|
| Greg's diary |
| May 2022 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Sunday, 1 May 2022 | Dereel | Images for 1 May 2022 |
| Top of page | ||
| next day | ||
| last day |
|
Zoom again
|
Topic: multimedia, technology, opinion | Link here |
Why couldn't Mark and I hear each other during our attempted Zoom conversation yesterday? The
Setup was relatively painful (or I have repressed the experience), and it even accepted the password that I chose (“Type 4771”), though once again it claimed that my newly invented email was a “frequently used email”. But when I tried to sign in, it wouldn't accept the password.
OK, new password, not overly different. Failed again. But then I discovered (how? I forget) that it had removed the space in the password and not even told me! I've had some bad password processing, including from Zoom, but this is the first time that I've found a program that changes the password, and that without telling me.
Sign in. How do I establish contact? Send email, it seems. OK, send an email, but of course I mistyped and got an invalid name. Didn't worry Zoom. But I didn't get a message when I typed in the correct email address.
Later I tried from my side and was told “Sorry, you cannot use this link to add friends”. OK, how about a phone number? The program (or maybe Android) is too stupid to know its own phone number, so I had to enter it. And then I had to type in a CAPTCHA! Into a mobile phone non-keyboard! It worked on the second attempt, and finally I was able to establish contact. But there was some issue with audio again. It was turned off, but I could turn it on—which promptly turned off video! Somehow, without any obvious reason, I had enabled “Safe driving mode”:
|
|||||||||||||||||||||||
More flailing around—I thought I would remember what I did, but it was so strange that I can no longer remember. Finally I had a connection with Yvonne. And the sound worked.
Problem solved? No, because I don't know how I got there. But my problems were not over yet: how do I leave the “meeting”? I didn't find out. I had to stop the app.
I've been growling about Android apps for nearly 10 years, but this is the strangest and most frustrating I've found so far.
OK, what alternative? By comparison WhatsApp is trivial. But Mark doesn't want it, so how about Skype? I've used that in the past, but not on euroa, my Microsoft laptop. OK, install and wait for the obligatory email.
It didn't come. More checking and discovered that it had made it here and been classified as spam by one of my ancient spam traps:
procmail: No match on "From:.*pain rel"
procmail: Match on "From:.*protection"
procmail: Locking "/home/grog/Mail/Caughtspam/spamsites.lock"
procmail: Assigning "LASTFOLDER=/home/grog/Mail/Caughtspam/spamsites"
procmail: Opening "/home/grog/Mail/Caughtspam/spamsites"
procmail: Acquiring kernel-lock
procmail: Unlocking "/home/grog/Mail/Caughtspam/spamsites.lock"
procmail: Notified comsat: "grog@18199734:/home/grog/Mail/Caughtspam/spamsites"
From account-security-noreply@accountprotection.microsoft.com Sun May 1 10:07:52 2022
Subject: Microsoft account security code
Folder: /home/grog/Mail/Caughtspam/spamsites 10673
How old is that? It could be up to 20 years old. And the message matched:
From: Microsoft account team <account-security-noreply@accountprotection.microsoft.com>
Subject: Microsoft account security code
Please use the following security code for the Microsoft account gr*****.
gr***** sounds right. Time to get rid of the filters. It's easier to just delete them.
|
Running with George
|
Topic: animals, opinion | Link here |
After walking with George last week, Larissa had symptoms of joint pain. They went away after a single dose of Meloxicam, but we've decided not to take her for walks for a few months.
And George? Well, he doesn't have to walk. How about bringing him here and letting him run around with the dogs? Arguably that's just at much of a problem, but we'll see.
We started off by letting him run free with Lena, who is less excitable, and then let Lara out on a lead. As I expected, she just lay down and let George examine her (run the cursor over an image to compare it with its neighbour):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Let her run free? I was for it, but both Yvonne and Peggy were worried what might happen if George suddenly got upset. So Yvonne went and found a muzzle:
|
|||||||||||||||||||||||
Then we let Lara loose. As I expected, no fireworks (again, run the cursor over an image to compare it with its neighbour)
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And then they set off into the wild blue yonder, stopped only by a gate:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Back again:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
We left it at that, clearly a success. And Lara didn't have any soreness today.
|
Lost kitchen utensils
|
Topic: food and drink | Link here |
Our “hair dryer” “air fryer” has a stand with a drip tray underneath:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
A couple of weeks ago the drip tray disappeared. Not that serious: we have a cake pan that is almost exactly the same size and which I've been using. But where did the real drip tray go? Today we found it:
|
|||||||||||||||||||||||
It's under the cutting on the left. I had idly wondered at the time that the tray was made out of metal and not out of plastic.
|
More security issues
|
Topic: technology, opinion | Link here |
While doing something completely unrelated, discovered today:
=== grog@eureka (/dev/pts/39) ~ 20 -> ssh freefall.freebsd.org
Permission denied (publickey).
Disconnected at Sun 1 May 2022 14:42:30 AEST
Huh? I've been connecting from eureka to freefall for decades, and “I haven't changed anything”. In fact, I'm still logged in from a different xterm:
=== grog@freefall (/dev/pts/54) ~ 27 -> w
3:44AM up 81 days, 1:15, 55 users, load averages: 0.20, 0.26, 0.23
USER TTY FROM LOGIN@ IDLE WHAT
...
grog pts/54 121-200-11-253.79c80b.mel.nbn.aussiebb.n 12Apr22 - w
Is there something wrong with my keys on eureka? How does it look from lax?
=== grog@lax (/dev/pts/2) ~ 2 -> ssh freefall.freebsd.org
agent key RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo returned incorrect signature type
grog@freefall.freebsd.org: Permission denied (publickey).
Disconnected at Mon 2 May 2022 03:49:03 UTC
Well, that gives me more information. I've seen that before, but on that occasion it wasn't a problem. But it's reasonable to assume that I should update my encryption keys.
How do I do that? I've had my current keys for ever, and all I recall is that there are newer, even more secure ways of doing it. But how? Asking Mark would be a good idea, so sent off a message to him.
In the meantime, how about ssh -v? Tried that,
=== grog@dereel (/dev/pts/3) ~ 10 -> ssh -v freefall
(much output omitted)
debug1: Next authentication method: publickey
debug1: Offering public key: /home/grog/.ssh/id_rsa RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo agent
debug1: Server accepts key: /home/grog/.ssh/id_rsa RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo agent
Authenticated to freefall.freebsd.org ([96.47.72.132]:22) using "publickey".
...=== grog@freefall (/dev/pts/18) ~ 1 ->
It works! But that was from dereel! Why? I've established that it's using the same key. Comparing debug output wasn't very helpful, especially as the messages are a little unclear. But this section seems relevant (- at the start of the line is dereel, + is eureka, space is both):
debug1: Next authentication method: publickey
-debug1: Offering public key: /home/grog/.ssh/id_rsa RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo agent
-debug3: send packet: type 50
+debug1: Offering RSA public key: /home/grog/.ssh/id_rsa
+debug3: send_pubkey_test
debug2: we sent a publickey packet, wait for reply
-debug3: receive packet: type 60
-debug1: Server accepts key: /home/grog/.ssh/id_rsa RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo agent
-debug3: sign_and_send_pubkey: RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo
-debug3: sign_and_send_pubkey: signing using rsa-sha2-512 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo
at this point, dereel authentication has succeeded
...
much later, after dereel has a prompt from freefall
+debug1: Authentications that can continue: publickey
+debug1: Trying private key: /home/grog/.ssh/id_dsa
+debug3: no such identity: /home/grog/.ssh/id_dsa: No such file or directory
Some of the differences are due to the different version of ssh, but somehow freefall has silently ignored eureka's key, while it accepted exactly the same key from dereel. This makes no sense at all to me.
Moving on, checked the handbook, which wasn't very helpful, and also found this page, which points me at keymaster.freebsd.org and keycheck.freebsd.org. And keycheck tells me that my key “works with the new sshd requirements”, but it doesn't when invoked with the same key from eureka.
|
More Academia strangenesses
|
Topic: technology, general, opinion | Link here |
Another message from Academia today:
19 N 01-05-2022 To acd@lem (1114) Academia.edu Connect with 2 co-authors of "Porting UNIX Software: From Download to Debug" on Academia.edu
What's wrong with this message? What's right about it? Firstly, I'm the only author of the work, but assuming that there had been others, wouldn't you think that I'd know them and not use Academia to connect with them, especially since they want me to pay money for the privilege?
And how do I get them to stop this false claim? I suppose the US way is to engage a lawyer, but they should have some way to contact them to correct incorrect claims on their part.
| Monday, 2 May 2022 | Dereel | Images for 2 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Alternate day syndrome
|
Topic: general, opinion | Link here |
Sometimes I get the feeling that things happen on alternate days: one day, like yesterday, I do a lot of stuff worth reporting on, and on the other day I report on it. Today I wasn't finished with my diary until 16:30, and as a result, I did almost nothing worth reporting on.
| Tuesday, 3 May 2022 | Dereel | Images for 3 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More Australia Post pain
|
Topic: general, technology, opinion | Link here |
Call on Yvonne's mobile phone today. Robyn, from Australia Post, clearly in relation to Yvonne's damage report. First questions about email, giving me ample opportunity to view my opinions about Australia Post's outsourced mail system. No, Australia Post doesn't outsource its mail. And of course they can't accept unsolicited mail: it needs to be an answer to an existing question so that they can associate it with an existing problem.
So I pointed out that the Subject: line contains the case number, which should eliminate any issues—she had no answer to that—and that email came from salesforce.com and not auspost.com.au. “That's us! Look at the TV ads and sports sponsoring! Everybody does it!”.
And why did she call on this number? She had Yvonne's last email in front of her, asking her not to use the phone number. Could she not read it? “Sorry, I can't tell you that. You're not authorized”. Yvonne wasn't there, so I couldn't do anything. Helen Miller was, so maybe I should have asked Helen to authorize her. After all, apart from the female voice she had nothing to go on.
Later she called back, still on the mobile phone, Yvonne spoke to her and brought the phone to me. Too fast. She hadn't understood that Yvonne had authorized her, and she had to do it all over again. But finally I got her to note the correct number to call on, and not to use Yvonne's mobile phone.
The discussion was interesting. Of all the mail sent, she only had two: the message they sent on 21 April, and Yvonne's reply on 29 April. No record of Yvonne's message of 21 April, none of the two messages I sent, none of the three copies of the receipts that we sent, not even a record of their own message of 27 April!
So finally she sent me another message so that I could reply to it, tripping over the obviously wrong australiapost@lemis.com, and finally (5th try or so) she got the information she wanted. No thanks to Australia Post's broken complaint handling.
Later Yvonne's phone went off again. Don, from Australia Post. No, no indication of another phone number to call. No indication that I was authorized to deal with him. So he had to hang up and go back and talk to Robyn.
Finally he called back on the right number and confirmed that the painting had been packed appropriately and still damaged. They're prepared to refund a maximum of $100, the default insurance sum. I suppose that's all he can do, though as I said, I didn't expect that the $100 was all that was applicable in case of negligence by Australia Post.
Still, it has been an experience:
Afterwards I took a look at other mail that I have received from Australia Post, like this notification:
Received: from mta15.notifications.auspost.com.au (mta15.notifications.auspost.com.au. [136.147.141.217])
by mx.google.com with ESMTPS id r12-20020a056214068c00b0044175bf603csi8336741qvz.213.2022.03.28.19.32.59
for <groggyhimself@gmail.com>
(version=TLS1_2 cipher=ECDHE-ECDSA-AES128-GCM-SHA256 bits=128/128);
Mon, 28 Mar 2022 19:32:59 -0700 (PDT)
So they do have some of their own email infrastructure. This, too, suggests a severely broken IT infrastructure.
|
The demise of national postal services
|
Topic: general, technology, opinion | Link here |
Is salesforce.com really a subsidiary of Australia Post? It certainly doesn't look like it:
# whois.markmonitor.com
Domain Name: salesforce.com
Registry Domain ID: 4623063_DOMAIN_COM-VRSN
Registrar WHOIS Server: whois.markmonitor.com
Registrar URL: http://www.markmonitor.com
...
Registrant Name: Domain Adminstrator
Registrant Organization: Salesforce.com, Inc.
Registrant Street: Salesforce Tower, 415 Mission Street, 3rd Floor
Registrant City: San Francisco
Registrant State/Province: CA
Registrant Postal Code: 94105
Registrant Country: US
Registrant Phone: +1.8006676389
Registrant Phone Ext:
Registrant Fax: +1.4159017040
Looking at their web site, they offer all sorts of services, so many that email doesn't even show on their first list.
