|
|
|
Friday, 1 August 2025 | Dereel | Images for 1 August 2025 |
Top of page | ||
next day | ||
last day |
Jane Ashhurst arrives
|
Topic: general, animals, food and drink | Link here |
Jane Ashhurst arrived today for a weekend. Yvonne had her first chance to drive her new car alone when she went to pick her up. Bruno was delighted:
|
Jane brought an interesting paste with her:
It's apparently Neri ume, a sauce so obscure that Wikipedia doesn't know it. Google Gemini tells me
Neri plum sauce, or Neri Ume, is a paste made from pickled Japanese plums (umeboshi). It's known for its distinct sour, salty, and sometimes subtly sweet flavor. It's incredibly versatile and can add a fantastic tangy kick to many dishes.
We had sushi for dinner tonight. Gemini had continued with
Sushi Rolls: Use Neri Ume as a filling or accent in maki sushi rolls.
So I tried some with that as a dipping sauce. Probably not the way to go; I'll investigate the alternatives. Jane uses it for a quick rice dish along with bonito flakes.
Next cardiac choice
|
Topic: health, opinion | Link here |
Call from Professor Peter Kistler today to discuss whether Yvonne should get a stent. Spent some time getting a laptop running Microsoft for a Zoom call, but this time he just called on the “phone” (the modern word for “mobile phone”, which at least omits the pain).
We thought it was about stents, but he thought it was about her medication and prescribed Sotalol instead. Somehow he hadn't been informed of the chest pain that Yvonne had had a few weeks back. He dismissed the idea that the pain was due to cardiac issue: she had had an MRI last year. That doesn't tally with my records, but it suggests that things can't get worse. But what could be done is to have a pacemaker installed, which would obviate the need for medication, something that Yvonne would like very much. And Dr Rodney Reddy can do the implant.
Downsides of a pacemaker? None. But he (Kistler) needs to do a minor ablation a month after implant to turn the thing on: if it stops, so does the heart. That sounds like a downside to me. I get the feeling that it's no longer his area, and that another talk with Reddy would be a good idea.
Nikkor 105/2.5
|
Topic: photography, opinion | Link here |
Saw a link to a video on the Nikon Nikkor 105 mm f/2.5 lens. I have one of them, one that I got with my Nikon FM2 five years ago. I was not overly interested in that lens—I found the 50 mm f/1.4 and 85 mm f/1.4 more interesting. But it seems that the 105/2.5 is also a classic lens, and the article came up with a lot more information, including pricing that I hadn't expected: he showed this particular model (AI-s aperture coupling) on sale in Japan for 48,800 ¥, currently rather over AUD 500, the upper limit of what I established 5 years ago. And that lens appeared not to have been in as good a condition as mine.
Then there is another article on the lens on the Nikon web site, which I must read. Other links are here and here.
The Nikon page also has articles on other lenses, including the 85/1.4. A lot more investigation to do if I ever find the time.
Saturday, 2 August 2025 | Dereel | Images for 2 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Signs of the times
|
Topic: history | Link here |
Seen in my calendar today:
3/4 life: Wednesday, 17 May 2006
May 17 Greg last flew in an aeroplane, 2006
At the time I didn't realize what a turning point that was.
Ossobuco
|
Topic: food and drink, general | Link here |
Margaret Swan has left us a truffle. What do we do with it? It's an obvious accompaniment for risotto alla milanese. And that's what you eat with ossobuco. So ossobuco it was today:
|
|
|
|
Google Translate fail
|
Topic: food and drink, language, technology, opinion | Link here |
It's been a while since I made ossobuco, so I checked on the web, including the the Italian Wikipedia site. Seemed straightforward enough, but to be sure I understood everything I put it through Google Translate. After that I understood nothing. The best mistranslation was (original on left, translation on right):
Ingredienti principali | Main Ingredients | |
Carne bovina | Beef | |
burro | donkey | |
“Carne bovina” means “bovine meat”, and the context makes it clear that it should be veal. But that's nothing compared to translating “burro” as “donkey”. « Burro » means donkey in Spanish, but not in Italian.
More playing around (I think a couple of backward and forward translations) led me to even stranger things:
The ''''''' ( (the (the (the ( : in Milanese. other handwriting; pronunciation: ) is a typical dish of obtained from the homonymous (the (the (the ).
It's been a long time since I've seen such a mess from Google Translate.
Opening cans
|
Topic: food and drink, opinion | Link here |
Once upon a time everybody had a can opener, but nowadays most cans have a ring pull. The tomatoes I used for the Ossobuco didn't, though. Nearly 10 years ago I bought a rather strange opener which removes the entire top, but somehow I don't like it. So today I found another conventional opener and used that. I wish I hadn't:
|
|
|
Of course, I should have known better. I had exactly the same pain five years ago with the same opener. It's gone now. I still have an electric one that I should try again.
Yet another Nikkor 105/2.5 article
|
Topic: photography, opinion | Link here |
Just before going to bed, found another article about the Nikon Nikkor 105 mm f/2.5. He wants it to be reintroduced. Why? Yes, it seems once to have been a particularly good lens, but that was half a century ago. Only five elements! Now even prime lenses have up to four times as many elements, and certainly for good reasons.
Sunday, 3 August 2025 | Dereel | Images for 3 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Too much cooking
|
Topic: food and drink, health, opinion | Link here |
Somehow yesterday's cooking exhausted me. I had planned huevos rancheros for breakfast and beef with orange and broccoli for dinner, but one after the other I found it too difficult. Instead I thawed out some lamb pilau biriani for the evening meal.
Problem: there were only 2.1 portions, according to my notes. OK, make some alu masala. That's not difficult. But then Jane complained of a cold, and though she did eat something, we didn't finish the biriani, and I didn't even finish cooking the alu masala. Not a very satisfactory day.
Monday, 4 August 2025 | Dereel | Images for 4 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Goodbye, Jane
|
Topic: health, general | Link here |
Jane was feeling no better this morning, and she had decided to return home today. She gave me just enough time to warn Yvonne that she needed to be at the railway station at 9:30. And almost before we knew it, she was gone.
More breakfast noodle experiments
|
Topic: food and drink | Link here |
Somehow I'm getting bored with my breakfast recipes. Today I had a number of leftovers to process, something that I normally do with nasi goreng, but I didn't feel like that today. So I made something like my fake pad Thai, just with a different ALDI spice paste and with some Beijing noodles::
|
|
|
The result was different, but not bad. I think it's time to start considering my noodle dishes as variable, like I have been doing for some time with nasi goreng.
lagune, finally
|
Topic: technology, general, opinion | Link here |
Without Jane, I finally had enough time to look at updating lagoon. It has taken me less than 6 weeks. Build a new world, update ports to the latest and greatest, fight the cables holding the thing in place, into Yvonne's office, fight the cables holding (the old) lagoon in place, install a 16 port switch because it's all I had apart from an old 10 Mb/s hub, and finally it was in place:
|
Things weren't perfect. Even before installing the new machine, I ran into cable problems with the old lagoon: the display no longer connected it. But probably we won't need that.
For the fun of it, I suspended the system with zzz(8) before moving it and watched what happened when I powered it on again. As I feared, it booted normally.
OK, fire it up. Why do the NFS file systems not get mounted?
Mount the file systems, start X. A good semblance of Yvonne's environment. Just the mouse didn't have middle button emulation, something that I need to automate:
# Find mouse index. This assumes that the last one in the list is the
# one we're looking for.
MOUSE=`xinput|grep -i mouse|tail -1|sed 's:.*id=::; s: .*::'`
# Look for the index of the enabler.
INDEX=`xinput list-props $MOUSE|grep 'Middle Emulation Enabled ('|sed 's:.*(::; s:).*::'`
# And enable
xinput set-prop $MOUSE $INDEX 1
Then the sizes and fonts of the xterms were wrong. That's a separate issue that I haven't got round to fixing for over a year now, since the transition to fvwm3. No /var/mail: that's another symlink to add.
Photo processing? We use distress (a Microsoft box) to run DxO PhotoLab and other stuff. But I can't wake it. wake(8) needs to be setuid. And then it came up with my credentials instead of Yvonne's, another configuration issue to look at.
By contrast, sound Just Worked: plug the plug into the output socket, and all is well. I hadn't expected that.
The big issue, of course, was with web-oriented stuff. Both firefox and Chromium refused to start. A new device! Chromium was the most helpful:
=== yvonne@lagune (/dev/pts/1) ~ 7 -> chrome
[1674:69546526969872:0804/130003.158863:ERROR:chrome/browser/process_singleton_posix.cc:358] The profile appears to be in use by another Chromium process (74466) on another computer (lagoon.lemis.com). Chromium has locked the profile so that it doesn't get corrupted. If you are sure that no other processes are using this profile, you can unlock the profile and relaunch Chromium.
[1674:69546526969872:0804/130003.158884:ERROR:chrome/browser/ui/views/message_box_dialog.cc:198] Unable to show message box: Chromium - The profile appears to be in use by another Chromium process (74466) on another computer (lagoon.lemis.com). Chromium has locked the profile so that it doesn't get corrupted. If you are sure that no other processes are using this profile, you can unlock the profile and relaunch Chromium.
OK, fool, how do I unlock the profile? I've seen this before, but this time I asked Google Gemini. Simple:
=== yvonne@lagune (/dev/pts/1) ~ 7 -> rm .config/chromium/Singleton*
But why can't the browser offer that function, or at least the information?
And then there was WhatsApp, which wanted her to involve her mobile phone and potentially its horrible glass keyboard. After about 10 minutes I was able to work out how to do it without the keyboard.
And mail? I installed Postfix, but not its configuration. And the system came up running dma, the DragonFly Mail Agent. Maybe that's a better choice. First I need to RTFM.
TRT: Still no access
|
Topic: technology, multimedia, animals, opinion | Link here |
One of the pressing reasons for the upgrade to lagoon was because Yvonne had signed up for a subscription to TRT method, an equine training site that is particularly fussy about what browsers it talks to. Now that we have the latest and greatest firefox and Chromium, it shouldn't be a problem, right?
Wrong. First we no longer had a login, which was relatively simple to fix. But when we did, the videos didn't display. What we got was:
|
![]() |
|
Are those padlock icons an indication that it doesn't like something? It's too polite to say. Sent off a message to support and got no answer by evening. I have been able to access the videos before, so it's presumably a bug in their web software.
Old spice
|
Topic: food and drink, opinion | Link here |
The alu masala that I made yesterday called for black mustard seed. Where is it? I couldn't find it. All I had was yellow mustard seed, which had expired about 11 years ago.
How much else has expired? It's time to dispose of really old stuff, say over 10 years old. And there was plenty of it:
|
|
It's gone now, though it hurts. Much of the stuff smelt perfectly usable. But in many cases we also have newer versions. And somehow the spice shelves don't look any emptier:
|
|
Tuesday, 5 August 2025 | Dereel | Images for 5 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Scale accuracy
|
Topic: general, opinion | Link here |
Last monh I found problems with my bathroom scales, and since ALDI had them on special, bought a second set to compare. The results were not good:
|
The new scales are the ones on the left, and they showed less than the ones I had, though not consistently. Here it's 700 g less, but it varies. So, since they're on special and can be returned, we bought two more of them. They're all broken! I've established that my existing scales show the same as Yvonne's, or maybe 100 g more. But what I had from the new ones was, compared to the existing scales:
|
|
In more detail:
Scale | Average | Max | ||||||||||
difference | ||||||||||||
1 | -0.5 | -0.5 | -0.3 | -0.8 | -0.53 | 0.5 | ||||||
2 | -0.8 | -0.6 | -0.2 | -0.4 | -0.5 | 0.6 | ||||||
3 | -2.8 | -2.4 | -1.6 | -2.5 | -2.33 | 1.2 | ||||||
So not only do they disagree with each other, they disagree with themselves from one reading to the next. Back they go, of course, but I'm really quite disappointed that they're that bad. I'm reminded of the German homonyms „Waage“ (weighing scales) and „vage“ (vague, inaccurate).
