|
| Greg's diary |
| March 2020 |
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Sunday, 1 March 2020 | Dereel → Ballarat → Dereel | Images for 1 March 2020 |
| Top of page | ||
| next day | ||
| last day |
|
Leonid in pain
|
Topic: animals, opinion | Link here |
Woken round 5:50 this morning by what proved to be Leonid howling in pain. Why? We investigated, but found nothing obvious, and he stopped.
Woke again round 8:00 and found him decidedly unhappy with himself. After some consideration thought that he might have a bone stuck in his throat. Called up the vet at 8:37—it would have to be a Sunday!—and got an answering machine with instructions to leave a message and I'd get a response within 15 minutes, and if I didn't, call back in half an hour.
Called again at 9:07 and got Louisa, who had registered the call, but their voice mail service had stated that there was no message, and of course it was too polite to tell her my phone number. In any case, she could see Leo immediately, and by 9:40 we were there in Ballarat. By this time it was clear that there was serious swelling in his mouth:
|
|
||||||||||
That red area under his tongue was of concern. Broken tooth? Something sticking in the gum? They'd have to anaesthetize him and investigate. Leave him here, should be ready in the afternoon.
That happened as predicted, including the hefty bill of not quite $800, half of which was callout fees. Leo, can't you get sick during weekdays? In any case, all went well, nothing sticking in the gum, though clearly there had been something. no teeth broken (as my photos showed). Louisa promised to send me a photo of the hole to my phone, but that, too, didn't work. Maybe that was my fault for not understanding the intuitive nature of Android? Finally she sent it to Yvonne via Facebook!
|
|||||||||||||||||||||||
In any case, he recovered well from the anaesthetic, and by evening was almost back to his normal self.
|
Carry a camera!
|
Topic: photography, technology, opinion | Link here |
In the 1960s some organization in the UK brought out the slogan “Carry a camera”, and I certainly did what I could to adhere to it. Things haven't changed that much, and when we went to the vets I brought two cameras—and left them in the car! So when I took the photos of Leonid, I had to use my phone. The results are correspondingly terrible. I could at least have brought my bag with the E-PM2.
|
More teevee upgrade pain
|
Topic: technology, opinion | Link here |
Finally the build for teevee is finished. Reboot. make installworld. Oh:
===> stand/i386/loader_lua (install)
installing DIRS BINDIR
install -d -m 0755 -o root -g wheel /boot
cc -target x86_64-unknown-freebsd12.1 --sysroot=/spool/obj/eureka/home/src/FreeBSD/svn/stable/12/amd64.amd64/tmp -B/spool/obj/eureka/home/src/FreeBSD/svn/stable/12/amd64.amd64/tmp/usr/bin -O2 -pipe -I/eureka/home/src/FreeBSD/svn/stable/12/stand/i386/btx/lib -nostdinc -I/spool/obj/eureka/home/src/FreeBSD/svn/stable/12/amd64.amd64/stand/libsa32 -I/eureka/home/src/FreeBSD/svn/stable/12/stand/libsa -D_STANDALONE -I/eureka/home/src/FreeBSD/svn/stable/12/sys -Ddouble=jagged-little-pill -Dfloat=floaty-mcfloatface -ffunction-sections -fdata-sections -DLOADER_GELI_SUPPORT -I/eureka/home/src/FreeBSD/svn/stable/12/stand/libsa/geli -DLOADER_DISK_SUPPORT -m32 -ffreestanding -mno-mmx -mno-sse -mno-avx -mno-avx2 -msoft-float -march=i386 -I. -Iinclude -I/eureka/home/src/FreeBSD/svn/stable/12/stand/common -I/eureka/home/src/FreeBSD/svn/stable/12/contrib/lua/src -I/eureka/home/src/FreeBSD/svn/stable/12/stand/common -I/eureka/home/src/FreeBSD/svn/stable/12/stand/liblua -DLUA_FLOAT_TYPE=LUA_FLOAT_INT64 -DLOADER_CD9660_SUPPORT -DLOADER_EXT2FS_SUPPORT -DLOADER_MSDOS_SUPPORT -DLOADER_UFS_SUPPORT -DLOADER_GZIP_SUPPORT -DLOADER_BZIP2_SUPPORT -DLOADER_NET_SUPPORT -DLOADER_NFS_SUPPORT -DLOADER_TFTP_SUPPORT -DLOADER_GPT_SUPPORT -DLOADER_MBR_SUPPORT -DLOADER_ZFS_SUPPORT -I/eureka/home/src/FreeBSD/svn/stable/12/stand/libsa/zfs -I/eureka/home/src/FreeBSD/svn/stable/12/sys/cddl/boot/zfs -Wall -I/eureka/home/src/FreeBSD/svn/stable/12/stand/i386 -DLOADER_PREFER_AMD64 -std=gnu99 -Wsystem-headers -Werror -Wno-pointer-sign -Wno-empty-body -Wno-string-plus-int -Wno-unused-const-variable -Wno-tautological-compare -Wno-unused-value -Wno-parentheses-equality -Wno-unused-function -Wno-enum-conversion -Wno-unused-local-typedef -Wno-address-of-packed-member -Wno-switch -Wno-switch-enum -Wno-knr-promoted-parameter -Wno-parentheses -Oz -Qunused-arguments -DLUA_PATH=\"/boot/lua\" -I/eureka/home/src/FreeBSD/svn/stable/12/libexec/flua/modules ERROR-tried-to-rebuild-during-make-install -c /eureka/home/src/FreeBSD/svn/stable/12/stand/common/interp_lua.c -o interp_lua.o
cc: error: no such file or directory: 'ERROR-tried-to-rebuild-during-make-install'
*** Error code 1
What's that? I've never seen that before. Could it be that the source tree got updated between build and make installworld? In any case, the difference in the binaries is probably not that bad, so I left it for the time being.