And why should Australia Post, the company responsible for delivering mail in Australia, outsource their mail delivery to any company? That's their core business, or at least it was in the last millennium before email became dominant. Surely other large postal services don't do that! Off looking at MX records, and found:
| USPS (USA) | usps-com.mail.protection.outlook.com. | |
| Deutsche Post (Germany) | mx1.dhl.iphmx.com. | |
| Hong Kong Post (Hong Kong) | mail.hongkongpost.hk | |
| (UK) gpo.uk | For you, my friend, a special price | |
| Royal Mail (UK) | royalmail-com.mail.protection.outlook.com. |
gpo.uk belongs to a domain squatter! And apart from Hong Kong Post, they all outsource their email! I thought that Germany might be an exception, since DHL belongs to them, but that's not the domain name:
Domain Name: iphmx.com
Registry Domain ID: 1530264571_DOMAIN_COM-VRSN
Registrar WHOIS Server: WHOIS.ENOM.COM
Registrar URL: WWW.ENOM.COM
Registrar: ENOM, INC.
Registrant Organization: REDACTED FOR PRIVACY
Registrant Street: REDACTED FOR PRIVACY
Registrant City: REDACTED FOR PRIVACY
Registrant State/Province: CA
What privacy? enom (their styling) makes a business out of domains and email. But clearly not a subsidiary of Deutsche Post.
What really gets me, though, is that so many can't even get DNS records for the servers through which their mail goes. That's trivial, and they should be able to do it better than they way they do. After all, I have:
=== grog@eureka (/dev/pts/18) ~ 29 -> nslookup -q=mx lemis.com
lemis.com mail exchanger = 10 mx0.lemis.com.
lemis.com mail exchanger = 20 mx1.lemis.com.=== grog@eureka (/dev/pts/18) ~ 30 -> nslookup mx0.lemis.com
Name: mx0.lemis.com
Address: 121.200.11.253=== grog@eureka (/dev/pts/18) ~ 31 -> nslookup 121.200.11.253
253.11.200.121.in-addr.arpa name = 121-200-11-253.79c80b.mel.nbn.aussiebb.net.
|
Exploring thesauruses
|
Topic: language, opinion | Link here |
What's this?
|
|||||||||||||||||||||||
If it weren't for the upper part, it would be something like a small sideboard. But with it? The problem is compounded by the certainty that people in different English-speaking countries use different words. Time for a thesaurus. There are plenty of them, of course.
First to the State Library of Victoria, who supply me free access to the Oxford English Dictionary. Yes, their suboptimal search engine returns an Oxford Thesaurus, but when I follow the links it divides like a hydra and gives me nothing useful, though I can purchase a paper copy—not exactly what I had in mind.
OK, try online: thesaurus.com. For “sideboard” it gave me “cupboard” (close), “buffet” (not here), “closet” (something that I can't really associate with anything like a sideboard), “credenza” (a word that I had not heard of before, proving my assumption that different countries have different vocabularies) or “cellarette”. Nothing appropriate.
Further investigate showed that “credenza” might even be an appropriate term. The OED tells me that it's “A sideboard, free-standing cupboard, or storage chest, originally Italian or of Italian style”. Problem: my guess is that nobody round here would know what I meant.
Today Helen Miller was here, so I asked her. “Dresser” or “hutch”. Words I had never heard of before in that connection. A “dresser” is a dressing table, after all. But OED knows better: a dresser is
2. A sideboard used for storing crockery, glasses, cutlery, and other dining items, typically consisting of several display shelves mounted above drawers and a cupboard.
Pretty much an accurate description. Why “dresser”? Ah, that comes from the first meaning.
1. A kitchen table used for preparing or dressing food prior to serving; (also) a side table in a dining room or hall from which food or dishes are served. Now historical.
And “hutch”? I automatically thought of rabbits, but OED tells me (years in italics):
1303 A chest or coffer, in which things are stored.
And rabbit hutches? That came much later:
1607 A box or box-like pen or ‘house’ in which an animal is confined, as a rabbit-hutch.
So what do I call the thing? I really don't know. “Dresser” is the closest, but if I didn't know it, there will be many others.
It wasn't until later that I looked up Wikipedia, which came up with even more terms, including madia and chiffonier,
In passing, discovered a Thesaurus link on the OED descriptions, sadly not very helpful. And thesaurus.com seems remarkably simplistic. I signed up for their ”synonym of the day” and got “match the word” questions, like:
Which one is a synonym for mean? intend, abscond or mend?Which one is a synonym for direct? order, chaos or confusion?
With a bit of thought, it's easy enough to find the answer, but I find it a little mean when the answer page states “intend is a synonym of mean”.
| Wednesday, 4 May 2022 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Scanning old photos again
|
Topic: photography, history, technology, opinion | Link here |
Mail from Patrick Cheah today, asking for permission to use this photo:
|
||||||||||||||||||||||||
Why that? It's for a Facebook group dedicated to the Straits Echo, which seems to have only just survived the time of this photo. And that's the building in the middle:
|
|
||||||||||
OK, can I improve on it? It's old enough that I still have the negatives, some of the first that I took with my Pentax SV. Get them out and scan them.
Oh. Not as simple as I had thought. As I had noted, I had had difficulties with developing the film, and large parts were undeveloped.
Next, where are my details about scanning photos, in particular negatives? I didn't find them until too late: they're here. Found out by trial and error that DIGITAL ICE (frozen fingers?) was a bad idea for black and white negatives, and finally scanned them in.
It proved that the photo that Patrick was interested in was one of a series of test photos that I took after buying a Pentax SV for Carmel Paull. That meant that I didn't have much in the way of details about the lenses, not even a clear name: what I wrote in my notes could be something like Kawanon or Komura. So I had to modify my createexif script to add appropriate information.
And then the (Epson) scanner “fixed” a problem that I had had with the old Canon scanner: it had almost no understanding of framing and cropping, leaving lots of space around the image. The Epson didn't: it carefully removed even the edges of the image. In this case, it meant that the roof of the building was trimmed. Here the results from Canon and Epson (run the cursor over an image to compare it with its neighbour):
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Things still weren't done. My next step, particularly important here, was to “optimize” the images. For that I use Perfectly Clear, which usually works well.
This time it didn't. It crashed, repeatedly. Why? Maybe it's not intended for monochrome images, but that's no reason to crash. Instead I had to go back to Ashampoo “Photo Optimizer”, with which I have had mixed results. But today I couldn't even remember how to run it! It only offered me one image at a time. It took several minutes and some RTFM to discover that the GUI was configured to hide the selection window, after which things ran relatively smoothly.
Much later I think I have identified my scrawl. The lenses I tested were the 28 mm f/3.5 mm Super Takumar (first two images, and the lens that I bought 5 months later), the 35 mm f/3.5 mm Super Takumar (third image), the Kawanon 35 mm f/3.5, the Komura 28 mm f/3.5 and the Komura 35 mm f/2.5. The unevenness in the first 4 images is due to the development problems.
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
It's interesting to notice how unsharp the corners are, including in the photo that Patrick chose (first one). That wouldn't happen nowadays.
|
|
||||||||||
|
|
||||||||||
|
Custom spam
|
Topic: language, technology, opinion | Link here |
Spam's a fact of life, but I find it interesting to try to guess the languages in which it arrives. Recently I got this one:
To undisclosed- ( 131) Mrs. Kristalina Kæri styrkþegi!
Icelandic? It had lots of text:
Date: Fri, 29 Apr 2022 09:28:05 +0000
From: "Mrs. Kristalina" <swiilasmith641@gmail.com>
Subject: Kæri styrkþegi!
Reply-To: info.ubabankofice02@gmail.com
Kæri styrkþegi!
Reyndar höfum við fengið heimild frá nýjum fjármálaráðherra og stjórnarráði
Peningamálaeiningar Sameinuðu þjóðanna til að rannsaka óinnheimta fjármuni
sem lengi er skuldað í ríkiskörfu Sameinuðu þjóðanna, sem hefur leitt til
þess að ruglaða eigendur trúðu því að þeir hafi verið blekkt af svindlarum
sem nota nafn Sameinuðu þjóðanna við rannsókn okkar. Samkvæmt
gagnageymsluskrá kerfisins okkar með netfanginu þínu er greiðslan þín á
lista yfir 150 styrkþega sem flokkaðir eru sem: Ógreiddur happdrættissjóður
/ Ógreiddur happdrættissjóður / Ófullnægjandi erfðir millifærslu /
Samningssjóðir.
...
Tengiliður: MR. KENNEDY UZOKA
Skaðabótadeild (United Bank for Africa)
Yes, no doubt: it's Icelandic, and some kind of phishing. But what's the reference to the “United Bank for Africa”? And given that there are something like 314,000 native speakers of Icelandic, too few to even figure in List of languages by number of native speakers, and that a large proportion of them speak English, it's not clear what this spam wants to target.
| Thursday, 5 May 2022 | Dereel | Images for 5 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More eBay surprises
|
Topic: technology, general, opinion | Link here |
Yesterday I received a second refurbished Braun electric toothbrush. I hadn't bothered mentioning the first (early last month), but it had been satisfactory, so I ordered another one.
It was at least interesting from the point of view of the tracking. It took a long time, and every time I checked, it showed “Received and ready for processing”, dated 19 April. But on Tuesday I received information that it was ready for pickup. What did the tracking look like? eBay told me that it was both “Sent” but not yet “in transit”, but also ready for pickup:
|
|
|||||||||
And the detail tracking? Take your choice of:
|
|
|||||||||
|
|
|||||||||
In principle it's the same, but they change the names to protect the innocent. On 27 April it was either processed in Chullora (according to eBay) or Sydney (Australia Post). On Tuesday it was processed at 7:21 either in Wendouree (eBay) or “Winter Valley” (AusPost, a name that neither I, Google nor Wikipedia know). The one thing that they had in common was that it seemed to have been left for 8 days before they did anything with it.
But it's defective! Doesn't charge! OK, return. I've done enough of them that eBay clearly thinks it's time to change the rules. Now I must return it first and print out the label that they gave me:
|
|||||||||||||||||||||||
What label? None, of course, and there was nothing I could do to produce one. Once again I had to fight the eBay web site to get a call back from a human being or close copy. First, after I replied to the call with “Greg Lehey”, he asked for my name. And then “thank you for being an eBay customer for 23 years”, which I suppose is something. But clearly it's a “follow the script”, and because my answer wasn't in the script, I ended up with questions like “have you already returned it?”.
And he wasn't sure either, putting me on hold for a considerable period of time. Finally sent he me a label to print out.
You'd think that eBay would have their act together by now. Why did I get an empty label?
|
More scanning
|
Topic: photography | Link here |
It's been 55 years since my father and I left on our Asia Trip by car from Singapore (my father) or Kuala Lumpur (myself) to London. I took a (far too small) number of photos of the drive and scanned them in years ago on my ill-fated Canon 9900F scanner. The results are frequently terrible:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's clearly an exposure issue at the time, something that I can't change. But since then I have not only changed scanners (now the Epson “Perfection” 4990, not perfect, but much better) and also found the exposure record for the photos. I have also refined my scanning technique. So, since I currently have the scanner set up: scan them and postprocess them all over again.
It's slow business, and once again I'm left with the feeling that there's something very wrong with the way I developed the images at the time, probably an issue with the “substitute” developer. It'll keep me going for a while.
| Friday, 6 May 2022 | Dereel | Images for 6 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More slide scanning
|
Topic: photography, technology, opinion | Link here |
Scanning slides is a slow process, and there are many issues to solve or work around. As I said yesterday, the Epson “Perfection” 4990 is not perfect, especially not its software. There are many things that could be done better, including cropping: there's no way to adjust the crop. I had that issue with the Penang Street photos that I processed on Thursday, and it hit me in a different guise today. It seems that it doesn't handle cropping portrait format slides. I have to put them in in landscape orientation and then rotate them. The difference is significant, though it wasn't immediately obvious that the landscape format also truncates (run the cursor over an image to compare it with its neighbour):
|
|
|||||||||
|
|
|||||||||
Apart from that, first indications are that the results are much better. I haven't finished processing the scanned images, but these two show, the ones I picked at random show a definite improvement:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
They're still not good: the first one was just plain underexposed, and it wasn't possible to recover more detail, though I'll try harder when I'm finished. And the second one has lots of dirt on it, made worse by overexposure. But the difference between the old and the new photos is very significant.
|
More temperature measurements
|
Topic: technology, food and drink, general, opinion | Link here |
One of our freezers is clearly defective. When I open it I frequently find that the temperature on the front of the baskets is round 0°: they're moist. But it seems to cool. Is it a thermostat issue? Hysteresis? Offset? For that I need to know how the cooling cycle works.
