always reads loader.conf from partition a? So I had to change /destdir/boot/loader.conf. Fortunately, it loads the modules from the correct partition, in this case partition d. Booted, and this time the drivers got loaded. And I couldn't mount the root file system. The device names had changed again! Fortunately the loader lists the known devices, so I was able to establish that the boot disk is now called /dev/ada0, and the root file system is thus /dev/ada0s1d. Only /dump, on the old PATA drive, has kept its device name:
But what a pain! And I still couldn't detect my eSATA drive. Further investigation of the jumpers gave me the answer. Isn't it a good idea to take photos of the cards?
|
|||||||||||||||||||||||
This chip has four ports, but for some reason only two work; I'll have to find out which two later.
In fact, this isn't correct. It's a 2 port board, but I was confused by the fact that the ata driver always allocates devices for master and slave, though SATA doesn't have slaves.
The jumpers are clearly set for internal SATA (which matches the cables they supplied). Clearly the manufacturer considers the idea of eSATA to be an afterthought. I had seen the legend on the board when I got it, but it didn't make much sense until I saw the texts CON1 to CON4 next to the connectors.
But changing that means climbing through the mess in my office again to remove the card, change the jumpers and put it back.
Instead, started backing up my photos to the disk via USB. It had run all night in teevee without any problems, taking pretty much exactly the 80,000 seconds that I had estimated. The real time was 21 hours, 12 minutes:
It's interesting to note that disk speeds haven't kept up with disk capacity. My first disk was a Seagate ST-225 (“25 Megabytes”, including not only the sectors, but the formatting; really it only managed 20 MB). Its rotational speed was 3600 rpm, and its geometry (the famous CHS) was 616 cylinders, 4 heads and 17 sectors per track, giving a total real capacity of 21,446,656 bytes, or 20.5 MB. Those 2,464 tracks (616 × 4) transferred at the rate of 60 a second, so at least in theory it would have been possible to transfer the entire disk contents in a little over 41 seconds. Presumably positioning overhead and controller limitations would have increased that, but it's still orders of magnitude faster than today. If I could transfer from the 2 TB disk at the 1.5 Gb/s that the original SATA standard allows, it would still take nearly 3 hours to transfer the entire contents of the disk.
Since the disk was completely empty, the first thing was to create a new file system on it. That wasn't as simple as I thought: teevee is still running an old version of FreeBSD, and the gpart command wanted lots of information that I didn't want to give it, like start offset and size. So brought it back and did it on dereel, running into more problems which proved to be due to my omitting a step in the description in this diary (now fixed). And then, once again, I got the messages:
They were serious, too: they're the same once that preceded the crash the other night. Callum Gibson suggested that it might be an issue with the USB cable, so changed that, and was able to do my photo sync with no further problems. But that was relatively little, only about 1 GB. On reflection, I had used a different cable when connected to cojones, and it happened there too. On the other hand, it's an issue that I've only seen with the one disk (the newest one with the eSATA interface), so it's possibly a compatibility issue there. As a result put the disk back on teevee and started backing up the 1 TB of data on my photo disk. Another overnight run, since now I'm transferring the data by 100 Mb/s Ethernet.
|
Scanning solid objects: not good enough
|
Topic: photography | Link here |
I was going to take some photos of the eSATA adapter for the sake of discussion, but it occurred to me that it's relatively flat, so tried it in the scanner. The result looks something like Google Maps “satellite” view:
|
||||||||||||||||||||||
So I took the photos with my E-30 instead, in the process running into more issues. It was all on a tripod, and I went to a lot of trouble to line up the card parallel with the viewfinder. But what I got was:
|
|||||||||||||||||||||||
How did that happen? That was nothing like what I saw in the viewfinder. It appears that there are alignment issues with the viewfinder. The obvious thing to do is to use live view, but I didn't think of that. Instead I used GIMP to reshape the image.
|
More firefox experiments
|
Topic: technology | Link here |
The firefox crashes continue unabated—if anything, they're getting more frequent. Today tried setting up a completely neutral profile to see if they still didn't happen, in the process discovering how many things you depend on in the profile. Some don't appear to be settable any more, and the continual change in font sizes really gets on my nerves. Why do so many web programmers think they know better about how big characters should be?
Anyway, it didn't crash. It hung. And when I tried to restart it, the profile no longer showed up in the startup menu. What a pain! How I wish I could find a web browser that fulfils even minimal reliability criteria!
| Saturday, 18 June 2011 | Dereel | Images for 18 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Facebook branches out
|
Topic: technology, opinion, general | Link here |
Yes, I have a Facebook account, for reasons I don't really understand. I look in about once every month or two to see who wants to be my “friend”, and then leave again. Today, though, I got a mail message:
OK, what did Max have to say? Was it really nothing, or did Facebook mess up? Went to take a look. It wouldn't tell me immediately: I had to press “View This Wall Post”. Did that and found:
|
|
||||||||
|
|
||||||||
So yes, "" was less than he said, it seems. But I had to click again. And this time I got a “sign away your life Yes OK?” message:
|
|
||||||||
Who's BranchOut? A third party company? I didn't accept that, of course, but went to my Facebook page, where I found nothing of interest.
So what is this all about? Subtle spam? Later on I came back and took another look. It seems that BranchOut is Facebook's reply to (or maybe satire of) LinkedIn, a network for Professionals. Took another look and went through some kind of initiation ceremony, where I was presented with 30 pairs of people and asked whom I would enjoy working with—not, it seems, more, but at all, though I suspect this wasn't the intention. I had to select one in each case. But how do you do that? I have many and varied interests, and my “friends” on Facebook reflect that. Here an example where I couldn't possibly make a choice:
|
|
||||||||
If I were hacking file systems, I'd choose Kirk in an instant. But if I were hacking Peruvian Paso bloodlines, I don't think he'd be of much help, and I'd choose Pam.