Kangaroo crossing
|
Topic: animals | Link here |
Walking the dogs today, saw this:
|
Nothing unusual, but a particularly clear view of how kangaroos move across soft ground.
Suffering Alyogyne
|
Topic: gardening, opinion | Link here |
Somehow our Alyogyne huegelii is not doing well. It has been windy recently, but I didn't think it enough to cause this kind of damage:
|
|
|
Should I maybe try to plant another one in a more protected position?
Ossobuco: worth the trouble?
|
Topic: food and drink, opinion | Link here |
Ate some more of Saturday's ossobuco today. Somehow it wasn't cooked enough (only 2 hours, as specified). Next time, if there is a next time, it will be 4 hours. But somehow it's boring. We should find better ways to eat truffles.
Wednesday, 6 August 2025 | Dereel | |
Top of page | ||
previous day | ||
next day | ||
last day |
80 years since what?
|
Topic: history, opinion | Link here |
Eighty years ago today, a nuclear warhead was dropped on a city in Japan, completely obliterating it. Have people become more civilized since then? I don't know. But it's sad to note the almost complete lack of mention of the incident in the world's newspapers.
Jane left something behind
|
Topic: health, general, opinion | Link here |
Woke up this morning with a scratchy throat, and Yvonne was also not feeling her best. It seems that Jane didn't leave soon enough to avoid infecting us. So far it's not serious, and hopefully it'll blow over soon.
Reducing orange juice quality
|
Topic: health, food and drink, opinion | Link here |
Things have quietened down in the USA, though their supreme leader continues to cause pain. But today I read an article in the Washington Post: they're planning to reduce the amount of sugar in orange juice in the USA. What? That would ruin it.
It seems that the sugar content of current US orange juices is round 24 g per 8 oz glass. Yes, that's their bizarre measurement, and clearly they're 8 US fluid ounces, 29.573 ml and thus 4% more than an Imperial fluid ounce. Reducing to sane measurements, that's rather more than 10% w/v of sugar. And they're talking about reducing the allowable sugar quantity from 18 g to 17 g in an unspecified serving (I'm guessing at 7.2% w/v):
The FDA sets a minimum standard for the sugar content in order to call a drink “pasteurized orange juice,” the kind commonly sold at grocery stores without added sugars. If manufacturers fall below that threshold, they are essentially no longer allowed to call their products “pasteurized orange juice.”
What a horrible thing to do to sugar-addicted US Americans!
And here was I thinking that pasteurization was a process to prevent things going bad and had nothing to do with the composition. I took a look at our fresh Australian orange juice. 9.8 g/250 ml, still a strange measurement, but which clearly corresponds to 3.92% w/v. That's OK: it's not pasteurized, and it's more than sweet enough for my taste. Why do they pasteurize orange juice in the USA? Pasteurization doesn't improve the flavour. Do they have unpasteurized juice with less sugar? The juice in our fridge has a use-by date of 27 September, and we didn't buy it all this week.
lagune: the next fix
|
Topic: technology | Link here |
Yvonne went shopping today, time to do some very little work on lagune. It's mainly working now, but mutt wasn't working correctly. It seems that the port didn't install /usr/local/lib/.muttrc (a strange place). Copying the file from lagoon fixed that, but we still have issues with the fonts.
And then there's this frustrating mail(1) program. It's part of the base installation of FreeBSD, but mailutils, a port dependency that I don't need directly, installs its own mail(1) in /usr/local/bin. And /usr/local/bin is in front of /usr/bin in my PATH. The result:
=== root@lagune (/dev/pts/3) /var/tmp 64 -> mail grog@lemis.com
Cc: fdafdasf
Subject:
(Interrupt -- one more to kill letter)
That Subject: prompt is the key. It shouldn't be there, and for reasons I have long forgotten, this version of mail(1) doesn't work well for me.
=== root@lagune (/dev/pts/3) /var/tmp 65 -> wh mail
15464781 -rwxr-xr-x 1 root wheel 216136 15 Jul 11:12 /usr/local/bin/mail
8976076 -r-xr-xr-x 3 root wheel 105296 4 Aug 11:40 /usr/bin/mail=== root@lagune (/dev/pts/3) /var/tmp 66 -> chmod -x /usr/local/bin/mail
=== root@lagune (/dev/pts/3) /var/tmp 67 -> hash -r
=== root@lagune (/dev/pts/3) /var/tmp 70 -> pkg which /usr/local/bin/mail
/usr/local/bin/mail was installed by package mailutils-3.19
Done!
Thursday, 7 August 2025 | Dereel | Images for 7 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Under the weather
|
Topic: health | Link here |
Yvonne was feeling decidedly seedy today and spent most of the day in bed. I wasn't feeling nearly as bad, though I also had sniffles and a cough. Hopefully it will soon be over.
What's that stuff?
|
Topic: general, opinion | Link here |
Found this in my washbasin in the course of the day:
|
|
What is it? And how did it get there? My best bet is that it's something off a plant, but what? And how did it get there? Until proof of the contrary, it's a cat.
cron on lagune
|
Topic: technology | Link here |
Did relatively little work on lagune today. On the one hand, there isn't much to do, and on the other hand Yvonne is not very interested in her computer at the moment. But I can't get cron to run! Well, it runs and writes to its log file, but it doesn't do what I put in crontab. How do I debug that?
Friday, 8 August 2025 | Dereel → Cape Clear → Dereel | Images for 8 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Health
|
Topic: health, general | Link here |
Yvonne was still feeling under the weather today, and she spent most of the day in bed. I wasn't feeling too bright myself. It started off alright, but by evening I was feeling decidedly unwell, although I didn't have any outward symptoms, and we went to bed 2 hours early.
Off to Pene
|
Topic: animals, general | Link here |
Yvonne had an appointment with Pene Kirk today to give Larissa one of her periodic injections. So I took her, seeing Pene for the first time in months. All over there and back in 45 minutes.
lagune: cron and dma gang up
|
Topic: technology, opinion | Link here |
So why doesn't cron work on lagune? Why, doesn't cron work on lagune? After some searching discovered that it did, but the mail system (dma, the DragonFly Mail Agent) wanted some configuration after all:
--- dma.conf 2025/06/26 10:49:45 1.1
+++ dma.conf 2025/08/08 02:21:45
@@ -1,7 +1,7 @@
#
# Your smarthost (also called relayhost). Leave blank if you don't want
# smarthost support.
-#SMARTHOST
+SMARTHOST mx0.lemis.com
# Use this SMTP port. Most users will be fine with the default (25)
#PORT 25
@@ -60,4 +60,4 @@
# MASQUERADE herb@ert will send all mails as herb@ert
# Directly forward the mail to the SMARTHOST bypassing aliases and local delivery
-#NULLCLIENT
+NULLCLIENT
After that, cron was able to send me its error messages, and I got things working. But it still doesn't explain why the @reboot line didn't work.
Preparing photos for printing
|
Topic: photography, opinion | Link here |
Once upon a time, all photos were printed, with the exception of slides and the ones I postponed for another day. But since we have computers and digital photos, I almost never do it any more. The last time was 3½ years ago, and it was enough fun in itself. But now Bev Smith wants some prints of the photos I took at the end of last month, and she wants to put them in a frame. That's more complicated: I need to crop them to a standard format. Which of these photos is best?
|
|
|
Arguably none of them, but I don't get much choice. The first is the ridiculously wide rendition from my mobile phone, the second a 3:2 ratio (like 35 mm cameras and their modern successors) which could be printed at 8"×12", and the third a 5:4 ratio suitable for printing at 8"×10". While it's surprising how little difference there is (the native Micro Four Thirds system format of 4:3 falls in between at 1:1.333), I need to make a decision on every single photo. And that was a real pain.
Cleaning up the bathroom cupboard
|
Topic: health, general, opinion | Link here |
Yvonne is coughing a lot, and I'm coughing too. She found some Strepsils in the medications Jane left behind, but they're for sore throats, not coughing. And in any case, I have some.
Do I? Where are they? Before I knew it, I was removing everything from the cupboard under the washbasin:
|
|
Lots of stuff that I had completely forgotten about, the cruft of 10 years, much of which interested Bruno:
|
|
Yes, two different kinds of Strepsils, varying only in flavour. And yes, two different kinds of cough mixture, one this horrible Senega & ammonia mixture that I got 6 years ago, and another one that I bought on the following day. We both tried it with uncertain results.
Saturday, 9 August 2025 | Dereel | Images for 9 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Gradual recovery
|
Topic: health | Link here |
Feeling a lot better this morning, arguably normal. Yvonne wasn't, and it took her all day before she started feeling relatively comfortable.
Jack Hua revisited
|
Topic: food and drink, opinion | Link here |
I've been cooking the Jack Hua rice noodles in advance for ever, but last time I tried cooking them immediately before incorporation into the dish, somewhat hampered by not knowing the correct cooking time. Now I know it: instead of one minute in advance, I need 5 minutes: clearly the pre-cooked noodles soften in the meantime.
Today I did it again, and discovered that the noodles have a completely different texture, softer and slightly sticky, somewhat like Kway Teow. To my taste they're much better like that. I liked the noodles before, but this way they're even better, and they don't disintegrate.
avidemux again
|
Topic: multimedia, animals, technology, opinion | Link here |
Yvonne was feeling much better in the afternoon, with the result that she went out and took some video of her horses.
Oh. That requires avidemux, which has caused me pain for over 10 years. I had seen it as the last frontier of bringing lagune up to speed.
OK, nothing for it, bite the bullet. It ran as slow as molasses—so did rdesktop to distress. But both ran without any further problems.
Hopefully the speed issues are related to running over NFS; I wouldn't put it past this kind of program to issue single sector I/O. Now that it's working in principle, I can move the photo disks from lagoon to lagune and then retire lagoon.
Dak galbi?
|
Topic: food and drink, language, opinion | Link here |
Yvonne was still feeling not hungry this afternoon and only wanted a soup. OK, time for me to look at one of the myriad spice pastes that I had bought without thinking that I would have to eat them alone. These three single-server Korean ones sprung to mind:
|
|
|
The fine print (and Korean text) show that the first is really Bulgogi, the second is Jeyuk Bokkeum, and the third is Dak Galbi. But look at that spelling: 닭갈비. I've never seen the bottom of the first syllable before. It looks like a ㄹ, variously an r (at the beginning of a syllable) or an l (at the end), followed by a ㄱ (g or k). So I would have read that syllable as dalk or dark, not dak. High time to brush up on my Korean lexicography.
The second two promise Gochujang and “hot and spicy”, just what I need. What else do I need? For the Da[rl]?k Galbi, 250 g of cabbage amongst other things. So it's the Jeyuk Bokkeum. 280 g of pork for one serving! I ended up using 240 g, and it still seemed far too much:
|
|
The result didn't look too bad:
|
And of course it was far too much for a single serve. But it didn't taste too bad.
Sunday, 10 August 2025 | Dereel | |
Top of page | ||
previous day | ||
next day | ||
last day |
Beef, broccoli and orange again
|
Topic: food and drink, opinion | Link here |
As planned for last weekend, finally got round to cooking beef with orange and broccoli. To my surprise, the broccoli, now 10 days old, was still in perfect condition. And I skipped deep frying the meat, which on the one hand saved some work, but it left the meat unevenly cooked. I think that deep frying is really an improvement. And there wasn't enough sauce. More seriously, Yvonne decided that she doesn't like the orange peel, which is the whole point of the dish. I wonder if we can salvage the rest. Or maybe it would work with thin strips of peel.
Part of the process was to grind a star anise. Where is it? Oh, there, expired in March 2010. That was one that got away from my purge last week. Fortunately I found some fresh star anise, and it's amazing how much stronger the aroma is. I wonder if I should dispose of other spices that are “only” 5 years old.