I've measured freezer temperatures in the past with a meat thermometer, but they show only the current temperature. I really need a min/max thermometer. Surely they're available on eBay?
Indeed, and not even expensive. You can get thermocouple probes on their own for about $13, but you can also get a dual thermometer (digital, of course) with min/max and two probes for $35. That sounds like a bargain.
But wait, there's more. How about one with WiFi connectivity? Four probes, and only $84. Oh. For reasons that I suspect are due to poor firmware design, the minimum temperature is 0°. It's not the probe: all the thermometers I found use type K probes, which have a range of -200° to 1350°, far beyond what I would ever want to measure.
Further searching became even more expensive. It seems that the most common WiFi-controlled thermometers, including the $84 one (IBBQ-4T) come from Inkbird (a name which, for some reason, I find hard to remember). And they have another model, Inkbird IBBQ-4BW that seems almost identical, except that it costs $122 and has a strangely specified temperature range:
Short Time Measurement Temperature Range: -30℃~+300℃/-22℉~+572℉
Continuous Monitoring Temperature Range: -20℃~+250℃/-4℉~+482℉
Now why that? In any case, it would do the trick, though it's not clear what the real difference is. Should I buy it? Yes. Apart from the GAS value and the convenience, I could use it to explore meat cooking times. That's what it seems to be designed for, as the BBQ in the product name emphasizes. It's not so much min/max as continual readings, and the app produces at least the kind of temperature graph that it thinks appropriate, which could certainly be useful.
|
eBay strikes again
|
Topic: technology, opinion | Link here |
The thermometer I was looking at, the Inkbird IBBQ-4BW, was advertised for about $122, with an additional $3 postage. But eBay offered 10% discount with a promotional code. OK, Buy It Now, apply the code, get a price round $112.
But for 50¢ more I could get express postage. Normally I don't bother, but that's a ridiculously low supplement. OK, select. And my discount went away!
Back to normal postage. Still no discount! Much examination with different browsers and systems, and finally discovered that they had cancelled the discount in mid-purchase! BAD eBay.
More looking and found an item that they hadn't seen fit to display before, although it matched the search criteria: IBBQ-4BW and an IHT-1P penknife-format temperature probe, for $115 including postage, somewhat less than $3 more. I don't really need the IHT-1P, but under the circumstances it seems to be the best choice.
Buy It Now. “Are you human?”. Yes, idiot, which is more than I can say for you. But no, I was presented with a never-ending sequence of these HORRIBLE CAPTCHAs before I could finally pay my hard-earned money. The whole matter took 48 minutes!
Looking back in my diary I see that I have been grumbling about eBay for decades. But why do they keep trying to annoy me?
| Saturday, 7 May 2022 | Dereel | Images for 7 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The riding arena: first
|
Topic: animals, general | Link here |
Troy Addicoat along with his mate in mid-morning, just as we had thought he wasn't coming, and did the first step in putting up the shed over the riding arena:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And that's all. Just holes and piles of dirt:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Now an inspector needs to come (probably Monday or Tuesday) and confirm, yes, there are holes, and they're wide and deep enough. After that he can fill them with concrete, probably on Thursday, and wait for it to cure, about a week. Then he can finally start erecting the thing.
|
The legacy of Bruce Evans
|
Topic: technology, history, opinion | Link here |
Bruce Evans died in mid-December 2019, just before the COVID-19 pandemic. We coopted Peter Jeremy to tidy up his computer remains, and he found these:
|
|||||||||||||||||||||||
Who wants them? Peter offered them to Warren Toomey and me. Warren wasn't interested. I thought that some of them could be useful, so I put my hand up. But how do you transport them? Shipping could be expensive, especially as one of the computers (HP Z800, not exactly a name I would have thought modern) is very heavy. So I went looking for people going by car from Sydney to Dereel. And there the problems started. Not only did nobody want to travel any more, there were threats that interstate travel would be prohibited.
Finally things were looking up, and about this time last year we had planned a clinic with Anke Hawke. Jane Ashhurst (who visited us last month) was coming by car. So her husband Scott picked up the computers, loaded them into the car, and they were on their way south when they heard the announcement: state borders are closed. They probably could have made it across before the closure was enforced, but then they wouldn't get home again for who knows how long? And we had to cancel the clinic, amongst a number of other things. It's now scheduled for next week after a delay of nearly a year.
And then Scott found his freedom too, taking his ute and caravan on a roundabout trip to the Simpson desert. Today he came by with the hardware:
|
|||||||||||||||||||||||
They're here, after nearly 2 years!
|
Running bde's computers
|
Topic: technology, opinion | Link here |
I really don't have time to look at bde's computers at the moment, but how difficult can it be to fire up a laptop? More than I thought: after the extended time, the batteries were completely discharged. OK, charge batteries and try:
|
|||||||||||||||||||||||
Hewlett-Packard has never been my favourite brand, but that message (apart from the problem it reports) really presses one of my buttons. How can a reputable maker print messages that substitute backslashes (\) for slashes (/)?
But what's wrong with the laptops? One of them claimed that there was some issue with the BIOS, presumably a discharged CMOS battery, but both of them wouldn't boot. OK, boot from a USB stick. And they both got stuck in a reboot loop!
That's certainly less successful than I had expected. What has gone wrong there? Do HP laptops require special TLC to boot? Mañana, if not pasado mañana,
| Sunday, 8 May 2022 | Dereel | Images for 8 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Suffering Buddleja
|
Topic: Stones Road house, gardening | Link here |
Since we've been here in Stones Road—concidentally 7 years ago today—our garden flowers have suffered greatly. In particular the Buddleja x weyeriana, which had flourished in Kleins Road, have barely survived. So I was pleasantly surprised when I managed to resurrect the last dying plant. Here photos from three months ago and last month:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And then on Friday I found this:
|
|||||||||||||||||||||||
What's wrong there? Water? I tried that (as the photo shows), and it seems to be the case. Today it looks like this:
|
|||||||||||||||||||||||
It hadn't seemed particularly dry before, but it seems that Buddlejas are particularly sensitive to soil water content: I think that my previous failures could have been due to too much water.
|
More bde computer stuff
|
Topic: technology, opinion | Link here |
So far my experience with Bruce Evans' computers hasn't been overly positive. I can't boot either laptop from disk or USB stick. How about a DVD? Dragged out an old DVD, which might be appropriate for these machines, one 10, the other 20 years old, and tried to boot.
The older one, a Compaq nx6325, booted from DVD, but then the display went blank and stayed blank. It's ancient (“Designed for Windows XP”); should I even bother? I have a Dell laptop of that era, and it's too old for anything useful.
The newer one, an HP EliteBook 8460p, came up with a particular surprise. I couldn't insert the DVD! The hub seemed too large for the DVD:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Looking at the photos, it seems that there are two issues: firstly, the hole in the DVD is really smaller than the hub mechanism, but it seems that it's supposed to fit half on top of the hub. But there's a spring-loaded catch, and it seems that one of the springs is missing. That wasn't visible until I took and processed the photos. I'll have to go back and try again.
But wait, there's more! Well, it didn't seem so: the accessories box just contained a few random cables and mice:
|
|||||||||||||||||||||||
Picked it up to put it away, but it was surprisingly heavy. Lifting the cables showed:
|
|||||||||||||||||||||||
An HP EliteBook 8570p, newer than the other two! Yes, I knew about this machine—it was in the photos Peter sent me—but I had assumed it had gone to a different owner.
OK. Boot. Failure! Again! And it seems that this one won't even try to boot from a USB stick, so I really needed the DVD. And it did boot from the DVD, showing me that the disk had no boot sector. How did that happen? Copied the content to eureka for further examination.
Somehow this is all much more work than I expected.
|
Wölferl and Constanze
|
Topic: general, opinion | Link here |
Yvonne's latest acquisition:
|
|||||||||||||||||||||||
She has dubbed them „Wölferl“ and „Constanze“. I don't know what to make of the fact that Constanze is rough and Wölferl is smooth.
|
Credit card renewal
|
Topic: technology, opinion | Link here |
My credit card expires at the end of the month, and I'm getting various reminders to update my details with people authorized to deduct from the card.
First Vonex, the company designed to make MyNetFone look good. Initially I couldn't even access the site. And when I could, it rejected my credit card details without explanation.
Then came ALDImobile. No problem accessing their web site—some things have improved—but I only had the option of removing my payment method, not updating it. Called them up and was connected to mumble, which proved to be a way to pronounce “Hazel”. She talked me through it: yes, first remove the payment method, then you can enter a new one. Obvious, isn't it? No. Asked her to register a complaint, but I doubt she will do it.
Then Vonex. This mumble made Hazel appear crystal clear. At the end of the sentence I wasn't sure what language she was speaking, but ultimately she spelt it out: “Gail”. Am I getting hard of hearing, or are phone “consultants” getting hard of speaking?
She first wanted me to identify myself. Account number? Don't know. How about a phone number? That should be on her display, of course. But then I saw that there was an account number on their non-functional update page. “OK, I'll see if I can pull that up”. Then she wanted my date of birth again. What do I know? I never registered a date of birth for them, so I suspect that I never told them one.
Finally we found that my address was another form of identification. She took down my payment details and told me that all would be well. Why couldn't I do it online? Others had reported that too; they're updating their web site. It seems that they've been doing that (or needing to do it) for at least 6 months.
| Monday, 9 May 2022 | Dereel | Images for 9 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Still more laptop problems
|
Topic: technology | Link here |
So now I have an image of the complete disk of the HP EliteBook 8570p on eureka. It should be possible to recover an overwritten master boot record by searching for the magic numbers of the file systems. Started off a superficial search with strings(1). But it found nothing! A quick comparison with the disk on eureka showed that it should have found lots of things. Is the disk encrypted?
During discussion on IRC, mentioned that I wasn't able to boot from a USB stick on that machine. Matti Kupiainen suggested that maybe the USB bus on one side of the laptop might be defective. Tried on the other (left) side. Bingo! But I ended up in the same boot loop that I had with the other EliteBook. Tried on the right side again. It booted, but of course went into the loop again.
OK, get an older boot image, though this machine is only about 3 years old, presumably purchased only shortly before Bruce died. The image was about 1.1 GB, just too big for a 1 GB card. But that's OK. I have this card:
|
|||||||||||||||||||||||
Put it in dereel, which reported:
da0 at umass-sim0 bus 0 scbus3 target 0 lun 0
da0: <Lexar LRWM04U 1.00> Removable Direct Access SPC-4 SCSI device
da0: Serial Number 201408282030
da0: 400.000MB/s transfers
da0: 3763MB (7706624 512 byte sectors)
da0: quirks=0x2<NO_6_BYTE>
Huh? That's 4 “GB”! How come the label only says 2? Who's lying? The obvious thing is to try to read it:
=== root@dereel (/dev/pts/0) ~ 67 -> dd if=/dev/da0 of=/dev/null bs=1m
3763+0 records in
3763+0 records out
3945791488 bytes transferred in 48.255269 secs (81769133 bytes/sec)
I've seen cards before with fake labels showing more content than they really have, but this is the first time I've seen one showing less. It's also much faster than the other cards I have: the 8 GB card manages only about 11 MB/s. Still, it works, and I tried an older image (FreeBSD 11) on it, unfortunately with the same results on all three machines. It seems that HP laptops have quirks that make it non-trivial (so far impossible) to boot from USB.
OK, I can boot the EliteBook 8570p from DVD. In principle that should work for the EliteBook 8460p too. Why can't I physically insert a DVD? On closer examination the drives look almost identical. Here the 8460p, then the 8570p (run the cursor over an image to compare it with its neighbour):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly the hub on the 8460p is damaged. How easy is it to swap DVD drives between the two machines? Or is it possible, with sufficient care, to put a DVD on the remains of the hub? Do I care?
|
Destroying Bruce Evans' legacy
|
Topic: technology, general, opinion | Link here |
Clearly one person who might shed some light on my issues with Bruce Evans' laptops is Peter Jeremy, whom we entrusted to pick them up and distribute them. So I asked him whether the disks were encrypted:
> I've taken a look at the laptops so far, and I've run into trouble
> with all of them. All of them claim to have no bootable system, ...
