|
| Greg's diary |
| August 2023 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Tuesday, 1 August 2023 | Dereel | |
| Top of page | ||
| next day | ||
| last day |
|
Yet another lazy day
|
Topic: general, health | Link here |
Yet another day where I have done almost nothing. Is this the result of my incident 2 months ago? Maybe I should stop taking the amitriptyline and see if I feel any brighter.
|
Little garden work
|
Topic: gardening | Link here |
Jess Mellington came along and helped me move the maple “Chantilly Lace” into position in front of the house entrance:
To do that she first needed to remove the Mirabilis jalapa growing in the pot that was in that position, and also a couple growing in the ground below. I had expected round 50 tubers, but in fact there were only 4, of considerably differing sizes:
|
|||||||||||||||||||||||
|
Complicating Ballarat city traffic
|
Topic: general, opinion | Link here |
Over a year ago I commented about the crazy idea of opening Bridge Mall in Ballarat to traffic. But now more details are emerging: they have started work, and part of it involves removing further thoroughfares in the vicinity. They published this image:
So it seems that eastbound traffic will need to go through Bridge Mall to get to Bakery Hill. This is quite literally the middle of Ballarat and the main east-west thoroughfare. Is this really the best option?
|
Panic!
|
Topic: technology, opinion | Link here |
Finished watching TV round 21:45. The usual prompt: keep or erase. Erase.
But then nothing. No new prompt. Not in that window, not in any other. Into the office to check. eureka dead, and bde also hung. Nothing for it: reboot eureka, which saved a dump. The accompanying core.txt tells me:
panic: cannot reassign paging buffer
The stack trace shows that that comes from reassignbuf, which seems reasonable. Tracing back takes me to softdep_process_journal, ffs_syncvnode and nfsvno_fsync. This was during the monthly backups, so it's clearly related. But why now? It was the first full dump since I bought a new backup disk. Is it related? At the very least I should consider removing soft updates from the disk.
| Wednesday, 2 August 2023 | Dereel | Images for 2 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The morning after the night before
|
Topic: technology | Link here |
Into the office to see how things were working after yesterday's panic. All went well, and even the backups that were halted by the panic completed afterwards, though I have my doubts as to their consistency. And an old friend resurfaced, the unprotected printfs:
tiwi.lemis.com kernel log messages:
+n<6f>sn nservfesrf ssnefrsveeu rs ersvere reuvrekea:r/he ka:e/ohuomer: enoktm ea: rneos:po/t hreondmineg:
+ not srponedinsg
+p
+onding
+r
+ eureka:/home: not responding
+nfs server eureka:/: not responding
|
Time to tidy up
|
Topic: general, opinion | Link here |
My office is in a mess. In fact, it has been in a mess for over 8 years, since we moved in.
|
|||||||||||||||||||||||
At that time I had put in a small table on the right for things that I needed to find a place for. But it's still there, though hard to find. And the table on the left is also full of papers, cameras, monitors and random junk.
|
|||||||||||||||||||||||
Where do I start? By taking a second-rate “before” photo. Who knows when the “after” photo will come?
|
Office representation
|
Topic: photography, opinion | Link here |
Taking the “before” photo of my office begged a question of perspective. The obvious view seemed to be clear, a 200° panorama at desk height:
|
|||||||||||||||||||||||
But somehow that seems funny. In the end, I chose the same view that I had used over 8 years ago, sort of:
|
|||||||||||||||||||||||
There must be a better way.
| Thursday, 3 August 2023 | Dereel | Images for 3 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Back recovery, next step
|
Topic: health, opinion | Link here |
Gradually I'm recovering from my slipped disk. During the day I have almost no pain, and I can walk normal distances with the dogs. But I've still been taking my amitriptyline tablets at night.
Time to stop? Last night I went without. Not good. For reasons I don't understand, my leg was “uncomfortable”. Not painful, but I kept looking for a more comfortable position, and I kept waking up.
What do I do? I think I'll have to try for a few days and see if things get better.
|
Random cursor navigation with firefox
|
Topic: technology, opinion | Link here |
For reasons I don't understand, the firefox instance running on eureka:0.2 has gone crazy. Press PgDn and it goes to the bottom of the page. Press Home and nothing happens. In the middle of a long display PgUp and PgDn seem to position randomly.
I've seen this before, and my workaround was to start a new firefox profile. But there must be a better solution. The main issue is how to search for the solution, but finally I found this page, which tells me that I have “caret browsing” enabled. In firefox-speak that apparently means “The caret is the name of the vertical "|" text cursor", and it allows you to navigate without a mouse. But if that's right, then I don't know if I even want to know.
Turn it off? F7. No, that doesn't work. But the support page also gave me marginally correct instructions to reset it:
Settings -> General: Browsing
remove checkmark: [ ] "Always use the cursor keys to navigate within pages"
And yes, it was set. How did that happen? Turned it off. And how do I save it? Automatically? The first time I tried it, it stayed enabled, but maybe I changed tabs the wrong way round. I've been using firefox for nearly 20 years, and I still can't stand it. If it weren't for the even worse chrome (or is that chromium?), I wouldn't use it.
|
Tanduri again
|
Topic: food and drink, opinion | Link here |
Another chicken tanduri tonight. Building on past experience, started with 430 g of chicken, cooked at 500 W for 7½ minutes. Previously I claim to have done it at 600 W, but the new microwave oven doesn't offer that possibility.
After that, grilled in the “hair dryer” at 230° for an intended 11 minutes, but after 9 minutes it was done. Can we leave it like that?
| Friday, 4 August 2023 | Dereel | Images for 4 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another night's pain
|
Topic: health, opinion | Link here |
Another restless night without amitriptyline. Should I start taking it again? At some point it might stop, so for the moment I'll continue without. After all, it has been nearly 2 months since the incident.
|
Checking CJ's “modem”
|
Topic: technology, opinion | Link here |
CJ Ellis along this afternoon as planned with his brand new Netcomm NF20MESH “modem”, really a home gateway with VoIP lines. Nicely packed with brochures from Aussie Broadband telling him how to connect the cables. That was necessary, too, because it has two inputs:
|
|||||||||||||||||||||||
And yes, that's about as clear as it can be. From left, 2 phone lines, DSL phone input, USB input (what for?), 4 Ethernet lines (yellow), and only then the “WAN” Ethernet input (red)
OK, connect it up to bde. Damn, I really must remap a key for Ctrl, but yes, it gave bde an IP address (192.168.20.2), and I could ping the Global Internet. Put in a phone line. Strange dial tone. Try to dial. Get a US American “busy” (engaged) signal.
What's wrong there? Where's the setup guide? There isn't one! Off to the web to search for the user guide (here). Much reading, discover that the name has changed to “CloudMesh Gateway”, and that the password is written on the base of the unit:
|
|||||||||||||||||||||||
At least that confirms the name change. Connect up to the device and marvel that it has an invalid certificate. Ignore that and wander through Yet Another Maze Of Twisty Little Menus, all different, and finally, after I had almost given up, found the line configuration page under the tastefully truncated SIP Basic Settin:
|
|||||||||||||||||||||||
Not configured! And I don't have the settings, so we're dead in the water. So I
CJ has paid a premium to get a pre-configured gateway, only for me to discover that it
wasn't configured at all.
Well, yes, it did have one configuration parameter, probably from the factory: the SIP gateway address. But that's not enough. And yes, it occurred to me later that CJ's “My Aussie” page includes the details. But CJ wouldn't have been able to do anything with the thing at all: he would have been dead in the water. Dammit, Aussie, that's not good enough. It goes back.
Spent some time looking for an alternative source for the device, with the surprising discovery that they were all more expensive. But why NetComm? They're a strange company, and the fact that even the box can't decide whether it's an NF20MESH or a CloudMesh, and that its security certificate is invalid, doesn't inspire confidence. And I was sure that Grandstream had a pretty straightforward replacement for the Cisco SPA112, despite what I had been told two weeks ago. Off looking. Where are Aussie's hardware prices? Only Google knows. And yes, there's a handy setup guide for the Grandstream HT812, which, according to their Internet hardware page (also found thanks to Google), they sell for $69, fully $110 less than the device CJ was sold. They want to connect it to a “modem”, whatever that may be; I'm assuming that an NTD (NBN) would do.
More searching, and yes, it seems that it would do the trick. But a “live chat” with somebody purporting to be from Aussie tells me that they don't support that configuration. So why sell the things at all? The HT802 would do just as well, but they don't seem to sell that any more.
| Saturday, 5 August 2023 | Dereel | Images for 5 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The nightly pain
|
Topic: health | Link here |
Another night without amitriptyline. Still not pleasant, but I get the feeling that things are improving.
|
Chat transcript
|
Topic: technology, opinion | Link here |
What happened to the transcript of the chat that I asked for yesterday? My best bet was that it came from a server that didn't have reverse lookup, which I still reject.
Today I was looking on the Aussie Broadband web site again, and another chat person offered himself. OK, what happened to yesterday's transcript? Usual answer: “Check your spam folder”, “Did you specify the correct address?”. So I asked what the MTA name was. Not Aussie at all: livechat, which I interpreted (correctly, as it happens) as livechat.com. Now how was I to know that? And looking through the mail backup, I found:
Date: Fri, 04 Aug 2023 04:58:03 +0000
From: LiveChat <support@livechat.com>
To: 7973a0fd-c017-4359-90c6-65c4050d2aa0 <groggyhimself@lemis.com>
Subject: Verify your email address
Message-Id: <204149cc-fe5e-4e4a-85c3-feffb3f49916@mtasv.net>
## Verify your email address ##
We'll send messages to this email address only if you verify it by clicking the link below.
And that was all. So obviously spam that I didn't even remember deleting it. And how about, after “clicking” on the “link”, I received:
Date: Sat, 05 Aug 2023 02:44:49 +0000
From: LiveChat <support@livechatinc.com>
To: groggyhimself@lemis.com
Subject: Chat transcript
Message-Id: <929b7d79-fa77-42a6-8ec5-6b4abe98f771@mtasv.net>
Reply-To: groggyhimself@lemis.com
Chat transcript
Name: Greg Lehey
E-mail: groggyhimself@lemis.com
Tristan Fri, 08/04/23 02:37:32 pm Australia/Melbourne
Hello, I'm part of our Australian based Sales Team - just wondering if there is anything I can help you with?
Greg Lehey 02:38:16 pm
I'm looking at the Grandstream HT812. It seems to be a gateway device, but your docco sells it as a VoIP only device. Why?
Tristan 02:40:35 pm
G'day Greg 🙂 The Grandstreams aren't a modem/router, they're just used to connect through them for a voip service. If you're looking at ordering a modem through us, our Netcomms have inbuilt voip, otherwise a grandstream would be needed for a locked BYO, or if you use a Google nest that doesn't have voip inbuilt 🙂
Greg Lehey 02:43:07 pm
There are various Grandstreams. I'm looking at the HT812. According to https://www.grandstream.com/hubfs/Product_Documentation/ht81x_user_guide.pdf it's a NAT router too.
Tristan 02:48:09 pm
Personally we've never configured them using a router setting, ours come pre-configured for a Voip Setup, not as a router option.
Greg Lehey 02:49:05 pm
So you're throwing the baby out with the bath water?
Tristan 02:50:26 pm
The Grandstream's don't have any inbuilt wifi capability, to if they were to be used as a router, you'd only be able to use them by an ethernet conenction, and would require a wifi switch or additonal modem.
Greg Lehey 02:51:36 pm
We weren't talking about Wi-Fi. The installation has one PC and two phones, only. It has been working fine with a Cisco ATA for 10 years.
Tristan 02:54:22 pm
In that scenario, they could be used. It wouldn't come configured to be used as a router, that'd need to be configured aftre it's arrived if you wanted to purchase one through us, otherwise the same would be if your purchased one loacally instead.
20m 29s
Chat has been opened from website: https://www.aussiebroadband.com.au/internet/home-phone/
E-mail from https://www.livechatinc.com/powered-by-livechat/?utm_source=chat_transcript&utm_medium=email&utm_campaign=www.aussiebroadband.com.au
What a mess! HTML markup in plain text, incorrect date (US numeric format, in this case to be interpreted as 8 April), ridiculous indentation—and no warning that I needed to accept mail from a system that I had never heard of, and which was too polite to tell me that it was for a transcript from Aussie Broadband. I suppose that's “modern”. Still, even if the information was incorrect, it shows an understanding of the material that I wouldn't have expected from a third party. My guess is that there's somebody (one person?) in the background who fields the more complicated questions.
| Sunday, 6 August 2023 | Dereel | Images for 6 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Pain?
|
Topic: health | Link here |
I've been off the amitriptyline for 4 days now, and gradually my sleep patterns are normalizing, though I still wake up in some discomfort a couple of times a night. The difference is that I can get back to sleep. About time, too.
|
Tidying up the office?
|
Topic: general | Link here |
I should really be doing something about tidying up the office and the room across the hallway (which I intended to call the “music room” until I moved nearly all the instruments into cabinets in the lounge room). Gradually I'm reaching a deadlock where I can't move anything because I have nowhere to put things. Here's the “music room”:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's not as bad as it looks. Most of the boxes are empty and intended to be used to put things in, mainly cables of different kinds. But first I need to know where to put them.