In any case, continued, and as a prize got:
|
|
||||||||
Wow. I am so proud. And now of course I'm getting messages like:
Isn't that nice? I always wanted an empire, especially with awe-inspiring people in it. I suppose I'll investigate this for a while and then forget it, like I've done with other such sites.
|
Buying a navigator on eBay
|
Topic: general, technology | Link here |
Spent more time today comparing GPS navigators. It's almost impossible: even last year I noted that the manufacturers think that their raison d'être is as MP3 players and games, and that trend has continued (now they're MPEG-2 players too). The descriptions I get hardly address my requirements at all. Here they are again:
The “Hungarian” is an indirect requirement: of course nobody tells you what applications software the thing runs (the best is “Windows” CE), but I'm guessing that any receiver that has a Hungarian language option uses the Nav N Go iGO 8 software, which comes from Hungary. Spent some time comparing things and came up with the following partial chart:
| Seller | Price | Memory | SD card | FM transmitter | Bluetooth? | Hungarian? | Speed cameras | Free update | Return | Location | ||||||||||
| helomolto | $149 | 128 MB | 4 GB | Yes | No | Yes | Yes | exchange | Melbourne | |||||||||||
| helomolto | $169 | 128 MB | 4 GB | Yes | No | Yes | Yes | exchange | Melbourne | |||||||||||
| gl_4ever_wll | $149 | Yes | Yes | Yes | Yes | exchange | Melbourne | |||||||||||||
| ozchoice | $149 | 4 GB | Yes | Yes | Yes | Yes | $10 | Melbourne | ||||||||||||
| bikerboy111111 | $146.99 | |||||||||||||||||||
| lgo_smiley | $139.99 | 128 MB | Yes | Yes | Yes | Yes | $0, 30 days | Hong Kong | ||||||||||||
And yes, there are two entries for helomolto with what appears to be the same item, but with a “Buy it now” price differential of $20. Further investigation shows that they offer them for increments of $10 between $149 and $189. Just what you need to inspire confidence.
Looking at the list, it looked like a choice between the vendors ozchoice and lgo_smiley. ozchoice had the best description, even a video of how to install the reversing camera. While looking through there, found (well hidden) second general demo describing how to use the multimedia functions, the calculator, calendar, but also, as an afterthought at the end of the 4 minute video, it showed a brief one-second view of the navigator window: iGo 8. That's the software I'm looking for, so I bought it.
But what a pain! I must have spent 5 hours trying to find information I needed. And of course, after I bought it, I found another one in Mildura for $10 less. Still, they don't have a return policy, and I suppose I should reward ozchoice for going to the trouble of setting up the videos.
|
Dilleys to dinner
|
Topic: general | Link here |
Saturday night is “good food night”, and Chris Yeardley is almost always here. Today we invited Peter and Vicky Dilley as well. Fun was had, and I took lots of photos, unfortunately nothing spectacular.
| Sunday, 19 June 2011 | Dereel | Images for 19 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
We don't need no steenking testing
|
Topic: technology, opinion | Link here |
From time to time bugs, even big ones, sneak into development projects. The biggest usually get caught immediately during testing. But in some cases it seems that not even the minimum of testing gets done. It's a commit to fix a single-character mistake:
This one is a variant on a classic UNIX mistake. One character wrong, and you delete your entire /usr hierarchy. It's clear that whoever did this one didn't test the script. And the comments on the web page are well worth reading.
|
BranchOut... and fall over?
|
Topic: technology, opinion | Link here |
More copious mail from BranchOut today. I'm obviously not professional enough to understand it. So far I haven't found the text of any of the messages I was informed of, not even the ones I wrote myself. In fact, I can't remember having written any messages, but they tell me I did. Maybe I'm just getting senile, which would also explain why I can't navigate this site—it looks different every time I go there. On one occasion I was asked if I wanted to get to know Doug Rabson (already one of my “friends” on Facebook), and on the same page, not far away, I was asked to endorse him. But I can't find that any more, just:
|
|
||||||||
That's the same Pam Hay I was asked to compare to Kirk McKusick yesterday. I'm left with the feeling that, like Bumblebee, they let the users do the testing.
|
Transplanting house plants
|
Topic: gardening | Link here |
Somehow my garden work has ground to a halt. Despite the best of intentions (really!) I seem to have spent the last few days in front of a computer. But dinner last night showed us that the house plants in the formal dining room needed repotting, so did that, with some help from Lilac. For some reason she wanted to sit in the pot and didn't leave even as I started to fill in potting mix. In the end I had to remove her:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Philodendron has expanded significantly, and we now have four pots, though I think I may have made a mess of dividing the plant. But I'm sure some parts will survive.
|
Youtube: the pain
|
Topic: technology | Link here |
For a couple of reasons today I had to deal with YouTube. Firstly, I wanted a link to the video of the GPS receiver that I mentioned yesterday, and secondly Yvonne wanted me to put some videos up on YouTube.
How do you reference a video on YouTube? I still don't understand completely, but you can embed a video player in a page like this:
It works, and it looks like this:
But the problem is that it's not valid HTML. That looks like another can of worms that I'll follow some other time.
I did that on 24 June 2011, and the image above, while looking identical, is now standards compliant.