Monday, 11 August 2025 | Dereel | Images for 11 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Tonkotsu? Zum Kotzen!
|
Topic: food and drink, opinion | Link here |
Tried another “instant noodle” dish today, ALDI's “Tonkotsu ramen":
|
|
OK, how do I cook it? It's described on the back in tasteful condensed 5 point white on pale green text. Even when I enlarge it, it's still difficult to read:
|
OK, identify the tonkotsu sauce and dehydrated vegetable sachets:
|
And yes, this is the side with the “text”. There's nothing at all printed on the other side. It's a process of elimination. Next, where do I put it? They supply a microwave oven proof container and lid:
|
|
But the instructions are explicit: put the ingredients in a “microwave safe” bowl. That doesn't sound as if they're referring to what they supplied. But more to the point, it makes sense to put them in the bowl from which they will be eaten:
|
It's not very much water, only 220 ml. Normally I'd put between 280 and 320 ml. And there's clearly not much in there apart from noodles and dried vegetables. OK, add half a boiled egg, some small cooked prawns and a bit of chopped spring onions:
|
How was it? It could easily have been the worst instant dish that I have had, but last month I had worse. This just tasted of nothing. After some soya and chili sauces it tasted better, mainly of soya and chili sauce. A revelation, to be sure: who eats this junk? And of course I find my word play appropriate: in German, „zum Kotzen“ means “to make you vomit”.
Printing photos
|
Topic: photography, technology, opinion | Link here |
Harvey Norman have a special on photo printing at the moment, finishing tomorrow. When tomorrow? They're too polite to say. So to be on the safe side, I should order Bev Smith's photos today. How hard can it be?
Terrible!
The problem? It started with me not being able to log in, though I have an account. Stupidly, I hadn't written down the password, though I was relatively sure what it was. And while I was fighting their password recovery system (“Can't reset password right now”), it came up with my login details. But by then I had created a second account.
Then choose one of two different fonts for the same “album”, one existing and one new:
![]() |
|
And that one just didn't upload the images:
|
“Failed”. Why, fool?
After half an hour of head-scratching, I looked for help. No human available, just another of these sites that only allow you to report problems that they have thought of. Sent off a request anyway, went out and thought it over. What if the image is too large for their system? The file I tried to upload was 20.16 MP or 6.4 MB in size, not inappropriate for an 8"×12" print, which would require 8.64 MP at 300 dpi. But try something smaller?
Yes! This stupid site can't accept normal sized images. The camera (OM System OM-1 Mark II) has a relatively low resolution. What about these “mine is bigger than yours” cameras with 80 MP or more, like the Fujifilm GFX100S with 102 to 408 MP? Sorry, we don't do Fuji?
Much more mouse pushing, and finally I had all the images in place.
Oh:
![]() |
|
After all the trouble I went to to crop the images to 4:5 or 2:3! This one was to be printed 8"×12", or 2:3. What do I have?
Size: 4417 x 2945 pixels (13.01 megapixels, 1:1.50)
But the exact aspect ratio would be either 4417 × 2944⅔ pixels or 4417½ × 2945 pixels. Is it really that fussy? Other images (the minority) got through without being “cropped”. By accident I put a 4:5 image in the 8"×12" section, and it was of course also cropped. But the display showed where it was cropped, so I'm relatively sure that these ones are correct.
And how did I delete the incorrect image? I didn't find out how. I just set the number of copies to 0, but it remained on the page.
Finally, check out:
![]() |
|
What does that have to do with payment method? I don't think they know either. And yes, I accepted both items to maintain my sanity. And it was done! We'll send a confirmation to brokenphotos@lemis.com. They didn't, but I have a confirmation number, and PayPal is my witness.
To be fair, this isn't all Harvey Norman's fault. The processing is done by Fujifilmimagine, a site too polite to reveal anything about itself—that's described on a different web site on a different domain: “Fujifilm Imagine is an award-winning multichannel platform successfully used around the world”.
![]() |
|
I wonder who supplied the award.
Fixing the gate
|
Topic: general, opinion | Link here |
We've had a really Heath-Robinson arrangement round the front gate to stop Larissa from escaping to the road:
|
And it didn't work, so we had to close it with a chain, rather detracting from the electric gate. Yvonne has been trying to get Paul Donaghy to come by for some months now: pull the dropper and replace it closer to where the gate travels, and put some battens between the dropper and the other side of the fence post. Straightforward enough. We have the dropper equipment (on loan from Chris Bahlo), and Yvonne arranged that Paul would bring some battens.
Paul came this afternoon, where I really had other things to do (this damn photo printing pain). No battens. Clearly a loss of communication. So we replaced the dropper and put some more electric fence wire on the other side of the post. No, says Yvonne, Lara can still get through. Another dropper. OK, now it works.
|
No, she still wants the battens. OK, just leave me out of it.
Yellow-striped black cockatoos
|
Topic: animals, photography, opinion | Link here |
Seen walking today:
|
|
That could really have done with a longer lens. I really should take the 12-200 lens on such walks.
Truffle omelette
|
Topic: food and drink | Link here |
We still have some of Margaret Swan's truffle left over. Nothing for it. An omelette, which I normally think of as being a breakfast dish:
|
Tuesday, 12 August 2025 | Dereel | Images for 12 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Sunrise
|
Topic: general, photography, opinion | Link here |
Up at 7 this morning to go to the toilet, and saw this outside:
|
Problem: I was in my underwear, and I didn't have time to get dressed. This photo worked, sort of. One that didn't was taken from my office with the Olympus M.Zuiko Digital ED 8 mm f/1.8 Fisheye PRO. The reflections kill it, but I can't make up my mind which of these two projections is worse:
|
|
No banking without app!
|
Topic: general, technology, opinion | Link here |
Last month I received notice from my accountants that Macquarie Bank wanted me to use an app to verify every transaction. Normally the accountants do that for me.
I've already made my objections clear, but at the time my adviser, Peter O'Connell, was on holiday. He's back now and called me up and discussed the matter. Yes, I must approve every transaction with their Macquarie Authenticator, which replaces their previous system of sending SMSes.
I've grumbled at length about people who associate mobile phones with security. But is this maybe the misguided way of the future? It seems that we don't have much alternative. I need to contact Bank Australia and see if they can help, but I doubt it. Can I use my old mobile phone, the one that no longer works as a mobile phone? To be investigated.
Or am I incorrect? Is it really possible to be secure with mobile phones? Are face and fingerprint recognition reliable enough? There seem to be too many potential attack vectors that I can't oversee.
But I have seen some that the banks haven't, so I'm not confident. One example is an email that I forgot, dated 9 October 2024:
You may already be aware that from late November, Macquarie Bank will require clients to approve all new payments initiated by PPT Financial via the Macquarie Authenticator App. This will replace the option of approving payments via SMS secure codes.
But that's doubly incorrect: I have never approved a payment via SMS, and this requirement didn't happen. Does that increase my trust level?
Laugengebäck again
|
Topic: food and drink | Link here |
Baked more Laugengebäck today. Gradually we're getting our act together:
|
|
That was after 18 minutes at 220°, rather longer than the recipe envisages. Are they dark enough? They're now very close to the image in that recipe:
But would more time in the lye help? Higher temperature? Longer baking time? Or is everything good as it is?
Wednesday, 13 August 2025 | Dereel | Images for 13 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Toyota: RTFM
|
Topic: general, technology, opinion | Link here |
Yvonne has had her new car for two weeks now, and she's still very happy with it. But that's at least because she doesn't use the radio. It's modern, but the controls seem straightforward enough:
|
|
Problem: pressing on the ⏼ does nothing useful. After much reading the detailed manual, which thoughtfully also provides information about models unrelated to her car, I discovered the solution. It's intuitive: press on the volume button underneath the symbol.
OK, try it out this morning. Oh. The radio was already off. OK, turn it on. Works as advertised.
|
Turn it off again. Nothing happens.
So I still don't know how to turn the radio on and off, after two weeks! Why is this so complicated?
As if that weren't enough, I thought it a good idea to check the oil level. Open the bonnet with the convenient catch under the steering wheel, where I expected it to be. But I couldn't find the secondary catch under the bonnet. There's almost always a catch that you can feel easily, but not here. More RTFM. Yes, it's where I expected it, and there are two conflicting drawings, politely not showing the orientation, that describe this lever, slightly offset from the middle:
|
![]() |
|
You need to pull the lever up. Once again, why is this so complicated?
Completing lagune
|
Topic: technology, opinion | Link here |
So Yvonne has been able to process videos on lagune with avidemux. And I don't see any great issues with photos with DxO PhotoLab. Time to move the (external) disks to lagune, showing in the process that I'm even less agile than I should be.
But it didn't take long. Update /etc/fstab and mount:
=== root@lagune (/dev/pts/3) ~ 135 -> mount /Photos/
mount: /dev/da0p1: Device busy
Huh? I've never seen that message on mount. It would make sense if the device were mounted elsewhere, or that the mount point were in use, but neither were the case, and FreeBSD doesn't care if you mount the same device or mount point twice.
=== root@lagune (/dev/pts/3) ~ 137 -> df /Photos/
Filesystem 1048576-blocks Used Avail Capacity Mounted on
/dev/nda0p2 198,102 63,658 118,596 35% /
Nope, that's the root file system.
=== root@lagune (/dev/pts/3) ~ 138 -> mount /Photos
=== root@lagune (/dev/pts/3) ~ 139 -> umount /Photos
=== root@lagune (/dev/pts/3) ~ 140 -> mount /Photos/
=== root@lagune (/dev/pts/3) ~ 141 -> df /Photos/
Filesystem 1048576-blocks Used Avail Capacity Mounted on
/dev/da0p1 7,630,061 4,123,279 3,430,481 55% /Photos
That looks right. But what happened the first time?
Lithium-ion AA batteries
|
Topic: general, technology | Link here |
Yvonne back from shopping with a set of 8 Lithium-Ion AA batteries, which I hope will last longer in her wireless microphones. We've established that Nickel–metal hydride batteries don't have enough voltage to hold out more than an hour or so, and non-rechargeable batteries are unpredictable, so we end up changing them before they're discharged. We can do that with Li-Ion batteries too, but they should last longer, and they're rechargeable.
The batteries came with a remarkably complicated-looking charger that shows red (charging) and green (charged), along with a dancing display of blue on the batteries themselves. And it even has a beeper to tell us when each battery was charged.
How do 3 V Li-Ion batteries supply the 1.5 V that AA batteries should supply? Unevenly, it seems. Before charging the batteries had between 1.50 and 1.58 V. After charging, they had between 1.50 V (not the same battery) and 1.62 V. I wonder if there's some kind of voltage regulator in there. Strangely, there's little about them on Wikipedia, which only refers to the 3 V versions.
Thursday, 14 August 2025 | Dereel | Images for 14 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
More lagune pain
|
Topic: technology, general, opinion | Link here |
So now lagoon has been completely migrated to lagune. But there was still this strangeness that I noticed yesterday. I had managed to mount /Photos without difficulty, but then I had the same problem with the backup disk:
Date: Thu, 14 Aug 2025 05:05:19 +1000
From: Cron Daemon <root@lagune>
Subject: Cron <root@lagune> /home/local/bin/syncphotos
mount: /dev/da1p1: Device busy
And what's more, /Photos was no longer mounted. A remount showed:
=== root@lagune (/dev/pts/3) ~ 150 -> mount /Photos/
mount: /dev/da0p1: Device busy=== root@lagune (/dev/pts/3) ~ 151 -> mount /Photos/
mount: /dev/da0p1: R/W mount on /Photos denied. Filesystem is not clean - run fsck. Forced mount will invalidate journal contents: Operation not permitted
You have new mail in /var/mail/grog=== root@lagune (/dev/pts/3) ~ 152 -> date
Thu 14 Aug 2025 09:36:21 AEST
Is there something wrong with the disk? Tried copying raw data:
=== root@lagune (/dev/pts/3) ~ 154 -> dd if=/dev/da1 of=/dev/null bs=1m count=1000
dd: /dev/da1: Device busy
0+0 records in
0+0 records out
0 bytes transferred in 18.072301 secs (0 bytes/sec)=== root@lagune (/dev/pts/3) ~ 155 -> dd if=/dev/da1 of=/dev/null bs=1m count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 5.032489 secs (208361330 bytes/sec)=== root@lagune (/dev/pts/3) ~ 156 -> dd if=/dev/da1 of=/dev/null bs=1m count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 5.040510 secs (208029758 bytes/sec)=== root@lagune (/dev/pts/3) ~ 157 -> date
Thu 14 Aug 2025 09:42:56 AEST
Once again this “device busy” on the first attempt, followed by a success on the second. What's in the logs?