Spent some time looking at a couple of boxes to the right of my desks. They proved to contain computer documentation up to 50 years old, including notes I made in a computer course in January 1973 (the green file). What a lot of clear, detailed notes I made at the time! I never thought I had ever gone to so much trouble. And sadly none of the documentation is of much interest. Underneath was a video recorder and a surprising number of CDs:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
At this rate I'll never get done.
|
Caret browsing: caught in the act
|
Topic: technology, animals, opinion | Link here |
Into the lounge room this afternoon, turned on the TV and saw:
|
|
|||||||||
That's the thing that I had to remove on Thursday. But this time I know how it happened: Bruno had stepped on the keyboard. I should ensure that it's upright when I'm not there.
|
News download problems
|
Topic: multimedia, technology, opinion | Link here |
I download news streams from Al Jazeera multiple times during the afternoon: the news (32 minutes) and the first 18 minutes of the news at 14:00, 15:00, 16:00 and 18:00. The first 18 minutes are in case we want to watch them before the news broadcast is finished: look at the first 18 minutes, then get the rest from the 32 minute version. I do this with the invocation
(sleep $2 && ps wwaux|grep $ALJAZ | grep -v grep|awk '{print "kill -INT " $2}' | sh) &
LC_ALL=de_DE.UTF-8 youtube-dl https://www.youtube.com/watch?v=$STREAM -o $ALJAZ
STREAM is the YouTube stream identifier, ALJAZ is the name of the file and $2 is the duration (1080 or 1920 seconds). This starts downloading the stream and then kills the download after the specified time. The result has some errors, so I recode it with ffmpeg:
# And recover the mess that YouTube delivers
ffmpeg -i $ALJAZ -c copy ${ALJAZ}.mp4
But today at 15:35 there was no 30 minute version, only the first 18 minutes. What went wrong? Read the almost indecipherable transcript:
ylist/index.m3u8: Immediate exit requested
Exiting normally, received signal 2.
[ffmpeg] Interrupted by user
ERROR: unable to download video data: [Errno 2] No such file or directory: '/spool/News/AlJaz-15.part'
So ffmpeg receives a signal. But how? Clearly in this batch job it's not really “Interrupted by user”. And it repeated at 16:32 and 18:32, while the 18 minute versions were OK. So was the 32 minute version of 14:00. What's causing that? Why did it only happen now, after 3 years? Is it repeatable?
|
Kangaroos: the one that got away
|
Topic: animals | Link here |
Yvonne into the lounge room in the late afternoon to tell me that there was a kangaroo on the verandah behind me. A quick look showed that he had moved on to the arches just outside the lounge room window, about 5 m away. Off to get my camera.
Too late:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
He must have seen me, and I was just able to catch him in mid-air (and out of focus) before he disappeared into the neighbour's garden.
| Monday, 7 August 2023 | Dereel | Images for 7 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More garbled log messages
|
Topic: technology | Link here |
Seen again this morning:
bde.lemis.com kernel log messages:
+Limiting closed port RST response from 213 to 193 packets/sec
+nnfs sefrver s nsfeeurrsv esre revuerre ka:e/u:r eknao:t/ :r ensopto nrdeisnpgond
+i
+ng
+e
+ka:/: not responding
+nfs sernverf s serveeurr ekeuraeka:/:: /: is ailive agains
+ alive again
+nfs server eureka:/: is alive again
+Limiting closed port RST response from 203 to 194 packets/sec
+Limiting closed port RST response from 209 to 216 packets/sec
+Limiting closed port RST response from 236 to 193 packets/sec
+nfs server eureka:/home: nfsno t sreesponrdivnegr
+
+eureka:/home: not responding
+n<6>nffss sseerrvveerr eeuurreekkaa:://hhoommee:: nnoott rreessppoonnddiinngg
+nfs server eureka:/home: is alive again
+nfs server nfse server ueurreka:/homee:k ais :aliv/e ahgaoinm
+e: is alive again
+nfs server eureka:/home: is alive again
We've seen that before, and there are ways to fix it, but I find it more amusing than anything else. The cause is simple: I had taken bde off the network to test CJ Ellis' hardware, and I forgot to reconnect it before the nightly backup. No harm done.
|
The eternal system upgrade
|
Topic: technology, opinion | Link here |
The refereence to the garbled messages above brought me back to a topic that I've been postponing for years: upgrade eureka. It seems that I went to some trouble to document that, only a little more than 12 years ago. Will I ever finish?
|
Chasing the YouTube issues
|
Topic: technology, multimedia, opinion | Link here |
After failing to find the cause of the news download issues, put more debug code in. Download the news twice in parallel: will both copies do the same thing? Also link the downloaded file to a second name so that I have a copy if ffmpeg deletes the first.
The results were interesting:
Date: Mon, 7 Aug 2023 15:30:00 +1000 (AEST)
From: Cron Daemon <groggyhimself@tiwi.lemis.com>
ln: /spool/News/AlJaz-15.part: No such file or directory
Where did that go? It should have been there at the time.
Date: Mon, 7 Aug 2023 15:00:02 +1000 (AEST)
From: Cron Daemon <groggyhimself@tiwi.lemis.com>
To: groggyhimself@tiwi.lemis.com
Subject: Cron <groggyhimself@tiwi> /home/local/bin/aljaz AlJaz-15 1920
Message-Id: <202308070500.377502kQ021903@tiwi.lemis.com>
[youtube] Extracting URL: https://www.youtube.com/watch?v=gCNeDWCI0vo
[info] gCNeDWCI0vo: Downloading 1 format(s): 96
[download] Destination: /spool/News/AlJaz-15
[hls @ 0x85e7be000] Skip ('#EXT-X-VERSION:3')
...
yt-dlp automatically appends a .part to the file name. But looking more carefully, the whole transcript (all 5200 long lines of it) consisted mainly of (my highlighting)
[hls @ 0x85e7be000] Opening 'https://rr2---sn-fpqxc5oq-hxal.googlevideo.com/videoplayback/id/gCNeDWCI0vo.1/itag/96/source/yt_live_broadcast/expire/1691406004/ei/VHrQZMNPk46DxQ-hr7mIBQ/ip/121.200.11.253/requiressl/yes/ratebypass/yes/live/1/sgoap/gir%3Dyes%3Bitag%3D140/sgovp/gir%3Dyes%3Bitag%3D137/hls_chunk_host/rr2---sn-fpqxc5oq-hxal.googlevideo.com/playlist_duration/30/manifest_duration/30/spc/UWF9f1O91g67FqBoSl9xCO_7Tuyb0-hB8o_WzVAHag/vprv/1/playlist_type/DVR/initcwndbps/1772500/mh/Dn/mm/44/mn/sn-fpqxc5oq-hxal/ms/lva/mv/m/mvi/2/pl/22/keepalive/yes/fexp/24007246,24362685,51000012,51000024/beids/24350017/mt/1691383976/sparams/expire,ei,ip,id,itag,source,requiressl,ratebypass,live,sgoap,sgovp,playlist_duration,manifest_duration,spc,vprv,playlist_type/sig/AOq0QJ8wRQIhAKAyQOf-qmeBIRVNRWV2Afl6yP-_pNJ7Xw7MHQzTn6kVAiAKwgPKwe7oeZ5g8PADwB-cnnj8VxRbUPTKYz1P17_yaA%3D%3D/lsparams/hls_chunk_host,initcwndbps,mh,mm,mn,ms,mv,mvi,pl/lsig/AG3C_xAwRgIhAPqdjFz1OrQPRT_30p_-N-TkdV5ZI6vSuZJ7PXN2ELFWAiEA_iCNFbsqqWDlITcjNs7w4V8oDbftwtmHpuki8TNsaj8%3D/playlist/index.m3u8/sq/2595483/goap/clen%3D81639%3Blmt%3D1691289247121575/govp/clen%3D1803186%3Blmt%3D1691289247121573/dur/5.000/file/seg.ts' for reading
[https @ 0x8618f2b00] HTTP error 403 Forbidden
[hls @ 0x85e7be000] Failed to open segment 2595483 of playlist 0
So for some obscure reason the entire session was refused, request for request. Nothing to do with ffmpeg at all: there was no file to convert. Why is that?
| Tuesday, 8 August 2023 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Two months of pain
|
Topic: health, opinion | Link here |
It's been 2 months since my slipped disk, 2 months of gradually subsiding pain. The current status? Almost back to normal, but not quite. There's still some discomfort at night, and a certain numbness from time to time in my lower left leg. Last night there were also a couple of short but painful sensations in the upper left leg.
I'm walking more, and since paying attention to my posture I find that Yvonne has dropped considerably in height. But it's also clear that my legs are weaker: muscles have degenerated, and I'll need to do a lot of walking to get back to the way I was.
And somehow it's difficult to separate these issues with earlier ones. Somehow this all started 9 months ago with my test for osteoporosis. Maybe I shouldn't have done so.
|
News downloads: solved?
|
Topic: technology, multimedia, opinion | Link here |
So it seems that my errors downloading Al Jazeera news are related to yt-dlp. That happens; the sites find new ways to trick it, and yt-dlp finds way to outsmart the tricks.
I found a solution for that a couple of months ago: simply get https://github.com/yt-dlp/yt-dlp-nightly-build . But that (already!) no longer works:
=== root@tiwi (/dev/pts/6) /usr/local/bin 34 -> fetch https://github.com/yt-dlp/yt-dlp-nightly-build
fetch: https://github.com/yt-dlp/yt-dlp-nightly-build: Not Found
Much searching around the yt-dlp site, which suddenly seems so much more opaque than last time. But finally I found it:
=== root@tiwi (/dev/pts/9) /usr/local/bin 20 -> yt-dlp --update-to nightly
Available version: nightly@2023.08.04.221808, Current version: nightly@2023.05.11.094900
Current Build Hash: 68cfb85d91da7da1a1c6f0d01b0f488b89c3275ca89fb893d6e573ba07a30d9d
Updating to nightly@2023.08.04.221808 ...
Updated yt-dlp to nightly@2023.08.04.221808
And so far it seems to work. But it was flaky before, so I'll wait a couple of days before declaring success. Daniel O'Connor also notes that the “standard” release is available at https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp, but so far the nightly releases have worked.
| Wednesday, 9 August 2023 | Dereel | Images for 9 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Updating the E-1 firmware, try 2
|
Topic: photography, technology, opinion | Link here |
Two years ago I tried to update the firmware on my Olympus E-1. I knew it would be a problem because the modern software no longer wants to know about it, so I found an old version of Olympus Studio 2, which should have done the job, but I ran into trouble: it couldn't even detect the camera.
OK, I currently have lots of things that I need to do, so why not revisit the update? After all, I don't really need it. First, though, see what a real computer says about the USB connection:
Aug 9 13:41:55 dereel kernel: ugen0.4: <OLYMPUS E-1> at usbus0
Aug 9 13:41:55 dereel kernel: umass0 on uhub1
Aug 9 13:41:55 dereel kernel: umass0: <OLYMPUS E-1, class 0/0, rev 2.00/1.00, addr 3> on usbus0
Aug 9 13:41:55 dereel kernel: umass0: SCSI over Bulk-Only; quirks = 0xc301
Aug 9 13:41:55 dereel kernel: umass0:3:0: Attached to scbus3
Aug 9 13:41:55 dereel kernel: da0 at umass-sim0 bus 0 scbus3 target 0 lun 0
Aug 9 13:41:55 dereel kernel: da0: <OLYMPUS E-1 1.00> Removable Direct Access SCSI-2 device
Aug 9 13:41:55 dereel kernel: da0: Serial Number 500015490
Aug 9 13:41:55 dereel kernel: da0: 40.000MB/s transfers
Aug 9 13:41:55 dereel kernel: da0: 3831MB (7847280 512 byte sectors)
Aug 9 13:41:55 dereel kernel: da0: quirks=0x3<NO_SYNC_CACHE,NO_6_BYTE>
Aug 9 13:42:12 dereel kernel: ugen0.4: <OLYMPUS E-1> at usbus0 (disconnected)
Aug 9 13:42:12 dereel kernel: umass0: at uhub1, port 8, addr 3 (disconnected)
Aug 9 13:42:12 dereel kernel: da0 at umass-sim0 bus 0 scbus3 target 0 lun 0
Aug 9 13:42:12 dereel kernel: da0: <OLYMPUS E-1 1.00> s/n 500015490 detached
Aug 9 13:42:12 dereel kernel: (da0:umass-sim0:0:0:0): Periph destroyed
Aug 9 13:42:12 dereel kernel: umass0: detached
That seemed reasonable enough, so plug it into distress and see what the Olympus software has to say about it:
|
|||||||||||||||||||||||
|
|
|||||||||
The first message appeared when I tried to check for updates, and the second when I tried to save the configuration. At least it saw that there was a camera there, so that's some kind of progress.
But then I looked more carefully. I had started the wrong Olympus software, “Workspace” instead of “Studio”. That's much more modern, and we know that it doesn't work. What does “Studio” say?
|
|
|||||||||
That's what I saw before.