That was otherwise fine for Yvonne's videos, but I had the Devil's own job finding the URL of the general demo of the GPS receiver that I was looking for. There was no mouse menu item for the URL, and the source of the page showed nothing that even remotely resembled a URL. Strangely, Jürgen Lock (on IRC) had no difficulty finding the URL of the clip (http://www.youtube.com/v/kfQyjqVKcos?fs=1&hl=en_US), maybe because he had NoScript on. The problem was, that was the wrong video: it's clear that the most important thing about GPS navigators is the reversing camera and not the navigation software, so that's what they show you on the main video. But when it's done, you get a menu, reshuffled every time, showing other videos that are available:
|
|
||||||||
How do you find the URL of the correct one? It shows the start in the tiny windows at top right, but I can't copy the location, and any attempt to scroll makes the current video restart. Again I found nothing in the text of the page. No youtube, no embed. So I fired up wireshark to observe the traffic. And sure enough, it found the URL and displayed it—and refused to let me copy it from the screen. Somehow wireshark has joined the non-UNIX interface crowd, and I couldn't find any way of getting the information out. So saved the file and looked through it with Emacs. More failure: I couldn't find the text that I saw in the wireshark data window. Why not? It wasn't encrypted. I suppose I should have followed up, but by now I was so frustrated that I gave up and typed in the cryptic URL. And with <embed> that gave me the video, nice and in my diary, probably in contravention of many copyright laws. But at least the link worked, and I was able to link to the YouTube site.
What's wrong with this picture?
The markup recommended to display the clips is invalid HTML.
It's almost impossible to find the URLs of the clips. Why? Is this some kind of protection, or just short-sightedness? Since it seems to be trying, I suspect it's a bug.
To find the additional video clips, you need to wait for the first to complete (or slide the position bar to the end, which still requires loading the whole clip).
This doesn't have to be YouTube's fault, but it makes things really painful to work with.
| Monday, 20 June 2011 | Dereel | Images for 20 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Moving the lounge room cupboard
|
Topic: general | Link here |
Over six weeks ago we moved our lounge room to the other end of the house, and then our rearrangements stalled because we had to move the lounge room cupboard, no easy feat. Today CJ and Sue came along, so we set to taking the thing apart and moving it. It was still no easy feat, and it shows how much junk we have:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In the process, found a mouse nest under the base:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I wonder how they even got in there.
Finally it was done, and all we have to do is tidy up the lounge room:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In the process I suppose we should throw out a lot of old junk. I have a serious issue with where to put things.
|
Firefox with tabs
|
Topic: technology, opinion | Link here |
My firefox problems are getting no better. People keep telling me I should be using tabs, not windows (Window manager? What's that?). I have my reasons, some of which are more philosophical than practical (why should a web browser want to be a window manager?), but it seems worthwhile reviewing the situation, especially in view of my suspicion that the crashes are related to the number of windows open. So today I reconfigured things for tabs, and I've been trying to live with it for a while.
The problems with tabs are real: on the one hand, I want to be able to have a page load in the background and work in another. You can do that with tabs, but there's lots of unnecessary mouse-pushing to do in order to be able to see it again. With windows, they're just there. And in one case I wanted to be able to compare the contents of two tabs. You can't do that; There Can Only Be One.
And then there's the identification. I typically have up to 20 firefox windows at any one time. I have my window manager set up to display the icons on the left hand side of the screen, one below the other. firefox displays them across the top, so that even with only 11 tabs I can barely identify the tabs:
|
||||||||||||||||||||||
I suppose my habit of starting my pages with “Greg's foo” doesn't help, but with only a few more tabs it wouldn't make any difference if I were to rename them. It just reinforces the point: a browser is not a window manager.
On the other hand, though: yes, firefox ran all day without crashing. And instead of growing to 1 GB in size, it settled down round 600 MB. So it does look like there's a problem with many windows. But image rendition is just as glacially slow with tabs.
| Tuesday, 21 June 2011 | Dereel | Images for 21 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Winter storms in
|
Topic: gardening | Link here |
Lots of rain and wind last night, and on the news we heard that there had been lots of fallen trees around Melbourne. Here we had one fallen Acacia branch, which conveniently fell on another much smaller acacia:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It doesn't look as if the little one has been harmed.
|
Upgrading the friends' computer infrastructure
|
Topic: gardening, technology | Link here |
Into town today to meet with some of the committee members of the Friends of the Ballarat Botanical Gardens. They've discovered my computer background and have asked to take their computer infrastructure in charge. That looks like it'll be fun, and hopefully not too much work.
|
GPS navigator: disaster
|
Topic: general, technology | Link here |
While in town, picked up my new GPS navigator. Compared to the old one it's enormous:
|
|||||||||||||||||||||||
Headed back home via the back roads, and the navigator happily told me to leave the main road and take a detour through a paddock, something that not even the old one did. Like the old navigator sometimes did, it didn't show a charge light much of the time—I wonder if there's a software problem there. The software itself looked almost identical to the software on the old navigator, and I had no difficulty using it. The larger format makes it much more readable. But there's much more detail on the screen—it's not just a scaled-up version of the old one.