I did wipe all the disks, so that's why none are bootable
WHAT? That's wanton destruction! I'm not exactly lost for words, but I'm amazed and deeply disappointed. What has Peter done with this action, which must have taken him some time? We discussed picking up Bruce's computers a couple of years ago, and I recommended asking Peter, because I considered him a trustworthy, reliable person. We discussed in some detail what to do with them and the data on them. Peter chose to put the whole thing, round 2.7 TB, on Google Drive, a storage format completely alien to most of us, but if it's still there, I can't find it. One thing that most definitely was not discussed was destroying the originals of the data.
So: it looks as if this information is gone forever, unless Peter can come up with backups. But why did he delete the information? We didn't discuss that, and he had no authority to do so. This is particularly concerning because I recommended him as a person of trust.
Bruce Evans was a remarkable man. He was also known for having patches “in a drawer” at home, presumably on one of these computers. Are they now gone forever?
And then there's the question of how he configured his systems. This is particularly relevant for the laptops, but my guess is that he had some clever ideas about how to configure and run FreeBSD systems. Also all gone.
What do I do with the machines? The two towers are probably worth salvaging, as is the EliteBook 8570p, but the other two laptops are now effectively junk: the effort required to get them running again is not worth it, especially for the older one.
| Tuesday, 10 May 2022 | Dereel → Ballarat → Dereel | Images for 10 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Blood test and toothbrushes
|
Topic: health, general | Link here |
Into Ballarat today for my twice-yearly blood test. On the way, yet another attempt to get sensible dental equipment. After my experience with the refurbished toothbrushes, decided to get a new one instead, possibly a better model. I had paid $35 for a refurbished brush holder only (no power supply), and I found the same model, Braun Oral-B Pro 500, brand new with brush and charger for $50. That seemed a better option.
But how about something better? There's a big jump to the next one they had in stock, for $105. What do I get for that? It seems that it has a pressure sensor that displays different colours in front of your mouth. Isn't that clever? Not in my book.
And why are there no angled interdental brushes any more? It seems that I can get things like this from name brands for about 80¢ per piece, or these for 7¢:
|
|||||||||||||||||||||||
To make it easier for the name brands, these (and all similar ones) have a mint flavour, which puts them out of the question. Why do people do these things?
|
Bruce Evans legacy, continued
|
Topic: technology, opinion | Link here |
So what happened to Bruce Evans' legacy? Looking back through the email exchange, it seems that at the time Peter Jeremy went to some trouble to save all the data, making his subsequent behaviour even more surprising. Found a number of what I think are disk images on Google Cloud, and started downloading the smallest of them, 15 GB compressed. And that's about all I did; I'll find what it is tomorrow.
But I've been in this kind of position before. 20 years ago I wrote
The time I spent working on this problem this week cost more than a mid-size PC.
|
Understanding my ssh problems
|
Topic: technology, opinion | Link here |
Why can I access freefall.freebsd.org from some machines and not from others, though I use the same ssh key in every case? Off to compare:. I added the colour highlighting later to make things more recognizable:
=== grog@eureka (/dev/pts/44) ~/Photos/19670427 74 -> cat ~/.ssh/id_rsa.pub ; uname -a; ssh-add -l; ssh freefall
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDVYUjpMuqVqp/JR8ex/fzRpTb+ygOQDU/aaS40TNdCtXgcGjzYfAt0N5QtUX2IAJlgYvLaao+hAoXXfvKTs0rybhWBKuApxTOB7gVEYtNUz3qIkHsBun4FB8zNYCvL2DVTlB46xkbNYZoZB+Yh5TrbdiXyvO917k2Rj6jUGlcc1oQn6xqlSE0/qSByhsGfPV99S72DJdhGOCwRPbt35Co85dlJ6zh7Mj2+Im0m41SblbJ2JBvf4IPAn+YumkFWt3FJy1zKRgl3kxfh8yDZxgqrBaylGhaJZ9vCBW7WICKKZMm+ewuxFX9nFtPvB4OirRxCLgN6m3keSFpSH3nqS2Rb grog@dereel.lemis.com
FreeBSD eureka.lemis.com 10.2-STABLE FreeBSD 10.2-STABLE #2 r290972: Wed Nov 25 11:38:38 AEDT 2015 grog@stable.lemis.com:/usr/obj/eureka/home/src/FreeBSD/svn/10/sys/GENERIC amd64
2048 3b:c3:00:e1:ea:57:43:7f:14:78:90:58:c9:61:bf:d8 /home/grog/.ssh/id_rsa (RSA)
Permission denied (publickey).
=== grog@dereel (/dev/pts/3) ~ 16 -> cat ~/.ssh/id_rsa.pub ; uname -a; ssh-add -l; ssh freefall
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDVYUjpMuqVqp/JR8ex/fzRpTb+ygOQDU/aaS40TNdCtXgcGjzYfAt0N5QtUX2IAJlgYvLaao+hAoXXfvKTs0rybhWBKuApxTOB7gVEYtNUz3qIkHsBun4FB8zNYCvL2DVTlB46xkbNYZoZB+Yh5TrbdiXyvO917k2Rj6jUGlcc1oQn6xqlSE0/qSByhsGfPV99S72DJdhGOCwRPbt35Co85dlJ6zh7Mj2+Im0m41SblbJ2JBvf4IPAn+YumkFWt3FJy1zKRgl3kxfh8yDZxgqrBaylGhaJZ9vCBW7WICKKZMm+ewuxFX9nFtPvB4OirRxCLgN6m3keSFpSH3nqS2Rb grog@dereel.lemis.com
FreeBSD dereel 14.0-CURRENT FreeBSD 14.0-CURRENT #3 main-n255108-9fb40baf6043: Fri Apr 29 09:53:15 AEST 2022 grog@dereel:/usr/obj/eureka/home/src/FreeBSD/git/main/amd64.amd64/sys/GENERIC amd64
2048 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo /home/grog/.ssh/id_rsa (RSA)
Last login: Sat May 7 23:04:30 2022 from ffm.lemis.com
=== grog@lagoon (/dev/pts/2) ~ 7 -> cat ~/.ssh/id_rsa.pub ; uname -a; ssh-add -l; ssh freefall
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDVYUjpMuqVqp/JR8ex/fzRpTb+ygOQDU/aaS40TNdCtXgcGjzYfAt0N5QtUX2IAJlgYvLaao+hAoXXfvKTs0rybhWBKuApxTOB7gVEYtNUz3qIkHsBun4FB8zNYCvL2DVTlB46xkbNYZoZB+Yh5TrbdiXyvO917k2Rj6jUGlcc1oQn6xqlSE0/qSByhsGfPV99S72DJdhGOCwRPbt35Co85dlJ6zh7Mj2+Im0m41SblbJ2JBvf4IPAn+YumkFWt3FJy1zKRgl3kxfh8yDZxgqrBaylGhaJZ9vCBW7WICKKZMm+ewuxFX9nFtPvB4OirRxCLgN6m3keSFpSH3nqS2Rb grog@dereel.lemis.com
FreeBSD lagoon.lemis.com 12.1-STABLE FreeBSD 12.1-STABLE r355358 GENERIC amd64
2048 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo /home/grog/.ssh/id_rsa (RSA)
agent key RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo returned incorrect signature type
grog@freefall.freebsd.org: Permission denied (publickey).
=== grog@teevee (/dev/pts/6) /spool/Series/Hubert-und-Staller/10 42 -> cat ~/.ssh/id_rsa.pub ; uname -a; ssh-add -l; ssh freefall
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDVYUjpMuqVqp/JR8ex/fzRpTb+ygOQDU/aaS40TNdCtXgcGjzYfAt0N5QtUX2IAJlgYvLaao+hAoXXfvKTs0rybhWBKuApxTOB7gVEYtNUz3qIkHsBun4FB8zNYCvL2DVTlB46xkbNYZoZB+Yh5TrbdiXyvO917k2Rj6jUGlcc1oQn6xqlSE0/qSByhsGfPV99S72DJdhGOCwRPbt35Co85dlJ6zh7Mj2+Im0m41SblbJ2JBvf4IPAn+YumkFWt3FJy1zKRgl3kxfh8yDZxgqrBaylGhaJZ9vCBW7WICKKZMm+ewuxFX9nFtPvB4OirRxCLgN6m3keSFpSH3nqS2Rb grog@dereel.lemis.com
FreeBSD teevee.lemis.com 12.1-STABLE FreeBSD 12.1-STABLE r359522 GENERIC amd64
2048 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo /home/grog/.ssh/id_rsa (RSA)
agent key RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo returned incorrect signature type
grog@freefall.freebsd.org: Permission denied (publickey).
=== grog@tiwi (/dev/pts/1) ~ 13 -> cat ~/.ssh/id_rsa.pub ; uname -a; ssh-add -l; ssh freefall
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDVYUjpMuqVqp/JR8ex/fzRpTb+ygOQDU/aaS40TNdCtXgcGjzYfAt0N5QtUX2IAJlgYvLaao+hAoXXfvKTs0rybhWBKuApxTOB7gVEYtNUz3qIkHsBun4FB8zNYCvL2DVTlB46xkbNYZoZB+Yh5TrbdiXyvO917k2Rj6jUGlcc1oQn6xqlSE0/qSByhsGfPV99S72DJdhGOCwRPbt35Co85dlJ6zh7Mj2+Im0m41SblbJ2JBvf4IPAn+YumkFWt3FJy1zKRgl3kxfh8yDZxgqrBaylGhaJZ9vCBW7WICKKZMm+ewuxFX9nFtPvB4OirRxCLgN6m3keSFpSH3nqS2Rb grog@dereel.lemis.com
FreeBSD tiwi.lemis.com 13.1-STABLE FreeBSD 13.1-STABLE #2 stable/13-n250531-6182c7881829-dirty: Fri Apr 22 12:53:34 AEST 2022 grog@tiwi.lemis.com:/usr/obj/eureka/home/src/FreeBSD/git/stable-13/amd64.amd64/sys/GENERIC amd64
2048 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo /home/grog/.ssh/id_rsa (RSA)
agent key RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo returned incorrect signature type
grog@freefall.freebsd.org: Permission denied (publickey).=== grog@tiwi (/dev/pts/1) ~ 15 -> ssh -V
unknown option: '-V'
OpenSSH_8.8p1, OpenSSL 1.1.1n-freebsd 15 Mar 2022
=== grog@ffm (/dev/pts/1) ~ 14 -> cat ~/.ssh/id_rsa.pub ; uname -a; ssh-add -l; ssh freefall
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDVYUjpMuqVqp/JR8ex/fzRpTb+ygOQDU/aaS40TNdCtXgcGjzYfAt0N5QtUX2IAJlgYvLaao+hAoXXfvKTs0rybhWBKuApxTOB7gVEYtNUz3qIkHsBun4FB8zNYCvL2DVTlB46xkbNYZoZB+Yh5TrbdiXyvO917k2Rj6jUGlcc1oQn6xqlSE0/qSByhsGfPV99S72DJdhGOCwRPbt35Co85dlJ6zh7Mj2+Im0m41SblbJ2JBvf4IPAn+YumkFWt3FJy1zKRgl3kxfh8yDZxgqrBaylGhaJZ9vCBW7WICKKZMm+ewuxFX9nFtPvB4OirRxCLgN6m3keSFpSH3nqS2Rb grog@dereel.lemis.com
FreeBSD ffm.lemis.com 11.1-STABLE FreeBSD 11.1-STABLE #0 r326456: Sat Dec 2 13:06:23 UTC 2017 grog@ffm.lemis.com:/usr/obj/usr/src/sys/GENERIC amd64
2048 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo /home/grog/.ssh/id_rsa (RSA)
Last login: Tue May 10 07:14:42 2022 from 121-200-11-253.79c80b.mel.nbn.aussiebb.net
=== grog@lax (/dev/pts/1) ~ 1 -> cat ~/.ssh/id_rsa.pub ; uname -a; ssh-add -l; ssh -V; ssh freefall
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDVYUjpMuqVqp/JR8ex/fzRpTb+ygOQDU/aaS40TNdCtXgcGjzYfAt0N5QtUX2IAJlgYvLaao+hAoXXfvKTs0rybhWBKuApxTOB7gVEYtNUz3qIkHsBun4FB8zNYCvL2DVTlB46xkbNYZoZB+Yh5TrbdiXyvO917k2Rj6jUGlcc1oQn6xqlSE0/qSByhsGfPV99S72DJdhGOCwRPbt35Co85dlJ6zh7Mj2+Im0m41SblbJ2JBvf4IPAn+YumkFWt3FJy1zKRgl3kxfh8yDZxgqrBaylGhaJZ9vCBW7WICKKZMm+ewuxFX9nFtPvB4OirRxCLgN6m3keSFpSH3nqS2Rb grog@dereel.lemis.com
FreeBSD lax.lemis.com 12.0-RELEASE-p8 FreeBSD 12.0-RELEASE-p8 GENERIC amd64
2048 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo /home/grog/.ssh/id_rsa (RSA)
unknown option: '-V'
OpenSSH_7.8p1, OpenSSL 1.1.1a-freebsd 20 Nov 2018
Disconnected at Tue 10 May 2022 07:20:07 UTC
agent key RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo returned incorrect signature type
grog@freefall.freebsd.org: Permission denied (publickey).