Aug 14 09:37:07 lagune kernel: (da1:umass-sim1:1:0:0): Polling device for readiness
Aug 14 09:37:09 lagune kernel: (da1:umass-sim1:1:0:0): READ(10). CDB: 28 00 00 00 00 00 00 08 00 00
Aug 14 09:37:09 lagune kernel: (da1:umass-sim1:1:0:0): CAM status: SCSI Status Error
Aug 14 09:37:09 lagune kernel: (da1:umass-sim1:1:0:0): SCSI status: Check Condition
Aug 14 09:37:09 lagune kernel: (da1:umass-sim1:1:0:0): SCSI sense: NOT READY asc:4,1 (Logical unit is in process of becoming ready)
Aug 14 09:37:09 lagune kernel: (da1:umass-sim1:1:0:0): Error 16, Retries exhausted
Oh. And when the backup was attempted?
Aug 14 05:05:18 lagune kernel: (da1:umass-sim1:1:0:0): SCSI sense: NOT READY asc:4,1 (Logical unit is in process of becoming ready)
Aug 14 05:05:18 lagune kernel: (da1:umass-sim1:1:0:0): Error 16, Retries exhausted
Aug 14 05:05:18 lagune kernel: g_vfs_done():da1p1[READ(offset=65536, length=8192)]error = 16
And immediately before it in the log file was:
Aug 13 21:02:32 lagune kernel: (da0:umass-sim0:0:0:0): SCSI sense: NOT READY asc:4,1 (Logical unit is in proce
ss of becoming ready)
Aug 13 21:02:32 lagune kernel: (da0:umass-sim0:0:0:0): Error 16, Retries exhausted
Aug 13 21:02:32 lagune kernel: g_vfs_done(): da0p1 converting all errors to ENXIO
Aug 13 21:02:32 lagune kernel: g_vfs_done():da0p1[READ(offset=313759268864, length=32768)]error = 6 supressing further ENXIO
Aug 13 21:02:32 lagune kernel: UFS: forcibly unmounting /dev/da0p1 from /Photos
At least that explains why /Photos wasn't mounted. But why is all this happening? Everything points to the system not waiting until the disk is up to speed. What has changed between FreeBSD 12 and 14? Still more searching to be done in an area where I had hoped not to have to do anything.
Web pages from the distant past
|
Topic: history, opinion | Link here |
While looking in my RCS control files today, found this:
-r--r--r-- 1 grog lemis 7,942 11 Nov 1996 work-history.html,v
-r--r--r-- 1 grog lemis 4,138 11 Nov 1996 early-history.html,v
What's that? They prove to be the earliest web pages that I ever stored in RCS, thus also reinforcing my assumption that my home page, dated February 1997, was in fact started much earlier. I've tidied up early-history.html here.
Sadly, they're not very interesting, and not surprisingly they're full of markup errors. work-history.html doesn't render at all. Only after looking at the source did I discover that it's in roff format.
Interestingly for the dating was this first line in the files:
<!- Hey, emacs! Edit this file in -*- html -*- mode!
That's completely incorrect, of course, and the end of the comment is missing, but it suggests that I already had Emacs macros to handle web pages. Unfortunately I can't find any reference in the Emacs revision history.
Many-coloured Schlumbergera
|
Topic: gardening | Link here |
The Schlumbergera that Yana brought us over 2 years ago now seem to have all taken, though one is still thinking about it:
|
But the others have a number of flowers in different colours:
|
|
|
|
I wonder what the first one has in store.
Friday, 15 August 2025 | Dereel | Images for 15 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
More lagune fun
|
Topic: technology, opinion | Link here |
So why does lagune time out trying to mount my external USB backup disks? Spent some time looking through man pages and myriad sysctls without finding anything obvious. Then, just for the fun of it, I asked Google Gemini: “How do I set the number of retries for a USB disk to become ready on FreeBSD?”
Now what's the chance of getting a useful answer to so specific a question? But I did! In fact, a surprisingly detailed description of the situation, along with what to do about it.
But it's so complicated that I can't help thinking that there must be an easier way. There is one, of course: just try to copy a sector from the disk before trying to mount it. It will probably fail, but by that time the disk will be up to speed.
Another thing I forgot: syslogd should log to eureka. Done, but it needs restarting syslogd. That shouldn't be necessary.
More nvidia fun
|
Topic: technology | Link here |
Display hydra:0.3 is dead! Not the monitor, just the display. Switch to server 1 and it's fine. Switch back and it's dead.
The background is understandable: I had used that monitor to look at some weirdness on distress, the Microsoft box. Clearly it had something to do with that. But why isn't it coming back again after correct connection? Fired up nvidia-settings and discovered:
|
![]() |
|
How did that happen? And how do I fix it? Spent some time pointing and clicking. It had a field Configuration, but I couldn't select anything. With a bit of moving around, it somehow came good, but this program always confuses me.
Macquarie bank security
|
Topic: general, technology, opinion | Link here |
On Tuesday I considered the alternatives to Macquarie Bank's requirement of authenticating via mobile phone. One possibility would be to use my old mobile phone, the one that can not longer work as a phone. My guess is that their Macquarie Authenticator app will require the mobile phone network, even though a direct Internet connection would be more appropriate.
But that's only a guess. To be sure, let's install it and see what it says.
|
Where is it? Ah, it's far too secure to say what it is. Just that black padlock “Authenticator”, also designed to make it difficult to find in the list of apps. OK, fire it up:
|
That's clever, isn't it? I never had one. So I sent mail to Peter O'Connell, asking for it to be sent by normal post to avoid the insecurity of email. His response included:
Note, Macquarie will send you your ID number, however it has to be sent to your email address.
Under no circumstances, will they send in the mail..
Somehow every step scares me and convinces me of my suspicion that they don't know what they're doing. Others seem to agree. Here the current ratings section on the Toyshop page:
|
Saturday, 16 August 2025 | Dereel | Images for 16 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
lagune: The last step?
|
Topic: technology, general | Link here |
So what's missing for lagune? Not much, just the issues with the USB disks and the connection with distress, the Microsoft machine. I really hate setting up Samba and Microsoft networking.
Finally got round to doing it. It wants a password! And I didn't write it down. OK, I can set one, since it's only for Microsoft to talk to lagune. Once again I dragged out my decades-old copy of The Complete FreeBSD. Time to add it to the new lagoon page.
Setting up on the Microsoft side worked without problems once I found their “share” page, which seems to wander from one unexpected place to another. I never will understand Microsoft.
With the disks, noted a reference to a port smartmontools with a program smartctl that, with the -a option, produces much output that I must investigate more closely.
The only other issue was for syncing to the web servers: she didn't have the ~/.ssh files for fra.lemis.com. That was easy enough to fix. Now all we need is for her to test it, which should come soon.
Sunday, 17 August 2025 | Dereel | Images for 17 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
lagune: done?
|
Topic: technology, photography | Link here |
So far the migration to lagune has gone relatively painlessly. In particular my
concerns about avidemux proved
unwarranted, and I was able to mount share file systems on Microsoft with
relatively little pain. So today Yvonne processed her first
photos.
All went well until I put them through my postprocessing scripts. They failed. No ExifTool. OK, run round in circles trying to find the name of the port (it's p5-Image-ExifTool-devel-13.33 or most of that string), install it. More missing ports. ImageMagick? Ah, you mean ImageMagick7. And then the one that completely caught me off guard:
for i in `ls *.jpeg *.gif *.png 2>/dev/null`; do echo $i; done | by 2 /Photos/Tools/mktinysmall.php /Photos/Tools `pwd` small 270000; MYDIR=`pwd`; IMAGELINKS=`pwd`/Imagelinks; if [ -f $IMAGELINKS ]; then (cd ~/public_html/Photos/`basename $MYDIR`/small && checkimagelinks $IMAGELINKS small); fi
sh: /Photos/Tools/mktinysmall.php: not found
sh: /Photos/Tools/mktinysmall.php: not found=== yvonne@lagune (/dev/pts/3) ~/Photos/20250816 13 -> l /Photos/Tools/mktinysmall.php
-r-xr-xr-x 1 grog lemis 2,025 1 Jul 2021 /Photos/Tools/mktinysmall.php
Huh? Ah, this is sh's inimitable way of saying “I can't find the interpreter in the shebang line”:
#!/usr/local/bin/php
OK, install PHP. But which?
=== root@lagune (/dev/pts/8) /home/grog 6 -> pkg search php | wc -l
3292
Over 3000 ports for PHP alone! Does it understand just php? Of course not:
=== root@lagune (/dev/pts/8) /home/grog 7 -> pkg install php
pkg: No packages available to install matching 'php' have been found in the repositories
Much searching brought me to php85, which installed and worked.
Somehow most of the pain here is in the Ports Collection.
In passing, it's interesting to note this message:
WARNING: The convert command is deprecated in IMv7, use "magick" instead of "convert" or "magick convert"
“IMv7” is clearly ImageMagick's way of referring to itself. And yes, that makes sense: “convert”, like many ImageMagick programs, breaks name spaces. I should read up on it, but even the brief help takes up several pages:
=== yvonne@lagune (/dev/pts/3) ~/Photos/20250816 18 -> magick | wc -l
magick: invalid argument for option @ error/magick-cli.c/MagickImageCommand/1358.
347
Reading books again
|
Topic: history, general, opinion | Link here |
A while back we watched “A Town like Alice” on TV. Yvonne was impressed, and wanted to buy the book.
What? What are public libraries for? So we borrowed it from the Ballarat library, and Yvonne read it and found it good.
Did I want to read it too? I have barely read any books in the last few years. They've been overtaken by digital media. But it sounded like a good idea, so I'm currently gradually working my way through the book, rediscovering in the process my liking of Shute's style. In this case it was particularly the geographical details, the part that Yvonne found boring. Sixty years ago my father and I drove through much of the area described in the march described in the book, and I was able to differentiate nearly all the real locations from the made-up ones. Only one detail was wrong: Shute wrote that the Japanese attacked Malaya after the Pearl Harbor attack; in fact it was before, in the early hours of 8 December 1941.
And also 60 years ago today I met my first real girlfriend, Lesley Cannings. I had planned not to mention this kind of anniversary, but Shute forced my hand
More app vulnerabilities
|
Topic: technology, opinion | Link here |
Another report of a mobile phone security breach. Some badly written app leaks sensitive information.
Sure, they'll (probably) fix it. But how many others are there out there that might do the same? How many will come in the future? I'm sure that there must be some testing facility, maybe one that's obligatory or will soon become so, but they can only ever prove the existence of vulnerabilities, never their absence. This app (“Dreamhome”) is on the toyshop, so that's not much help. The entry currently includes:
Data safety
Safety starts with understanding how developers collect and share your data. Data privacy and security practices may vary based on your use, region, and age. The developer provided this information and may update it over time.
No data shared with third partiesLearn more about how developers declare sharing
No data collectedLearn more about how developers declare collectionData is encrypted in transit
You can request that data be deleted
It's worth savouring these statements. No data collected, but it's encrypted in transit. You can request that it be deleted, but there's no guarantee that the request will be honoured. Doesn't that fill you with confidence? And it's all the more stupid because this is a vacuum cleaner. There's no reason on earth why you should need any credentials to run the thing.
Roast pork again
|
Topic: food and drink, opinion | Link here |
Roast pork for dinner this evening. I've always had difficulty getting the crackling right, so this time I followed not one, but two recipes: the one on the packaging and this one from Australian Pork. Or at least, that's what I thought. Australian Pork, too, seem to have changed their minds.
In principle the recipes are similar, but the devil's in the detail. First oil and salt, then put in a hot oven for a while, then (maybe when the crackling starts to be evident) drop the temperature.