Back home, backed up the memory card, not helped by Apple. I could access it with the reader, and it happily auto-mounted it, but it wouldn't let me copy the card itself while it was mounted. And when I umounted it, it destroyed the device node! So there's no way, it seems, to copy a raw device with Apple. Used tar instead and found, to my amazement, that the “latest 2011 maps” were anything but:
That is all the more unusual because it's older than the map data in the old navigator, which I bought well over a year ago:
The New Zealand maps are nearly 2 years out of date. There was also no word of the free update service they offered. That wasn't the only problem. The application itself was older. Here the “new” unit:
=== grog@dereel (/dev/pts/7) /src/GPS/ScanFast-7020/iGo8 7 -> ls -l iGO8.exe
=== grog@dereel (/dev/pts/7) /src/GPS/ScanFast-7020/iGo8 8 -> strings iGO8.exe | grep Build
And, by comparison, the “old” one:
=== grog@dereel (/dev/pts/8) /src/GPS/ResidentFlash/NNG 6 -> l NNGNAVI.EXE
=== grog@dereel (/dev/pts/8) /src/GPS/ResidentFlash/NNG 7 -> strings NNGNAVI.EXE | grep Build
This also means that there's no text-to-voice, despite the claims. If that wasn't enough, parts of it are in Chinese:
|
|||||||||||||||||||||||
|
|
|||||||||
“Wince” indeed! The real show-stopper, though, was that I couldn't charge the damn thing. I tried the car charger, the supplied AC adapter, a couple of USB cables, but the thing doesn't charge. I checked for the promised manual in PDF format, but it wasn't there. The battery is now nearly discharged, and thus as useless as the old one it was supposed to replace. My good-will towards the seller has completely evaporated. Sent an exasperated message to the seller asking them to fix things. We'll see what happens.
In the meantime, did even more searching for GPS navigators. It's beginning to look like many of them use the Nav N Go iGO 8—certainly this one does:
|
|
||||||||
|
|
||||||||
I wonder how they crashed it. It shows an even older (April 2009) version, though the information suggests that the photo itself is nearly 12 months old, so it doesn't necessarily mean that that's the version they're currently shipping. To be investigated.
|
Aligning camera for photos of flat objects
|
Topic: photography | Link here |
Taking the photos of the navigator had an interesting side effect. Clearly I wanted to get the device centered on and perpendicular to the lens axis. In the past I've done that mainly by eye and the viewfinder. But the screen of the navigator was reflective, and I saw the reflection of the lens in the viewfinder. Bingo! Get that exactly in the centre, and things are aligned correctly. It's trivial to use a small mirror for other objects, such as Yvonne's paintings.
|
Rearranging furniture: the background
|
Topic: general | Link here |
Callum Gibson looked with interest at the photos of the furniture rearrangement, but it seems I didn't include enough context for him to understand where things had moved.
To recapitulate from when we moved in to now: there were two rooms that we could use as a lounge room, and we've swapped them: the second from the top on the left in this plan (“room 3”), and the second from the bottom on the right (“room 10”):
|
||||||||||||||||||||||
After moving in to the house, rooms 3 and 10 respectively looked like they did in the left column. Now they look like on the right:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Wednesday, 22 June 2011 | Dereel | Images for 22 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Navigator problems: cause identified?
|
Topic: technology | Link here |
Somehow it seems funny that the problems I'm having with both GPS navigators are so similar. In each case, the only problem is charging them. And in each case they charge via the micro-USB port. Went to take a look at the port itself. The results were informative but somewhat disappointing. Here's the socket on the new navigator:
|
|||||||||||||||||||||||
One pin missing, one bent. I wasn't sure how many pins were supposed to be there—it's at least hypothetically possible that pin 5 isn't always needed, so I bent the bent pin back to shape and tried again, still with no success. Later I checked: the outside pins are ground and +5V, so they're really needed.
But how did that happen? Took a look at my old navigator and the cable I used to charge it, which was even more upsetting:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But, again, how can that happen? It's not as if I am particularly rough with my connectors, and I've never seen this kind of damage before. My best bet is that somewhere some component must have been damaged before I got it. No idea any more which, but the cable looks funny.
|
More reception problems
|
Topic: multimedia | Link here |
TV reception has been good for a while, but lately it's got bad again. I can't see any obvious causes. A little rain, maybe, and two recordings at the same time were both affected, and later the reception was good again. My guess is that setting the “fine-tune” flag has reduced or eliminated one of multiple problems, but not all. It's interesting to note that radio reception at the moment is terrible; I wonder if the TV problems are related.
| Thursday, 23 June 2011 | Dereel → Melbourne → Dereel | Images for 23 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Off to the Big Smoke
|
Topic: general | Link here |
Off to Melbourne today to do some shopping and pick up a new GPS navigator. It's only been a year since I got the old one, but it's clear I've come to depend on it. Despite its lack, found my way to the Queen Victoria Market without too much difficulty, helped by the surprisingly low level of traffic, and got in and out after spending only 55 minutes and $300 odd.
|
New navigator
|
Topic: technology | Link here |
Then down to Clayton to pick up the new navigator, not helped by Google Maps' complete lack of understanding of the street numbers in the Prince's Highway, just one of Australia's main highways; it put the address in the middle of a main highway junction.
This is no longer the case. The link above shows the correct location, a reason to use screen shots rather than links.
It's interesting to see the other side of an eBay operation: a couple of young blokes in a room in an industrial area. Changed the unit pretty quickly, and confirmed, yes, the power connector is OK. The Microsoft “Windows CE” is also the English version, and one of the blokes (Maurice; the other is Ken) showed me how to get the spoken street names. It seems that it's a property of the individual voice profile, and the ones with Text to Voice have the text “(TTV)” in the menu entry.
They still didn't have any more modern maps, and it seems they didn't know there are any. They're not expecting new maps for another 8 months. Still, they offered the latest ones, and that's part of the contract, so I left with a warning that my feedback would depend on them finding the latest (or at least recent) maps.
And now that I can use it? The spoken street names are there, but they're clearly at least partially synthesized, and sometimes barely intelligible. “Bannockburn” (town) was correct, but “Ballarat“ was incorrectly intoned and barely intelligible, as was “Grattan (St)”.
And the comparison with the old navigator? There were more differences than I expected. I had already noted that when I left a suggested route, the old navigator would prefer to find a new route in the direction I was going, even if it was much longer, rather than get me to turn around. The new navigator tells me to turn around, repeatedly on the way home (it thought I should go via Ballarat, but via Geelong is much faster). It also seems to react more slowly than the old one, but maybe that's the voice profile. On the whole, the only real advantage is the increased size, which makes things a lot easier to see.