To chase yesterday's mpv errors, rebooted the old system. And they're still there! Damn! Did I just miss them before? Rebooted to the new system again, but after a while it still hadn't responded. Into the lounge room to see what was going on:
|
|||||||||||||||||||||||
Huh? It was repeatable, of course. Is this something to do with the failed installworld? The function in question was the messenger rather than the culprit:
local function defaultframe()
if core.isSerialConsole() then
return "ascii"
end
return "double"
end
OK, what happens if I just type boot into the prompt? It boots! But I wasn't out of the woods yet. X didn't start, and while typing things in, I got many repeated characters. From my comments for this diary written at the time:
Can't boot: lua craps out.
boot works
Multiple charaaaacter echoooos.
Can't start X
Can't su.
And the “can't su”? I discovered that I could log in as root on a different vty, so I didn't investigate it until later.
OK, that leaves X. Xorg.0.log didn't help much:
[ 32.844] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[ 32.844] (EE) NVIDIA: system's kernel log for additional error messages and
[ 32.844] (EE) NVIDIA: consult the NVIDIA README for details.
And of course there was nothing in the log that helped. For some reason I tried another boot, during which I saw:
|
|||||||||||||||||||||||
Huh? Who installed the 440 version of the Nvidia driver? It must have been a misguided pkg upgrade. But why did it only bite me now? Simple, I rebooted. While the system was running, I was using the old module. OK, remove the driver and reload the old one (390). Start X. Hang.
Damn, what's wrong there? According to the vomit on the vty from which I started it, some Java issue. Shot it down and tried again.
[ 619.925] (WW) NVIDIA(0): The NVIDIA GeForce GT 730 GPU installed in this system is
[ 619.925] (WW) NVIDIA(0): supported through the NVIDIA 390.xx Legacy drivers. Please
[ 619.925] (WW) NVIDIA(0): visit http://www.nvidia.com/object/unix.html for more
[ 619.925] (WW) NVIDIA(0): information. The 440.59 NVIDIA driver will ignore this
[ 619.925] (WW) NVIDIA(0): GPU. Continuing probe...
That wasn't what I had in the version that hung. I had already tried it out on the console:
|
|||||||||||||||||||||||
And a cursory inspection of the log file for the hung X session confirmed that the correct driver had been detected. And I hadn't reloaded everything. Where did it get this idée fixe of 440? Tried several times without success, and finally gave up and rebooted the old system. There are other things to do.
In the evening, discovered something else: now there were no error messages running mpv. I can't be completely sure, but it's probable that all my attempts in the afternoon were over the network, so that might have a bearing on the problem, possibly even for the new system.
|
eBay notifications
|
Topic: technology, opinion | Link here |
While in the lounge room, logged in to eBay, apparently for the first time in this instance of teevee. That required me to find the email address and the (unflattering) password.
And then I got this message:
|
|
|||||||||
What nonsense is that? Apart from the fact that it doesn't tell me which email address it's talking about (I had to check the mail headers to confirm), I have had that particular address since 6 July 2003. I've been with eBay for over 20 years; when will they finally get their act together?
In passing, the headers of my very first registration with eBay, with a different (here falsified) email address:
From reg@ebay.com Sat Jul 3 10:20:38 1999
Received: from moccasin.ebay.com (hognose.ebay.com [216.32.120.65])
by allegro.lemis.com (8.9.1/8.9.0) with SMTP id KAA14677
for <groggyhimself@lemis.com>; Sat, 3 Jul 1999 10:20:35 +0930 (CST)
Date: Sat, 3 Jul 1999 10:20:35 +0930 (CST)
From: reg@ebay.com
Message-Id: <199907030050.KAA14677@allegro.lemis.com>
Received: (qmail 12520 invoked from network); 2 Jul 1999 16:51:24 -0000
Received: from bloodsucker.ebay.com (216.32.120.73)
by hognose.ebay.com with SMTP; 2 Jul 1999 16:51:24 -0000
Welcome to eBay! We're so happy to have you as our newest member.
Somehow host names like that belong to a bygone millennium, more's the pity.
|
Lasagnes au saumon
|
Topic: food and drink, opinion | Link here |
Today was Yvonne's birthday, so I offered to make salmon lasagne for her, for once with fresh pasta instead of this horrible tasting instant lasagne that is the only kind we can find in Ballarat. It's been years since we last did that, and the first experience (with a brand-new pasta machine) weren't encouraging. The dough stuck to the rollers, and I had great difficulty removing it:
|
|||||||||||||||||||||||
I had never had that before. Problems with the machine? No, it proved to be a too-moist dough. This time I used harder pizza flour for the first time, and it seems to require less moisture. Add flour and all was well.
| Monday, 2 March 2020 | Dereel | Images for 2 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More grid power issues
|
Topic: Stones Road house, general | Link here |
Another brief grid power failure this morning at 3:10:09. And another at 20:40:16. Again, not long enough for an “Off-grid” status from the inverter, but the second time round it upset the office UPS. So whatever the duration of the outage, it was enough to cause concern beyond the inverter.
|
Upcoming events
|
Topic: technology, general, opinion | Link here |
Received (barely) yesterday evening:
From bounces+wikitree=lemis.com@bounce.geni.com Sun Mar 1 23:59:44 2020
Date: Sun, 1 Mar 2020 04:59:36 -0800
From: Geni <no-reply@geni.com>
Latest Lehey Family updates February 23 - 29
Sunday, February 23
* Yvonne Lehey turns 71 on March 1.
Nothing like advance warning, is there? No, this is nothing like advance warning.
|
Daily teevee pain
|
Topic: technology, opinion | Link here |
At least one of yesterday's problems with the teevee upgrade, the lua errors, could have been that the source tree was updated between building world and installing world, so built another world today. To boot the new system cleanly, copied the old /boot/lua directory to the new system, and it booted cleanly. But I still had these repeated characters from the keyboard, to the point that I couldn't log in as rooott at all. I had to go to eureka and log in via the net.