What do I see here?
Off to look for the message on the web. This page tells me
This message means that the SSH connection negotiated a connection using an RSA key with a different signature algorithm, either SHA-256 or SHA-512. However, the SSH agent, when asked to make the signature for that connection, provided an SHA-1 signature, which isn't in compliance with the agent protocol.
But looking at the output, it seems clear that ssh-agent is returning an SHA-256 signed key.
Then this page comes up with a similar, less restrictive answer:
For this connection, your client and server have negotiated that a SHA512-based signature should be used instead. However, when the client asked ssh-agent to create the signature, ssh-agent ignored the new extension and just returned a SHA1-based signature instead. Usually this happens when the agent simply does not support the updated protocol.The solution is to upgrade the SSH agent. Latest versions of OpenSSH and PuTTY both have agents compatible with the new protocol; GnuPG's gpg-agent should also support this as of v2.2.6.
That's referring to Linux, of course. With FreeBSD this kind of thing should be impossible, just like upgrading only part of the ssh software is not easily possible. But looking more carefully, it doesn't seem to be correct. In the example discussed on that page, ssh -vvv returns:
debug3: sign_and_send_pubkey: RSA SHA256:[hash]
debug3: sign_and_send_pubkey: signing using rsa-sha2-512
agent key RSA SHA256:[hash] returned incorrect signature type
debug3: sign_and_send_pubkey: signing using ssh-rsa
That's signed with rsa-sha2-512, not SHA1. And that's what I see on tiwi too:
debug1: Offering public key: /home/grog/.ssh/id_rsa RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo agent
debug3: send packet: type 50
debug2: we sent a publickey packet, wait for reply
debug3: receive packet: type 60
debug1: Server accepts key: /home/grog/.ssh/id_rsa RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo agent
debug3: sign_and_send_pubkey: RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo
debug3: sign_and_send_pubkey: signing using rsa-sha2-512 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo
agent key RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo returned incorrect signature type
debug3: sign_and_send_pubkey: signing using ssh-rsa SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo
Why is that happening? How do I fix it?
|
Horse transport pain
|
Topic: animals | Link here |
Some time ago Emma Jefferson offered Yvonne a gift horse, but there were problems with the transport. Two months ago Yvonne wanted to go to Kybeyan to pick her up, but that too fell through. But finally the transport works—maybe. Only in the meantime Emma has decided to ask Real Money for the horse (named Dana). She was supposed to arrive tomorrow, but that transport also fell through, as I discovered after answering a call on Yvonne's phone, which she never has with her.
Now the transport will occur in three steps. The first is done: she's in Yass. Tomorrow she will go to Melbourne. And on Thursday if we're lucky, Friday if we're not, she'll finally be here.
But that's not all. When on Friday will she arrive? It would be better for her to have a few hours of daylight on arrival, and in any case, Anke Hawke is arriving in the evening to finally hold the clinic that should have taken place a year ago. I can see conflicts already.
| Wednesday, 11 May 2022 | Dereel | Images for 11 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Power fail!
|
Topic: Stones Road house, general, technology, opinion | Link here |
Shortly after 4:00 this morning it occurred to me that it had been quiet for some time. Normally I would expect to hear the air conditioner or the freezers. Power?
Yes! Once again a total power failure. OK, where's my torch? I really need to have something in a known place. Tried some likely places, but there was nothing there. Finally it occurred to me: my phone has a torch function, and I know where it is. One of the few times where it's really of use.
OK, check the PV inverter: no grid power. Out to check the meter. No problems! Back inside. No switches had tripped. Take the inverter out of circuit: power came back. So the inverter had failed.
What do I do now? Put the inverter back in circuit and see if things fail again? A good idea, if only to confirm the problem. But the inverter came back as normal and started charging. And it didn't fail again.
What causes that? After getting the machines back up, took a look at the inverter log and saw:
|
|||||||||||||||||||||||
Just after midnight the battery charge dropped to 20%, which should have triggered a recharge. But the inverter did nothing, and ultimately, round 0:41, it dropped to 5%, at which point the inverter cut out supply. For reasons that aren't obvious, it started charging again at 1:23, but only made it to 20% before giving up again. At 4:13 I took the inverter out of circuit, shown by the dip in the Grid/Genset Voltage curve (magenta). And when I turned it on again, the inverter functioned normally and continued to do so
Clearly there's a problem with the inverter here. It's not the circuitry (bypass switch, for example), since the inverter reported normal grid voltage the whole time. It just didn't charge.
There's more information in the inverter log, in fact, more than in the data that I suck out of the inverter, including things like inverter status and inverter alarms and grid supply frequency. The inverter alarms. The latter raised a 02000000 (is that octal?) when the battery discharged completely, but I think that's all it was trying to say. Sent mail off to Tomas Kucera and Fyodor Togornikov. Hopefully they can do something about it.
|
Power fail recovery, next attempt
|
Topic: technology, opinion | Link here |
Recovering from a power failure is always a pain, though it's getting easier. For reasons I still don't understand, eureka's hardware doesn't recover from a power outage: it requires manual operation of the power switch. And then it doesn't boot automatically: it stops at the Boot: prompt and waits for me to confirm with Boot. But then it came up without trouble.
But there are still so many loose ends. I started writing all this stuff in a HOWTO page, but it still seems to be greatly out of date. I really should keep it up to date, and also with enough information for me to be able to apply it. How do I start the weather station software? It needs to run as root, probably because I haven't set the device permissions correctly. And the link check script? Also.
|
Inspecting the holes
|
Topic: Stones Road house, animals, general | Link here |
Round 11:00 this morning Gagan Deep of Wyndham Building Permits came along to take a look at the foundation holes for the shed cover. With him he brought a tablet with the old planning permit for the old location. And he noticed that something was wrong: now the holes were only 1 m from the boundary, whereas in the old permit they were 2 m away (from the internal fence, but he doesn't seem to have noticed that). I pointed out that we had an amended planning permit, and even brought it out for him, but no, he needed an amended building permit. So he noted that the holes were acceptable in themselves, but we needed to amend the building permit.
GRR! Will our problems with this damned shed never end? Tried calling Donna Morris, who had sent us the original building permit, but ran into trouble both with the phones and with the fact that she wasn't available. Spoke to Kylie, who sounded very helpful and sympathetic, but who really could only forward the information, and who told me that Donna won't be available again until Friday.
|
Identifying USB disks at boot
|
Topic: technology, opinion | Link here |
The real issue with booting, still unsolved, is: how do I mount my USB disks correctly? eureka has 4 external backup disks, and on every boot they come up in as different devices, as I have noted before. What do I do? I've thought about it from time to time, wondering in particular if I haven't missed some more obvious method.
In principle I should be able to identify file systems by a label. The label is optional, but it would be a good start. But how do I find the label? One option would be to use gpart show, but the format is so confusing that I can't think of an easy way to parse it. Here an example:
=== root@teevee (/dev/pts/5) ~ 31 -> gpart show
=> 40 15628050976 ada0 GPT (7.3T)
40 128 1 freebsd-boot (64K)
168 83886080 2 freebsd-ufs (40G)
83886248 41943040 3 freebsd-swap (20G)
125829288 83886080 4 freebsd-ufs (40G)
209715368 15418335648 5 freebsd-ufs (7.2T)
=> 40 11721045088 da0 GPT (5.5T)
40 11721045088 1 freebsd-ufs (5.5T)
=> 40 11721045088 ufsid/5c1b040802cb9906 GPT (5.5T)
40 11721045088 1 freebsd-ufs (5.5T)
=> 40 11721045088 ufs/videobackup GPT (5.5T)
40 11721045088 1 freebsd-ufs (5.5T)
=> 40 11721045088 diskid/DISK-NA8TZEWG GPT (5.5T)
40 11721045088 1 freebsd-ufs (5.5T)
That shows two disks (sorry, “providers”): ada0 and da0, which almost coincidentally correspond to the devices /dev/ada0 and /dev/da0. But those are only part of the story: the devices are:
=== root@teevee (/dev/pts/5) ~ 32 -> ls -l /dev/ada* /dev/da*
crw-r----- 1 root operator 0x7d 11 May 04:13 /dev/ada0
crw-r----- 1 root operator 0x7f 11 May 04:13 /dev/ada0p1
crw-r----- 1 root operator 0x81 11 May 04:14 /dev/ada0p2
crw-r----- 1 root operator 0x83 11 May 04:13 /dev/ada0p3
crw-r----- 1 root operator 0x85 11 May 04:13 /dev/ada0p4
crw-r----- 1 root operator 0x87 11 May 04:14 /dev/ada0p5
crw-r----- 1 root operator 0xc2 12 May 10:35 /dev/da0
crw-r----- 1 root operator 0xc3 12 May 10:35 /dev/da0p1
Where are the others? /dev/ada0p* are represented as:
=> 40 15628050976 ada0 GPT (7.3T)
40 128 1 freebsd-boot (64K)
168 83886080 2 freebsd-ufs (40G)
83886248 41943040 3 freebsd-swap (20G)
125829288 83886080 4 freebsd-ufs (40G)
209715368 15418335648 5 freebsd-ufs (7.2T)
There's no heading, and you have to refer to the man page to understand the format:
The default output includes the logical starting block of each partition, the partition size in blocks, the partition index number, the partition type, and a human readable partition size.
No explanation of the => at the beginning.
Somehow I find this particularly unpleasant. Wouldn't this be a better format?
Start Size Device Type
40 15628050976 (7.3T) /dev/ada0 GPT
40 128 (64K) /dev/ada0p1 freebsd-boot
168 83886080 (40G) /dev/ada0p2 freebsd-ufs
83886248 41943040 (20G) /dev/ada0p3 freebsd-swap
125829288 83886080 (40G) /dev/ada0p4 freebsd-ufs
209715368 15418335648 (7.2T) /dev/ada0p5 freebsd-ufs
This still doesn't show the label; for that you use the -l option, which replaces the partition type with the label:
=> 40 15628050976 ada0 GPT (7.3T)
40 128 1 boot (64K)
168 83886080 2 rootfs (40G)
83886248 41943040 3 swap (20G)
125829288 83886080 4 rootfs-2 (40G)
209715368 15418335648 5 spool (7.2T)
Wouldn't that look better like this, including both partition type and label?
Start Size Device Type Label
40 15628050976 (7.3T) /dev/ada0 GPT
40 128 (64K) /dev/ada0p1 freebsd-boot boot
168 83886080 (40G) /dev/ada0p2 freebsd-ufs rootfs
83886248 41943040 (20G) /dev/ada0p3 freebsd-swap swap
125829288 83886080 (40G) /dev/ada0p4 freebsd-ufs rootfs-2
209715368 15418335648 (7.2T) /dev/ada0p5 freebsd-ufs spool
But that's only half the story. What does all this mean?
=> 40 11721045088 da0 GPT (5.5T)
40 11721045088 1 (null) (5.5T)
=> 40 11721045088 ufsid/5c1b040802cb9906 GPT (5.5T)
40 11721045088 1 (null) (5.5T)
=> 40 11721045088 ufs/videobackup GPT (5.5T)
40 11721045088 1 (null) (5.5T)
=> 40 11721045088 diskid/DISK-NA8TZEWG GPT (5.5T)
40 11721045088 1 (null) (5.5T)
The man page doesn't help there. Both the offsets and the index number show that these are four different ways of looking at the same partition. But where do the diskid/DISK-NA8TZEWG and ufs/videobackup come from? ufs/videobackup particular looks very close to what I would have labeled it. Currently the file system gets mounted on /VB2, but years ago it was mounted on /videobackup. The man page doesn't help there. I need to understand what's going on before I try modifying any code (which looks as if it wouldn't be too difficult).
|
Bruce Evans' disks analysed
|
Topic: technology, opinion | Link here |
With some difficulty I managed to navigate Google Drive and find some of the many disks that Peter Jeremy saved before he decided to destroy everything on the computers. This morning I had a 15 GB file ad0.img.xz waiting for me. How do I access it? file tells me:
=== grog@eureka (/dev/pts/19) /src/Downloads 30 -> file ad0.img
ad0.img: DOS/MBR boot sector; partition 1 : ID=0xb, active, start-CHS (0x0,1,1), end-CHS (0x3e8,254,63), startsector 63, 16081002 sectors; partition 2 : ID=0xa5, start-CHS (0x3e9,0,1), end-CHS (0x1e4,254,63), startsector 16081065, 57512700 sectors; partition 3 : ID=0xa5, start-CHS (0x1e5,0,1), end-CHS (0x3e0,254,63), startsector 73593765, 57512700 sectors; partition 4 : ID=0xc, start-CHS (0x3e1,0,1), end-CHS (0x11,80,63), startsector 131106465, 103335183 sectors
OK, how do I look at it? I could copy it to a physical disk, of course, but don't we have better tools than that? I had some recollection of a “vnode disk”. There!