Back home, backed up the micro SD card, and confirmed that yes, indeed, the maps and software are just as old as on the other unit. That remains to be solved.
|
More shopping
|
Topic: general, food and drink | Link here |
After picking up the navigator, on to Casa Iberica in Fitzroy, looking for Masa harina. Got that, of course, but not the kind I was looking for, Maseca: they're out of stock. Instead got two different kinds of masa from Minsa, one of them apparently blue—I've never even heard of that before. I hope the quality is better than their web site.
Then looking for lunch. There's a Mexican restaurant across the road, so off to repark the car nearby, taking an inordinate amount of time to do so, and then discovered the place was closed. There were plenty of other restaurants in the area, but most of them were closed too, and the others didn't look particularly attractive. Off, via Lygon St (too expensive) and Grattan St, and finally ended up near the Queen Victoria Market again, where we know a Korean restaurant. And it was closed too! Things seemed very quiet in town today, but it's still surprising. Finally to a Chinese restaurant nearby, “Hui's kitchen” or some such, where we got some less-than-excellent Calamari (yup, that's what they call them here), reminding me of the German expression “Außen hui, innen pfui”. People say that it's easy to find good eating in Melbourne, but my experience is different.
|
Sardines and other fish
|
Topic: food and drink | Link here |
Back home, Yvonne had the pleasant task of gutting the 2 kg of sardines that we bought at the market. Or at least, we thought they were sardines. There are clearly two different kinds of fish, shown here alternately. We're guessing that the longer, thinner ones are the sardines. The others have an interesting stepped pattern on the side:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I wonder if they accidentally got mixed, or whether they were caught at the same time.
| Friday, 24 June 2011 | Dereel | Images for 24 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Getting the Youtube videos right
|
Topic: technology | Link here |
Last week I finally got Youtube videos to display in this diary, using the <embed> tag. It was very clear that that's not the way to do it: <embed> is not standards compliant. The W3 validator gave explanations and pointed me to a page that claimed to make “flash satay”, though I don't understand the analogy. It's a mess, as the author confirms, made necessary (it seems) by Microsoft “Internet Explorer”. I've simplified things, which, it seems, means that people using “Internet Explorer” will have to wait until the whole file has been loaded before it starts playing. See if I care. Here's the new code for the example I showed last week:
Alas, this too didn't cut it. I finally found a better solution on the YouTube site on 28 June 2011.
|
External viewfinder monitor
|
Topic: photography, technology | Link here |
Earlier in the week I read an article describing a Sony monitor for displaying the live viewfinder or monitor images of a DSLR. In itself, not very interesting (and the links are in German), but what got me was the size and resolution of the device: 5", 800×480 pixels. That's the resolution of my new GPS navigator, which is 2" larger, and it has an aspect ratio of 15:9 (or 5:3), which no video or still camera has. Clearly the thing wasn't originally designed for cameras, which have aspect ratios like 3:2 (most SLRs and 35 mm cameras), 4:3 (compact, Olympus), or 16:9 (wide screen video). So I assume it's an adaptation of a GPS navigator display. And maybe I could use my navigator (which has an AV input) for the same thing.
Spent a little time in the afternoon confirming that yes, the camera can display on a TV. But I need an adapter cable for the navigator, which has a 4 connector 2.5 mm jack. People had suggested they could be expensive. Went looking on eBay and found the cheapest thing I've ever found: $1.50, including postage from Hong Kong! How can they possibly even break even with that kind of price? I suppose it'll take forever to come, but at that price I can wait.
Things aren't always like that. The Sony monitor costs € 400. My navigator cost the equivalent of € 110, so I can save € 290, gain 2 inches, a GPS navigator and a reversing camera. About the only thing I need to do is to find a mount for the thing. Despite my issues with the navigator, that can't be bad.
|
The end of the world, part 4711
|
Topic: technology, opinion | Link here |
Chris Yeardley sent me a URL for a description of why the world will end at the end of next year, being swallowed up by a black hole. I don't believe a word of it, of course, but there was nothing (apart from typical lack of substantiation) that I could point to to disprove it. It's not until you go to another page on the site that you realize that yes, indeed, the information on the site is correct. It'll be interesting to see what effect this has.
| Saturday, 25 June 2011 | Dereel → Melbourne → Dereel | Images for 25 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Off to the big smoke again
|
Topic: general, opinion, brewing | Link here |
Peter Dilley over far too early this morning for our planned trip to Melbourne—again!—to see a Brew in a Bag demonstration—again!. This was also only the second time to try out my new GPS navigator, and Peter was second-guessing with his BlackBerry.
That was interesting: he had real-time information about speed cameras. My navigator found no speed cameras between Bannockburn and Melbourne; the BlackBerry had up-to-date information and found not just dozens of speed cameras (which we confirmed visually) but also police car wait locations. We didn't see any police cars, but the places were paved, so it seems plausible. Clearly a great advantage of a device with network coverage. But also a disadvantage of a system with a closed data format and infrequent updates.
Other issues with my navigator also became apparent. I now understand why they have both Text-to-voice (TTV) and non-TTV voice profiles: the TTV texts are glacially slow—pauses in the middle of an announcement—and they slow the whole application down to a point where I missed a couple of turnoffs. On one occasion I got the message “You are over the speed limit” while stopped at a traffic light. Set a non-TTV profile and continued, and didn't have any further problems, but by then we were out of the CBD, so maybe that's just a coincidence.
|
Brew in a bag revisited
|
Topic: brewing, opinion | Link here |
The brew demonstration was at Grain and Grape done by Dan Walker and Geoff Hammond (Geoff is the one with the beard). They addressed a number of issues that I had had two weeks ago. In particular, they did a pretty thin mash, and they added the grist slowly, stirring all the time:
|
|||||||||||||||||||||||
That's easy if there are two of you, of course.