But that's where things differ. The instructions on the packaging want the meat cooked at 220° for 30-40 minutes, then 30-35 minutes/kg at 180°. No mention of meat temperature.
The Australian Pork recipe wanted it cooked at 250° with heat from above for 12 minutes, then at an unspecified lower heat until the crackling develops, then at 180° until the meat temperature reaches 80°.
That's what I recorded 3½ years ago. But since then, Australian Pork have changed their recipe completely. No top heat any more, only 220° to start with for 40 minutes or “until the rind crackles”. Then drop to 160° and cook for 30-35 minutes per kg, until the meat temperature reaches 65°.
I suppose the good news is that it's no surprise that I'm confused. In the last 3½ years Australian Pork has changed its recipe significantly. 65° is quite low for pork. 80° is too high, as I accidentally discovered today.
According to the recipes, my joint (1.17 kg) would take about 80 minutes to be done. My progress:
Start: meat at 14°, put into 240° oven.
|
12 minutes:
|
20 minutes:
|
24 minutes: skin “pops”:
|
|
Drop temperature to 230°.
40 minutes: meat temperature 59°, drop oven temperature to 160°.
|
51 minutes: meat temperature 68°, change heat to top only
|
63 minutes: meat temperature 78°, open oven door, leave on top heat to brown crackling:
|
66 minutes: meat temperature 80°, stop cooking.
That's very different from what the instructions suggest. Yes, the third recipe wanted 220°, not 240°, but described it in a very misleading way: “preheat the oven to 240°C/220°C fan- forced.” What they clearly (now it's clear) meant was “240° in a conventional oven, 220° in a fan-forced oven”. But even that wouldn't explain the great difference in the times and temperatures.
While resting, the meat temperature went up to 89° before dropping again. That's far too much. And though the crackling was nice in places, it didn't go through to the meat:
|
How do I get it right? Next time I'll read this entry again and take it from there.
Bruno
|
Topic: animals | Link here |
Bruno is continually on top of cabinets:
|
|
That's a studio flash next to him, but it doesn't seem to interest him:
|
Monday, 18 August 2025 | Dereel → Ballarat → Dereel | Images for 18 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Dentist again
|
Topic: health | Link here |
Finally my sniffles have died down, and I went for my dentist's appointment originally due two weeks ago.
It was supposed to be a routine checkup, but somehow the issues discovered in February aren't over. Hopefully that's not another molar to be extracted.
New Indian grocery
|
Topic: food and drink, opinion | Link here |
On the way into town, in to visit a new Indian grocery, Patel Convenience Ballarat, another place with only Facebook web presence.
My first consideration was: why have they opened now? It has only been 5 months since the Indo-Asian Groceries opened a branch in Sebastopol, and I don't see the owner being run off his feet.
This place is enormous! I don't think I have ever seen as big an Indian food shop anywhere. And it's full of produce. There must be several tons of rice and atta, and an amazing number of other things, many of which I haven't even heard of. There were even over $1000 worth of Ganesh idols:
|
|
I hadn't intended to buy anything, but I ended up with some beans (for frijoles de la olla) and tanduri masala. But I couldn't find a chatni that I liked. Somehow the food is very much tailored to Indian people. Are there enough in Ballarat to keep them running?
Picking up the photos
|
Topic: photography, general, language, opinion | Link here |
Are my photos ready? They promised to send me email when they were done, but that seems to be an idle promise. Fighting my way through the Harvey Norman web site, I discovered that the status was “IK DCM Set To Ready”. What does that mean? Called them up, and Chris wanted to know my name and whether the photos were normal size.
Normal size? What's normal for Chris? Less than 8"x10", anyway. Yes, they're ready. What does “IK DCM Set To Ready” mean? He doesn't know either.
In the afternoon went there and had to ask a cashier where the photo centre was. Vague pointing to the west. Over there, where another salesperson took me to the real place, only a few metres to the north of the first cashier. Again, name and size, please. Picked them up with no identification whatsoever, a far cry from five months ago, where only I was allowed to pick up the item, though I didn't record whether they asked for identification on that occasion.
The packaging was interesting, too. Quite substantial envelopes with the text “Photo's DO NOT BEND”. Is that in keeping with the rest?
And the photos? They don't look bad, though it looks as if they have cropped 1% to 2% off the edges. As far as I can see, that's nothing to do with the cropping they claimed on some images.
Tuesday, 19 August 2025 | Dereel | Images for 19 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
lagune: done?
|
Topic: technology, opinion | Link here |
My attempts to bring the USB disks up on lagune have been less than completely successful. Once again I ended up with a forcibly (and dirtily) umounted /Photos file system. The information from Google Gemini was useful, but it seemed to be over the top. It did, however, point to the fact that the timeout handling was in the CAM layer. Are there no sysctls that control that? What does this give me?
=== root@lagune (/dev/pts/8) /home/grog 29 -> sysctl kern.cam.da
...
kern.cam.da.default_timeout: 60
kern.cam.da.retry_count: 4
kern.cam.da.poll_period: 3
There we go!
=== root@lagune (/dev/pts/8) /home/grog 30 -> sysctl kern.cam.da.retry_count=6
kern.cam.da.retry_count: 4 -> 6
We'll see if that works, but I have a good feeling about it.
What else needs to be done to fully cut over to lagune? In principle nothing: lagoon, located physically under lagune, is no longer doing anything except maintaining its uptime:
lagoon up 438+11:00, 0 users, load 0.31, 0.23, 0.21
OK, build a new world on lagune before rebooting. Oh. It failed. Why that? But it's not critical for the work I have to do, so I just continued with the current system, dating back only 2 weeks. Remove the display cards from both machines and put the card from lagoon in lagune: the one originally in lagune had a noisy fan. Bring it up again.
Yes, it works for the most part. A surprising number of minor details, though. My xe script, which runs xearth, failed with an unexpected error:
Error: Can't open display: unix:0
Oh. unix is an old, worn-out magic word, and this particular incarnation of the script dates to
=== root@lagune (/dev/pts/5) ~ 5 -> l ~yvonne/xe
-rwxr-xr-x 1 yvonne home 4,783 6 Sep 1997 /home/yvonne/xe
Savour that: not quite 28 years old! Nowadays we just write :0. I suppose this has bitten me before, but I didn't chase it up.
Indeed, almost exactly five years before. I even came up with the same description.
But it seems that it was no longer in use. The real code is in the .xinitrc file. No idea why that didn't start; I'll follow up next time.
And once again the mouse didn't come up with middle button emulation. Why? Forgot to set it for this particular machine name. But while searching for the correct incantation, found (thanks again to Google Gemini) a great simplification. Here from the code for hydra:
# This one is the J.Burrows from Officeworks, also with side buttons
MOUSEID=$(xinput | grep -m 1 "Telink 2.4G Mouse" | sed 's/^.*id=\([0-9]*\)[ \t].*$/\1/')
if [ "$MOUSEID" != "" ]; then
# One of these should work
xinput set-prop $MOUSEID 286 -0.5
xinput set-prop $MOUSEID 296 -0.5
fi
Here I have tried set-prop on two different numbers to save going looking for them. But it would be so much easier the new way:
xinput set-prop "Telink 2.4G Mouse" "libinput Accel Speed" -0.5
And I still have issues with syslogd! /etc/syslog.conf is correct, but it doesn't log to eureka until I restart it. Is this another instance of the race condition that stops me mounting NFS file systems at boot time?
A federal Ukraine?
|
Topic: politics, history, multimedia, technology, opinion | Link here |
How do we solve the issue of the Russo-Ukrainian War? It's not helped by people like Donald Trump, who (currently) wants to award Volodymyr Putin parts of Ukraine.
That's stupid, of course. Russia committed to the territorial integrity of Ukraine in 1997. They have broken that treaty. And now they want to be rewarded for it?
But one way out might be to restructure Ukraine as a federal state, where each state (oblast) had more autonomy. Might that not be a good start to appeasing Russia?
But a memory came back to me. Didn't something like this happen in Servant of the People? Each oblast had a delegation that wanted something for themselves, maybe independence. But in which episode? The IMDB page wasn't very helpful. How about AI bots again?
Google Gemini came up with a clear statement:
The episode of "Servant of the People" that explicitly shows a delegation from various oblasts is Season 3, Episode 1.
In this episode, Ukraine has supposedly "disintegrated into small separate principalities" following the overthrow of Holoborodko. It depicts different regions being given new, often satirical, names (e.g., Luhansk and Donetsk oblasts united as "USSR," Lviv Oblast as the "Galician Kingdom," Crimea as the "Crimean Emirate," etc.). This is where the idea of delegations from these "new" oblasts would come into play, as they are portrayed as distinct entities.
OK, take a look. No, not what I was looking for. How about Chat AI?
Hello! In the TV series "Servant of the People," the episode featuring a delegation from the various oblasts is Season 1, Episode 3.
Not that one, either. Claude? That's easy. It has no idea. DeepSeek?
One notable episode that features a delegation from various oblasts is Season 2, Episode 6 ("The Minister"). In this episode, representatives from different regions come to discuss their problems with the central government, including issues like funding, infrastructure, and political disputes.
Another relevant episode is Season 3, Episode 3, where regional leaders and local officials interact with the president's administration, sometimes in a confrontational manner.
Now doesn't that give you the feeling that they know what's going on? It seems that the episode I was looking for was the last one, both in this list and in the series: season 3, episode 3, first broadcast on 28 March 2019, 3 days before the first round of the 2019 Ukrainian presidential election. And in that connection, it's interesting to see the results by oblast:
The areas that Russia claims are also the ones with the highest support for Servant of the People, something that I hadn't expected. Would a federal system help solve the current situation? It's hard to say. But why haven't I heard anybody come up with the idea?
What's that egg?
|
Topic: animals | Link here |
Yvonne found another egg in the dam paddock today, roughly here.
|
That's the white one; the brown is a large (69 g) hen egg. This one weighs 52.2 g. What kind of egg is it? How did it get into the middle of the paddock? My guess is that it's a wild duck egg. Is it?
Wednesday, 20 August 2025 | Dereel | |
Top of page | ||
previous day | ||
next day | ||
last day |
lagune: tying up loose ends
|
Topic: technology, opinion | Link here |
Another attempt to build a world for lagune today, this time with j -4 instead of j -8. And it worked.
Preparing for rebooting lagune applied my new-found information about mouse configuration:
=== yvonne@lagune (/dev/pts/2) ~ 2 -> xinput set-prop "Logitech USB Receiver" "libinput Accel Speed" -0.5
Warning: There are multiple devices matching 'Logitech USB Receiver'.
To ensure the correct one is selected, please use the device ID, or prefix the
device name with 'pointer:' or 'keyboard:' as appropriate.
unable to find device Logitech USB Receiver
Oh. Yes, xinput shows me that there's a keyboard with that name too. But where do I put the pointer:? Directly in front, without a space, it proves:
=== yvonne@lagune (/dev/pts/2) ~ 3 -> xinput set-prop "pointer:Logitech USB Receiver" "libinput Accel Speed" -0.5
Put that in the ~/.xinitrc, and after reboot it worked as expected.
So lagune is up and running, and things seem mainly OK. But there are still a few things that I've been keeping my eye on.
First the external USB disks. Did changing the number of retries work? Yes! Here the kludge in /home/local/bin/syncphotos:
--- /home/local/bin/syncphotos 2021/02/12 23:17:55 1.18
+++ /home/local/bin/syncphotos 2025/08/18 00:07:12
@@ -35,6 +35,11 @@
+if [ `hostname` = "lagune" ]; then
+ # Work around slow spin up of backup disk
+ dd if=/dev/da0 of=/dev/null count=1
+ dd if=/dev/da1 of=/dev/null count=1
+fi
Previously we had:
dd: /dev/da0: Device busy
0 bytes transferred in 18.057944 secs (0 bytes/sec)
dd: /dev/da1: Device busy
0 bytes transferred in 18.050228 secs (0 bytes/sec)
But now we have:
512 bytes transferred in 19.116480 secs (27 bytes/sec)
512 bytes transferred in 19.074229 secs (27 bytes/sec)
So it seemed it timed out one second too early. Now all is well.