OK, could it be the cable I was using? It's a micro-USB connector, and it seems rather thin. Try with the E-300. The FreeBSD system sees pretty much the same thing:
Aug 9 14:03:59 dereel kernel: ugen0.4: <OLYMPUS E-300> at usbus0
Aug 9 14:03:59 dereel kernel: umass0 on uhub1
Aug 9 14:03:59 dereel kernel: umass0: <OLYMPUS E-300, class 0/0, rev 1.10/1.00, addr 4> on usbus0
Aug 9 14:03:59 dereel kernel: umass0: SCSI over Bulk-Only; quirks = 0xc105
Aug 9 14:03:59 dereel kernel: umass0:3:0: Attached to scbus3
Aug 9 14:03:59 dereel kernel: da0 at umass-sim0 bus 0 scbus3 target 0 lun 0
Aug 9 14:03:59 dereel kernel: da0: <OLYMPUS E-300 1.00> Removable Direct Access SCSI-2 device
Aug 9 14:03:59 dereel kernel: da0: Serial Number 623531061
Aug 9 14:03:59 dereel kernel: da0: 1.000MB/s transfers
Aug 9 14:03:59 dereel kernel: da0: 3831MB (7847280 512 byte sectors)
Aug 9 14:03:59 dereel kernel: da0: quirks=0x3<NO_SYNC_CACHE,NO_6_BYTE>
Aug 9 14:04:11 dereel kernel: ugen0.4: <OLYMPUS E-300> at usbus0 (disconnected)
Aug 9 14:04:11 dereel kernel: umass0: at uhub1, port 8, addr 4 (disconnected)
Aug 9 14:04:11 dereel kernel: da0 at umass-sim0 bus 0 scbus3 target 0 lun 0
Aug 9 14:04:11 dereel kernel: da0: <OLYMPUS E-300 1.00> s/n 623531061 detached
Aug 9 14:04:11 dereel kernel: (da0:umass-sim0:0:0:0): Periph destroyed
Aug 9 14:04:11 dereel kernel: umass0: detached
And “Studio” had no problems, and it confirmed that the firmware was up to date. So it really looks as if the claims are incorrect. Maybe I need a still older version of “Studio”; this one is version 2.30, and it seems to date from 2007.
|
The decline of digital SLRs
|
Topic: photography, opinion | Link here |
It has been nearly 10 years since I bought my Olympus OM-D E-M1 Mark I, and immediately after I got it it was clear that the days of the digital SLR are over. But of course, the dead live longer, and beginners are still looking for them.
But today I saw an article in DPReview, showing this image:
163,526,230 interchangeable lens cameras produced in the first half of 2023. Of these, 144,616,439 (88.4%) were “mirrorless”, and only 18,909,791 (11.6%) were DSLRs. I'm sure that this isn't a sudden change, but I only noticed today.
|
Apologies from Aussie
|
Topic: technology, opinion | Link here |
Our “experience” upgrading CJ Ellis' home networking has been less than positive. Aussie Broadband agrees. The necessary configuration details, including specific information for various routers, and also profuse apologies, and an offer to provide the router free of charge, along, of course, with help setting it up. We don't really need that unless there's something unexpected, but it's a nice gesture.
| Thursday, 10 August 2023 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Understanding SCSI return codes
|
Topic: technology, opinion | Link here |
One of the things that irritate me about the nightly backups are the “error” messages that I get from some disks:
(da0:umass-sim0:0:0:0): READ(10). CDB: 28 00 00 00 01 68 00 00 40 00
(da0:umass-sim0:0:0:0): CAM status: SCSI Status Error
(da0:umass-sim0:0:0:0): SCSI status: Check Condition
(da0:umass-sim0:0:0:0): SCSI sense: NOT READY asc:4,1 (Logical unit is in process of becoming ready)
(da0:umass-sim0:0:0:0): Polling device for readiness
Clearly that's a feature, not a bug. I'm planning to suppress it unless verbose boot has been specified. But there are others:
(da1:umass-sim1:1:0:0): READ(10). CDB: 28 00 00 00 00 a8 00 00 10 00
Is that an error condition or something similar? What do the codes mean? There's a surprising dearth of information. Google searches don't help: this decoder wants you to take the individual hex bytes and paste them into individual boxes. Sorry, I don't go for that kind of masochism. And then I found a FreeBSD man page SG3_UTILS(8) which seems to do exactly what I want. But it refers to a program sg_decode_sense (along with over 50 other similarly named programs, not all described).
Oh. I didn't know about that. Where are they? Not on my system. Ah, they're a port, and the web site is too polite to say so, let alone which port. OK, probably time to upgrade dereel to the latest and greatest source. After that, went searching for the program, which proved to be in sg3_utils-1.47, or for people like me who still believe in path names, /usr/ports/sysutils/sg3_utils. Installed that.
=== root@dereel (/dev/pts/6) ~ 8 -> sg_decode_sense 28 00 00 00 00 a8 00 00 10 00
Probably uninitialized data.
Try to view as SCSI-1 non-extended sense:
AdValid=0 Error class=2 Error code=8
Well, that helps, doesn't it? Time to RTFM?
|
TV series: oldies but goodies?
|
Topic: multimedia, opinion | Link here |
For some reason, most of the TV series we watch are in German, not so much because of the language as because English language series seem particularly unfriendly. And of course we're gradually running out of series to watch.
One German series that is not too stupid is „In aller Freundschaft - die jungen Ärzte“ (“In all friendship - the young doctors”). It's not as silly as the title makes it sound. And it's a spin-off from an older series called, well, „In aller Freundschaft“. Off looking for it today, and found about 150 out of 1000 episodes. Downloaded the oldest I could find, episode 272 „Versprechungen“ (“promises”) from June 2005.
It was horrible! Stupid topics, poor recording quality, implausible surroundings. Is that an indication of the age? We'll look at one of the most recommended ones before discarding the series altogether.
| Friday, 11 August 2023 | Dereel | Images for 11 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
dereel hang
|
Topic: technology | Link here |
Into the office this morning to discover no windows from dereel open. But that's to be expected, since I rebooted (from the lounge room) and hadn't been back into the office since.
But I couldn't contact it! And then I discovered that, according to ruptime, dereel had been down since about 21:00 yesterday evening. Powered on the monitor and saw a whole lot of
last message repeated 10 times
last message repeated 9 times
last message repeated 9 times
Tried to log in, but it hung. Tried to reboot, but it hung syncing buffers. Nothing for it, the Big Red Button.
It came back normally enough, but the log file was interesting. The “last message: repeated” was trivial:
Aug 10 23:25:04 dereel ntpd[949]: error resolving pool 0.freebsd.pool.ntp.org: Name does not resolve (8)
But once again I had garbled NFS messages:
Aug 10 21:09:10 dereel kernel: <<6>n6>nfs server fs server eueruerkeak:a/:d/ump: not responding
Aug 10 21:09:10 dereel kernel: d
Aug 10 21:09:10 dereel kernel: ump: <6>nnfos server teureka:/dump: not responding
Aug 10 21:09:10 dereel kernel:
Aug 10 21:09:10 dereel kernel: responding
Aug 10 21:09:10 dereel kernel: nfs server eureka:/dump: nnot respfosn dsieng
Aug 10 21:09:10 dereel kernel: r
Aug 10 21:09:10 dereel kernel: ver eureka:/dump: not responding
Aug 10 21:09:10 dereel kernel: nfs server eureka:/dump: not responding
Aug 10 21:09:10 dereel kernel: n
Aug 10 21:09:10 dereel kernel: fs server eureka:/dump: not responding
Aug 10 21:09:10 dereel kernel: nfs server eureka:/dump: not responding
Aug 10 21:09:10 dereel syslogd: last message repeated 3 times
Aug 10 21:09:10 dereel kernel: nfs server eureka:/dump: not respondnfs server iennfs server eurekg
Aug 10 21:09:10 dereel kernel: a
Aug 10 21:09:10 dereel kernel: :/dump: not responding
Aug 10 21:09:10 dereel kernel: nfs server eureka:/dnufmsp :s enrver eureka:/dump: not responding
Aug 10 21:09:10 dereel kernel: o
Aug 10 21:09:10 dereel kernel: t responding
Aug 10 21:09:10 dereel kernel: nfs server eureka:/dump: not rnfs server eureka:/dump: not responding
Aug 10 21:09:10 dereel kernel: e
Aug 10 21:09:10 dereel kernel: sponding
Aug 10 21:09:10 dereel kernel: nfs server eureka:/dump: nnofts responding
Aug 10 21:09:10 dereel kernel: s
Aug 10 21:09:10 dereel kernel: erver eureka:/dump: not responding
Aug 10 21:09:10 dereel kernel: u
Aug 10 21:09:10 dereel kernel: reka:/dump: not responding
Aug 10 21:30:00 dereel syslogd: sendto: Host is down
Aug 10 21:30:00 dereel syslogd: last message repeated 1 times
This time, of course, the messages are by no means harmless. What caused it? The NTP messages were useful up to a point, as was the hang on login: everything points to issues with the network interface. But after rebooting everything went smoothly. Now to see what happens in the evening.
|
More VoIP pain
|
Topic: technology, opinion | Link here |
We had just settled down to breakfast when the doorbell rang. CJ Ellis. I had asked him to send me email before he came. But, it seems, that wasn't possible: his “Internet” froze while he was writing an email message, and so he didn't have any phone either.
He came with old ATA and new Netcomm NF20MESH gateway. OK, finish breakfast and set the thing up. That's fairly clear: go to the SIP Basic Settin tab and fill in the details:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But it didn't work! The Authentication name is an 8 digit number, but after 5 digits I couldn't enter any more! That makes no sense. Off to check the Basic Setup tab.
Bad idea. It seems that it's a “wizard”, which means that it resets the settings to default without so much as a “by your leave”. And suddenly I had a device that wanted to talk ADSL!
|
|||||||||||||||||||||||
OK, nothing for it. I'm in trouble. Yes, I know that the thing connects via Ethernet, but what else? Call Aussie Broadband support and ask for a call back.
While messing around, discovered that there was something wrong with the keyboard on bde, the laptop. The digit 5 didn't work! Find another keyboard, and I was able to enter the information without any difficulty beyond this stupid insistence on hiding the passwords.
And then I got a call back from Emm of Aussie Broadband, who clearly had documentation with the help of which he walked me through the reconfiguration of the device with the aid of an app on my phone that displayed a photo of the screen, not an easy thing to do when holding it in one hand. Still, he seemed to see what he needed, though it didn't quite agree with his rule book.
Finally we were done. Does my “Internet” work now? Well, no. My link did work, but it's connected to my local network. Please connect to the NTD. No, that's not going to work. Firstly I'd lose the phone connection, and secondly, in all probability, I'll have to wait for a DHCP timeout on the link, which can take up to 30 minutes. So I'll try that by myself. In the meantime, please check the SIP parameters that I entered. They're for CJ. Since CJ was sitting next to me, I got him to confirm that I could speak on his behalf. Then, again, check the parameters on the screen. They're CJ's, not mine. Sorry, I need CJ's details. About all he said was that There Could Only Be One SIP line. Sorry, that's not correct. After all, this is a two line device.
At this point I gave up and messed around a little more, unfortunately with no results after 2 hours. Sent CJ off with his ATA, leaving the gateway for me to look at when my frustration level dropped sufficiently. And then I discovered that “it works better if you plug it in”. I had disconnected the WAN cable from the gateway at some point, but left it so that it looked as if it was plugged in. Plug it in properly and I got my DHCP lease from eureka and could connect to the net (“My Internet worked again”).
So, what about the SIP parameters? Still no joy. Check Every Single Parameter. Oh. There's a note in the setup instructions:
Note: If you are setting up VoIP on a NetComm NF18 (MESH or ACV) or a NL19, please leave Outbound Proxy blank/disabled.
But this is an NF20MESH, not NF18 or NL19. And the outbound proxy wasn't set. Since the other parameters were set, I had assumed that they were all correct. Set it. Success! Well, partially. Line 1 registers, and I could make calls from it. But line 2 doesn't register
What's special about line 2? It's the one with CJ's real phone number. Wrong password? How can I tell? Not only does the Netcomm replace it with bullets, it later sets it to 6 bullets, regardless of the real length. OK, program it into my ATA. Works, sort of. It can call out, and my called phone shows the correct number, but I can't call in. That's presumably a second issue: Vonex has officially given up the number, but it still displays when CJ calls me from the Vonex ATA. But that's a separate issue.
Much messing around. Yes, I can swap the two SIP credentials, and I always get line 1 working and line 2 not working. Could it be that Emm was right that There Can Only Be One? That would be a real catastrophe in view of what has gone before. But that was enough for one day; later it occurred to me that it would be interesting to disable line 1 without any other changes and see what happens. Mañana.
Other random things: the device is set up to log out every 10 minutes, and for some reason it hides the password field with helpful information, making it very difficult to log in again. But finally I found a page:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Set timeout between 300 and 86400 seconds! My guess is that many people won't recognize that high limit, and clearly it's Just Plain Stupid. Why not offer the option to disable logout altogether? Of course, maybe it's there, just well hidden.
Then a random snippet I got from the hold dialogue while Emm was looking for documents: “Need your Wi-Fi password? It's on the back of your modem”. Apart from the fact that it's not a modem, there's no password on the back of this device, one of only 2 that Aussie currently offer (the other is the Grandstream HT812, which doesn't do Wi-Fi in the first place).
As if that wasn't irritating enough, I heard a continuous crackling from the left loudspeaker. Where did it come from? No idea, but not from a computer. After checking and reseating all cables, it stopped after I disconnected and reconnected the left input to the amplifier, which wasn't even connected at the other end.
And then in the evening CJ sent me email. He had called Aussie Support (how?), and they had told him to take the ATA out of circuit and connect his computer directly to the NTD. And that worked. But how could he call them without the ATA?
And then another detail: CJ has paid another $30 to Vonex. He won't get that back, so he might as well maintain the status quo until he uses it up. It takes the pressure off me, anyway.
|
Where's my phone?
|
Topic: technology, general, opinion | Link here |
While messing around with CJ's hardware, Yvonne came in and asked me if I had seen her (mobile) phone. No, but there's a way to find it, of course: call it from another phone and follow the ring. But this time we didn't hear anything.