While I was there, ran make installworld. This time it ran cleanly. Reboot. Those bloody lua messages again! So it seems that they weren't related to the errors I got last time. But what are they? Clearly not everybody's getting them. Could it be an incompatibility with the existing boot blocks? Should I update them just for the fun of it? What if I can't boot at all afterwards?
OK, try the other issues. I still couldn't start X. I had moved the Nvidia 440 drivers out of the way, but I still got the message:
[ 1042.999] (II) NVIDIA dlloader X Driver 440.59 Thu Jan 30 00:58:52 UTC 2020
Where the hell did that come from? Any further information in the log file? Yes, of course:
[ 1042.998] (II) LoadModule: "nvidia"
[ 1042.998] (II) Loading /usr/local/lib/xorg/modules/drivers/nvidia_drv.so
[ 1042.999] (II) Module nvidia: vendor="NVIDIA Corporation"
[ 1042.999] compiled for 1.6.99.901, module version = 1.0.0
[ 1042.999] Module class: X.Org Video Driver
[ 1042.999] (II) NVIDIA dlloader X Driver 440.59 Thu Jan 30 00:58:52 UTC 2020
That appears to have been installed with X! OK, copy the old version of the driver from the old partition, and how about that, X started!
So: somehow the new X installation has a driver that isn't compatible with the display hardware. That would at least explain some of the
That wasn't the only issue. Mouse remapping (buttons 1 and 3 to simulate 2) didn't work, and neither did the keyboard remapping. Both of those need to be looked at more carefully. When I tried to run xmodmap again to remap the keyboard, all I found was that this failed:
keysym Caps_Lock = Control_L
More gratuitous changes that break my scripts? I'll look at it later. In the meantime noted that the brightly coloured error messages were there again. Maybe they're right: my hardware is too old for the 2020s. I wouldn't even mind replacing it if I could find graphics cards that weren't so big and noisy, but for the while I'm stuck with them.
In summary, though, I have the following issues:
So: I've been using FreeBSD for over 25 years now, and upgrading is still an unsolved problem. If anything, things have got worse: where's the make world of the Good Old Days?
|
Shotcut: book with seven seals
|
Topic: multimedia, technology, opinion | Link here |
Yvonne spent some time this afternoon trying to use shotcut to trim a video clip, with spectacular lack of success. Apart from the fact that the use of the interface is completely at variance with anything I have ever seen, she wasn't even able to save the result. When she tried, all she got was a small XML file.
OK, see what I can do. FOOL! You don't save files, you export them. OK, tried that, after finding out how to trim files (simple: take the start and end markers and move them to the point you want—if the markers are displayed, which isn't all the time). And how about that, it saved a file. Did it have the correct format?
=== grog@teevee (/dev/pts/5) ~ 7 -> mplayer /eureka/home/var/tmp/shotcut.mp4
Playing /eureka/home/var/tmp/shotcut.mp4.
libavformat version 58.29.100 (external)
libavformat file format detected.
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x801b6b620]moov atom not found
LAVF_header: av_open_input_stream() failed
libavformat file format detected.
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x801b6b620]moov atom not found
LAVF_header: av_open_input_stream() failed
Exiting... (End of file)=== grog@teevee (/dev/pts/5) ~ 8 -> mpv /eureka/home/var/tmp/shotcut.mp4
[ffmpeg/demuxer] mov,mp4,m4a,3gp,3g2,mj2: moov atom not found
[lavf] avformat_open_input() failed
[ffmpeg/demuxer] mov,mp4,m4a,3gp,3g2,mj2: moov atom not found
[lavf] avformat_open_input() failed
Failed to recognize file format.
What went wrong there? Gradually I don't know if I care. For a last-ditch attempt, downloaded the latest version and installed it on dischord, my Microsoft box. It ran, but there was no display. I didn't bother checking why, though potentially it can't handle rdesktop, and is too polite to complain.
OK, lagoon is frustrating to use. Yvonne has a German keyboard, and other issues about the tactile interface irritate me. It doesn't run on teevee, one of my current issues with upgrading. Install on eureka? Unlikely to work: it's years out of date, waiting for me to find a way to update it. But it's worth an attempt:
/!\ WARNING /!\
Ports Collection support for your FreeBSD version has ended, and no ports are
guaranteed to build on this system. Please upgrade to a supported release.
Now doesn't that suggest that I would get support for more recent versions? Guarantees! I wish! I was able to bypass that, but when it wanted to rebuild python, I decided that it was too intrusive for such a fragile system.
It's quite possible that shotcut is a good program, but I have had nothing but trouble with it:
| Tuesday, 3 March 2020 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
teevee build pause
|
Topic: technology, opinion | Link here |
So how do I handle my issues with upgrading teevee? On the face of it, the upgrade process is broken. But why does nobody else complain? Obvious reasons could be that they maybe think it's modern, or that they don't want to make fools of themselves complaining in public: I belong (currently, at least) to the latter group.
Other reasons could be:
Clearly the best thing to do would be to check out the source tree freshly from the FreeBSD repository and try again, and also to reinstall the Nvidia driver. Mañana.
|
Historic photos
|
Topic: general, history, photography, opinion | Link here |
Gabi Reichert (the same person whose mail system has caused me untold grief) has returned a number of letters and photos to Yvonne. I thought that that was typical of people breaking up, but it seems that she's just trying to tidy up her household, something that we should probably also be doing.
OK, scan them in. Did that, with only moderate success. Process them with DxO PhotoLab and Ashampoo optimizer, also with moderate success. But where do I put them? By chance a single photo had a date on it, and another was of Yvonne's and my wedding, so we can easily put a date to that. And a number had month stamps from the processing, so I guessed round the beginning of those months. One, however, was confusing: a photo of Yvonne on Majia Negra:
|
|||||||||||||||||||||
The photo has a date stamp on the back: 04/96. But Yvonne is sure that it was taken Maryborough, Victoria in November 1997. I can't see how that can work.
And the rest? Guesswork.