And the NFS file systems? Now that mail is working properly, I get a message:
Date: Wed, 20 Aug 2025 14:42:38 +1000
From: Cron Daemon <root@lagune>
Subject: Cron <root@lagune> mount -t nfs -a
mount_nfs: eureka: Name could not be resolved at this time
mount_nfs: eureka: Name could not be resolved at this time
mount: /eureka/home: No such file or directory
...
That's the cron entry for @reboot, so it does work after all. Comparing that with the log messages, we have:
Aug 20 14:42:39 lagune ntpd[1389]: error resolving pool 0.freebsd.pool.ntp.org: Name could not be resolved at this time (2)
Aug 20 14:42:39 lagune kernel: re0: link state changed to UP
So for whatever reason (something I still need to check), interface re0 is down at that time. And so the mount commands also fail, and that's probably the reason that the syslogd logging fails too. OK, I can run a script @reboot that first waits a few seconds before mounting the NFS file systems and restarting syslogd.
How I hate git!
|
Topic: technology, opinion | Link here |
I haven't committed anything to the FreeBSD source tree for nearly a year, though there's plenty to do. To avoid the attention of the Grim Reaper, it's time for another one. Simple update to du.1. Commit, no problems. “Push”, problems, discretely commented in yellow on unknown background:
=== grog@hydra (/dev/pts/7) /src/FreeBSD/git/main/usr.bin/du 53 -> git commit du.1
hint: Waiting for your editor to close the file... Waiting for Emacs...
[main fc4f6c2cb3fd] Describe relationship between block count and EOF.
1 file changed, 8 insertions(+)
You have mail in /var/mail/grog=== grog@hydra (/dev/pts/7) /src/FreeBSD/git/main/usr.bin/du 54 -> git push freebsd
To gitrepo.freebsd.org:src.git
! [rejected] main -> main (fetch first)
error: failed to push some refs to 'gitrepo.freebsd.org:src.git'
hint: Updates were rejected because the remote contains work that you do not
hint: have locally. This is usually caused by another repository pushing to
hint: the same ref. If you want to integrate the remote changes, use
hint: 'git pull' before pushing again.
hint: See the 'Note about fast-forwards' in 'git push --help' for details.
OK, do what it says:
=== grog@hydra (/dev/pts/7) /src/FreeBSD/git/main/usr.bin/du 55 -> git pull
remote: Enumerating objects: 60, done.
remote: Counting objects: 100% (14/14), done.
remote: Compressing objects: 100% (3/3), done.
remote: Total 60 (delta 11), reused 11 (delta 11), pack-reused 46 (from 1)
Unpacking objects: 100% (60/60), 84.08 KiB | 2.33 MiB/s, done.
From https://git.freebsd.org/src
3ff0231c87f3..dc5f8c344299 main -> freebsd/main
hint: You have divergent branches and need to specify how to reconcile them.
hint: You can do so by running one of the following commands sometime before
hint: your next pull:
hint:
hint: git config pull.rebase false # merge
hint: git config pull.rebase true # rebase
hint: git config pull.ff only # fast-forward only
hint:
hint: You can replace "git config" with "git config --global" to set a default
hint: preference for all repositories. You can also pass --rebase, --no-rebase,
hint: or --ff-only on the command line to override the configured default per
hint: invocation.
fatal: Need to specify how to reconcile divergent branches.
Ha ha, only annoying you. In the illegible text it now says that you need --ff-only:
=== grog@hydra (/dev/pts/7) /src/FreeBSD/git/main/usr.bin/du 56 -> git pull --ff-only
hint: Diverging branches can't be fast-forwarded, you need to either:
hint:
hint: git merge --no-ff
hint:
hint: or:
hint:
hint: git rebase
hint:
hint: Disable this message with "git config advice.diverging false"
What's this merge nonsense? This was a virgin tree until I changed a single file. But who knows? Try again:
=== grog@hydra (/dev/pts/7) /src/FreeBSD/git/main/usr.bin/du 57 -> git merge --no-ff
hint: Waiting for your editor to close the file... Waiting for Emacs...
Merge made by the 'ort' strategy.
UPDATING | 7 +++++++
lib/libc/net/gai_strerror.c | 44 ++++++++++++++++++++++++--------------------
lib/libifconfig/Symbol.map | 2 --
lib/libifconfig/libifconfig.h | 1 -
share/man/man4/iwlwifi.4 | 4 ++--
share/man/man5/src.conf.5 | 9 ++++++---
share/mk/src.opts.mk | 2 +-
share/mk/src.sys.mk | 2 +-
sys/arm64/arm64/elf32_machdep.c | 2 +-
tools/build/options/WITH_CLEAN | 4 ++++
10 files changed, 46 insertions(+), 31 deletions(-)
create mode 100644 tools/build/options/WITH_CLEAN=== grog@hydra (/dev/pts/7) /src/FreeBSD/git/main/usr.bin/du 58 -> git pull --ff-only
Already up to date.
The first sign of success!
=== grog@hydra (/dev/pts/7) /src/FreeBSD/git/main/usr.bin/du 59 -> git push freebsd
Enumerating objects: 15, done.
Counting objects: 100% (12/12), done.
Delta compression using up to 32 threads
Compressing objects: 100% (7/7), done.
Writing objects: 100% (7/7), 947 bytes | 947.00 KiB/s, done.
Total 7 (delta 5), reused 0 (delta 0), pack-reused 0 (from 0)
remote: ================================================================
remote: Currently only allow merge from vendor/*
remote: ================================================================
remote: FATAL: VREF/MERGE-CHECK/src: helper program exit status 256
remote: error: hook declined to update refs/heads/main
To gitrepo.freebsd.org:src.git
! [remote rejected] main -> main (hook declined)
error: failed to push some refs to 'gitrepo.freebsd.org:src.git'
And that? That looks like a current situation, though it's not what release engineering says.
And in passing, this requirement for pushing from an up-to-date tree looks like a recipe for disaster on a tree that is updated frequently (several times a minute, for example). How can you keep your tree up to date long enough to perform the push?
HOW I hate git!
Thursday, 21 August 2025 | Dereel | Images for 21 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Leftover KL Hokkien Mee
|
Topic: food and drink | Link here |
We still have leftovers from Sunday's roast pork. The meat itself is no issue: cut it into 3 mm slices and freeze it for Abendbrot. But there's the crackling.
And then we have some cabbage left over from a planned meal of enchiladas, scrapped after I read my diary comments from previous occasions.
What can I do with them? KL Hokkien Mee? That uses crispy pork fat, not quite the same thing as crackling, but close. It also uses choi sam, not that far from cabbage. OK, fake it:
|
|
How was it? Quite edible. But I think I will stick to my normal recipe next time.
Preparing for spring
|
Topic: gardening, opinion | Link here |
Jesse Walsh along to do some tidy-up work in the garden, including completely removing the contents of one herb bed and half of the other one:
|
The last tomatoes of winter
|
Topic: gardening, food and drink | Link here |
It's 10 days before the first day of spring, and Jesse brought me this:
|
Yvonne tried one. It was bitter. What can we do with the rest?
Still more cabbage
|
Topic: food and drink, opinion | Link here |
Another way of using up the cabbage we have is as Weißkrautsalat, “white cabbage salad”. After discussion with Yvonne I used my version. And I don't think I like it much any more. As it stands (200 g of apple to 500 g of cabbage), there seems to be too much apple. And somehow it's not interesting enough. I put in 2.5 times as much caraway as in the recipe (which calls for 2.5 g for 500 g of cabbage), but I think it could be even more. And maybe it could do with more lemon juice.
Friday, 22 August 2025 | Dereel | Images for 22 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Fried cabbage
|
Topic: food and drink, opinion | Link here |
We still have lots of cabbage left over. How can I cook a breakfast from it? Once again Google Gemini to my aid, and once again it came up with some useful leads: Chinese Fried Cabbage (手撕包菜 - Shǒu Sī Bāo Cài), “Thai-Style Garlicky Cabbage with Fish Sauce” and “Hmong-Style Fried Cabbage with Bacon and Oyster Sauce”. I didn't follow up the last two, but the Shǒu Sī Bāo Cài sounded interesting, and in the end I made up my version, with the help of this recipe and this one.
quantity | ingredient | step | ||
60 g | minced pork | 1 | ||
15 g | garlic | 2 | ||
3 g | hot chili | 2 | ||
250 g | green cabbage | 3 | ||
2 g | sugar | 4 | ||
20 g | light soya sauce | 5 | ||
10 g | water | 5 | ||
7 g | Chinese wine | 5 | ||
10 g | Chinese vinegar | 6 | ||
15 g (1) | spring onion | 6 | ||
Preparation
Fry the pork in oil until slightly cooked. Reserve.
Fry the garlic and chili gently.
|
Tear the cabbage into medium-sized strips:
|
|
Add meat and cabbage to pan and fry lightly:
|
Add sugar and liquids and sauté until cabbage has softened a little. It should not be completely soft.
Add vinegar and chopped spring onions.
|
Reheat, then serve.
Rangefinder camera history, yet again
|
Topic: photography, history, opinion | Link here |
The centenary celebrations of the Leica continue. Today I looked at this video, by a Leica employee:
The breakthrough: the Leica II, a camera with an inbuilt rangefinder.
But was it? My understanding was that the first 35 mm camera with a built-in rangefinder was the Contax I. Both cameras were introduced in 1932, but when? Previously I had been given to understand that the Contax was first, but I didn't have a date. Off to ask the usual AI bots, which replied:
The Leica II, a 35mm rangefinder camera, was introduced by Leica in 1932. However, similar to the Contax I, there is no information available on the exact day or month of its release.
Now doesn't that help? Of course, the video clip didn't mention Contax at all, just that the Leica M3 was released “nearly 25 years later” (really 1954, 22 years later). It had great improvements! A rangefinder coupled with the viewfinder! A bayonet mount! Both what the Contax I had 22 years earlier. But unlike the Contax, it still had a horizontal travel cloth focal plane shutter.
I'm still puzzled that the Contax didn't run rings round the Leica. Looking through the technical details of the photos in „Das Gesicht des Krieges“ (1942), I find that out of an estimated 80 photos, only two were taken with Leica, about 20% with other cameras, including 6×6, and the rest (a good two-thirds) were taken with Contax. Why have things changed since then? Was it thanks to the Soviet theft of the plant?
Of course, the date isn't that important. It seems unlikely that Leitz and Zeiss weren't aware of what the other was doing, even if it didn't amount to espionage: a coupled rangefinder was such an obvious requirement. And indeed the Contax I showed significant signs of being rushed to the market.
Film rewound!
|
Topic: photography, opinion | Link here |
While investigating the Leica anniversary, played around with my FED 1. It still had a film in it: four months ago it was finished, but I was unable to rewind it.
But this thing has such a tiny rewind knob. Here on the left, compared to the Diaxette on the right:
It must require twice the turning that the Diaxette (the last camera I used without a rewind lever) did. Can it be that I just hadn't turned it enough, and that the stiffness had some other cause?
Yes! I was finally able to rewind the film. What is it?
|
Oh. A reusable (and from its appearance frequently reused) cassette. So now I need to find a way to identify the film. Open the cassette in a darkroom and cut off the leader? That might help decide whether it's colour or black and white, but not much more.
To be fair to the Leica/FED, the tiny rewind knob was not limited to that model. A number of my other cameras have such tiny rewind knobs, including the Asahiflex.
Saturday, 23 August 2025 | Dereel | Images for 23 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Garden flowers in late winter
|
Topic: gardening, general, opinion | Link here |
It's a month before the equinox, time for the monthly garden flower photos.
As I noted last month, this winter has even fewer flowers than normal. The spring bulbs are there, but nothing else seems to have noticed the coming spring, due in only 9 days. I had expected some signs of life from the Box Elder, Ginkgo and oak:
|
|
|
Hopefully that's also the reason why my Clematis “Edo Murasaki” hasn't shown any signs of life yet. It did flower last month, so hopefully it's just late.