=== root@echunga (/dev/ttypb) ~ 73 -> mdconfig -a -t vnode -f /src/ISOs/6.1-RELEASE-i386-disc1.iso
md0
Only problem: that file is a file system. What I have is a disk image containing file systems. Here a somewhat truncated output from fdisk:
=== root@dereel (/dev/pts/0) /mnt 84 -> mdconfig -a -t vnode -f Downloads/ad0.img
md0=== root@dereel (/dev/pts/0) /mnt 85 -> fdisk /dev/md0
Information from DOS bootblock is:
...
The data for partition 1 is:
sysid 11 (0x0b),(DOS or Windows 95 with 32 bit FAT)
start 63, size 16081002 (7852 Meg), flag 80 (active)
The data for partition 2 is:
sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
start 16081065, size 57512700 (28082 Meg), flag 0
The data for partition 3 is:
sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
start 73593765, size 57512700 (28082 Meg), flag 0
The data for partition 4 is:
sysid 12 (0x0c),(DOS or Windows 95 with 32 bit FAT (LBA))
start 131106465, size 103335183 (50456 Meg), flag 0
In passing, that's an interesting partitioning scheme for so small a disk. Not the way I would have done it, and I hadn't really expected Bruce to use Microsoft. But then, one of the reasons I wanted to look at these files was to discover that kind of information. But how do I access the file systems? Untidily, it proves, but it's possible.
The program needed is gnop, which apparently stands for “control utility for NOP GEOM class”, and is basically a test utility. Another man page that could be clearer. I found (and then lost again) an example
# gnop create -o 41255k md0
The -o is the offset from the beginning of the “provider”, in this case /dev/md0. There seems to be no provision for specifying a length. And clearly 41255k means 41255 kB, though that's not documented anywhere.
In my case, the offset for the interesting partitions was not an even number of kilobytes. What's the default scale factor? Sectors? No:
=== root@dereel (/dev/pts/0) /src/bde 91 -> gnop create -o 16081065 md0
gnop: Invalid offset for provider md0.=== root@dereel (/dev/pts/0) /src/bde 92 -> gnop create -o 16081065s md0
gnop: Invalid value for 'o' argument: Invalid argument=== root@dereel (/dev/pts/0) /src/bde 93 -> gnop create -o 8233505280 md0
About the only thing that works is an absolute byte count. I suppose I should UTSL, but for the moment I have more interesting things to do.
Moving on,
=== root@dereel (/dev/pts/0) /src/bde 97 -> l /dev/md0*
crw-r----- 1 root operator 0xa5 11 May 16:50 /dev/md0
crw-r----- 1 root operator 0xac 11 May 16:59 /dev/md0.nop
crw-r----- 1 root operator 0xad 11 May 16:59 /dev/md0.nopa
crw-r----- 1 root operator 0xae 11 May 16:59 /dev/md0.nopb
crw-r----- 1 root operator 0xaf 11 May 16:59 /dev/md0.nopd
crw-r----- 1 root operator 0xb0 11 May 16:59 /dev/md0.nope
crw-r----- 1 root operator 0xb2 11 May 16:59 /dev/md0.nopf
crw-r----- 1 root operator 0xb3 11 May 16:59 /dev/md0.nopg
Clearly those are the devices relating to the individual partitions
=== root@dereel (/dev/pts/0) /src/bde 98 -> bsdlabel /dev/md0.nop
# /dev/md0.nop:
8 partitions:
# size offset fstype [fsize bsize bps/cpg]
a: 131072 8192 4.2BSD 1024 8192 49152
b: 8225280 144585 4.2BSD 2048 16384 28512
c: 57512700 0 unused 0 0 # "raw" part, don't edit
d: 8225280 8369865 4.2BSD 2048 16384 28512
e: 8225280 16595145 4.2BSD 2048 16384 28512
f: 8225280 24820425 4.2BSD 2048 16384 28512
g: 24466995 33045705 4.2BSD 2048 16384 28512
Still more partitions! But those are the /dev/md0.nop?. So, I should be able to mount them:
=== root@dereel (/dev/pts/0) ~ 55 -> mount -o ro /dev/md0.nopa /mnt=== root@dereel (/dev/pts/0) ~ 56 -> cd /mnt/etc=== root@dereel (/dev/pts/0) /mnt/etc 57 -> cat fstab
/dev/ad0s2a / ufs rw,async,noatime 1 1
/dev/ad0s2b /cbak ufs ro,async,noatime,noauto 1 2
/dev/ad0s2d /usr ufs rw,noatime 1 2
/dev/ad0s2e /home ufs rw,noatime 1 2
/dev/ad0s2f /cvsbak ufs ro,noatime,noauto 1 2
/dev/ad0s2g /e ufs ro,noatime 1 2
...
There's a lot more in the fstab—in fact, one of the longest I have seen, with many comments. But now I at least know how to mount the partitions:
=== root@dereel (/dev/pts/0) /mnt/etc 59 -> cd ..=== root@dereel (/dev/pts/0) /mnt 60 -> mount /dev/md0.nopb cbak=== root@dereel (/dev/pts/0) /mnt 61 -> mount /dev/md0.nopd usr
mount: /dev/md0.nopd: R/W mount of /usr denied. Filesystem is not clean - run fsck.: Operation not permitted=== root@dereel (/dev/pts/0) /mnt 62 -> mount -o ro /dev/md0.nopd usr=== root@dereel (/dev/pts/0) /mnt 63 -> mount -o ro /dev/md0.nope home=== root@dereel (/dev/pts/0) /mnt 64 -> mount -o ro /dev/md0.nopf cvsbak/=== root@dereel (/dev/pts/0) /mnt 65 -> mount -o ro /dev/md0.nopg e
More interesting stuff in there, but first I need to understand it. So far, though, the contents seem even more interesting than I had thought.
| Thursday, 12 May 2022 | Dereel | Images for 12 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Irritating blood tests
|
Topic: health, opinion | Link here |
My blood test on Tuesday wasn't completely without consequences:
|
|||||||||||||||||||||||
That rectangular red patch on the left was where the plaster over the cotton wool had been. What causes that? It itches a lot, and I don't think it was the first time. Could it be the way I ripped it off, or that I left it on all day? That shouldn't be an issue. By contrast, the needle hole on the right was no problem at all. Next time I'll remove the plaster after a couple of minutes and see what happens.
|
GPS fail
|
Topic: technology | Link here |
One of the best uses I can find for a mobile phone is as a GPS navigator. It's not perfect, but I've decided that it probably works better than a dedicated GPS navigator. But today I got an unexpected message:
|
|||||||||||||||||||||||
Why that? How do you enable or disable GPS? It's not where you'd expect it on the right downwards “swipe”. After much searching I found this:
|
|||||||||||||||||||||||
That's under Settings/Location, at the very bottom of the Settings strip, after enabling it. But how did it get disabled in the first place?
|
Arena cover: the next step
|
Topic: Stones Road house, general | Link here |
I was so frustrated by yesterday's problems with the riding arena cover (dammit, what should I call it?) that I thought of reminding Yvonne that one of the conditions was that I should have nothing to do with the bureaucracy involved. But in the end I relented, and when I got a call from Brenton Carter of Wyndham Building Permits, I took it. Fill out this form, send in the plans (what plans? The only ones I have are from the planning permit, and they should already have them), and in a few days we should have an updated building permit (and, as it proved, also an invoice for the procedure). Hopefully we can get that done next week.
|
Another building interruption
|
Topic: Stones Road house, general | Link here |
Just as Yvonne had put the eggs into the pan for breakfast, we got another call. Will Tatnell along to look at tidying up the remains of the bracken in the house forest, and also to consider laying a pipe to a new horse trough. That took a while, but it looks as if we'll get it sorted out pretty quickly.
|
More ssh mysteries
|
Topic: technology, opinion | Link here |
Why does freefall.freebsd.org accept the same ssh key from some systems and not from others? On Tuesday I had established that the ones that were rejected were replying with the incorrect signature type:
debug3: sign_and_send_pubkey: RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo
debug3: sign_and_send_pubkey: signing using rsa-sha2-512 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo
agent key RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo returned incorrect signature type
debug3: sign_and_send_pubkey: signing using ssh-rsa SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo
But why? This was tiwi, running an almost completely new version of FreeBSD 13.1-STABLE. Clearly the ssh components were compatible. In preparation for a bug report, built a new world with today's latest and greatest:
=== grog@tiwi (/dev/pts/3) ~ 8 -> uname -a
FreeBSD tiwi.lemis.com 13.1-STABLE FreeBSD 13.1-STABLE #0 stable/13-n250745-1db9d7b7751c: Thu May 12 15:10:04 AEST 2022 grog@tiwi.lemis.com:/usr/obj/usr/src/amd64.amd64/sys/GENERIC amd64
And then:
=== grog@tiwi (/dev/pts/3) ~ 1 -> ssh freefall
agent key RSA SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo returned incorrect signature type
grog@freefall.freebsd.org: Permission denied (publickey).
Disconnected at Thu 12 May 2022 12:35:40 AEST=== grog@tiwi (/dev/pts/3) ~ 2 -> ps aux|grep ssh-agent
grog 56964 0.0 0.0 12868 2236 3 S+ 12:36 0:00.00 grep ssh-agent
Huh? No ssh-agent running? What's my key?
=== grog@tiwi (/dev/pts/3) ~ 3 -> ssh-add -l
2048 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo /home/grog/.ssh/id_rsa (RSA)
Where did that come from? OK, start an ssh-agent and populate it:
=== grog@tiwi (/dev/pts/3) ~ 4 -> eval `ssh-agent` >/dev/null=== grog@tiwi (/dev/pts/3) ~ 5 -> ssh-add < /dev/null
Identity added: /home/grog/.ssh/id_rsa (/home/grog/.ssh/id_rsa)=== grog@tiwi (/dev/pts/3) ~ 6 -> ssh-add -l
2048 SHA256:S7sZHLcY4dgw53/rF70vrScdPuGef3enHdJzuYA1WDo /home/grog/.ssh/id_rsa (RSA)
Not surprisingly, it's the same key. But the proof of the pudding is:
=== grog@tiwi (/dev/pts/3) ~ 7 -> ssh freefall
Last login: Fri May 13 00:26:32 2022 from 121-200-11-253.79c80b.mel.nbn.aussiebb.net
FreeBSD 14.0-CURRENT (CLUSTER) #0 main-n252623-389844c058e: Tue Jan 25 12:45:38 UTC 2022
...=== grog@freefall (/dev/pts/23) ~ 1 ->
So: somehow the system is responding without an ssh-agent, and that's part of the problem. And that's because I started the xterm from eureka, which also provides the key (I think). And looking at my .bashrc, I get a first inkling of differences. Greatly trimmed,
if [ "$HOSTNAME" = "teevee" -o "$HOSTNAME" = "tiwi" ]; then
eval `ssh-agent` >/dev/null
DISPLAY=:0 ssh-add < /dev/null
startx -- -listen tcp &
PATH=$PATH:/src/Samba/tivo/vplay/i386
else
eval `ssh-agent` >/dev/null
DISPLAY=:0 ssh-add < /dev/null
fi
In the process I note this modification to the PATH environment variable. What was /src/Samba/tivo/vplay/i386? The name gives the lie: it must be decades old, and I no longer have a directory /src/Samba. The line is from revision 1.26 of 23 May 2008, 14 years ago, and the comment suggests that even then I no longer knew what it was for.
But back to the story: teevee and tiwi are special, which could account for part of the problem. And yes, checking dereel, it does have ssh-agents running.