The date allocation isn't as straightforward as it seems. If at a later date we find that we were wrong, and I move them to the correct date, I'll be leaving a whole lot of corpses behind. But I'll cross that bridge when I come to it. More to the point is: how do I link to the “new” photos? A list of dates would be a start, but maybe I should just create a page with all the photos. They all have enough information that they can link back to the date under which they were stored. So that's this page.
| Wednesday, 4 March 2020 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
No toilet paper!
|
Topic: general, health, opinion | Link here |
In my mail this morning, found this article: People Panic-Buying Toilet Paper Are Most Likely To Die From Coronavirus, New Research Confirms. OK, that's The Shovel, an Australian satirical site, but there must be a grain of truth in it.
Yvonne came back from shopping with the news that toilet paper was sold out at Woolworths in Sebastopol. She wasn't looking for any, but the empty shelves reached out and grabbed her, presumably something like this image from the Shovel article:
But she came up with a good alternative, paper that we've been storing for five years since another emergency:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That doesn't answer the question of “why”, though. Clearly it's a case of panic buying, but why specifically toilet paper? This article explains somewhat, but not the specific product. And the closest known case of COVID-19 appears to be 150 km away in Melbourne.
|
Nokia 3 ring tones die
|
Topic: technology, general, opinion | Link here |
Six weeks ago, after upgrading the operating system on my Nokia 3 mobile phone, I discovered that it no longer produced a ring tone for an incoming call. That's basic telephone functionality! It proved only to have disabled my custom tones, so I growled and installed one of their particularly emetic standard tones.
But today Yvonne called me from town. Only vibrate. Fortunately I had it in my shirt pocket, so I received it. But now no tone sounded, not even when I tried to select them, something that had previously worked.
A bit of searching on the web proved this is not specific to Android 9 (or whatever the silly name is). This article and this one both confirmed it on multiple OS releases and multiple phone models. Some claimed that the standard Microsoft solution, reboot, would work.
So I tried that. Yes! And now my custom tone is back too, at least for the moment. But somehow there's something basically wrong in producing many phone models and many software releases with such a basic bug.
|
teevee progress
|
Topic: technology, multimedia, opinion | Link here |
The daily teevee pain today. First, check out a pristine copy of the source tree and
build a new world, again, this time with a kernel that divulges when it was built. I've run
into this before, but I thought people were going to fix the bugfeature. It took
me a while to find this completely non-intuitive setting, only six months ago: create a file with an equally unlikely name /etc/src.conf and
add:
# Don't drop build date, dammit!
# What a silly name for the knob.
WITHOUT_REPRODUCIBLE_BUILD=dammit
Was that the correct syntax? Yes. Here before and after:
FreeBSD teevee.lemis.com 12.1-STABLE FreeBSD 12.1-STABLE r355358 GENERIC amd64
FreeBSD teevee.lemis.com 12.1-STABLE FreeBSD 12.1-STABLE #0 r358594: Wed Mar 4 13:20:11 AEDT 2020 grog@teevee.lemis.com:/spool/obj/usr/src/amd64.amd64/sys/GENERIC amd64
This #0 is still silly. Once it showed the build iteration of the kernel, but nowadays it's automatically reset to 0 on every build.
Next, as planned, reinstall the boot blocks, cross fingers, and reboot. It worked! But this lua problem is still there! I later took a more careful look. I had already established that the error was in a function defaultframe(), referring to a function core.isSerialConsole(). But where does that come from? Simple, from /boot/lua/core.lua, and presumably this line in drawer.lua (where the error was reported) includes it:
local core = require("core")
So why did it fail? It all looks fine to me. Still, there's a simple workaround: remove the invocation. It's only necessary, as the name suggests, if I'm running a serial console.
Next, reinstall the correct Nvidia drivers. Before, pkg info showed:
=== root@teevee (/dev/pts/7) ~ 3 -> pkg info | grep nvidia
nvidia-driver-440.59 NVidia graphics card binary drivers for hardware OpenGL rendering
nvidia-settings-440.26 Display Control Panel for X NVidia driver
nvidia-xconfig-440.26 Tool to manipulate X configuration files for the NVidia driver
OK, check for the exact name of the old driver:
=== root@teevee (/dev/pts/7) ~ 4 -> pkg search nvidia
pkg: wrong architecture: FreeBSD:12.0:amd64 instead of FreeBSD:12:amd64
pkg: repository FreeBSD contains packages with wrong ABI: FreeBSD:12.0:amd64
What's that? Coincidentally nothing to do with me. I had seen some talk of the bug (database corruption) on the -CURRENT mailing list a few days ago, apparently one of two similar bugs. But nobody had reported it for 12-STABLE. Why not? Sent out a message asking for a workaround, got no answer by evening.
OK, reboot (clean!), and X also started, still with broken key mapping:
=== grog@teevee (/dev/pts/8) ~ 5 -> diff -wu xmodmap-before xmodmap-fail
--- xmodmap-before 2020-03-04 17:57:01.342966000 +1100
+++ xmodmap-fail 2020-03-04 17:42:37.232907000 +1100
@@ -1,9 +1,9 @@
xmodmap: up to 4 keys per modifier, (keycodes in parentheses):
shift Shift_L (0x32), Shift_R (0x3e)
-lock Caps_Lock (0x42)
-control Control_L (0x25), Control_R (0x69)
-mod1 Alt_L (0x40), Alt_R (0x6c), Meta_L (0xcd)
+lock Control_L (0x42)
+control Meta_L (0x25), Control_R (0x69)
+mod1 grave (0x40), Alt_R (0x6c), Meta_L (0xcd)
mod2 Num_Lock (0x4d)
mod3
mod4 Super_L (0x85), Super_R (0x86), Super_L (0xce), Hyper_L (0xcf)
xmodmap-before was the output before trying to remap, and xmodmap-fail afterwards. What does that all mean? Somehow these things seem backwards. But what I see is that the keycode 0x42 was successfully changed from Caps_Lock to Control_L, but it seems that it continued to serve as a lock function. There's one level of variables too many here.