The bag was also much wider than Peter's, and Geoff stirred almost constantly, thus addressing my main concern about uniform mash temperature:
|
|||||||||||||||||||||||
The other issues: it seems that there isn't a big difference in brewhouse efficiency, since the mash is thin, and they aerate in the fermenter after cooling (in fact, exactly what I do, so this isn't an objection after all).
And multi-step infusions? Dan didn't want to talk about that, and Geoff didn't want to talk about wheat beer (“Life's too short”). But I suppose the way they do it, with plenty of space to stir, it's no different from a normal mash. I should try it out some time.
|
Shopping and lunch
|
Topic: general, food and drink | Link here |
Peter took me to a couple of food shops that he knew when he lived in Melbourne, about 4 years ago. Both are in Victoria St Brunswick, not far from the Mediterranean Wholesalers where we sometimes buy Italian food. The first is Basfoods Direct, at 419 Victoria St, and the other is NSM Wholesalers at 381 Victoria St.
It seems that both have changed significantly since Peter was last there. At Basfoods found some cheaper olive oil than at Casa Iberica, and also some cheap Turkish Sambal Ulek, but there was nothing much at NSM. Still, a good place to know.
Peter then wanted to take me to one of the many Middle Eastern restaurants down Sydney Road, but they were gone. So was the restaurant where Yvonne and I ate a few years back. Took a look in at Mediterranean Wholesalers, but didn't find anything of interest, and then back to the car to investigate what my navigator thought of restaurants in the area. It found a few in Lygon St Brunswick (the better-known part of Lygon St is in Carlton), so down there. Most were shut, but finally we went into a Vietnamese place called Anh Minh at 44 Lygon St, which wasn't in my navigator. Had a reasonable Phở, though Peter—who seems to have some Vietnamese connections—wasn't impressed. He promised to tell me if he ever finds a good Vietnamese restaurant in Australia.
But what's wrong with this story? This is the second time in 3 days that we've had real difficulty finding food in a town that prides itself on good food and multi-ethnic culture. There are lots of Vietnamese people in Melbourne. They must eat in restaurants somewhere.
|
Cheap olive oil
|
Topic: food and drink | Link here |
On Thursday I was toying with buying some olive oil at Casa Iberica, but didn't because it looked too expensive ($30 for 4 l of “Extra virgin” cold-pressed oil from Carbonell). The Greek oil I bought at Basfoods cost only $23, again for 4 l. But on the way home we stopped in at Safeway for Peter to buy some stuff, and I found the Carbonell oil on special for $18—normally it costs $20. Pricing is strange. Clearly specialty shops have their price.
| Sunday, 26 June 2011 | Dereel | Images for 26 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The next firefox crash
|
Topic: technology, opinion | Link here |
Since switching from multiple windows to mainly tabs, my firefox processes have become much more stable. It ran from Wednesday until today, and I noted that the memory use was much less, only about 700 MB instead of 1 GB. But today it reached 1 GB and crashed again, though in a different place from usual. So it's still not clear whether the improvement was due to lack of race conditions in window handling, or some issue with the memory size itself.
And how do I like tabs? I still hate them. Yet another instance of the “there can only be one” mentality. There are so many times when I want to look at more than one window at a time, and tabs don't allow that. Sometimes I want to follow a number of links from a single page, which requires that the source page remain where it is. With tabs you have to go back to the page after every tab is created. Yes, you can open a separate window in those circumstances, but that's just damage control. There's no reason why firefox should pretend to be a window manager. And it doesn't do a very good job of it. My fears proved unfounded that the entries on the tab bar would become completely illegible; instead they reach a minimum size and then scroll, so you can't even see all of the tabs.
|
||||||||||||||||||||||
There's a way round that: the little inverted triangle at the right of the bar produces a drop-down menu with useful content:
|
|
||||||||
But to use that you need more mouse motion. I suppose a permanent window like that would not be a bad idea. But that would violate the principle of “there can only be one”.
|
1000 days and counting
|
Topic: technology | Link here |
One of my external servers (not http://www.lemis.com/) has now been up for 1000 days, the longest I've ever had a machine up and running:
That's a virtual machine, what's more, and both host and guest are FreeBSD. I worked with and for Tandem computers for 15 years, but I've never seen an up time like that. And http://www.lemis.com/, a shared machine, doesn't come close, though today is a bad time to compare; typically it's up for months at a time:
|
Garden flowers
|
Topic: gardening | Link here |
This time last year I started my monthly series of photos of flowers in the garden. The purpose wasn't so much to get photos of nice flowers, though I tried, but of what was in the garden, which can also mean some pretty sad-looking specimens.