So that could be part of the problem. But it's only part of it, because it doesn't explain why my external servers lax and ffm behave differently. I really need to understand better how these things work. What part of ssh does the signing? Clearly it's not ssh-agent, as I had thought.
|
Borzois barking?
|
Topic: animals | Link here |
In mid-evening, after it was completely dark, Lena and Larissa were outside in the front of house. And I heard frantic barking.
Borzois don't bark. What's going on there? Lara! It wasn't just a couple of “woofs”: it was continuous, loud barking, something that I have never heard before in 40-odd years of living with Borzois, apparently at something in the street. And then somebody said out of the darkness: “Don't worry, Gregory, everything's all right”.
Gregory? Who calls me Gregory? But I didn't see anybody, only reflections of what could have been a car parked in Grassy Gully Road, but which I now think were reflective road markers. But Lara wasn't satisfied: even inside she continued grumbling to herself for some time.
What can that be?
| Friday, 13 May 2022 | Dereel → Napoleons → Enfield → Napoleons → Dereel | Images for 13 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Superstition and Friday the 13th
|
Topic: general, opinion | Link here |
Friday the 13th again today, as every 7 months on average. An unlucky day, they say, clearly a superstition. But today we could almost believe it. We had multiple unexpected issues. On other days they'd be just coincidences. But today it's clearly because of the date.
|
Next horse clinic issue
|
Topic: animals, general | Link here |
So tomorrow is the start of the horse clinic postponed from a year ago—the one that also delayed the arrival of Bruce Evans' computers. Call from Anke on the way to the airport: they had cancelled her flight!
Considerable flurries of attempts to rebook on another flight, which ultimately came to nought. The clinic is cancelled at the last minute again! At least I hadn't got round to thawing out the food I had started last year.
Bloody Friday the 13th!
|
New toy
|
Topic: technology, general, opinion | Link here |
Today the thermometer that I ordered last week arrived in Napoleons. By this time it was clear that Yvonne wouldn't be going into town to pick up Anke, so off to Napoleons to pick it up.
They couldn't find it! In the process discovered that they don't have any network access to check tracking. Tried with my mobile phone, but the tracking number on the pickup slip was wrong, so it didn't help. Finally, after 15 minutes of searching, left with a request to call back when they found it.
I hadn't quite made it home when I got a call from Yvonne to tell me that they had found it almost immediately, but that she hadn't been able to get me on the phone: this was the 5th attempt. Back to Napoleons and picked it up: the sender had run the address together, so what was on the label was “GREG YVONNELEHEY”, and it had been stored under Y, not L.
More Friday the 13th? At least the difficulty calling me could have been because of flaky mobile network coverage between Napoleons and here.
|
Icebird thermometer
|
Topic: technology, general, opinion | Link here |
Back home, unpacked my toys: an Inkbird IBBQ-4BW multithermometer with WiFi and Bluetooth connectivity, and also an IHT-1P “penknife” thermometer.
How do I set it up? Very much RTFM, which was terrible. The instructions didn't match the behaviour, and for the WiFi connectivity they wanted me to sign up with an email address and password! Did all that, received a 6 digit confirmation number, which the app didn't accept. Asked for another one, but clearly the Inkbirds think that one is enough.
Dammit, how about Bluetooth? How about that, for the first time ever Bluetooth paired immediately, bringing a particularly stupid display:
|
|
|||||||||
Maybe it reflects the pain that even the developers had with the software.
OK, and now try the “Asssisted connection” for WiFi. But that didn't work: it was already connected! So clearly this app is particularly broken.
That's about all I did today. I'll check out the freezer tomorrow.
And the penknife? I don't really need it, since I already have a much smaller thermometer that does the same thing:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Still, the display is much better, and since it didn't cost anything, I can barely complain.
|
Dana arrives
|
Topic: animals | Link here |
Call from Cooper, apparently the last horse transporter for Dana's marathon journey from Kybean to here. He arrived at Chris Bahlo's place just as it was getting dark:
|
||||||||||||||||||||||||
The light from the headlights shows how dark it really was. Fortunately she was relatively easy to handle, and she was soon unloaded:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Put her in a yard for the night:
|
||||||||||||||||||||||||
Tomorrow, in daylight, we'll bring her to our place. These photos don't make it clear that it was very dark all the time.
| Saturday, 14 May 2022 | Dereel | Images for 14 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Dana comes home
|
Topic: animals | Link here |
Early breakfast today, then over to Chris Bahlo's place to pick up Dana.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
||||||||||||||||||||||
She's home! After only a little more than 5 days for 963 km, an average of not quite 8 km/h.
|
Measuring freezer temperature
|
Topic: technology, general, opinion | Link here |
Spent a bit of time with my new WiFi thermometer. It certainly shows Android apps from their worst perspective.
One of my concerns when buying it was whether the cables were long enough. I shouldn't have worried: they're 1.5 m each, and they're rather unwieldy. And how do I arrange the colours? They're coloured red, green, blue and something that may be intended to be yellow. RGB sounds right, leaving Y for fourth place:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But that turns out not to be the way the writers of the app think. They have implicitly assigned to colours in the graphs: red, yellow, green, blue (cyan).
Working with the app is an exercise in frustration. By default, it can't even display all the temperatures at once:
|
|||||||||||||||||||||||
It wastes screen real estate with silly pictures and overly large graphics. The cow head was a result of my playing around with this menu, before I found out how to remove it again:
|
|||||||||||||||||||||||
I suppose that could be useful for a thermometer designed for cooking meat, but does it have to be a picture of the head?
Pressing on the four squares at top right of the temperature display improves things, but only marginally:
|
|||||||||||||||||||||||
Now only half the screen is occupied, and all the controls are gone. So I have to alternate between the two. The second gives an overview, the first a marginal measure of control.
So what can I do with the thing? So far, it seems, I can display graphs of temperature against time:
|
|||||||||||||||||||||||
And that shows that there are discontinuities. Why? The app claims to be connected both via Bluetooth and WiFi. Is it trying only Bluetooth? The WiFi access point is only about 40 cm away, on top of the freezer.
And then sometimes when I select the graph module, it tries and then stops again. Is the graph module crashing? It sounds about right. The only way I have found to get it to work again is to stop and restart the app.
But how do I get at the data? All the thing offers is graphs with hard-to read axis markings. It must store the data somewhere. How do I get at it? How about sniffing the traffic between thermometer and phone? It has a rather sane-looking status list on the settings page:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's a rather strange address. It should be in the range 192.109.197.0/24. Can I ping it? Yes!
=== grog@eureka (/dev/pts/23) ~ 180 -> ping 121.200.11.253
PING 121.200.11.253 (121.200.11.253): 56 data bytes
64 bytes from 121.200.11.253: icmp_seq=0 ttl=64 time=0.012 ms
64 bytes from 121.200.11.253: icmp_seq=1 ttl=64 time=0.015 ms
And that, again, is a very short ping time.
=== grog@eureka (/dev/pts/23) ~ 170 -> host 121.200.11.253
253.11.200.121.in-addr.arpa domain name pointer 121-200-11-253.79c80b.mel.nbn.aussiebb.net.
Huh? That's the address of the Internet gateway. What's my real address?
=== grog@eureka (/dev/pts/23) ~ 172 -> arp -a
121-200-11-253.79c80b.mel.nbn.aussiebb.net (121.200.11.253) at 00:50:da:cf:07:35 on xl0 permanent [ethernet]
...
dhcp-225.lemis.com (192.109.197.225) at e8:db:84:b6:18:b1 on em0 expires in 1196 seconds [ethernet]
This terminally broken app has reported the wrong IP address! OK, is it listening to anything?
=== grog@eureka (/dev/pts/23) ~ 173 -> nmap dhcp-225
...
PORT STATE SERVICE
6668/tcp open irc
IRC? Why should that port be open? It isn't, really. An attempt to connect to it with irc just timed out. But telnet was more successful:
=== grog@eureka (/dev/pts/23) ~ 178 -> telnet dhcp-225 6668
Trying 192.109.197.225...
Connected to dhcp-225.lemis.com.
Escape character is '^]'.
Uk3.3Lº
3Ww×.]í±ñ;BÈlªUUk3.3LoøJÉãú;.Ñ F
ºs
tĵÝóZeýÏ$X´âi{įã`n
Now how do I go about analysing that?
|
Freezer temperatures
|
Topic: general, technology, opinion | Link here |
To my surprise, the temperatures in the freezer were not only acceptable, but slightly low. Here a typical view:
|
|||||||||||||||||||||||
The temperatures range between -19° and -22.8°, a little cool. But then there's this:
|
|||||||||||||||||||||||
That's the culprit! It seems to happen 2 or 3 times a day, and it lasts for nearly 2 hours (this graph shows a time span of 6 hours). My guess is that it's related to the automatic defrost. When I have more data, I'll compare it with the other freezers.
Opening the freezer makes a difference too, but it's nothing by comparison. This graph (time span 1 hour) shows three openings in relatively quick succession:
|
|||||||||||||||||||||||
Here the cyan trace is the sensor in front of the drawers, which is open to the outside when I open the door.
|
|||||||||||||||||||||||
The red trace is at the top, also unprotected by the drawer fronts. And the other two briefly go up to about -16°, but quickly recover, within minutes. The previous graph shows the temperatures inside the baskets above -16° for round an hour.
So: what is this? My guess is that there's something wrong with the defrosting mechanism.
| Sunday, 15 May 2022 | Dereel | Images for 15 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Early in the morning
|
Topic: animals | Link here |
Got out of bed at a normal time today. No sign of Yvonne. Ah, there, 200 m away:
|
|||||||||||||||||||||||
Dana didn't survive the transport completely unscathed. Yesterday we saw a scratch on her leg, already partially healed:
|
|||||||||||||||||||||||
And today she had a problem with her right eye, which was almost closed in the early morning, but which gradually improved in the course of the day.
|
What's this?
|
Topic: language, food and drink, technology, opinion | Link here |
Found in the fridge:
|
|||||||||||||||||||||||
When I found it, is was in this opened plastic bag:
|
|||||||||||||||||||||||
Now doesn't that tell me what it is? But never mind, Google Lens to the rescue:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The trouble with Google Lens' “translate” facility seems to be that it is to eager to please, even when the image is poor (here due to reflections). But it also makes it pretty useless. Running Google Lens against the photos I took later consistently gives me “Haidilao seasoning” for the red text and and “soup bag” for the black text. And of course, that doesn't help much.
What is it? It's oily and tastes vaguely salty. I should remember where I got it from and what I used it for.
|
Death of a journalist
|
Topic: politics, general, opinion | Link here |
Last Wednesday, Shireen Abu Akleh, a reporter for Al Jazeera, was shot dead in the Jenin refugee camp. We found out about it almost immediately, since we watch Al Jazeera news on a daily basis. They reported the death at 4:25 UTC, with the result that for the next 24 hours Al Jazeera news brought no other content, and we had to find other news sources to find out what was going in the world. Even now, four days later, nearly a third of the content relates to the incident, and even in the rest—here Annalena Baerbock, Germany's foreign minister, talking about Sweden's decision to join NADO, as she put it—the entire right side of the screen relates to Shireen, and the short news items on the strip at the bottom relate only to her death and subsequent treatment:
|
|||||||||||||||||||||||
Yes, it's sad that journalists get killed, but how many people were killed in Ukraine over the same couple of days? Admittedly, the Israeli occupation forces have helped keep the issue going. A Palestinian shot her! The police who attacked the funeral procession only did so because the coffin had been stolen, according to the Times of Israel:
The announcement came in the wake of international condemnation after officers charged Palestinians who were holding and surrounding the casket at the funeral on Friday, and beat them with batons, nearly toppling the coffin. Police have claimed that the casket was seized by a mob, against the family’s wishes, outside Jerusalem’s St. Joseph’s Hospital, prompting the cops’ intervention. Abu Akleh’s brother, however, said the family and mourners had hoped to hold a “small procession” but were “bombarded” by officers as they left the hospital.
Somehow this all reminds me of the lies that Russia has been telling about Ukraine. Does anybody really believe them? And is it really just a coincidence that they also destroyed the Al Jazeera office building in Gaza?
Can anybody really claim that Israel is a respecter of human rights? Yes, the Palestinian leadership has a number of issues, but isn't it time, after nearly 55 years, for Israel to retreat from the areas they conquered and leave the administration to a neutral body (United Nations, maybe)?