Started X again, this time without trying to remap the keys, and ran it from an xterm. Success!
=== grog@teevee (/dev/pts/8) ~ 6 -> diff -wu xmodmap-before xmodmap-success
--- xmodmap-before 2020-03-04 17:57:01.342966000 +1100
+++ xmodmap-success 2020-03-04 17:58:26.751187000 +1100
@@ -1,9 +1,9 @@
-xmodmap: up to 4 keys per modifier, (keycodes in parentheses):
+xmodmap: up to 5 keys per modifier, (keycodes in parentheses):
shift Shift_L (0x32), Shift_R (0x3e)
-lock Caps_Lock (0x42)
-control Control_L (0x25), Control_R (0x69)
-mod1 Alt_L (0x40), Alt_R (0x6c), Meta_L (0xcd)
+lock
+control Control_L (0x42), Control_R (0x69)
+mod1 Meta_L (0x25), grave (0x40), Alt_R (0x6c), Meta_L (0xcc), Meta_L (0xcd)
mod2 Num_Lock (0x4d)
mod3
mod4 Super_L (0x85), Super_R (0x86), Super_L (0xce), Hyper_L (0xcf)
Now Control_L is control, not lock, and there is no lock key. That's what I wanted. But what's the difference? I ran exactly the same command, using the same definition file, once in the startup script (failed) and once from an xterm (success).
One possibility was another problem that I discovered: the Page_Down key no longer worked. Examination with xev showed that it was keycode 117, Multi_key. That's explicitly mentioned in the configuration file. But it wasn't intended to be Page_Down (or Next, as X prefers to call it). It seems that, for some reason, the keymap has changed since my last build. Why? How? More rebooting needed to find out.
And then the simulated middle button on the mouse didn't work. Again no reason why not. Same moused configuration. But while reading moused(8), I discovered that I hadn't told it to simulate the middle button (-3 option). So previously it had done it anyway, a benign bug on which I had learnt to rely, and now they had fixed it. Add -3 and all was well.
And the error messages I got from mpv and shotcut? Gone! So it really was a driver issue. Despite all the other pain, that's particularly relieving.
But we're not done yet. Sometimes the X server started automatically on login, as configured, and sometimes it didn't. Finally managed to save the output from startx and discovered:
/usr/local/bin/X: Only console users are allowed to run the X server
xinit: giving up
xinit: unable to connect to X server: Connection refused
xinit: server error
Huh? Since when? And why? And why does it only fail sometimes? Found some information online:
If you are sure about the security implications (anyone can start the X-Server with root rights, which, since it has privileged hardware access and accesses a lot of files can open up a big security hole) and care only for functionality, not for security, for
So when i try to use the Xorg command as a normal user, this is the error that it gives me :
/usr/lib/xorg/Xorg.wrap: Only console users are allowed to run the X server
the following can help:
You can add to
/etc/X11/Xwrapper.config
the line
allowed_users = anybody
I didn't get round to trying that, and I'm not sure it's relevant (a “console user” isn't root).
Enough pain for one day? Definitely! But there was more. I've had issues with the screen saver for a long time. Sometimes it works, sometimes (especially on lagoon) it doesn't. And now it doesn't on teevee either. This command shows no obvious effect:
=== grog@teevee (/dev/pts/3) ~ 6 -> xtset dpms force off
Why not? It seems that I now need to turn off the monitor.
One problem seems not to be new: switching from an X display to a vty scrambles the screen. I've seen this from time to time over the years, and I still don't know why it only happens sometimes, though it seems reasonable to suspect the driver. But it happens on the old teevee system too, so clearly I haven't been too worried about it. On the other hand, it doesn't happen on lagoon, which uses the same display hardware and software. What's the difference?
I've been trying what I thought would be a relatively trivial upgrade for 5 days now, and I'm still not done. Some of the bugs I've encountered can't be my own fault:
Does anybody use this stuff any more?
| Thursday, 5 March 2020 | Dereel | Images for 5 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Winter on its way
|
Topic: general, food and drink | Link here |
We haven't had much of a summer, and now it's autumn. It shows:
|
|||||||||||||||||||||||
And we have still only harvested less than 10 tomatoes.
|
More NBN outages!
|
Topic: technology, opinion | Link here |
It's been over two weeks since the perpetual threat of National Broadband Network outages finished. And since then we haven't had any outages!
Until today:
| Duration | from | to | ||
| (seconds) | ||||
| 494 | 5 March 2020 14:20:21 | 5 March 2020 14:28:35 | ||
| 1234 | 5 March 2020 14:35:43 | 5 March 2020 14:56:17 | ||
| 24 | 5 March 2020 14:58:21 | 5 March 2020 14:58:45 | ||
Unregistered maintenance? No, I received no less than three emails on the subject:
Date: Thu, 5 Mar 2020 15:10:14 +1100
From: Aussie Broadband <support@aussiebroadband.com.au>
NBNCo has let us know that your service at 29 STONES RD, DEREEL VIC may currently be affected by an outage.
You can check the current status of this outage in the MyAussie App or Portal.
Date: Thu, 5 Mar 2020 15:30:14 +1100
NBNCo has let us know that your service at 29 STONES RD, DEREEL VIC may currently be affected by an outage.
NBN comments
05/03/2020 03:24pm AEDT - nbn is aware of an incident impacting end users on the Fixed Wireless network in Linton, VIC.
Alarms show a total loss of link. Remote troubleshooting is underway.
Date: Thu, 5 Mar 2020 16:00:19 +1100
NBNCo has let us know that the outage affecting your service at 29 STONES RD, DEREEL VIC has been resolved.
I suppose that's better than nothing, but look at those times. The first one was sent over 10 minutes after the problem was resolved, and the last one over an hour after. Presumably not Aussie Broadband's fault, based on the NBN (12 hour) timestamp. Still, it's nice to know that there's some kind of reporting, and it seems that there's some kind of monitoring of the network, something that I had doubted in the past.
|
Daily teevee update
|
Topic: technology, opinion | Link here |
The updated teevee is now working well enough for me to be able to use it, but it's still not fixed. What's the story with the keyboard mapping? And why can't I start X from the startup script?