OK, plan B: ask Google Maps. I track Yvonne's location, clearly really the location of her phone. And sure enough, Google Maps said that it's here. Enlarge the map: by the shed, maybe? Off to the shed, where Yvonne had just found it in her saddle bag. Still, a good use for the location feature.
| Saturday, 12 August 2023 | Dereel | Images for 12 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More fun with dereel
|
Topic: technology, opinion | Link here |
How did dereel survive the night with the new kernel? Well, it seems. All was still up and running this morning. But then, while I was doing something across the net with firefox, it hung. I couldn't even ping it from eureka. It's beginning to look more like a network interface problem than an NFS problem.
OK, build a new kernel, see how that goes. Not well: once again the system hung trying to do the nightly backups:
Aug 12 21:09:47 dereel kernel: nfs server eureka:/dump: not responding
Later eureka had its own problems:
Aug 12 21:35:48 eureka kernel: em0: Watchdog timeout Queue[0]-- resetting
Aug 12 21:35:48 eureka kernel: Interface is RUNNING and ACTIVE
Aug 12 21:35:48 eureka kernel: em0: TX Queue 0 ------
Aug 12 21:35:48 eureka kernel: em0: hw tdh = 398, hw tdt = 649
Aug 12 21:35:48 eureka kernel: em0: Tx Queue Status = -2147483648
Aug 12 21:35:48 eureka kernel: em0: TX descriptors avail = 773
Aug 12 21:35:48 eureka kernel: em0: Tx Descriptors avail failure = 0
Aug 12 21:35:48 eureka kernel: em0: RX Queue 0 ------
Aug 12 21:35:48 eureka kernel: em0: hw rdh = 6, hw rdt = 5
Aug 12 21:35:48 eureka kernel: em0: RX discarded packets = 0
Aug 12 21:35:48 eureka kernel: em0: RX Next to Check = 6
Aug 12 21:35:48 eureka kernel: em0: RX Next to Refresh = 5
Aug 12 21:35:48 eureka kernel: em0: link state changed to DOWN
Aug 12 21:35:52 eureka kernel: em0: link state changed to UP
Aug 12 21:35:52 eureka devd: Executing '/etc/rc.d/dhclient quietstart em0'
I'd have to guess that the two are related, but it's not clear how.
|
Revisiting sweet and sour fish
|
Topic: food and drink, opinion | Link here |
I've been cooking sweet and sour fish from time to time for over 50 years now, and what used to be a harrowing experience to cook has dwindled to something similar to what I do for breakfast every morning. This evening I did it again and discovered that my choice of vegetables has changed: 50 years ago there were bamboo shoots and water chestnuts in almost every recipe, and now I hardly use them.
Checking, I found that we did have some water chestnuts, open and refrigerated since early May, and completely mouldy. No bamboo shoots. OK, what can I do instead? Spent the best part of an hour looking for alternative recipes, during which time I discovered that most recipes either want a whole fish or fish cubes deep fried in batter, compared to my intention of just frying it. Too late for that approach today, but I did find this recipe, which suggested pineapple pieces, and also red onion instead of the spring onions I used. That seemed a good compromise for today, so instead of the spring onions and bamboo shoots I used red onions (80 g) and pineapple (40 g).
That worked relatively well, but somehow the quantities need improvement. Yes, I used less fish (250 g rather than 300 g, because that's what we had), but the other ingredients were roughly the same. But there was too much sauce:
|
|||||||||||||||||||||||
I later discovered that I had been this way before, 12 years ago, and I came to similar conclusions then. But then as now I don't have a real improved recipe to put on the recipe page, so I'll just put part of this discussion on the recipe page.
In passing, I cooked the (normal long grain) rice with more water this time. In the past I had used 1.5 times as much water as rice. Today I used 1.7 times as much. The difference is obvious, but I can't make up my mind which is better. Given that some people use up to 2 times as much, I'll leave it like this for a while.
| Sunday, 13 August 2023 | Dereel | Images for 13 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Cold!
|
Topic: general, technology, opinion | Link here |
Winter is drawing to a close, but last night was still one of the coldest, with a minimum temperature of -0.3° at the surprisingly late time of 7:01. And the air conditioner couldn't cope. It didn't really get the house warm until after 9:00. Hopefully we won't have too many cases like that.
|
More dereel pain
|
Topic: technology, opinion | Link here |
After yesterday's renewed failure with dereel, I started looking at other things. And the firefox that I ran on dereel, displaying on eureka, was ridiculously slow. And saving mail messages across NFS took so long that I aborted the attempt. What does ping say?
=== root@eureka (/dev/pts/5) /var/log 40 -> ping -c 100 dereel
PING dereel.lemis.com (192.109.197.135): 56 data bytes
64 bytes from 192.109.197.135: icmp_seq=0 ttl=64 time=15.803 ms
64 bytes from 192.109.197.135: icmp_seq=1 ttl=64 time=15.551 ms
...
Huh? 15 ms? This is a 1000 Mb/s interface. Other systems?
=== root@eureka (/dev/pts/5) /var/log 41 -> ping -c 100 tiwi
PING tiwi.lemis.com (192.109.197.177): 56 data bytes
64 bytes from 192.109.197.177: icmp_seq=0 ttl=64 time=0.330 ms
64 bytes from 192.109.197.177: icmp_seq=1 ttl=64 time=0.338 ms
^C
That looks better. So at least I now have some difference. Reboot to the old kernel and carry on as before.
Oh. Still 15 ms ping times! A bit more investigation, and it seems to be related to firefox! When it's not running, ping times are normal, When it's running and doing something, the ping times go up to marginally over 15 ms. It seems that something is putting in a 15 ms delay. But how can firefox do that?
|
Anniversary breakfast
|
Topic: food and drink, opinion | Link here |
Gradually my breakfasts are getting boring. But on the other hand I have a number of spice pastes in the pantry that I've been meaning to try. The thing is that they're not really appropriate, but then most of my breakfasts abuse the original intention.
So how about this paste? As it happens, today is the 5th anniversary of its use-by date:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
OK, how about faking a prawn and squid noodle soup with it?
Per portion, except for the broth.
| quantity | ingredient | step | ||
| 1 sachet | Teans Gourmet Tom Yam paste | 1 | ||
| water to 1250 ml | 1 | |||
| 2 | fish balls | 2 | ||
| 60 g | pok choi, choi sam or similar, chopped | 2 | ||
| 50 g | squid rings, | 3 | ||
| 60 g | prawns | 3 | ||
| 40 g | bean sprouts | 3 | ||
| 150 g | Cooked noodles | 4 | ||
|
Cooking, cooking, cooking
|
Topic: food and drink, opinion | Link here |
But why only one dish today? How about some enchiladas? First, of course, we need the salsa. Off to make that, and discovered that I didn't have enough time to complete everything. So we had a omelette for dinner. Tomorrow the enchiladas.
|
Cooking: all good things are three
|
Topic: food and drink, opinion | Link here |
So, fake Tom Yam, enchiladas, ... All good things are three. And sure enough, I had forgotten to bake a loaf of bread yesterday, so today's the day.
But that starter:
|
|||||||||||||||||||||||
Is that OK? I don't know. It dates from 21 May, before my slipped disk. But it should be able to last that long. Still, I have a newer one, so why take the risk?
| Monday, 14 August 2023 | Dereel | Images for 14 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Air conditioner problems
|
Topic: Stones Road house, general, technology, opinion | Link here |
Funny noises from the air conditioner round 6:00 this morning. Not the loud noise of last month, just a too-short on and then off cycle.
Out to take a look. Once again the temperature in the lounge room was a couple of degrees too low, but there was nothing obvious in the displays. “Soft” power cycled the unit (from the control panel), and it displayed error code E14. From the error code list that I photographed on December 2019, that means:
|
|||||||||||||||||||||||
Not enough refrigerant? The unit still ran, but poorly. The outside temperature was round -1°, so that in itself wasn't unusual. But at 8:30, though it was running, the temperatures were 15.6° in the office and 18.3° in lounge, both with set temperature 21°. But the output from the heating vents was only 25°, so that's not surprising. Later it went up to 35°, after which I didn't measure any more. In the course of the morning temperatures were up to normal.
OK, time to call Tony Nesci, who has enough on his plate this week, it seems. He'll try to get here sometime this week, but it could well be the next week. About the only good things about the matter are that it's relatively certain what the problem is, and that if it doesn't get overly cold at night, we should be able to weather it.
|
dereel: I give up
|
Topic: technology, opinion | Link here |
So what's causing my network problems with dereel? I give up. But it looks serious, and the last version of the kernel I bought had the version string FreeBSD 14.0-ALPHA1. That's potentially a release show-stopper. Sent off a message to the FreeBSD developers' list.
Responses from Tomoaki AOKI and Kevin Bowling, the latter the developer responsible. He had added code for TSO to the driver, and with a little checking found that it was enabled:
=== root@dereel (/dev/pts/1) ~ 14 -> ifconfig em0
em0: flags=1008843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
options=4e527bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,WOL_MAGIC,VLAN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6,HWSTATS,MEXTPG>
ether 6c:0b:84:a5:c7:1c
inet 192.109.197.135 netmask 0xffffff00 broadcast 192.109.197.255
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
Disable it:
=== root@dereel (/dev/pts/1) ~ 15 -> ifconfig em0 -tso=== root@dereel (/dev/pts/1) ~ 16 -> ifconfig em0
em0: flags=1008843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
options=4e524bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,LRO,WOL_MAGIC,VLAN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6,HWSTATS,MEXTPG>
ether 6c:0b:84:a5:c7:1c
inet 192.109.197.135 netmask 0xffffff00 broadcast 192.109.197.255
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
And left that to run.
Kevin also asked for the sysctl output for em0 in case of failure. I didn't get that far, but looking at the output is amazing:
=== root@dereel (/dev/pts/1) ~ 19 -> sysctl dev.em|wc -l
156
I didn't get round to looking at it, but I wonder if we shouldn't be doing more to document these features.
|
Enchiladas, day 2
|
Topic: food and drink, opinion | Link here |
Continued today with yesterday's enchiladas. It still took nearly an hour.
First problem: the tortillas (made with the Colombian Doñarepa that Yana brought with her last month) were not flexible enough. I'm not sure that this has anything to do with the masa, but in any case Yvonne (who did the wrapping) needed to use toothpicks to hold the enchiladas together, and even then a couple broke:
|
|||||||||||||||||||||||
Probably I need more water with the masa. So far I have been using 1.6 times as much water; for the last 3 tortillas I used 1.8 times as much instead, which made for thinner tortillas.
And then the cheese. It should be queso cotijo, but I don't know what that is, nor where to get it from. I tried with fake Australian Mozzarella instead:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The result? Barely edible. The “Mozzarella” is definitely wrong, and I ended up scraping it off the enchiladas. I'm wondering if I shouldn't stop eating it altogether. But even the rest didn't taste as good as I recalled. That may be failing memory: 10 years ago I noted that I didn't like enchiladas very much, and 7 years ago I wasn't overly impressed either.
| Tuesday, 15 August 2023 | Dereel | Images for 15 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More air conditioner fun
|
Topic: Stones Road house, general, technology, opinion | Link here |
To quote the Ballarat Courier:
You were not the only one who struggled to get out of bed this morning, nor the only one who was frozen in their car or with pipes that just refused to work.Ballarat recorded its coldest morning for the year, with the Bureau of Meteorology's weather station at the airport hitting -3.6 degrees at 7am.
It wasn't quite that cold here, for some reason, only -1,4°, but it was far too cold for our ailing air conditioner. When I got up round 8:00 the temperature was still below 18°, and the unit was working only intermittently. At 9:30 I did a hard power cycle (turned off the mains power to the unit), after which it displayed the dreaded E14 during normal operation. It ran until 10:10 and then stopped, though the two zones that I had selected (set for 21°) had only 19.2° and 17.1°. Out to look. Although the outside temperature was up to 7°, the inlet temperature to the condenser was only round 0°, so presumably that was the reason. By midday inside temperatures were back to normal. One more cold night tonight, and then it should be milder. Hopefully Tony will make it here before the next cold snap.
|
More mobility issues
|
Topic: health, opinion | Link here |
Making breakfast today, I needed some garlic chives. No worries, there are plenty in the garden. Cut some and... I couldn't get up again! On the first attempt I nearly fell over, and I needed a nearby chair to pull myself up again. Later, walking the dogs, I tripped and nearly fell on my face, only stopping by grabbing a wall.
Clearly there's still a long way to go here. Somehow it all started with my DEXA test last November. Can I recover?
| Wednesday, 16 August 2023 | Dereel | Images for 16 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Backing up bde
|
Topic: technology, opinion | Link here |
bde.lemis.com (the laptop) has been connected to CJ Ellis' Netcomm NF20MESH router since last week, so it didn't do the nightly backups. But why not? The other (WAN) side of the router is connected to our local LAN. Can't I just mount the file systems across the router?