Later another report plausibly contradicted Israeli police's claims.
|
More Inkbird insights
|
Topic: technology, general | Link here |
Mail from Malcolm Caldwell this evening:
I have also setup freezer/fridge temperature logging. I looked at inkbird but in the end went with govee devices.
But I did see that the protocols for inkbird are known. (As are the govee ones).
For example you can find plugins for inkbird that work with open source home automation software like home assistant. https://github.com/stelford/home-assistant_inkbird
Also this looks promising: https://github.com/tobievii/inkbird
More fun to try out, even if they appear to be a layering violation. Certainly this display looks better than what Inkbird provides:
.
But I really should finish upgrading my computers first.
| Monday, 16 May 2022 | Dereel | Images for 16 May 2022 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Baked beans revisited
|
Topic: food and drink, opinion | Link here |
Time for another load of baked beans today. That's pretty straightforward, but today I added more pork skin, a total of 250 g for 750 g of (dried) beans. It'll take a while to find out whether that's a good quantity or not.
Also paid more attention to the cooking time: 5 hours, at which point the beans were still barely firm. But it takes hours for them to cool down, so my guess is that this is the best time to stop cooking.
|
HTML validation problems
|
Topic: technology, opinion | Link here |
Part of writing these diary entries is checking the PHP output for validity. I do this once for the HTML version, using the W3 validator, and once for the XML version, using this validator.
But lately the W3 HTML validator has been lazy. I find a bug, fix it, and it revalidate—and the bug's still there. Not in my source, but in the validator's view of the world. How do I report this problem?
|
More gpart/geom insights
|
Topic: technology, opinion | Link here |
Mail from Daniel Nebdal suggesting alternatives to rewriting gpart show. Doesn't gpart show -p do exactly what I want? It replaces the index numbers with “provider” names.
Surely I've seen that? Looking at the example I wrote last Wednesday, what I want is:
Start Size Device Type Label
40 15628050976 (7.3T) /dev/ada0 GPT
40 128 (64K) /dev/ada0p1 freebsd-boot boot
168 83886080 (40G) /dev/ada0p2 freebsd-ufs rootfs
83886248 41943040 (20G) /dev/ada0p3 freebsd-swap swap
125829288 83886080 (40G) /dev/ada0p4 freebsd-ufs rootfs-2
209715368 15418335648 (7.2T) /dev/ada0p5 freebsd-ufs spool
gpart -p -l gives me:
=> 40 7814034976 ada0 GPT (3.6T)
40 128 ada0p1 (null) (64K)
168 83886080 ada0p2 (null) (40G)
83886248 41943040 ada0p3 (null) (20G)
125829288 83886080 ada0p4 (null) (40G)
209715368 7604319648 ada0p5 (null) (3.5T)
Not quite the same. The type is gone, and if (null) is supposed to represent the label, it doesn't look right.
Ah, that's clear: the (null) means that there is no label! Let's add one:
=== root@dereel (/dev/pts/0) /mnt/etc 135 -> glabel label -v cbak md0.nopb
glabel: Can't store metadata on md0.nopb: Operation not permitted.=== root@dereel (/dev/pts/0) /mnt/etc 136 -> glabel label -v cbak /dev/md0.nopb
glabel: Can't store metadata on /dev/md0.nopb: Operation not permitted.
RTFM time. Following the instructions in glabel(8), I need
=== root@dereel (/dev/pts/0) /mnt/etc 138 -> tunefs -L cbak /dev/md0.nopb
tunefs: /dev/md0.nopb is not clean - run fsck.=== root@dereel (/dev/pts/0) /mnt/etc 139 -> df
Filesystem 1048576-blocks Used Avail Capacity Mounted on
...
/dev/md0.nopb 3,952 3,427 208 94% /mnt/cbak=== root@dereel (/dev/pts/0) /mnt/etc 140 -> umount /mnt/cbak/=== root@dereel (/dev/pts/0) /mnt/etc 141 -> tunefs -L cbak /dev/md0.nopb=== root@dereel (/dev/pts/0) /mnt/etc 146 -> gpart show -r
...
=> 0 148597079 md0.nop BSD (71G)
144585 8225280 2 7 (3.9G)=== root@dereel (/dev/pts/0) /mnt/etc 147 -> gpart show -l
...
=> 0 148597079 md0.nop BSD (71G)
144585 8225280 2 (null) (3.9G)=== root@dereel (/dev/pts/0) /mnt/etc 148 -> gpart show -p -l
...
144585 8225280 md0.nopb (null) (3.9G)
Now that's nothing like what I expected. First, why tunefs and not glabel? The man page doesn't say so, but my guess is because it's a label in a UFS file system, not the GEOM structures, and the tool for that is tunefs.
Next, why does it not work if the file system is mounted? There's no mention in the man page, so either the man page needs updating if there's a good reason, or it needs to be allowed in the program if there's no good reason. And finally, gpart show doesn't work there. It's presumably looking for a different label.
What about the label added by glabel? What if the message was really saying “can't change label while the file system is mounted”? OK, it's not mounted now. Try again.
=== root@dereel (/dev/pts/0) /mnt/etc 164 -> glabel label -v cbak md0.nopb
Metadata value stored on md0.nopb.
Done.=== root@dereel (/dev/pts/0) /mnt/etc 165 -> gpart show -l
...
144585 8225280 2 (null) (3.9G)
=== root@dereel (/dev/pts/0) /mnt/etc 166 -> gpart show -l -p
...
144585 8225280 md0.nopb (null) (3.9G)
It seems that the only thing that will display the labels is geom -t:
=== root@dereel (/dev/pts/0) /mnt/etc 180 -> geom -t | grep LABEL
...
md0.nopb LABEL ufsid/574c22de02cc6de6
md0.nopb LABEL ufs/cbak
md0.nopb LABEL label/cbak
So clearly there are more than one label, though the ufsid one surprised me. But how many more are there if gpart show -l still displays (null)? This is a real can of worms.
|
Time to reconsider VoIP?
|
Topic: technology, general, opinion | Link here |
Lately I've had a lot of trouble understanding people on the phone. Am I going deaf? No, it just seems to be on the VoIP line, and Yvonne has noticed it too.
Has somebody changed the codec behind my back? Off to look for the current configuration:
|
|
|||||||||
What's the best codec to use? According to this page, G.711, G.722, G.729 and also rans. Where's PCMU? PCMA? The best I can find on this ATA is G.722. Spent some time reconfiguring before I discovered that PCMU and PCMA are varieties of G.711, so I had to replace it again. Reboot the ATA.
But should I even be using these old Uniden phones any more? Can't I run VoIP on my mobile phone? I haven't had any issues with the sound quality there, and at home it doesn't cost anything. First off to find what the best VoIP apps are. Surprise, it's still ZoiPer. From this page:
Zoiper is one of the few SIP and VoIP applications left which is why it is so unique. It focuses mostly on phone calls and includes support for Bluetooth, IAX support, the ability to make phone calls over the Internet, and other additional protocols.
Too old-fashioned, it seems. Now you use Skype, Facebook Messenger, Google foo or WhatsApp, all of which have the extreme disadvantage that the other end also needs to have them installed. OK, install ZoiPer, which now wants me to pay money. The interface is just as emetic as it was a year ago. At the end I had this interface:
|
|||||||||||||||||||||||
What's that? A photo of Yvonne. OK, are we connected? No!
|
|||||||||||||||||||||||
|
|
|||||||||
What an amazing difference between the simplistic and useless home display and the status showing me that I wasn't even connected, and giving me no better explanation than “(60)”. To make up for it, it tries continually to get me to upgrade to services that cost money. A bit of searching shows that the code means “connection in use”, which makes sense. But why is everything to do with mobile phones so stupid?
|
Doctor, heal thyself!
|
Topic: health, technology | Link here |
I had an appointment with my doctor, Paul Smith, this afternoon. Problem: he has COVID-19. Well, he's recovering, but not safe for a physical meeting. So we did it over the phone. Nothing much to report: blood test results as expected, though he's still puzzled by the high MCV (106 fl) and fasting glucose (7.2 mmol/l). But the HbA1c value was 6.0 and that's more important, though it seems higher than it has been. As he said “All in all you're not dying any faster than anybody else”.
On the positive side, I had no difficulty with the voice quality on the VoIP call. Were the other problems coincidences, or did rebooting the ATA help?
| Tuesday, 17 May 2022 | Dereel | Images for 17 May 2022 |
| Top of page | ||
| previous day |
|
More inkbird and freezer insights
|
Topic: technology, food and drink, opinion | Link here |
I've had my Inkbird IBBQ-4BW up
and running stumbling for a couple of days now. During that time I've
discovered:
The horrible app is almost useless. It drops data, it doesn't save raw data, so I'm limited to what little information it supplies in the way of graphs, and it continually crashes. It's interesting to note that the reviews of the app almost all seem to agree with me.
The thermometer can run while it's being charged, but what I've seen so far suggests that it doesn't work well like that: the temperature display jumps up and down 5° or more in a matter of seconds.
The freezer clearly has significant problems with its defrosting cycle (if that's what it is). It seems to cool normally most of the time, but today I had a particularly bad example, where the temperature in front of the drawers rose to about 7.8°:
|
|||||||||||||||||||||||
The temperature in the top drawer (red curve) rose to nearly 3°, but the others weren't so bad at -8° or so. Still unacceptable, of course.
How do the temperatures compare on the other freezer? Moved the system there; we'll see how it compares.
|
Reverse engineering Inkbird
|
Topic: technology, opinion | Link here |
So the app for the Inkbird IBBQ-4BW is almost useless. Reverse engineer the protocol and collect the data myself? This article discusses the details. It seems to broadcast something. OK, For a first attempt, I took a look at what it had to say:
11:24:09.324696 IP inkbird.lemis.com.49154 > 255.255.255.255.6667: UDP, length 188
11:24:14.325785 IP inkbird.lemis.com.49154 > 255.255.255.255.6667: UDP, length 188
11:24:17.322125 ARP, Request who-has inkbird.lemis.com tell inkbird.lemis.com, length 46
11:24:19.326255 IP inkbird.lemis.com.49154 > 255.255.255.255.6667: UDP, length 188
11:24:24.324427 IP inkbird.lemis.com.49154 > 255.255.255.255.6667: UDP, length 188
11:24:27.321882 ARP, Request who-has inkbird.lemis.com tell inkbird.lemis.com, length 46
...
11:24:37.321858 ARP, Request who-has inkbird.lemis.com tell inkbird.lemis.com, length 46
11:24:47.321888 ARP, Request who-has inkbird.lemis.com tell inkbird.lemis.com, length 46
There are the UDP packets described in the article, though they're a different length and apparently encrypted. Do I want to know? But look at those ARP requests. ARP is used to find the MAC address of another system, like this:
11:59:17.433713 ARP, Request who-has tiwi.lemis.com tell eureka.lemis.com, length 28
11:59:17.434042 ARP, Reply tiwi.lemis.com is-at cc:52:af:41:c3:55 (oui Unknown), length 46
Unix systems check every 20 minutes, but Linux seems to be much more frequent. But what earthly use is an ARP request every 10 seconds asking for the systems' own IP address?
But wait, there's more:
11:24:56.506706 IP inkbird.lemis.com.59056 > ec2-3-124-225-12.eu-central-1.compute.amazonaws.com.8886: Flags [P.], seq 69:282, ack 70, win 3690, length 213
11:24:56.773039 IP ec2-3-124-225-12.eu-central-1.compute.amazonaws.com.8886 > inkbird.lemis.com.59056: Flags [P.], seq 70:139, ack 282, win 3752, length 69
11:24:56.901644 IP inkbird.lemis.com.59056 > ec2-3-124-225-12.eu-central-1.compute.amazonaws.com.8886: Flags [.], ack 139, win 3621, length 0
...
Who is it talking to there? And why?
The idea of using third-party reverse-engineered software is good, but it seems that every device that Inkbird makes is different, and the chance of being able to connect to it easily seems minimal. Even their own apps have difficulty recognizing different devices. So I'll just accept it with its extreme limitations. It still does better than a min/max thermometer, but what a waste of potential!
|
Groceries: back to the future
|
Topic: general, history, opinion | Link here |
Read on Statista today:
Grocery delivery has come into focus in the coronavirus pandemic, expanding what was long considered the “last frontier of e-commerce” to a much larger number of households.
What does grocery delivery have to do with E-commerce? Over 60 years ago grocery deliveries were the norm for us in Kuala Lumpur. Naina Mohamed had a delivery service for half the town, and we'd fill out order slips for what we wanted on the following day. Looking back it's amazing that it worked so well. But suddenly that's new?
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 |