For the X problem I tried a solution suggested by this page. Create /etc/X11/Xwrapper.config with this content:
allowed_users = anybody
Then to investigate the keymap. Clearly I needed the old system to compare with, so rebooted there and saved the keymap with xmodmap -pk. Boot back to the new system. X still won't start automatically, so either the suggestion was wrong or the location of the file was wrong. On the face of it, it seems likely that the location was correct. It's a relatively benign error (just means that I need to start X manually), but I'd like to know what the problem is. At least it gave me the opportunity to confirm that the vty keyboard “bounce” is still with us.
And the keymaps? Beyond the standard QWERTY keys there was almost no correlation. There are 255 (256?) possible key codes, but the diff between the two maps was 320 lines long. In particular, the old keymap had:
KeyCode Keysym (Keysym) ...
Value Value (Name) ...
105 0xff56 (Next) 0x0000 (NoSymbol) 0xff56 (Next)
117 0xff67 (Menu) 0x0000 (NoSymbol) 0xff67 (Menu)
147
And the new keymap keymap has:
117 0xff56 (Next) 0x0000 (NoSymbol) 0xff56 (Next)
130 0xff31 (Hangul) 0x0000 (NoSymbol) 0xff31 (Hangul)
131 0xff34 (Hangul_Hanja) 0x0000 (NoSymbol) 0xff34 (Hangul_Hanja)
135 0xff67 (Menu) 0x0000 (NoSymbol) 0xff67 (Menu)
147 0x1008ff65 (XF86MenuKB) 0x0000 (NoSymbol) 0x1008ff65 (XF86MenuKB)
And to make things even more confusing, xev reported different keycodes again. The Menu key (oxff67), which I remap to Multi_key, reported as keycode 147, not keycode 135. In any case, I was able to fix things with:
xmodmap -e "keycode 117 = Next"
xmodmap -e "keycode 147 = Multi_key"
But why these changes? The new keymaps have a number of keycodes that the old ones don't, like the Hangul and Hangul_Hanja that I included in the list above. It's been nearly 30 years since I delved into X keymaps, but it looks like I'm going to have to again.
On the positive side, the screen saver now works. What was wrong yesterday? How will I ever know?
|
A bone! A bone!
|
Topic: animals | Link here |
Leonid is recovering well from his injury over the weekend, but he's still on special diet, in particular no chicken frames. But what's that to a Real Hunting Dog? He found this in the scrub on the side of the road:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Sadly (except from Nikolai's perspective) he had to put it down:
|
||||||||||||||||||||||||
|
Dying Buddleja
|
Topic: gardening, opinion | Link here |
Despite all the attention, our last Buddleja × weyeriana appears to be dying, though it's trying to produce new growth:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Why? Once again I'm wondering if there isn't something in the soil. But it's clear that we need to do something else, so I made 5 cuttings and planted them. Hopefully they'll take; then I can plant them somewhere else.
|
PIXIO fail
|
Topic: photography, multimedia, technology, opinion | Link here |
Yvonne wanted to take another sequence of video clips today using the PIXIO “Robot Cameraman”. Set it up in place, turned on the beacons... nothing. No power.
Further examination showed that it was only beacon 1, which appeared to be completely discharged. Aborted the effort, with considerable growling on Yvonne's part, and put the things on charge. Yes, everything was charged except for beacon 1, which charged normally.
What's the issue? Apart from the documented bug with the controller itself (after charging you must power it on, then off again, to avoid immediate self-discharge), I've seen the beacons spontaneously turn on when removed from the charger. Was that the reason? Hopefully it's not a dying battery.
|
Chicken pinwheels
|
Topic: food and drink, opinion | Link here |
Yvonne bought something new when shopping yesterday, “chicken pinwheels”. After baking in the oven (190° for 25 minutes), they looked like this:
|
|||||||||||||||||||||||
And the verdict? Edible, but boring. Yvonne is now considering doing something similar with mincemeat.
| Friday, 6 March 2020 | Dereel | Images for 6 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another grid failure
|
Topic: Stones Road house, general | Link here |
Another brief grid power failure this morning at 1:53:26.
|
Garden flowers: early autumn
|
Topic: gardening, opinion | Link here |
My garden photos are a monthly thing, now towards the end of the month, but this has been a strange year, and it's worth a few more photos. For example, the tomatoes are still not ripe, though the chilis are getting there:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
The Clematis “Edo murasaki” has brought forth a single flower:
|
||||||||||||||||||||||||
It's still, after over two years, looking pretty sickly. Maybe the epazote, flowering vigorously on the other side of the lattice, is at least partially to blame:
|
||||||||||||||||||||||||
I have a feeling that we will never need to sow epazote again.
Then there are these lilies, which I think should have been flowering a month ago:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
The Strelitzia nicolai is also still flowering, though the writing's on the wall:
|
||||||||||||||||||||||||
Still, it has been flowering now for over two months.
And then there's the dying Buddleja × weyeriana. Despite all attempts, it seems to be fading fast:
|
|||||||||||||||||||||||
On the other hand, the cuttings I took look surprisingly good:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Why is that? Normally cuttings look quite sickly for a while, until they manage to produce a few roots.
|
PIXIO: success!
|
Topic: photography, multimedia, opinion | Link here |
Yvonne had another go taking videos with her PIXIO “Robot Cameraman” today. It worked. So it seems that the issue with beacon 1 was really that it had turned itself on, probably when being disconnected from the charger, and Yvonne hadn't noticed.
But there's another issue: framing. The PIXIO includes software for controlling the focal length of some very few Sony video cameras, but modern video-capable stills cameras, both with and without mirrors, are too tasteful to offer such functionality. I considered buying a video camera instead of the Olympus OM-D E-M5 Mark III—after all, it would have been cheaper—but came to the conclusion that framing is too complicated an issue to leave to the PIXIO.