This month the sun was shining for once, and things like the Xerochrysums and Gazanias were flowering:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
A Canna and a ginger were also trying, though I don't think the Canna will make it:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Hebes are also blooming slowly but continuously. I wonder if they need more fertilizer. I've noted that the ones on the east side of the garden are significantly smaller than the ones in the north (where I took these photos), and they get less fertilizer.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I've now hung the petunias in the greenhouse, severely limiting headroom, but they seem to be enjoying it. There's plenty of new growth, and of course they're still flowering:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
So is the red Mandevilla, which has been there for some time. I expect it will continue to flower through the winter:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Comparing with last year, the roses aren't looking as happy. Only the pink rose in the north bed has buds, with the exception of a solitary bud on the Iceberg rose on the south side of the verandah, which is otherwise looking quite unhappy.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Some flowers we didn't have last year, like the Tradescantia and the Oenothera glazioviana, and this Iris-like plant that I still haven't identified:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And despite my expectations, the spring-flowering bulbs are not as far advanced as last year. About the only one I found is a small Narcissus, while last year yellow daffodils were on their way:
|
|||||||||||||||||||||||
Last year we also had Arum lilies and Senna aciphylla, both of which aren't showing any inclination to flower yet.
|
More playing with GPS navigation systems
|
Topic: technology, general | Link here |
Spent some time playing around with files on my GPS navigator. Added the newer maps and some voice files from my old navigator, and to my surprise they worked. The maps I bought last September work in the new navigator, but they didn't work in the old one, and I had suspected that there was some incompatibility.
Just how slow is the navigator? Took a look at the test of motorbike navigators in c't, where they tested the time to calculate two journeys: one from Hamburg to Königsdorf, and the other from Tromsø to Cadiz. Now that I have European maps in the system, I could compare it.
My navigator took 42 seconds to calculate the route from Hamburg to Königsdorf, slower than most of the devices (between 13 and 36 seconds, only the Tom Tom UrbanRider took 56 seconds). The distance (“fast” route) was 845 km, the time 7:53 (incidentally an average speed of 107 km/h, which would be unsafe and therefore illegal in Australia). That compared with distances between 822 and 845 km from the test candidates and times of between 6:58 (121 km/h!) and 7:49 for the others. Google Maps comes up with 844 km and 7:37.
The test results in the article for the route from Tromsø to Cadiz were interesting: neither of the Garmin units nor the Tom Tom unit were able to calculate it, though the Garmin zūmo® 220 took 5 minutes before it failed. The others took between 17 and 92 seconds to calculate distances between 5,388 and 5,492 km and times between 52:29 and 70:40. Google maps calculated 5,309 km and 58 hours. Mine took 110 seconds and calculated a distance of 5,322 km and a time of 58:01. So yes, it's relatively slow, but it works. I'm surprised that the big names failed. It's also interesting to note that the fastest GPS in the test was from ALDI.
|
Spam from Oracle
|
Topic: technology, opinion | Link here |
Spam is getting bad enough as it is without big companies getting in on the act. I recently received the following message:
My first impression was that this was somebody posing as Oracle, but it seems that the domain oracleeblast.com is really registered by Oracle. Sent a message to the abuse people asking why they're stooping to spam, but of course got no reply. The name “oracleeblast” sounds like it's intended for spamming. I used to have a positive opinion of Oracle, but their behaviour in the last few years has destroyed any positive thoughts. I wish they'd go away.
|
The dragon kills
|
Topic: gardening | Link here |
Nele Koemle came by today to look at some horses. Back in the summer I had promised her a sucker from our Paulownia kamikawii (“Sapphire dragon”), but it had to wait until winter and it had lost its leaves. That's the state now, so out to dig it out.
I barely got it out. The entire surface of the soil is a maze of twisty little roots, all different. No wonder the Gazanias in the vicinity died.
| Monday, 27 June 2011 | Dereel | Images for 27 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More GPS navigator games
|
Topic: technology | Link here |
More playing around with my GPS navigator today. One bug relates to finding petrol stations:
|
|||||||||||||||||||||||
|
|
|||||||||
And yes, the message is correct: the file doesn't exist, and the navigation application stops and you're back in the start menu. I thought I could fix that by using the POI (Point of Interest) file from my old navigator, which worked correctly. Copied it across, but it didn't make any difference. What's causing that?
Also did a bit of playing around with the platform software, Microsoft “Windows” CE. Despite my dislike of Microsoft, it's interesting to have a relatively familiar interface to the box. In the process, found that I could set it to present itself as a mass storage device on the USB interface; that's a great advantage over the old one, which required Microsoft ActiveSync 4.5 and thus a Microsoft machine to access it. With the new one, it looks like this:
That's familiar; the old GPS navigator also had the PROTOTYPE identifications. So now I can access the thing directly from my own machine, but only the micro SD card. That's a relief, though: the thing is pretty flimsy, and I don't like moving it around. Now I can run rsync against it, though the first time presented a surprise: it copied everything, even though most of the files were there already. That proved to be due to Microsoft's lack of understanding of time zones. As a result, the timestamps on my computer and the GPS file system were offset by 10 hours. For example:
=== grog@dereel (/dev/pts/20) /src/GPS/ScanFast-7020-current 41 -> l -T Ebook/ /GPS/Ebook/
Still, that will only happen once.
|
Winter flowers: straggler
|
Topic: gardening | Link here |
Walking through the garden today, found a flower I hadn't expected:
|
|||||||||||||||||||||||
That's one of the Gazanias that we picked up in Stawell last September. It has flowered before, but always looked a little unhappy.
| Tuesday, 28 June 2011 | Dereel | Images for 28 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Blue tortillas
|
Topic: food and drink | Link here |
Huevos rancheros for breakfast today, and tried out the new blue masa. It really looks strange:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And how does it taste? Hard to say. Normally I use 25 g masa per tortilla, but on the packet I saw the nutritional information assumed 6 tortillas per 100 g, so I made 3 instead of 2 out of my 50 g. That worked—at least this masa, from Minsa, holds together much better than the previous Maseca—but of course the tortillas were very thin. I think I'll stick to my 25 g next time. From a flavour point of view, it didn't taste as strong, but otherwise OK. I'll compare it with the yellow masa next time.