Yes, sort of. The router has its own view of the DNS world:
=== root@bde (/dev/pts/10) ~ 21 -> cat /etc/resolv.conf
search home
nameserver 192.168.20.1
“home”. What kind of domain is that? Yes, I could replace it with the correct resolv.conf, but it doesn't seem to be worth the trouble when any reset would also replace it with this nonsense. So just mount the file systems with fully qualified domain names:
=== root@bde (/dev/pts/10) ~ 22 -> df /dump
Filesystem 1048576-blocks Used Avail Capacity Mounted on
eureka.lemis.com:/dump 9238060 3601610 4897404 42% /dump
And that worked, and also enabled me to start an xterm on eureka, greatly simplifying things.
|
VoIP setup, next step
|
Topic: technology, opinion | Link here |
Mail from a Jeremy from Aussie Broadband regarding the complaint about CJ Ellis' Netcomm NF20MESH router. I hadn't done anything more since Friday. He's the successor to Karen, who left at the end of last week.
OK, time to do so again. Called up Aussie at 11:24, and almost immediately got a call back from Andina, if I got the name right. This time I started the authentication with CJ's name, explaining that I was authorized, and that I was setting up the router for him. This time she wanted the post code for Berringa, something that they otherwise don't think important enough to put on their pages. Got past that. No, the router is down! I had to explain (several times in the course of the next 75 minutes) that I wasn't at CJ's place, and that I had two problems: the first line works, but I can't call in, and nothing that I could do would get the second line to work.
Off she went, putting me on hold, presumably to address the first issue. When she came back, she didn't have much to say about it. She just said that the router wasn't connected to the NBN. No, as I said, it was on my LAN. How about this photo app that Emm used last week? OK, tried that, but she couldn't read anything. Finally she accepted my suggestion that I tell her what was on the screen, including those fields (such as SIP Proxies) that worked with the first line. I had to remind her several times that the problem was limited to the second service, and that there was no need to replace a password, because both worked on the first service.
More discussion, and it seems that the device has the ability to be accessed from the WAN, normally disabled. She asked me to enable FFTP (and not FTP or TFTP), which proved to be her way of saying HTTP:
|
|
|||||||||
But then of course there's the firewall. That's straightforward enough. With only a couple of problems I added:
=== root@eureka (/dev/pts/5) ~ 76 -> ipfw add 11 allow ip from 192.109.197.243 to any
00011 allow ip from 192.109.197.243 to any=== root@eureka (/dev/pts/5) ~ 77 -> ipfw add 11 allow ip from any to 192.109.197.243
00011 allow ip from any to 192.109.197.243
And that worked. But in the meantime she decided that the device is defective, and she'll send a new one. I have great doubts; the symptoms don't sound like a defective unit. But it's a possibility, even if an expensive one. I reminded her several times that the unit should be configured and tested, and she repeated that it would be. She said that she did look at the device configuration, but unless she had the unpronounceable password, I don't know how she got in. Do they maybe have a master password?
Waited in vain for a mail message confirming that the new device was under way. What I did get was a mail from Jeremy which appeared to suggest that this was all my fault:
I have reviewed the latest update. It seems that our device was reset and the ACS details for remote access were wiped. Our team has facilatated a replacement. Please ensure that you have called in when changing over the devices so that we may ensure the second VoIP number is correctly configured through remote access.
Our customer support team is providing excellent care in addressing you in the moment. Please allow our customer support team to triage your concerns. If you have any concerns with our handling of your support ticket, please reach out to me for further assistance.
I have difficulty understanding that. What does he mean by ACS? Certainly not the Australian Computer Society. Maybe Access Control Service? That's obsolete. In any case, I sense the implicit accusation that I had misconfigured the device. And while remote access is an excellent idea under these circumstances, it's not standard, and it has potential for abuse. Is he suggesting that they deliver these devices with a remote access loophole without telling people? Surely not.
And under the circumstances the claim of “excellent care” is doubly wrong: first, they have admitted serious issues, and secondly that's not the way to treat a customer who is clearly upset. I wonder what training the complaints people get.
Finally got the notification that it had been sent:
From: Aussie Accounts <accounts@aussiebroadband.com.au>
Date: Wed, 16 Aug 2023 at 12:27
Subject: Re: [Ticket#2023081610010017] Warranty Replacement Confirmation
To: Cyril (CJ) Ellis <cjellis@gmail.com>
Hi Greg,
As per our discussion today, we're sending out a warranty replacement unit.
Sent to the wrong email address, and no mention that it would be configured and tested, of course, just a lot of threats of what would happen if the old unit was not returned quickly and in a clean and hygienic condition. Somehow I don't get the impression that they're trying to be user-friendly.
In passing, it's amazing how much traffic went to the router:
=== root@eureka (/dev/pts/5) ~ 83 -> ipfw show 11
00011 525617 27886283 allow ip from any to 192.109.197.243
00011 810348 1202346708 allow ip from 192.109.197.243 to any
Those are gigabytes of data. Is it correct? Yes, there's SIP traffic there, but no active phone calls. And somehow the device is pinging eureka.
|
A new dog?
|
Topic: animals, opinion | Link here |
While I was on the phone with Aussie Broadband, Yvonne called in to ask if we wanted a new dog, and to tell me that (while we were speaking) somebody had rammed her car.
The ramming proved not to be serious: if any damage was done, it was to the other car. But another dog? Why? And a greyhound/whippet cross? When she got back, she gave me the details, one of which was that she was likely to kill any cat with which she came in contact. So taking her is out of the question. But the reason for considering her in the first place? We don't need another dog. But it seems that the owner can't keep her, and he can't find anybody to take her either. Given the ridiculous prices that people charge for animals nowadays, that's really puzzling
|
Bruno does a runner
|
Topic: animals | Link here |
Somehow the dogs were particularly irritating this evening, and towards 22:00 Yvonne tried to put them out in the small yard behind her bedroom. And they refused to go. OK, I came along and “helped” them. But I forgot to shut the bedroom door, so Bruno came out too and disappeared into the garden.
Off with torches to find a cat running around at top speed. Somehow we managed to catch him, but I had hoped that that kind of antic was a thing of the past. When I let the dogs out the front door, he sits and watches, but doesn't try to escape.
Could we have left him out overnight? Possibly, but there's no guarantee that he wouldn't have gone out to the road. A good thing we caught him.
| Thursday, 17 August 2023 | Dereel | Images for 17 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
History?
|
Topic: history, opinion | Link here |
Should I recall what happened 58 years ago, like I've done almost every year? No.
|
New computer, finally?
|
Topic: technology, opinion | Link here |
It's been over 9½ years since I built eureka, my main machine, and over 7 months since I last started planning a replacement for it. I really need to do something. Back to looking at hardware choices, and it occurred to me that I hadn't really considered Intel processors.
One of the criteria was power consumption, also one of the factors limiting the speed of my chosen processor. 105 W I could accept; 280 W I couldn't. And 170? One of the reasons I hadn't made much progress.
But why hadn't I looked at Intel? Did so today and came up with the Core i9 13900F. 65 W TDP! Here an updated excerpt from January's table:
| Processor | CPU Mark | Single thread | Cores | Threads | TDP | |||||
| Ryzen 9 7950X | 63513 | 4314 | 16 | 32 | 170 | |||||
| Core i9-13900KS | 62044 | 4784 | 24 | 32 | 150 | |||||
| Core i9-13900F | 53513 | 4514 | 24 | 32 | 65 | |||||
| Ryzen 9 7900X | 52076 | 4319 | 12 | 24 | 170 | |||||
| Core i9-13900 | 49164 | 4404 | 24 | 32 | 65 | |||||
| Ryzen 9 5950X | 45881 | 3467 | 16 | 32 | 105 |
The difference between 65 and 170 W is 2.5 kWh per day, or $0.88 at current prices ($0.35/kWh). That's $322 per year, or over $3000 in 9½ years. And the exact performance figures also keep changing.
So what's the story? The processors have two different kinds of core: 16 “efficient” cores with one thread each, using 65 W, and 8 “performance” cores with 2 threads each. And they can use up to 219 W. But how often do you need the “performance” cores?
Another interesting thing is that the single thread performance of even the cheapest Intel chip is better than the best Ryzen in this comparison. Potentially that's good; a lot of software, such as enblend, uses only one core.
Other issues: the Intel chips can only address 128 GB RAM, compared to 192 GB for the Ryzens. Is that important? And they have 120 MB of cache memory, if I have understood this correctly. It's not that long ago that my computer main memory was less than that.
Of course, there's yet another question: how does FreeBSD handle two different kinds of cores? Scheduling needs to know about their relative power. Still more stuff to investigate. There's also this review that I should read.
|
Don't fire people because of their medication!
|
Topic: health, opinion | Link here |
I've been a member of the Australian Legalise Cannabis Party for nearly a year now. Why? I don't consume drugs. But it's clear that drug consumers aren't going to go away, and governments around the world spend inordinate amounts of money chasing criminals who abuse the illegality of the substances. Better to try to control it, like most governments around the world do with alcohol.
The party made it into parliament in the last state elections, and since then they have done things that I found unexpectedly good. Now they're drawing attention to a complete nonsense in the laws: while recreational use of cannabis is still illegal in Australia, it is widely used in medicine. But if somebody is being treated with cannabis and is subjected to a drug test at work, he is liable to be dismissed. That's completely stupid, and it's in stark contrast to the widespread legal use of opioids.
Why is the law like that? My guess is that it's because it hasn't been brought up to date in accordance with other legislation. Will it get fixed? Given the prevalent opinions, I'm not sure. But I have certainly sent off my letters to the lawmakers that be.
| Friday, 18 August 2023 | Dereel → Ballarat → Dereel | Images for 18 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Next Aussie Broadband blunder
|
Topic: technology, general, opinion | Link here |
Mail from CJ Ellis this morning: his new router is waiting for him at the Napoleons post office.
On the one hand, that's not bad. They really did get it sent immediately. But aren't the details clever? That's 32 km from his home, so he's involved in a 64 km journey over and above bringing it to me. And I had specifically requested and received confirmation that it would be sent to me.
As it happened, I was in town in the afternoon, so on the way back I dropped into the post office. Yes, the parcel was there. But the address I gave (Derwent Jacks Road, Berringa) was wrong! It proved that it had been addressed to CJ at my house!
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
With some difficulty I managed to persuade the postmistress to give me the package; I suspect that Aussie Broadband doesn't want to hear her opinion. Chris Bahlo was also there by chance, just about to jump in and give me a character reference.
This has been going on for nearly a month now! Can it get any worse?
|
More east Asian food
|
Topic: food and drink | Link here |
To town this afternoon, first to the Fruit Shack, where they seem to have given up on noodles. Once they had a whole shelf of them, but now there are so few (maybe only 20% of what they used to have) that I couldn't find anything I liked the look of.
On to the Pilipino shop next door, where to my delight I found not one, but both brands of caramel for KL Hokkien Mee. I've never seen the other one (Yuen Chun) before:
|
|||||||||||||||||||||||
|
Parkinson's?
|
Topic: health, opinion | Link here |
On to the doctor, getting caught up in enough traffic to arrive 4 minutes late, something that I hate. I've been having involuntary shaking in my right leg, and I suspect it was related to my slipped disk. Heather Dalman had already looked at them and decided that it wasn't restless legs syndrome, but what is it?
Paul didn't know either, but he took it seriously enough to send me to have a Doppler imaging of the veins in my right leg. He didn't think it was Parkinson's disease, but he didn't rule it out either. More fun.
| Saturday, 19 August 2023 | Dereel | Images for 19 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Aussie router: proof of the pudding
|
Topic: technology, opinion | Link here |
Aussie Broadband didn't crown themselves in glory by sending the new router to the wrong address. But I didn't have time yesterday to look at what they had sent. I managed to wait until after breakfast to ensure that I didn't get a stomach upset, and that proved to be a good idea.
The agreement was to configure and test a new router and send it to me. They failed on the second by putting in my address but CJ's name. How about the first? In particular, does the default configuration really have open ports, as I could have deduced from Jeremy's mail yesterday:
We supply our routers preconfigured for service and our ACS details are included by default with the configuration. They are able to be removed by the subscriber at any time. We do not access the router without authority from our subscribers.
Despite my request, he did not explain what he meant by ACS. But given that statement and that Andina had asked me to open port 80 in order to access the machine, it looks as if Aussie really ships the device with open ports.
But no, it seems OK:
=== root@eureka (/dev/pts/5) ~ 101 -> nmap 192.109.197.242
PORT STATE SERVICE
5060/tcp open sip
Nmap done: 1 IP address (1 host up) scanned in 8.12 seconds
OK, what about the SIP configuration? Only the first line showed an “active” LED, not what we agreed. Fight my way into the device, this time looking more carefully at the invalid certificate that so upset firefox:
|
|||||||||||||||||||||||
|
|
|||||||||
example.com? This certificate is just a sample! That's not Aussie's fault, at least directly. But what kind of second-rate vendor can't even get a valid certificate? And why does Aussie recommend them anyway?
Finally got to the configuration pages:
|
|||||||||||||||||||||||
|
|
|||||||||
So not only have Aussie not tested the thing, they haven't even completely configured it.
So, does the one line work? I can call out, but I can't call in. That could be related to this misconfiguration:
|
|
|||||||||
“Use SIP Outbound Proxy” isn't set. The instructions I received 10 days ago implied that I needed to set this parameter, and I didn't get it working until I did.
OK, this is just plain stupid: an Aussie support person decided, with little evidence, that the router I had was defective. Thus the inability to call in to one line (not the one configured here) and the inability to configure the second line. So they sent a second one, promising several times to ensure that it is not only configured, but also tested. Also agreed to send it to me, not to CJ.
What did they do? Sent a partially configured device which clearly hadn't been tested, and that to any invalid person/address combination. You'd think that they were trying to annoy me. Even in regard to this sad story, this is a low point.