One case in point became evident today. The PIXIO follows a “badge”, which normally the actress wears on her wrist like a watch, barely visible on this screen shot:
|
|||||||||||||||||||||||
Why is it overexposed? That's deliberate. For reasons I don't understand, Yvonne likes it that way, and nothing I can do can convince her otherwise.
But the real issue is that the image is incorrectly framed. I could easily fake it to centre the image 2 m to the left of the “badge”, but then things would be even worse when the horse was on her other side. I suppose we're going to have to find a way to attach it to the horse, possibly hidden in its mane.
|
Chicken wings revisited
|
Topic: food and drink, opinion | Link here |
Time to cook some more breakfast chicken wings today, the ones I eat with nasi lemak. Tweak the recipe? Sure, why not. I do them sous-vide, and so far the recipe called for:
| quantity | ingredient | step | ||
| 1 kg (about 10) | chicken wings | 1 | ||
| 20 g | garlic | 1 | ||
| 10 g | ginger | 1 | ||
| salt | 1 | |||
| sesame oil | 1 | |||
Today I made 2 kg, but reducing to the quantities above, I added:
| quantity | ingredient | step | ||
| 1 kg (about 10) | chicken wings | 1 | ||
| 20 g | garlic | 1 | ||
| 10 g | ginger | 1 | ||
| 15 g | soya sauce | 1 | ||
| 2.5 g | dark soya sauce | 1 | ||
| 5 g | sesame oil | 1 | ||
How many bags? There were about 22 wings, and one bag seemed not enough. But I think I misjudged. They don't all need to be in a single layer—in fact, it's difficult to keep them that way when evacuating the bags. I ended up with (half the wings):
|
|||||||||||||||||||||||
So: next time put all the wings in one basket.
The results? I won't know that for a day or two, until I actually eat them. But I think next time I might try twice the quantity of marinade.
| Saturday, 7 March 2020 | Dereel | Images for 7 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another garage sale
|
Topic: general | Link here |
Julie Donaghy had another garage sale this week, so of course Yvonne had to go, and it seemed prudent for me to come with her to avoid ending up with too much junk.
As it happened, there wasn't too much. We would have got away for under $10, including some brandy balloons, if it hadn't been for a case that reminded me of a bassoon case:
|
|||||||||||||||||||||||
It proved to be a foldable table and four seats:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And that cost us the princely sum of $20.
I was also rather amused by the presence of Paris Hilton, Julie's sow:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
She's rather like a pet dog.
| Sunday, 8 March 2020 | Dereel | Images for 8 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
This can take a _long_ time
|
Topic: multimedia, technology, opinion | Link here |
Julie Donaghy had had some DVDs of Heartbeat for sale yesterday, but we didn't take them: we already have the episodes. But then it occurred to me that the quality of the copies I have is pretty terrible, so for $5 for 17 DVDs (5 seasons) it seemed to make sense. So today Yvonne went off to pick them up.
17 DVDs? She came back with 54! Julie had lent her the complete DVDs for Lost, of which we had never heard, but which Julie says is excellent—IMDB agrees.
OK, the usual problem: we don't have a DVD player, so I needed to read them in to disk. How do you do that? As I discovered 9 years ago, it seems that the easiest way is to use mplayer:
$ mplayer dvd://1 -dumpstream -dumpfile $1
OK, tried that.
libdvdread: Encrypted DVD support unavailable.
Yes, I've seen that before too. I need to build libdvdcss locally. For bizarre reasons it's legal to distribute the source for the decryption routines, but not the binaries.
Damn. How much danger is there that I'll run into trouble if I rebuild libdvdread? Tried and ran into trouble:
===> libdvdread-6.0.2 depends on shared library: libdvdcss.so.2 - not found
*** Error code 1
Well that's clever, isn't it? After all, that's the dependency I wanted to build. I don't understand why the system didn't just build it. But OK, do it manually, which was quick and easy.
Do I even need to build libdvdread now? No! Just building libdvdcss was enough. Then off to read the DVD, running in to an old message:
libdvdread: Attempting to retrieve all CSS keys
libdvdread: This can take a _long_ time, please be patient
libdvdread: Get key for /VIDEO_TS/VIDEO_TS.VOB at 0x0000017f
libdvdread: Elapsed time 0
...
The _long_ time is so short that you can't see any delay at all.
Read in the files, not without surprises: somehow the total length was wrong. Playing with mplayer showed a length of roughly 25 minutes instead of the 45 it should have been. With mpv it was even worse, only round 10 seconds. When it reached the “limit”, it just changed it to a higher limit, and (later) we were able to watch the first episode in its entirety with mpv. The only issue was with seeking: try to seek beyond the current idea of EOF and it just stops.
“Heartbeat” was a different problem. For the first time ever, it did take a _long_ time to copy:
libdvdread: Get key for /VIDEO_TS/VTS_02_1.VOB at 0x000039b2
libdvdread: Error cracking CSS key for /VIDEO_TS/VTS_02_1.VOB (0x000039b2)!!
libdvdread: Elapsed time 1342
Is it even worth continuing if it can't crack a key? It did continue, and it copied 3 episodes to disk. Take a look at them from eureka, and they looked fine. It wasn't until I wiped the vomit off the xterm that I saw:
AV: 00:00:10 / 00:51:39 (0%) A-V: 0.000
Error decoding audio.
AV: 00:01:11 / 00:51:39 (2%) A-V: -0.000
Error decoding audio.
That was withmpv. mplayer was even more emphatic: it couldn't find any sound track at all.
It happened on two different DVDs, so it looks as if it's a genuine encryption issue, not poor quality media (which, in fact, looked pristine). So have I finally run into the limitations of DVD copying software? Put one of the DVDs into euroa, my Microsoft laptop, where “Windows” media player ran into issues probably unrelated to my problem (“try selecting a different display size”). Also tried installing mplayer and mpv on euroa, but at first I couldn't start it. No icon, and when I tried entering mpv.exe it didn't want to know:
|
|
|||||||||
I had to search for the thing to start it, and when I did, I couldn't get my head around the navigation:
|
|||||||||||||||||||||||
What does that mean? How do I “drop” a DVD into that big, empty box? It didn't seem to be worth the trouble to find out.