|
Embedding YouTube videos, continued
|
Topic: technology, opinion | Link here |
After some discussion with Callum Gibson, found that YouTube is not consistent in its suggested markup. When I uploaded this clip, I got a URL http://www.youtube.com/watch?v=cajZMd9HkVM and this code to embed it (after reformatting):
The problem here is that it contains the dreaded <embed> tag, which isn't standards conformant. Callum told me this wasn't so. It turns out that we're both right; it just depends on how you ask. After uploading, and once you've fought your way through a maze of twisty little menus, all counterintuitive, you can find the opportunity to “share” the video. First you need to start watching it (thus downloading what you've just uploaded, without even wanting to watch it). Then select Share. It gives you an Embed option with code like this:
This is one and the same video, just shown in two different places, and with two different source URLs. But the second one is standards-conformant, and as the markup suggests, it allows you to set full screen mode, something that my previous code didn't. Wrote a PHP function that takes a YouTube ID and creates the markup. So <?php youtube ("cajZMd9HkVM", 560, 349); ?> gives:
But why should this be so difficult? YouTube is an enormous operation. You'd think they would get their basic menu structure and markup right and consistent.
|
Getting back into the garden
|
Topic: gardening | Link here |
I've been neglecting the garden lately, and it shows. Once again the weeds are on the increase, notably on the paths, where I'm going to have to spray them. But it's difficult at this time of the year. Did some weeding, but it's discouraging work, and didn't do much.
| Wednesday, 29 June 2011 | Dereel | Images for 29 June 2011 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
YouTube markup: Still more issues
|
Topic: technology | Link here |
So yesterday I finally thought I had got my markup for YouTube videos right. But the W3 validator disagreed with me:
Line 5187, Column 31: "allowfullscreen" is not a member of a group specified for any attributeallowfullscreen>
So I removed that, and the page validated. And, surprise! I still had the full screen functionality. So that invalid keyword is also unnecessary. And both YouTube markup suggestions are invalid.
That wasn't all, though. Callum Gibson read the description, of course, and told me that my second example was wrong. After a lot of discussion, it proved that what he saw in his newsfox display wasn't what I wrote. Specifically, I had this text, using iframe:
But what newsfox displayed was this. iframe has become embed:
I never wrote that. Something changed the entire markup, even the example, which really looks like this:
It's not in the RSS feed; other clients of the feed show it correctly. After a considerable amount of thought, it seems that it's newsfox itself that is changing the markup, in the process making it invalid.
|
More dawdling in the garden
|
Topic: gardening | Link here |
A very mild frost this morning—I saw a couple of dead leaves with ice crystals on them, but nothing on the live plants or in the bird bath.
Somehow my garden work has dragged to a halt. It's not the weather this time: it's currently not too bad. I really should take advantage of the few times when I really feel motivated to work in the garden.
|
ANZ: Use credit card even for cash payments
|
Topic: general, opinion | Link here |
A couple of days ago I received a debit card from ANZ bank to replace the ATM card that I have used for years. It includes VISA access, but as a debit card the sum is debited immediately, like normal ATM cards. The accompanying brochure told me to always select “Credit” when paying unless I wanted cash paid out.
What's wrong with that? Many vendors charge extra for credit. And I don't want cash out. Called up ALDI, who confirmed that they would charge extra even for debit cards. Called up ANZ and spoke to Josh, who told me it would not cost anything more. When I told him about ALDI, he finally told me that yes, I could still pay with “Savings”.
Asked to speak to his supervisor—why do I have to do that so often?—and spoke to David, who explained why the documentation told me to use “Credit”: Because of the “Falcon”, a brand name which he pronounces the American way, /ˈfælkən/, and not /ˈfɔːlkən/. It seems that this bird gives me additional protection against fraud. I told him I hoped it would work better than “Verified by Visa”. He laughed and said yes, much better. So it seems that even in ANZ they know and accept how bad VBV is.
But what's the point? I have a credit card. The /ˈfælkən/ protects me there too. And I don't have to pay immediately. And why is the documentation so bad? On reading the details on the web, it seems that I now have to inform ANZ when I am travelling overseas, or my credit card may be blocked. Nothing about that in the documentation they sent me. Just at the very end they say “If you're travelling overseas, it's a good idea to let us know beforehand”, on a different number, and without warning of possible consequences. I'm coming closer and closer to changing my bank.
| Thursday, 30 June 2011 | Dereel | Images for 30 June 2011 |
| Top of page | ||
| previous day |
|
Heavy work in the garden
|
Topic: gardening | Link here |
CJ showed up this morning with his tractor to address some of the heavier jobs we have to do in the garden: remove the root system of the Aloysia citrodora (lemon verbena) that we've been planning to remove for over two months, dig the planting holes through the gravel for the Agapanthus that I had also tried to plant round the same time, and cut down the remaining trees in the Cathedral.
That was all fun, but we got it done. The Aloysia root was even bigger than we thought:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Digging the holes for the Agapanthus was mainly a question of manœvering the tractor into position; holes that took me half an hour took the tractor about 10 seconds:
|
|||||||||||||||||||||||
The big Cathedral tree was the biggest problem, because of its height (about 10 m, I suppose). I raised CJ up onto a platform on the fork lift of the tractor, and he was able to get off a number of the higher branches:
|
|||||||||||||||||||||
|
|||||||||||||||||||||
|
|||||||||||||||||||||
But by evening we still had the one on top that we don't know how to get down. Here before and after:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Winter flowers: more stragglers
|
Topic: gardening | Link here |
While in the garden, found some more flowers that have bloomed since Sunday, the first Arum, the first Camellia and the first Acacia baileyana:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Things have still not progressed as far as this time last year.
Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.
| Top of page | Previous month | Greg's home page | Today's diary entry | Next month | Greg's photos | Copyright information |