So, what do I do? I could try configuring it myself, but why? I already have a functional device. And this message from Jeremy on Wednesday seems to be trying to blame me for their myriad fault:
I have reviewed the latest update. It seems that our device was reset and the ACS details for remote access were wiped.
So I won't give him any room for passing the buck. They can call in with their mysterious ACS and fix the problem, or I can return it to them. Started writing a long message, which I may finish tomorrow. But what's really sad is that once Aussie were the Good Guys. They have certainly “wiped” that impression with me. It'll be a long time before I can trust them again. I'm reminded of Simon Hackett's experience with what became of Internode after he left the company. Ironically, that drove him to Aussie. Would he still do so after an experience like mine?
Later in the afternoon CJ turned up, unannounced. I suppose I could give him the old device. Checked that it still worked, and yes, even call-in worked on his normal number, one of the things that I had reported on Wednesday. So finally the porting is over! He took the router with him, connected it up, and it worked. So modulo the second line, which he doesn't plan to use anyway, things are fixed from his side. And the fact that call in works on this device also means that the failure on the other device is not related to the somewhat unorthodox network topology.
Well, almost. He received a postage charge of $15 for one of the devices. That wasn't agreed upon. Can Aussie do nothing right any more?
| Sunday, 20 August 2023 | Dereel | Images for 20 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
15 years of panoramas
|
Topic: photography, technology, opinion | Link here |
Fifteen years ago today I did my first panorama shots, of the house in Kleins Road:
|
|||||||||||||||||||||||
I wasn't overly happy with the results even at the time. The gradation was bad, the cropping was bad. But since then I've refined my techniques, made scripts of them all. How difficult can it be to repeat and compare?
The biggest issue, amusingly, was that each view was an individual shot, not the 3 bracket HDR shots that I did nowadays. But without too much difficulty I came up with this:
|
|||||||||||||||||||||||
From a technical viewpoint, that's quite acceptable.
|
Bruno and Elena come closer
|
Topic: animals | Link here |
Bruno gets on quite well with the dogs, particularly with Larissa, with whom he plays surprisingly roughly. But Elena is another matter. Things are improving, though:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Honest scammers
|
Topic: technology, general, opinion | Link here |
Read in my mail inbox today:
401 N + 20-08-2023 To grog@lemis.c ( 248) Australia Post Servi N + RE: Final warning: your package cannot be delivered. ! ! !
It wasn't until later that I saw these silly trailing ! ! !, but even without it something rang a bell. What does the content look like? HTML only, of course, but it started:
<!--
Here we go some texte here to fuck security 3f326964bc9f8685b63713420ca2e1bf
what do you think about it51e26b3b627b4187657d5f7ec4030caf
--><!DOCTYPE html>
Well, at least they're honest, even if their spelling and punctuation leave something to be desired.
|
Paella again
|
Topic: food and drink, general, opinion | Link here |
Paul and Suzy Kirby and Graeme and Linda Swift along for dinner this evening, for which we did gazpacho andaluz and paella valenciana. For the latter I made some changes: less chorizo (200 g instead of 400 g) and more squid (200 g instead of 100 g).
Suzy had had reservations about eating paella, apparently because of a bad experience in the past, so Yvonne had cooked something else for her (incongruously, berner Rösti), but when she saw it, she tucked in just as much as the others. Graeme explained: there are many kinds of paella, most without seafood, and like the previous one Suzy had, frequently just with lots of tomato.
Clearly people enjoyed it:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And based on what's left over, I think I can change squid and chorizo yet again: 250 g of each next time.
|
Ghostly music
|
Topic: music, technology, opinion | Link here |
It seems that Paul plays the Didgeridoo, and he was interested in the instruments I have in the cabinets, particularly the bassoons
So I had to pull one out (the oldest, the Sautermeister), and of course Suzy and Linda had to pull out their mobile phones and take video of the matter. And how about that, it's haunted! Lots of orbs in the recordings! It seems that all four guests are interested observers of the paranormal, and this bassoon clearly has ghosts associated with it. Certainly the fact that both phones recorded them, while neither did when I brought out my Simiot-Tabard flute.
So: what it it? It couldn't be an issue with the video recording software, could it? It must be magic.
| Monday, 21 August 2023 | Dereel | Images for 21 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
A hollow voice croaks “Interoperability”
|
Topic: general, technology, photography, opinion | Link here |
Yesterday Suzy Kirby and Linda Swift took videos of me playing the bassoon to illustrate the “orbs” that it created. And they promised to send them to me.
I still don't have them. It's not clear what happened to Suzy's, but Linda sent hers to Yvonne via Facebook. And for whatever reason, I can't access it. Yvonne is investigating, but so far I haven't been able to liberate them.
Part of this is due suboptimal choices, of course. It should be possible to send videos to people's phones, though I recall having had issues with that in the past. Somehow this whole mobile phone scene is fraught with incompatibilities. The app is in the foreground, not the function.
|
Goodbye Marriotts
|
Topic: general, opinion | Link here |
Saw Garry Marriott, our neighbour to the south, for the first time in weeks this afternoon. They've sold their house! Well, the actual date is not yet clear, but they're hoping to be out of here before Christmas. And this was the man who said, 10 years ago, that he would only leave feet first.
|
Out of saffron!
|
Topic: food and drink, opinion | Link here |
Yesterday's paella finished off almost all our saffron. Off to look for some more online, greatly impeded by inaccurate pricing. Finally found 9.2 g for $45, post included. But the most expensive stuff was 0.1 g packs for $10, postage extra. They only had 10 packs, half of what we need for a paella, but if they had had enough, it would have cost $205.50! And what we bought would have left little change from $1000. I wonder if they sell any of it.
|
Another bassoon resource
|
Topic: music | Link here |
While looking at old musical instruments yesterday, came across this page about Heckel bassoons, which I don't think I have seen before. They describe details of my Heckel number 3620 that I hadn't known:
Early prototype back thumb bridge for C# trill key. This is an early example of a design that would later return as a seldom used factory opinion.
They also note that it was shipped on 30 April 1891, somewhat earlier than the 1892 that I had been lead to believe.
| Tuesday, 22 August 2023 | Dereel | Images for 22 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Garden flowers in late winter
|
Topic: gardening | Link here |
It's a month before the equinox, time for the monthly garden photos. The combination of winter and lack of attention to the garden shows. There's not much to see, though a comparison with last year is of interest. In particular, the Hibiscus rosa-sinensis “Uncle Max” seems to have recovered, though it didn't flower all year. Here this time last year and now:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Alyogyne huegelii is still crooked, but doesn't seem to be suffering:
|
|||||||||||||||||||||||
The Buddleja x weyeriana on the south side of the house has finished flowering, but it has also grown almost beyond recognition since being planted there in November (first photo):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And both Abutilons are doing well:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The first is the one over Piccola's grave, which clearly needs weeding.
It's far too early for Clematis to flower, but there are signs of preparation:
|
|||||||||||||||||||||||
The Magnolia × soulangeana that we received in March seems to be preparing buds:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And as ever, the last roses of summer still haven't given up:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Even a Strelitzia reginae seems to be showing signs of recovery:
|
|||||||||||||||||||||||
It has only been there 8 years.
It's a little early for tulips, and most seem to have died, anyway, in stark contrast to the Narcissus. But one is flowering anyway:
|
|||||||||||||||||||||||
|
Upcoming politician
|
Topic: politics, opinion | Link here |
Seen in the Ballarat Courier: Clarendon teacher resigns after expletive-laden rant to students. It seems that students made video recordings of her ranting:
In the video, the teacher berates students for the "f---ing appalling mistakes" in their essays, describes some of the work as "absolute dog sh-t" and statements that are "so f---ing sh-t". "I nearly killed myself as I was reading them," she said in frustration at the work some of the students handed in as draft essays.
What happened? She was dismissed from her position. She should have listened to Volodymyr Zelenskyy and got her students to crowdfund launching a political career.
|
Bruno and Yvonne
|
Topic: animals | Link here |
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Wednesday, 23 August 2023 | Dereel | Images for 23 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Digital driver license
|
Topic: technology, opinion | Link here |
Message from VicRoads today: I qualify for a “digital driver license”. This is a pilot project that seems to be limited to Ballarat, but it sounds interesting. It even seems to be relatively well done by Australian government standards: install the app, enter details including my VicRoads password, and choose a 6 digit pin. Since my phone has the feature, also have the option of using the fingerprint reader.
How secure is that? As secure as a physical driver license or more so, I think. I (physically) lost my license, probably in April this year, though I didn't find out until May. Anybody who found it could have used it for identification purposes. So this might actually be more secure. I wonder who accepts it.
|
Excuses for domestic violence
|
Topic: fiction, general, opinion | Link here |
Yvonne has a bruise on her chin:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
How did that happen? We've decided that it must have happened during a siesta, when she nodded off and the corner of her hard-cover book bored itself into her chin. But who would believe that?
| Thursday, 24 August 2023 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Panic!
|
Topic: technology | Link here |
In the office this morning, handling the overnight backups, discovered that the lagoon windows on eureka:0.1 were gone. ruptime told me that lagoon had an up time of 5 hours. Yesterday it was 124 days.
What happened? No dump:
=== root@lagoon (/dev/pts/2) ~ 7 -> dumpon -l
/dev/null
I should fix that. But /var/log/messages contained the clue: hundreds of USB error messages for the photo backup disk. They didn't go away, and to add insult to injury the two devices had changed device numbers after boot. Power down and see what happens. After reboot, all was well.
It's been a while since I've had USB hell, but it doesn't seem to have gone away.
|
Which CPU, continued
|
Topic: technology, opinion | Link here |
It's been well over 7 months since I started thinking about building a new machine. And I'm still not much further.
One of my issues was power consumption, and that made Intel look interesting. But I was only going by TDP, and that's a very inaccurate measurement. Today I found this page, which paints a very different picture. The AMD Ryzen 7950X, with a TDP of 170 W, was pretty much the fastest I was looking at. But the review shows it using 72 W idle, 132 W with one thread fully loaded, and 332 W with all cores loaded. I can live with that: 98% of the time the machine will be effectively idle.
And the Core i9-13900F? Not in this comparison: it hadn't been released at the time. But the previous model Core i9-12900K uses similar power, only it's about ⅔ of the processing power.
And then? There are also chip sets, and it seems that the X670E has the best performance, though it's not clear how much advantage it has for me over the X670. So can I finally move on and put something together?
|
CJ's VoIP setup, yet again
|
Topic: technology, opinion | Link here |
Continuing mail with Aussie Broadband's complaint service. It's tough going. It's clear that the consultant isn't reading what I said. But today's message took the cake:
I have previously advised to you that the matter is with our L2 support team, and you should address them directly, as they have been working to address Mr. Ellis' concerns in the background.
Well, no. This is a complaint, and he's left holding the buck. But more to the point, he has advised me of no such thing, and how do I address L2 support? The people I've spoken to there all seem to be competent, and every time I have been in contact with one, we have had quick results. But my last complaint (in April) was about exactly the difficulty in reaching L2.
But before that, he wrote:
There is no ability to ship the devices out with both the VoIP services preconfigured. This is a limitation of our software for configuration before ship. We are not able to circumvent at this time but we would enjoy an ability to do so. We must add it manually.
WHAT? NOW he tells me? We've tried twice to get exactly that result, a pre-configured router. Why did they even bother if they knew they can't? The answer, I suspect, is because nobody knew, at least nobody with whom I had contact. My guess is that my complaints contact had finally contacted L2 support, and they had told him. And I think that underscores Aussie's real basic problem: lack of communication.
He also stated that they have configured the router and both services are active. Spoke to CJ, and with some confusion discovered yes, he has two phones, so he could plug one in. He also came over to bring me the old Cisco ATA, with which I want to play. But no power supply! With some difficulty, I discovered that I had forgotten to give him the power supply for the router, so he used the Cisco one, which has the same kind of plug. Problem: the Netcomm router wants a 12 V power supply, but the Cisco had only 5 V. How could that work?
He went back home and picked up the power supply. What voltage? Clearly the makers don't want you to know. This label is reflective (see top right) and 3.5 cm on a side. Even with a magnifying glass I couldn't read it. But yes, in characters about 1 mm high, we have:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Is it even from Cisco? Yes, it has the name well hidden in a couple of places:
|
|||||||||||||||||||||||
But that can't drive a 12 V device. Can it? Plugged it into the other router. Yes! It really works on 5 V. Somehow that blows away my whole concept of voltages. Decades ago when I got my first Germanium transistors (“red spot” and “green spot” if I recall correctly), the slightest overvoltage would destroy them. Now we have devices that run on less than half their rated voltage. I wonder what's going on inside.
But does the second line work? CJ called me on it, but I couldn't call back in. Will this never end?
|
HT802 noise
|
Topic: technology, opinion | Link here |
All the VoIP devices I have seen so far have logging capability, though what they say varies considerably. The current Grandstream HT802 seems particularly useless. I had hoped for something that would log incoming and outgoing calls, like I had 27 years ago. But if the information is there, it's completely flooded by error messages:
Aug 24 05:34:25 homephone HT802 [c0:74:ad:37:66:d8] [1.0.21.4]TR069 WARNING [1.0.1.141] UDP Connection Request Changed event been droped, tr111 dormant: 0
Aug 24 05:34:26 homephone HT802 [c0:74:ad:37:66:d8] [1.0.21.4]CallRecord::writeCDRFile, No space! current file size =51158bytes, need extra 80 bytes.