Things continue to deteriorate. As I feared last month, my Camellia japonica has had it. Here two months ago, last month and this month:
|
|
|
It would be gone by now had Jesse had the time to remove it on Thursday.
Less unexpected is the situation in the north paddock, where Yvonne had a number of seedlings planted two months ago. Then we had:
|
|
Now we have:
|
|
|
Somehow our Alyogyne huegelii is continually beset by damage. It's still alive, but another branch has broken off:
|
|
|
The Strelitziae are looking relatively good, though. The Strelitzia nicolai has two flowers on their way, and the Strelitziae reginae also seem to be preparing something:
|
|
|
|
And though the Hibiscus rosa-sinensis “Uncle Max” in the garden is not looking as good as it could, one in the lounge room is flowering happily:
|
|
|
And the cutting I made six months ago is also growing steadily:
|
A Town Like Alice continued
|
Topic: general, opinion | Link here |
I'm still reading “A Town Like Alice” at a snail's pace, checking lots of things as I go along. I find my guess about the location where the women ended up to be quite close, somewhere between Kuala Dungun and Kuala Terengganu. But the geographical accuracy is suffering, and it seems that Shute had never heard of Kuala Terengganu. The women spent a lot of time between Jerantut and Maran, though they had come from the south. Jerantut was never close to anything there, and I wonder why Shute mentioned it at all. Presumably he had been there at some time, though it must have been really tiny. It was barely on the map when I was last there.
And then the sea journey from London to (ultimately) Kota Bharu. All well as far as Colombo, but then the sea journey went via Rangoon (way off course) and Penang to Singapore, from whence she caught a plane to Kota Bharu. That makes no sense at all: she could have disembarked in Penang and caught a much shorter flight to Kota Bharu. Still, an interesting book from my point of view.
Sunday, 24 August 2025 | Dereel | Images for 24 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Nothing doing
|
Topic: general | Link here |
Somehow I've been too active lately, but today was an exception: I did almost nothing.
Vietnamese Phở bò
|
Topic: food and drink, opinion | Link here |
I've been making Phở bò) (Vietnamese beef soup) from a Thai paste for years now, though I doubted its authenticity. And I also bought some cubes that do come from Viet Nam.
|
I've eaten this version before—this was the last cube in the pack. On the previous occasion I was irritated by the quantities (1 cube is intended to give 500 ml of broth), and as a result didn't stick to the instructions. Today I just made 320 ml. And it didn't taste bad. Not very similar to the Por Kwan paste, but almost certainly more authentic.
Still more new spices
|
Topic: food and drink, opinion | Link here |
In the evening, chicken tanduri, made with this mixture:
|
And yes, it's better than the Patak paste that we've been using. That's no surprise. For once, I stuck exactly to the recipe:
|
Monday, 25 August 2025 | Dereel | Images for 25 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Macquarie security
|
Topic: general, technology, opinion | Link here |
I've finally got round to doing something about the Macquarie Bank authenticator. I had been warned that I could be in for a long wait, so I waited until I had a couple of hours to spare.
The call started well, with an on-hold announcement giving me URLs with “forward slashes”. But my concerns about the time were unfounded. I was speaking to Angelica, a non-native English speaker, after 3 minutes, and despite some intelligibility issues the whole call only took 17 minutes.
What did I want? My Macquarie ID. But somehow they had ended up with me being registered with Yvonne's phone number, and since modern communication and authentication revolves around phones, I first had to change that. And basically that's all I managed.
In preparation I had brought Yvonne's phone for identification, but no, they didn't need that. My phone number, full name and date of birth. Not even my address!
But which phone number? I have a new ALDImobile SIM specially for the authenticator, and it's currently waiting for activation, so I currently don't have a number. That so completely baffled Angelica that I had to repeat it twice after a suitable timeout. OK, give my normal number, I can change it later.
She gave me other information that I haven't seen elsewhere: the authenticator is only needed for transfers over $2000. How many of them are there? There was also talk of my online bank account, but what do I need that for? And if this is an account for the Lehey Family Superannuation Fund, why should I have direct access at all? Why is the account (apparently) in my name?
And that was basically the call. I was to wait for an SMS with the URL of an app that would complete the authentication. That involved taking photos of my driver license and of myself (“smile, please”). And the data that it gleaned was less than convincing:
|
What's wrong with this picture? At the very least, they ignored one potential source of identity: the old phone number. Yes, it can maybe no longer exist or have been given to somebody else, but that could have been confirmed. More to the point, how have they positively identified me? The photo of the driver license might help, but the incorrect identification suggests that the identification program isn't very clever. And the “smile please” is presumably to thwart the use of a photo for identification purposes.
And that's all that I see. Somebody could have stolen my driver license, and if they look like me or use makeup techniques to look like my driver license photo, they could pass that test too. That alone, along with name and date of birth, is completely insufficient.
What else can they do? Owner lookup for the phone number, of course, and presumably they'll do that. Is that sufficient?
What next? Wait for email from Macquarie with details of how to get the coveted ID. That could take up to a week.
The good news seems to be that the authenticator is really just an authenticator. So it doesn't really seem to pose the kind of security threat that I had feared. But we'll see when they have done their due diligence.
New clothes dryer
|
Topic: general | Link here |
Our clothes dryer is getting tired: it requires several attempts to power on before it stays on. Nothing for it, says Yvonne: a new dryer.
Today Glen Castleman came by with a new, modern one with LED display, which thus cost $200 instead of the $150 for the last one:
|
Hopefully it will last longer: we had the last one for less than 3 years.
The revenge of Corsair
|
Topic: technology, opinion | Link here |
Now that lagune is up and running, I need to look at hydra. I have had it for nearly two years, and I've never finished setting it up. In the meantime we have a new release of FreeBSD. Time to build a shadow machine on dereel, Bruce Evans' old zetaplex machine, which gave:
=== root@hydra (/dev/pts/38) /usr/src 17 -> uname -a
FreeBSD hydra.lemis.com 13.4-PRERELEASE FreeBSD 13.4-PRERELEASE #1 stable/13-n258139-35bde8969eb6-dirty: Sun Jul 28 08:42:07 AEST 2024 grog@hydra.lemis.com:/home/obj/home/src/FreeBSD/git/stable/13/amd64.amd64/sys/GENERIC amd64=== root@dereel (/dev/pts/0) /usr/src 12 -> uname -a
FreeBSD dereel.lemis.com 14.3-STABLE FreeBSD 14.3-STABLE #5: Mon Aug 25 14:54:35 AEST 2025 grog@dereel.lemis.com:/home/obj/hydra/home/src/FreeBSD/git/stable/14/amd64.amd64/sys/GENERIC amd64
First, though, I wanted more memory. One of the things I want to test is VirtualBox, which uses ridiculous quantities of memory. And dereel only had 16 GB, with two memory slots free.
That's OK: I have had two 8 GB SIMMs floating around for some time. Just what I need. But the ones already in the machine have heat sinks, so it's difficult to get a normal SIMM in between the two. Remove one. But somehow it didn't work as expected. On (much) closer inspection, it seems that these SIMM sockets have levers only at the top, and at the bottom they're fixed:
|
|
OK, finally get the chips in. Power on. Nothing. Much investigation that pointed to the SIMM I had removed:
|
That's one of those horrible violent “Vengenance” SIMMs. But more to the point, it claims to be 16 GB!
|
Why is it showing as only 8 GB?
After some messing around, got the original two SIMMs working again. I'm reminded of this xkcd cartoon:
Put the other SIMMs into the old lagoon, where they worked fine, but it only had 8 GB to start with, so I still haven't reached my goal of 32 GB. But dereel is now up to date.
Tuesday, 26 August 2025 | Dereel | Images for 26 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Instagram: You are not worthy
|
Topic: technology, opinion | Link here |
Received a message today with a link to an image on Instagram. Probably not interesting, but it's straightforward enough to take a look.
Or so I thought. First I had to log in. But that wasn't straightforward:
![]() |
|
![]() |
|
Why? To the best of my knowledge I have never posted anything on Instagram. Did somebody else, maybe, breaking my (weak) password? Followed up the offer to receive my “content” and found:
|
![]() |
|
30 days to send it to me? OK, Instagram, see if I care.
And yes, I did get a warning nearly a year ago. Only it didn't make any sense, and they didn't say why.
Macquarie: the next step
|
Topic: technology, general, opinion | Link here |
As promised, mail from Macquarie Bank today, in HTML-only format:
As requested your phone number has been updated.
Punctuation is original. No indication of the promised Macquarie ID. Sent mail to Tori McLean, who arranged it so quickly that both her reply arrived only a minute later than her reply. And yes, I was able to set it up. But then more obsolete security measures:
To assist in fraud prevention, you are required to register for 2-Step Authentication, an extra layer of security used to identify you by using something you have (e.g. mobile phone) or something you know (e.g. secret questions and answers).
That proved to be secret questions: In what city/town did you grow up? What is the name of the first school you attended? Who was your childhood hero? And no, I didn't have the option of selection my own question (“What do you think of our antediluvian security?", for example), and the answers had to be in a limited character set that at least excluded an apostrophe. And the help file was interesting, containing things like
3. Go to Step 3 to continue
But finally it was set up. I wonder if I will ever need it.
Later a survey about how I found it. And another indication that this is a real Australian bank:
Date: Mon, 25 Aug 2025 23:25:57 -0700 (PDT)
From: Macquarie Bank Customer Experience <invitation@feedback.macquarie.com>
Subject: Gregory, we'd love to hear from you
I suppose that's what people get from interfaces that can't tell the time.
VirtualBox revisited
|
Topic: technology, opinion | Link here |
The main reason for updating dereel was to test the latest version of VirtualBox. Tried to start it, first loading the kernel module:
=== root@dereel (/dev/pts/0) /eureka/Photos 26 -> kldload /boot/modules/vboxnetadp.ko
kldload: can't load /boot/modules/vboxnetadp.ko: module already loaded or in kernel
/var/log/messages gave the clue:
Aug 26 14:15:31 dereel kernel: KLD vboxdrv.ko: depends on kernel - not available or version mismatch
Aug 26 14:15:31 dereel kernel: KLD vboxnetadp.ko: depends on vboxdrv - not available or version mismatch
dereel is running FreeBSD 14.3, but the packages system had loaded kernel modules for 14.2. And if you save the 75 pages of log messages that normally just run off the screen, and you read it, you can discover that this is a feature, not a bug:
Message from virtualbox-ose-kmod-71-7.1.12.1402000:To avoid crashes due to kernel incompatibility, this module will only load on FreeBSD 14.2 kernels.
A lot of help that is! When am I going to get past pain with the Ports/Packages collection? Nothing for it: I had to build the package from source, requiring a surprising number of dependencies. And they have changed the menus that they still pop up in the middle of a port. Once they were in colour and had some kind of graphical interface. Now we're back to curses, and I can't navigate them:
![]() |
|
Maybe that's just because I save the log (see above), but it doesn't seem to be an improvement.
After barely more than 100 minutes and 620 pages of output, it was done. And it installed with no problems, producing a surprising amount of information in the process:
VirtualBox was installed.
You need to load the vboxdrv kernel module via /boot/loader.conf:
vboxdrv_load="YES"
You also have to add all users to your vboxusers group in order to use vbox.
% pw groupmod vboxusers -m jerry
Reboot the machine to load the needed kernel modules.
Bridging Support:
=================
For bridged networking please add the following line to your /etc/rc.conf:
vboxnet_enable="YES"
USB Support:
============
For USB support your user needs to be in the operator group and needs read
and write permissions to the USB device.
% pw groupmod operator -m jerry
Add the following to /etc/devfs.rules (create if it doesn't exist):
[system=10]
add path 'usb/*' mode 0660 group operator
To load these new rule add the following to /etc/rc.conf:
devfs_system_ruleset="system"
Then restart devfs to load the new rules:
% /etc/rc.d/devfs restart
Troubleshooting:
================
Running VirtualBox as non-root user may fail with a fatal error
NS_ERROR_FACTORY_NOT_REGISTERED. In this case delete /tmp/.vbox-*-ipc file.