Aug 24 05:34:35 homephone HT802 [c0:74:ad:37:66:d8] [1.0.21.4]CallRecord::writeCDRFile, No space! current file size =51158bytes, need extra 78 bytes.
Today was unusual because of the USB error messages, but a couple of days ago the file had 7840 entries, of which 5353 were these nonsense messages from the ATA. So goodbye, logging.
Of course, these are error messages. But I don't believe that they're genuine or important, just another example of sloppy device programming, like the nonsense messages that all web browsers seem to generate, only to write them to /dev/null.
|
What use certificates?
|
Topic: technology, opinion | Link here |
Last week I ranted about the invalid security certificate in the Netcomm NF20MESH router. But Jamie Fraser went further: what kind of certificate would work for a device that has continually differing IP addresses?
Further investigation shows that the device is quite happy to answer on port 80 (http), which of course doesn't require a certificate. Of course, the correct response on https would be a redirection to the http port. At least with firefox, if I enter simply 192.168.20.2, I get https appended. If I enter http://192.168.20.2, I have no problems.
But ultimately xkcd has the obvious observation:
| Friday, 25 August 2023 | Dereel | Images for 25 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
CJ's VoIP: done!
|
Topic: technology, opinion | Link here |
After all the fun with CJ Ellis' VoIP setup, I have finally got round to looking at my own setup. I have four lines, two of them accidents. The other two are the one I normally use and the one that's in the phone book. The latter one used to divert automatically to voice mail, along with a suggestion to call the other number if you know it.
That worked well with MyNetFone, but Aussie doesn't inform you when a message diverts to voice mail. So why not divert to the answering service on the phone itself? One reason is that the phone doesn't have a configuration option to divert immediately, just after 3, 6 or 9 rings. But then there's this “Toll Saver” setting. What does that mean? Maybe what I'm looking for, divert immediately? Set it to find out.
And today I found out at least part of the answer. A call came in on that line, and it rang. But this time the number looked plausible, so I answered. Sam, from Aussie Broadband.
Why was he calling on this line? He was even more confused than I. He called on the wrong number, he couldn't repeat the pass code that I had given them, and he didn't have my mobile phone number, making him even more suspicious. But he still wanted to send me a 2FA code, by email because of the missing phone number. And that message didn't come. I asked him what address he sent it to: CJ's!
After the call was over, it eventuated that he had already spoken to CJ, who asked him to call me, and he went by the old information on CJ's records, not mine. He simply wanted to say that the router was finally configured, Call in on the second line? Tried it. Yes!
So, after only 37 days, they have finally solved the technical issues. Well, almost. What did they need to do to get the second line working? I'd really like to know.
And what do I do about the rest? Somehow I feel too close to the whole issue. I had really hoped that Aussie would be the good company, but what I'm seeing here is an incipient Telstra: excellent technical personnel running the network, and substandard customer relations. What should I do? I can just close the case, I can publish my experiences in more convenient form than here, or I can try to get the attention of Aussie management—if they care.
| Saturday, 26 August 2023 | Dereel | Images for 26 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
I read it in the news. It must be true.
|
Topic: politics, general, opinion | Link here |
Two articles from the Kyiv Independent today:
Lukashenko says Wagner Group will stay in Belarus. Belarusian dictator Alexander Lukashenko denied reports of Wagner mercenaries leaving Belarus following the reported death of the group’s founder, Yevgeny Prigozhin, according to the state-controlled news agency Belta.
Border Guards: Number of Wagner fighters in Belarus decreases. The number of Wagner Group mercenaries stationed in Belarus is gradually decreasing as they depart for Russia, State Border Guard Service's spokesperson Andrii Demchenko said on Aug. 25.
Arguably the difference is the reporter, but somehow it highlights the confusion that's going on at the moment.
|
Steak and kidney pie in hair dryer
|
Topic: food and drink, opinion | Link here |
Steak and kidney pie for dinner today. For some reason I have always cooked them in the oven in the past, but what speaks against the “air fryer”, specifically the “hair dryer”? Tried that today. In the oven I had cooked for 20 to 30 minutes at 210°. Here's how things looked at the start and after 7 minutes in the “hair dryer”:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
By comparison, in the oven they looked like this after 20 minutes:
|
|||||||||||||||||||||||
I turned the temperature down to 160° and cooked for another 6 minutes, but it didn't seem to make much difference:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And mine was far too salty. How did that happen? Still, it looks like the “hair dryer” is the way to go.
| Sunday, 27 August 2023 | Dereel | Images for 27 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Magnolia blooms
|
Topic: gardening | Link here |
Winter is coming to an end, and our new Magnolia × soulangeana is starting to flower:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The flowers aren't very big. I wonder if that's the way they're meant to be, or that they'll grow, or that it's maybe been because it has only been in the ground for five months.
|
Alternative video sources
|
Topic: multimedia, technology, opinion | Link here |
ORF, the Austrian TV broadcaster, has just repeated all episodes of Stockinger, a series that I had been looking for. Today I finished downloading the remaining episodes, which they had provided out of order. Only one episode was missing: the first!
Did I miss it? No, it seems. It was broadcast on ORF 3 on 19 August, but I couldn't find any reference to it in my lists. More searching suggested that I could find it on Serienjunkies. Yes! They're really trying to sell their services, of course, but they offer a free download service, and to make the paid version more attractive, they make the free version really difficult to use. Dozens of CAPTCHAs, slow download speed, in this case 2 files that can't be downloaded without a pause of several hours (or, of course, a change of web proxy), and in RAR format! Tried it anyway, and discovered that the thing is in low (640x480) resolution. Still, for some things it could be useful.
| Monday, 28 August 2023 | Dereel | Images for 28 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More ports hell
|
Topic: technology, opinion | Link here |
Nearly 30 years ago Warren Toomey published scans of John Lions' A Commentary on the UNIX Operating System on the Usenet alt.folklore.computers newsgroup. I formatted it and subsequently published it. It contains a number of scan errors.
Recently I received updated versions of the sources from Conway Yee, and I've finally got round to incorporating them. But:
=== grog@eureka (/dev/pts/21) /home/Book/Lions 116 -> make
latex lionc.tex
make: exec(latex) failed (No such file or directory)
*** Error code 1
Oh. Has it been that long? OK, let's install. On dereel, since eureka is so down-rev.
=== root@dereel (/dev/pts/4) ~ 24 -> pkg search latex
pkg: Warning: Major OS version upgrade detected. Running "pkg bootstrap -f" recommended
pkg: Repository FreeBSD missing. 'pkg update' required
Ah, right, I have just upgraded dereel to FreeBSD 15-CURRENT. OK,
=== root@dereel (/dev/pts/4) ~ 25 -> pkg bootstrap -f
pkg: Warning: Major OS version upgrade detected. Running "pkg bootstrap -f" recommended
pkg(8) is already installed. Forcing reinstallation through pkg(7).
The package management tool is not yet installed on your system.
Do you want to fetch and install it now? [y/N]: y
Bootstrapping pkg from pkg+http://pkg.FreeBSD.org/FreeBSD:15:amd64/latest, please wait...
pkg: Error fetching http://pkg.FreeBSD.org/FreeBSD:15:amd64/latest/Latest/pkg.txz: Not Found
A pre-built version of pkg could not be found for your system.
Consider changing PACKAGESITE or installing it from ports: 'ports-mgmt/pkg'.
A bit more checking showed that yes, there's a directory http://pkg.FreeBSD.org/FreeBSD:15:amd64/, but it's empty. 15-CURRENT is brand new, and it seems that we don't have any packages for it yet.
OK, we have other computers. How about tiwi? But first, what's the package? I got a total of 38 hits, of which these seemed the most likely:
=== root@tiwi (/dev/pts/16) ~ 4 -> pkg search latex
latex-beamer-3.57_1 LaTeX class to create a beamer presentation
latex-biber-2.19,1 BibTeX replacement for users of biblatex
latex-jlisting-0.2 Support for Japanese text within lstlisting environment
latex-ltablex-1.0_2 Combines the features of both the tabularx and longtable packages
latex-mk-2.1_5 Collection of makefile and scripts for LaTeX documents
latex-resume-20010823_3 Resume class for LaTeX
latex-service-0.1_11 LaTeX service for GNUstep
latex2html-2016_1 Convert LaTeX documents to HTML
latex2rtf-2.3.17_1 Translator from LaTeX to RTF
latexdiff-1.3.3 Determine and mark up significant differences between latex files
But none of those look right. Under TeX, maybe? That way madness lies:
=== root@tiwi (/dev/pts/16) ~ 5 -> pkg search tex | wc -l
466
466 packages! After some searching, whittled it down to “only” 21:
tex-basic-engines-20230313 Basic TeX Engines
tex-dvipdfmx-20230313 DVI Previewer(kpathsearch) for X
tex-dvipsk-2023.1 Convert a TeX DVI file to PostScript
tex-formats-20230313 Formats for basic TeX engines and the 'latex' command
tex-jadetex-3.13_5 TeX backend for Jade, DSSSL processor for SGML/XML documents
tex-kpathsea-6.3.5 Path searching library for TeX
tex-libtexlua-5.3.6_1 Lua library for TeXLive
tex-libtexluajit-2.1.0_2 LuaJIT library for TeXLive
tex-luatex-1.16.0 Extended version of pdfTeX using Lua
tex-ptexenc-1.4.3 Library for Japanese pTeX and its tools
tex-synctex-2.0.0_2 Synchronization TeXnology parser library
tex-web2c-20230313 TeX implementation translating WEB to C
tex-xdvik-22.87.06_2 DVI Previewer(kpathsearch) for X
tex-xetex-0.99993_4 TeX System with Unicode and Modern Font Technologies
tex-xmltex-1.9_4 Non-validating XML parser, written in TeX
texlive-base-20230313_3 TeX Live Typesetting System, base binaries
texlive-docs-20230313 TeX Live Typesetting System, documentation
texlive-full-20230313 TeX Live, Full Version
texlive-texmf-20230313 TeX Live Typesetting System, texmf Tree
texlive-texmf-source-20230313 TeX Live Typesetting System, texmf Tree (source)
texlive-tlmgr-20230313 TeXLive manager modules
Why, oh why, can't there be a port just called tex? After a quick web search, found this article, which recommended:
You should be able to install the FreeBSD package pkg install print/texlive-texmf. Installing the UNIX version of TeX Live from TUG should also work.
Now isn't that an immediately obvious choice? Armed with that information, tried a more promising name:
=== root@tiwi (/dev/pts/9) ~ 38 -> pkg install texlive-full
The following 52 package(s) will be affected (of 0 checked):
Installed packages to be REMOVED:
firefox: 103.0.2,2
New packages to be INSTALLED:
adobe-cmaps: 20051217_4
...
Number of packages to be removed: 1
Number of packages to be installed: 40
Number of packages to be upgraded: 11
The process will require 6 GiB more space.
4 GiB to be downloaded.
Proceed with this action? [y/N]:
Remove firefox? On my TV? No way. I thought that those days were over. Sure, I can probably recover if things go wrong, but tiwi is sacred. Back to dereel:
=== root@dereel (/dev/pts/4) /usr/ports 8 -> mailme git pull --ff-only
From https://git.freebsd.org/ports
0b552432e344..0d7e6419057d main -> freebsd/main
ed89d160df93..e0ae7a6c4af6 2023Q3 -> freebsd/2023Q3
hint: Diverging branches can't be fast-forwarded, you need to either:
hint:
hint: git merge --no-ff
hint:
hint: or:
hint:
hint: git rebase
hint:
hint: Disable this message with "git config advice.diverging false"
fatal: Not possible to fast-forward, aborting.
Sigh. Can nothing go right? Why did I end up with a conflict in a hierarchy that I haven't changed? Yes, the merge worked, but by that time I had decided that the last thing I wanted was to fight my way through a maze of twisty little ports, all irritating.
So I have two choices:
“Wait” sounds like a good idea.
|
Microphone battery life
|
Topic: photography, technology, opinion | Link here |
Yvonne doesn't use her wireless microphones much, but she has been doing so more recently. In the past we had had issues with battery life with rechargeable batteries, so we used non-rechargeable batteries. But they don't last for ever, and the devices have particularly inaccurate battery state displays. Turn one on, note that it shows a fully charged battery—for 10 seconds or so. Then the display changes to “discharged”, and unlike most devices I know, it immediately stops working.
How long do rechargeable batteries last? They really only need to last the maximum of 2 hours or so that the PIXIO “Robot Cameraman” lasts. Today I chose 3 pairs of 2 rechargeable batteries: the Sony batteries I got with my Nikon “Coolpix” L1 on 19 March 2006 (1600 mAh), two ALDI batteries about 4 years old (2100 mAh) and 2 Eveready batteries of about the same age (1300 mAh). For some reason the Eveready batteries took far longer to charge than the others. Fully charged, put them in the devices, the Evereadys in the receiver, which I think uses less power.
The results? The Sonys lasted 3⅓ hours, the Evereadys 4½ hours, and the ALDIs 6 hours. That's more than adequate, though it might be an idea to get some newer batteries.
|
Bruno
|
Topic: animals | Link here |
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Tuesday, 29 August 2023 | Dereel → Ballarat → Dereel | Images for 29 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
New web site scam?
|
Topic: technology, opinion | Link here |
The 404 document on http://www.lemis.com/ checks the referrer of a page and sends me an email if it's also from my web site. That helps find broken local links, along with lots of URLs with embedded JavaScript in the URLs. But lately I've been getting a surprising number of faked URLs of a completely different nature:
984 N + 26-08-2023 To grog@lemis.c ( 6) World Wide Web Owner N + FAILURE: /collections/downloads/inga-juuso+h-moll+mendelssohn <- http://www.lemis.com/grog/Albums/Animals/Bruno.php?dirda
990 N + 26-08-2023 To grog@lemis.c ( 6) World Wide Web Owner N + FAILURE: /de/service/downloads <- http://www.lemis.com/grog/diary-apr2023.php?article=D-20230428-004023&subtitle=Aussie+i
991 N + 26-08-2023 To grog@lemis.c ( 6) World Wide Web Owner N + FAILURE: /collections/downloads/inga-juuso+merging-clock-recording+matangi-quartet <- http://www.lemis.com/grog/diary-apr
992 N + 26-08-2023 To grog@lemis.c ( 6) World Wide Web Owner N + FAILURE: /collections/downloads/inga-juuso+wim-kegel+matthew <- http://www.lemis.com/grog/diary-apr2023.php?dirdate=20230
995 N + 26-08-2023 To grog@lemis.c ( 6) World Wide Web Owner N + FAILURE: /collections/downloads/inga-juuso+franz-liszt+matthew-passion <- http://www.lemis.com/grog/Albums/Animals/Bruno.
996 N + 26-08-2023 To grog@lemis.c ( 6) World Wide Web Owner N + FAILURE: /collections/downloads/jan-van-der-crabben+floris-mijnders+martin-fondse <- http://www.lemis.com/grog/Albums/Ani
998 N + 26-08-2023 To grog@lemis.c ( 6) World Wide Web Owner N + FAILURE: /collections/downloads/jan-van-der-crabben+staes+matthew-passion <- http://www.lemis.com/grog/Albums/Animals/Bru
1074 N + 26-08-2023 To grog@lemis.c ( 6) World Wide Web Owner N + FAILURE: /collections/downloads/jan-van-der-crabben+symphony-no-6+masseurs <- http://www.lemis.com/grog/Albums/Animals/Br
None of those URLs exist, and they're not even close. But many of the “referring” pages are /grog/Albums/Animals/Bruno.php, and the references are frequently to something musical. Are we looking at some application of the current “Artificial Intelligence” technology?
|
More git pain
|
Topic: technology, opinion | Link here |
So yesterday I had problems with git which I fixed with a merge --no-ff. And today?
=== root@dereel (/dev/pts/4) /usr/ports 29 -> mailme git pull --ff-only
From https://git.freebsd.org/ports
648c8874180a..c196d1f5ffb8 main -> freebsd/main
hint: Diverging branches can't be fast-forwarded, you need to either:
hint:
hint: git merge --no-ff
hint:
hint: or:
hint:
hint: git rebase
hint:
hint: Disable this message with "git config advice.diverging false"
fatal: Not possible to fast-forward, aborting.
That didn't last long, did it? How about the other alternative?
=== root@dereel (/dev/pts/4) /usr/ports 30 -> mailme git rebase
Successfully rebased and updated refs/heads/main.
You have mail in /var/mail/grog=== root@dereel (/dev/pts/4) /usr/ports 31 -> mailme git pull --ff-only
Already up to date.
That at least had the advantage that it was much faster. But what will it look like tomorrow?
|
Doppler ultrasound
|
Topic: health | Link here |
Into town in the afternoon for a Doppler ultrasound investigation of my right leg. Long wait, long procedure. The whole thing took 40 minutes. The ultrasound device was much more modern than anything I have seen before, with a second display on the wall showing flow rates and directions. Quite impressive.
And the result? Nothing obvious. Looks like we'll have to look elsewhere when I see Paul Smith next week.
|
More PoS terminal fun
|
Topic: general, technology, food and drink, opinion | Link here |
After the ultrasound treatment I had an appointment for an eye test. But I had 40 minutes to wait, so off to Big W to look for plants.
That was simple: they don't do that any more. But while looking around I found a kitchen mixer, something in which I failed in January. But here I found one for $29, as opposed to the $60 to $500 range in January. So I bought that, with some difficulty. Self checkout:
|
|||||||||||||||||||||||
Do I need a bag? No, of course not. I can carry the box as it is. WRONG answer. At the end I couldn't pay. I also could not longer lie and say that I had a bag. I had to select “Assistance”. One was standing almost directly behind me, but it seems that the device is too polite to inform her. Instead I just got a login screen:
|
|||||||||||||||||||||||
I didn't watch what she did, but it only took a couple of seconds. Then I was moved to the other (smaller) terminal on the right:
|
|||||||||||||||||||||||
“You can't buy here unless you have a mobile phone”? I couldn't get rid of it: the selection window had moved to the other display again. What a mess!
|
The demise of Bridge Mall
|
Topic: general, opinion | Link here |
After that down to Bridge Mall, about which I have ranted a few weeks ago. They're really tearing it up:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But never mind. Now at least half of the shops are shut:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I wonder what it will look like when they have finished their “improvements”.
|
Eye test
|
Topic: health, opinion | Link here |
Then to Specsavers for an eye test, mainly encouraged by an email I received recently offering me two pairs of glasses, one coated, for free (well, my health insurance pays 100%). First the preliminaries, including once again the auditory test that I ranted about last year. The results looked similar. Here last year and this year:
|
||||||||||||||||||||||||
I didn't find that until I got home, of course. In the meantime I had made an “appointment” (for a phone call) with an audiologist in a couple of weeks' time. Maybe he can explain to me how they get these results,
Things were slow, but I had a particularly good optometrist who established that my eyes had hardly changed since the last time, but also that I have beginning cataracts in both eyes. Got a referral to an opthalmologist, where I'll presumably find more details.
Then to order the glasses. They knew nothing of the offer! I promised to forward the email to them, but they preferred a printout. Somehow we're not quite in the computer age yet. In any case, as a result I can defer the whole thing until next week.
| Wednesday, 30 August 2023 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Overnight pain again?
|
Topic: health, opinion | Link here |
Restless night. More pain in my left leg of the kind I had a month or two ago. But why? Shouldn't it be gone by now? It can't have anything to do with yesterday's Doppler ultrasound. That was on the other leg, and there was nothing obvious to suggest a reversion. But I had hoped that all traces of the slipped disk would be gone by now.
|
git pull, day 3
|
Topic: technology | Link here |
So, yesterday and the day before I failed in a real update to my ports tree with git pull. git merge --no-ff worked for a day, then I was back to the original problem. git rebase worked yesterday. And today?
=== root@dereel (/dev/pts/4) /usr/ports 35 -> mailme git pull --ff-only
From https://git.freebsd.org/ports
54b3112625f8..ec04036bb656 main -> freebsd/main
hint: Diverging branches can't be fast-forwarded, you need to either:
hint:
hint: git merge --no-ff
...
It still doesn't work! OK, give up. Does merge still work?
=== root@dereel (/dev/pts/4) /usr/ports 36 -> git merge --no-ff
hint: Waiting for your editor to close the file... /usr/local/bin/emacsclient: can't find socket; have you started the server?
/usr/local/bin/emacsclient: To start the server in Emacs, type "M-x server-start".
/usr/local/bin/emacsclient: No socket or alternate editor. Please use:
...
Huh? What's that? I didn't start an edit session, and that last message confirms that there's no editor. Try again?
=== root@dereel (/dev/pts/4) /usr/ports 38 -> git merge --no-ff
fatal: You have not concluded your merge (MERGE_HEAD exists).
Please, commit your changes before you merge.=== root@dereel (/dev/pts/4) /usr/ports 39 -> mailme git rebase
error: cannot rebase: Your index contains uncommitted changes.
error: Please commit or stash them.
But there were no changes! I'm reminded of this xkcd cartoon:
OK, move the tree away and start again from scratch, remembering that this one silly command wants to be run in the parent directory, and it wants to create the implicit directory itself:
=== root@dereel (/dev/pts/4) /src/FreeBSD/git 50 -> Log git clone -o freebsd --config remote.freebsd.fetch=+refs/notes/*:refs/notes/* https://git.freebsd.org/ports.git
===== Wed 30 Aug 2023 15:11:46 AEST on dereel: git clone -o freebsd --config remote.freebsd.fetch=+refs/notes/*:refs/notes/* https://git.freebsd.org/ports.git
Cloning into 'ports'...
Updating files: 100% (156765/156765), done.
...
git finished at Wed 30 Aug 2023 15:41:06 AEST
And that seemed to work, sort of. Only somewhere in the back of my head I have this recollection that I now need to kick it hard to get it to work only marginally sanely. Some config file?
How I hate git!
| Thursday, 31 August 2023 | Dereel | Images for 31 August 2023 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Modern addresses
|
Topic: general, opinion | Link here |
Ordered some wine online today. But it didn't like my address:
|
|
|||||||||
Now admittedly writing the name of the town in CAPS isn't a good idea (and wasn't mine; a previous modernity). But collapsing “Road” to “Rd”?
|
More security strangenesses
|
Topic: technology, opinion | Link here |
I have two Microsoft machines, dischord.lemis.com (“Windows” 7) and distress.lemis.com (“Windows” 10), which I use for running photographic software. I only access them via rdesktop Today I decided to check something from tiwi.lemis.com, my TV computer. Nothing happened. OK, try from a shell:
=== grog@tiwi (/dev/pts/5) ~ 247 -> /home/local/bin/dordesktop distress 1870x1030+0
wake: Cannot open bpf interface: Permission denied
User grog sword scabbard!
Connecting to distress
ATTENTION! The server uses and invalid security certificate which can not be trusted for
the following identified reasons(s);
1. Certificate issuer is not trusted by this system.
Issuer: CN=distress
Review the following certificate info before you trust it to be added as an exception.
If you do not trust the certificate the connection atempt will be aborted:
Subject: CN=distress
Issuer: CN=distress
Valid From: Thu Jun 1 15:24:39 2023
To: Fri Dec 1 16:24:39 2023
Certificate fingerprints:
sha1: 1a1a935412aa47988d3d5fa8e0c4a9032812e016
sha256: 21fb8eb0192bc513ba589215ccb46d085d2fadb3e5d1e7d865176e6db4fe1299
Do you trust this certificate (yes/no)? yes
Core(error): tcp_tls_connect(), TLS handshake failed. GnuTLS error: The operation timed out
Failed to connect, SSL required by server.
Disconnected from distress, status 76
Connecting to distress
Core(warning): Certificate received from server is NOT trusted by this system, an exception has been added by the user to trust this specific certificate.
Connection established using SSL.
Clipboard(error): xclip_handle_SelectionNotify(), unable to find a textual target to satisfy RDP clipboard text request
Protocol(warning): process_pdu_logon(), Unhandled login infotype 1
OK, the wake invocation is clear. /usr/sbin/wake needs to be setuid, and I fixed that. But what's the rest of this stuff? It suggests that I scratched my head too long, but after that it worked. I really wish I understood this stuff better.
|
Build me a new machine
|
Topic: technology, opinion | Link here |
More investigation of the configuration for my new machine. The most important discovery was that the Computeralliance configurator is not very clever. It will happily offer me an AMD Ryzen CPU, an Intel motherboard and an el-cheapo processor fan. Clearly I can't rely on that.
But why should I? I know what I want. Let the people who build it suggest the exact components. What I have is:
AMD Ryzen 9 7950X processor.
Appropriate fan.
192 GB RAM.
X670-based motherboard with at least 4 PCIe slots.
Normal case that doesn't look too silly.
High efficiency power supply (Antec EarthWatts or similar).
2 TB M.2 SSD.
DVD drive.
Maybe: Two NVIDIA display cards with HDMI, DVI and D-Sub outputs. I don't want to change over all my monitors, but the cards I have found are not much more modern than what I can get out of eureka.
And that's really all. That shouldn't be too difficult.
|
Bloody CapsLock! Reboot!
|
Topic: technology, opinion | Link here |
Later in the afternoon I was reading mail on tiwi.lemis.com, and it suddenly refused to obey my instructions. In fact, the whole X display behaved strangely. Finally found a shell window and discovered that CapsLock was set.
On most computers that's not a problem. Press the CapsLock key again and it goes away. Only: I don't have a CapsLock key! I have xmodmapped it out of existence (the key with that inscription is Ctrl, like God intended). How do I get rid of it? Did some thinking, but couldn't come up with anything. Switch to ttyv0 and back again? All that happened was that the ttyv0 display was corrupt, like far too frequently on modern machines.
OK, stop X. How I hate doing that! It had been running for months, probably the 106 days since I last booted tiwi. But nothing for it. Then back to (invisible) ttyv0 and run startx again. Nothing obvious happened. To eureka and try to start it from there. Various non-obvious messages, but when I finally got a display, I couldn't create any windows on it. It looked as if I had started server 1 when I wanted server 0, but there was no indication that server 0 was still running.
Dammit, what can I do? Normally I would have looked further, but this was cutting into our TV watching time. Reboot! After 106 days, 3½ months! How I hate that! And all for a bug that somehow set my CapsLock status!
Somehow there are too many buglets that have been around for ever:
I won't get them fixed in a hurry, but what happens if this bites me again? Clearly I need a CapsLock key for emergencies. Somewhere up the top right of the numeric keypad, maybe?
Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.
| Top of page | Previous month | Greg's home page | Today's diary entry | Next month | Greg's photos | Copyright information |