If you experience "Network: write Failed: Cannot allocate memory" errors
try to increase net.graph.maxdata in /boot/loader.conf
If you are using AIO, then increase these limits (https://bugs.freebsd.org/168298):
vfs.aio.max_buf_aio=8192
vfs.aio.max_aio_queue_per_proc=65536
vfs.aio.max_aio_per_proc=8192
vfs.aio.max_aio_queue=65536
To check if AIO is used use: kldstat -v | grep aio
If you are experiencing VMs freezes with an error in VBox.log such as:
"
00:01:29.590192 AssertLogRel /usr/ports/emulators/virtualbox-ose/work/VirtualBox-6.1.44
/src/VBox/VMM/VMMR3/PGMPhys.cpp(5148) int PGMR3PhysAllocateHandyPages(PVM): RT_SUCCESS(rc)
00:01:29.590221 87/128: idPage=0x3d400 HCPhysGCPhys=000000027eaed000 rc=VERR_NO_MEMORY
00:01:29.590247 Changing the VM state from 'RUNNING' to 'GURU_MEDITATION'
00:01:29.590261 Console: Machine state changed to 'GuruMeditation'
00:01:29.590695 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
00:01:29.590696 !!
00:01:29.590697 !! VCPU0: Guru Meditation -8 (VERR_NO_MEMORY)
"
and see a lot of free RAM, then increase sysctl vm.max_user_wired
(https://bugs.freebsd.org/252227).
For dedicated VM servers, a good start is:
("RAM in GB" - max(1, "RAM in GB" / 32)) * 1024*1024*1024 / PAGE_SIZE
PAGE_SIZE on x86 is 4096.
Examples with different RAM sizes in GB:
16: ( 16-max(1, 16/32)*1024*1024*1024/4096 = 3932160 => 15GB
32: ( 32-max(1, 32/32)*1024*1024*1024/4096 = 8126464 => 31GB
48: ( 48-max(1, 48/32)*1024*1024*1024/4096 = 12189696 => 46.5GB
64: ( 64-max(1, 64/32)*1024*1024*1024/4096 = 16252928 => 62GB
128: (128-max(1,128/32)*1024*1024*1024/4096 = 32505856 => 124GB
256: (256-max(1,256/32)*1024*1024*1024/4096 = 65011712 => 248GB
When assigning IP addresses in host-only mode from ranges outside
of VirtualBox's defaults, the respective ranges need to be listed
in /usr/local/etc/vbox/networks.conf (https://bugs.freebsd.org/259399).
Check the wiki page for known issues and troubleshooting:
https://wiki.freebsd.org/VirtualBox
But that was enough pain for today. Next I need to find how to import my existing virtual machines.
Wednesday, 27 August 2025 | Dereel | Images for 27 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
Mobile telephony pain
|
Topic: technology, general, opinion | Link here |
Wednesday is the day Yvonne goes shopping. She frequently calls me to ask about something she has found. What I need is a photo. But that's trivial nowadays with video calls. All I need to do is to show her how to do it. But what app should I use?
What do the twins recommend? WhatsApp. OK, that could even have the advantage of being cheaper, since—I think—it uses the data network rather than the mobile network. How hard can it be?
Almost impossible. I failed. Well, I established contact, but for our purposes we need the rear camera. How do you switch to it? Slowly and unsurely, it proved. A popup appeared in the bottom right, including a symbol that suggested that it might switch cameras. Press on it. Nothing happens. Press on it again. It switches to the rear camera and back again: I think that I hadn't waited long enough after the first press. And out of annoyance the emblem disappeared.
All this was accompanied by an extremely irritating low-frequency acoustic feedback between the cameras, which I couldn't get rid of even by turning the volume down to minimum. By this time Yvonne was getting quite upset, and I was only just able to discover that I could possibly retrieve the switch icon from the main display.
Yvonne went shopping, and after cooling down a bit I tried another experiment: does WhatsApp really work without a phone? Played around with hirse, which doesn't have a SIM. Yes, it does work, and I can register with the phone number from albo, my “real” mobile phone. It disabled access on albo, which is fair enough. Now we just need to find out what is cheaper. Arguably hirse could still have a useful life ahead.
Yet another experiment. When Yvonne called, using the standard Android app, I selected “video call”. Bad idea. We could see each other, but there was no audio. And after hanging up, I couldn't call her; after the normal 30 second delay with no ringing, it went to voice mail.
Back home, I came to the conclusion that my selection didn't switch the call to video: it started a second call, thoroughly confusing both of us. My guess is that the first call had been put on hold, along with the audio. But even back home I still couldn't call the phone. I had to reboot to do this.
In passing, as so often, the display on my phone went blank while talking to Yvonne. I suppose that's a feature (save power), but it's particularly irritating. I had to press the power button to get any display again. Potentially it's related to the Raise To Answer app, which never seems to have done anything useful, but which I suspect is responsible for the displays I sometimes see when trying to answer, something like “don't cover the microphone”, and obscuring the answer display. Goodbye,Raise To Answer. We'll see if at least that pain goes away.
WHAT a pain this Android is!
Mona attacks camera
|
Topic: animals, photography | Link here |
I need to be more careful where I put my cameras. Today Mona walked around my desk and knocked my OM System OM-1 Mark II along with the 12-200 lens onto the ground.
The damage? The lens cap came off. The lens itself shows no sign of damage, but the back of the body—presumably where the camera hit the ground—has some paint missing from one corner of the body. But it's so small that it could have happened earlier.
|
|
And the flash trigger on top of the camera has been cracked open:
|
|
That's harmless. Apart from the fact that it's old and I have more than I need, it still works, and the lid just clipped back into position.
Thursday, 28 August 2025 | Dereel | Images for 28 August 2025 |
Top of page | ||
previous day | ||
next day | ||
last day |
How to answer a mobile phone call
|
Topic: technology, general, opinion | Link here |
Yesterday's mobile phone pain was at least partly due to the fact that Android, or maybe Xiaomi, have changed the way they want calls answered. I don't understand how that could happen, since I haven't updated anything. But they seem to do that, either from one brand to another, one Android release to another, or maybe the phase of the moon. Looking back through the years, I have:
March 2018, Samsung phone: swipe right.
|
January 2020. Two different ways:
![]() |
|
![]() |
|
![]() |
|
February 2022:
![]() |
|
![]() |
|
But that was years ago. Time to annoy people again. Now, to answer a phone, you need to “swipe” in a different direction:
|
|
Or do you? Just to make it more fun, they don't say.
By chance, while searching the phone setup options, I came across this:
|
|
My first reaction was to think that the wrong digit was raised, but yes, it really wants to say “Accept incoming call” or some such. And yes, I can just press on it:
![]() |
|
No “Vertical swipe”, of course. That's an old, worn-out magic gesture.
Why “swipe”?
|
Topic: language, opinion | Link here |
I've grumbled for years about the use of the term “swipe” to describe the motions needed to get the attention of a mobile phone. The Oxford English Dictionary has a number of definitions: “to drink hastily and copiously”, “to strike at with the full swing of the arms”, “to deal a swinging blow or hit at”, “to steal”. But for some never reason I never checked a US American dictionary. Did that today in my Webster's New Encyclopedic Dictionary (“all new 1994 edition”). To my surprise, the much less detailed definition was a noun meaning “a strong sweeping blow” and a verb meaning “to strike or wipe with a sweeping motion”, presumably strong.
But on revisiting the OED, I find that they also have another definition “To pass (a credit card, identity card, etc.) through an electronic device in order to read and process data magnetically encoded on it”. Somehow this action with mobile phones is crying out for a different term.
Still more mobile phone pain
|
Topic: general, technology, opinion | Link here |
For reasons almost completely unrelated to my pain answering phones, I decided to put a second SIM in Yvonne's phone. We're currently both using ALDImobile's $15 card with 365 day expiry. But it charges a lot for data, and currently we're using a lot. Yvonne's card lasts for about 2 months before needing recharging. A second card for data? I've been doing that for a while, and I think that it just comes in cheaper. I had the SIM card that I had planned for hirse. Put it in, activate it. That's simple, says the packaging.
They lie, like they have lied before. Their web site has bugs, and after entering my code the NEXT button did nothing. For some reason I called their support line before trying a different browser, and of course that's what they asked me to do. But though I got further with Chromium, it still kept erasing things. Finally I had to identify myself, and for some reason I chose my driver license. Don't do that! They want the number written on the back in 3 point text partially obscured by the larger text.
I should have known better. I've seen this before, and last time it took me three days.
This time was different. It just said “authentication failed”, and that I would have to pay $1 by credit card to complete the authentication. No option to go back and try again, no refund of the $1. I still had <mumble> on the phone, and asked him to do something. He did: he hung up.
Dammit, $1 is a small price to pay if it solves the problem. So I paid, and how about that, it was activated. But why is everything such a pain nowadays?
Friday, 29 August 2025 | Dereel | Images for 29 August 2025 |
Top of page | ||
previous day |
Google POLA
|
Topic: technology, general, opinion | Link here |
I seem to have spent the whole week fighting “modern” technology. Yesterday I established that the method of answering a mobile phone had changed with neither warning nor explanation. Today, while searching through the maze of twisty little menus, all different, I found that Google had updated the phone app yesterday. Admire the formatting in the description:
|
I still don't understand why they do this. If there has ever been a violation of POLA, this is it. What do they hope to achieve?
Do Google Gemini know? Yes:
Q: Why has the Google phone app changed the way you answer an incoming call?
A: Google has recently rolled out a major redesign for its Phone app, which includes a new way to answer incoming calls. This change is part of a broader "Material 3 Expressive" update, aimed at refreshing the look and feel of several Google apps.
Here are the key reasons and details behind the change:
New Gestures: The most noticeable change is the move from a vertical swipe (up to answer, down to decline) to a horizontal swipe (right to answer, left to decline). This gesture is more similar to the one used on iOS, which some users may find more intuitive. Preventing Accidental Actions: Google has stated that this change is intended to help reduce accidental answers or declines that can happen when a phone is taken out of a pocket. The horizontal gesture is less likely to be triggered by mistake. Aesthetic Redesign: The change is part of a larger visual overhaul. The new interface features a cleaner, more streamlined design with bigger, more defined buttons (often described as "pill-shaped"). This includes a prominent red "End Call" button. User Feedback: While some users appreciate the new, modern look and find the horizontal swipe easier to use, the change has been met with mixed reactions. Many users have expressed frustration with the sudden, unannounced shift and the new design's "oversized" and "blocky" appearance, especially since they didn't manually update the app.Can you change it back?
While you can't revert the entire app redesign, you can change the incoming call gesture back to the old style:
Open the Google Phone app. Go to Settings. Look for an option like "Incoming call gestures" or similar. Change the setting from "Horizontal swipe" to "Single tap."This will replace the swipe gesture with the more traditional "Answer" and "Decline" buttons.
There's this “intuitive” word again. Clearly “some users” are the people migrating from Apple, since it's anything but “intuitive” to people who are already using the app.
So what is this really saying?
No wonder they gave up their motto “Don't be evil”.
Authenticating with driver license
|
Topic: technology, general, opinion | Link here |
Another pain I had yesterday was activating the new SIM card. I could barely read the partially obscured tiny text on the back of my driver license, even with a magnifying glass, so I ended up having to pay $1 to bypass it.
But I have this digital driver license. Does that show the second number as well? Yes! It hides it behind a misleading layer of false security, but it's easily legible.
And it's clearly not the same as the one on my physical driver license! Never mind, I have two, in case one fails. Look at the other one. The text is in a different place, much more legible, and the same as on the digital license.
How do I have two licenses? I lost it two years ago and applied for a replacement, which received the same number. Then only a month later I found the same one, identical in every respect.
Well, no. This one detail was different, and it seems I gave my best guess at the old number. Somehow this all still seems really silly.
A whole week fighting “intuitive” modern technology!
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 | Greg's photos | Copyright information |