What next? Well, there's the old Digitrex GKX-9000 that I bought 16 years ago, but it's not connected up, and I don't know if it works any more. But at the very least it should be able to play the media.
|
Lack of ram?
|
Topic: food and drink, opinion | Link here |
Last year Chris Bahlo had a couple of sheep appear on her property, this time not belonging to David Yeardley. In the end she had them slaughtered, and we received a considerable amount of meat from her.
Problem: Yvonne doesn't really like lamb. But we need to eat it, so today we tried the rack (20 minutes at 200°, heating from above in the oven to internal temperature 54°). It didn't taste bad, though it's a bit messy. But Yvonne didn't like it, so it looks as if I'll have to share the other two racks that arrived the following month with somebody else.
|
The curry tree returns
|
Topic: gardening, opinion | Link here |
Three months ago I took my Curry tree outside to try to cure it of mites.
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It also cured it of leaves:
|
|||||||||||||||||||||||
But since then it seems to have recovered well, and today Chris Bahlo helped me bring it back inside again. It has many small leaves:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Presumably they just haven't finished growing yet. At least it looks nice and healthy.
| Monday, 9 March 2020 | Dereel | Images for 9 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More video processing pain
|
Topic: multimedia, technology, opinion | Link here |
More investigation of the files that I had ripped yesterday. The good news: the audio worked after all; the error messages were just confusion on the part of the player. The bad news: some of the episodes had not been decrypted.
As planned, conected up my ancient Digitrex GKX-9000 DVD recorder to the TV monitor today and put in one of the Heartbeat DVDs. It worked and behaved correctly, including providing me with the titles, which none of the software had extracted.
OK, how about copying the entire DVD to disk? It proves that vobcopy has a corresponding option:
vobcopy -l -m
The -l option shouldn't be necessary, but otherwise vobcopy thinks that files can't be larger than 2 GB. That worked, and I ended up with effectively a copy of the DVD file system. And I still have issues with the time information, which clearly relates to the way the files are split on DVD. A typical DVD (the one I'm using) has something like:
=== grog@teevee (/dev/pts/5) /spool/Series/Heartbeat/01/DVDVOLUME 143 -> l -R
total 1
drwxr-xr-x 2 grog home 1,024 9 Mar 13:55 VIDEO_TS
./VIDEO_TS:
...
-rw-r--r-- 1 grog home 1,073,709,056 9 Mar 12:50 VTS_02_1.VOB
-rw-r--r-- 1 grog home 1,073,709,056 9 Mar 12:56 VTS_02_2.VOB
-rw-r--r-- 1 grog home 377,356,288 9 Mar 12:59 VTS_02_3.VOB
That's track 2 (the 02 in the file names). The maximum file size is 2³⁰-2¹⁵ bytes, and DVDs automatically concatenate them. But it seems that both vobcopy and mplayer misinterpret the time in different ways.
There must be something better. The last version of vobcopy was ported to FreeBSD in June 2009. Is there a newer version? Hard to tell: http://www.vobcopy.org/ has an internal server error, a suggestion that it's not exactly bleeding edge.
OK, I'm still looking for a video editor. I think I can give up on shotcut. What else is there? Looking on the web, it seems that I should give kdenlive another chance. 110 packages to install! At least it didn't want to remove Emacs. Finally it was done:
=== grog@lagoon (/dev/pts/11) ~ 4 -> kdenlive
"Session bus not found\nTo circumvent this problem try the following command (with Linux and bash)\nexport $(dbus-launch)"
Ah, Linux... What it's trying to say is “we don't do no steenking network X”, but it's too polite to say that. I'll have to go to a display connected to lagoon. Mañana.
Also discovered that avidemux is a video editor. I'm sure I've used that before in some capacity, but didn't bother to look it up. Instead installed it. Only 3 packages!
New packages to be INSTALLED:
avidemux: 2.6.11_15
gcc9: 9.2.0
mpc: 1.1.0_2
Why gcc9? That's a compiler. OK, installed it.
=== grog@lagoon (/dev/pts/11) ~ 5 -> avidemux
bash: avidemux: command not found
OK, what clever name have they thought out for the executable? Any info in pkg-descr? No. What did it install in /usr/local/bin? 9 files with 19 names like gcov, cpp, c++, gcc and so on. Just the compiler!
OK, what's in pkg-plist? Nothing! This port doesn't seem to install an executable. What does the official documentation say? The “Quickstart” starts:
When you open a video file using the Open button (Ctrl+O) on the main toolbar, there are 3 basic things you can do with it in Avidemux:
Now that's helpful, isn't it? How do you run the bloody thing? Checked the “how to build”, but that was mainly workarounds for lack of dependency information, and it didn't say how to run it either. Another item for tomorrow.
| Tuesday, 10 March 2020 | Dereel | Images for 10 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Alternative multimedia editors
|
Topic: multimedia, opinion | Link here |
Yesterday I made very limited progress with video processing software. I established that vobcopy was so out of date (over 10 years) that its web server no longer works. That has since been fixed: it now belongs to a domain squatter. And avidemux installed libraries but no executable, not exactly an encouraging situation.
There must be good software out there. Did a bit of look on the web and came up with Top 12 Linux Video Editors to Edit Videos with Ease, written by Liza Brown on Feb 20, 2020, less than a month ago. Went through them and found:
Wondershare Fimora9, the product that this site sells, and which they find best. Problem: it doesn't run on Linux! Now isn't that silly?
The top editors were:
kdenlive, which doesn't want to work in a network environment. Still noted to try out.
Shotcut. Been there, done that, finished screaming.
pitivi: Pitivi is a Non linear video editor and also open source software that freely available. For intuitive video editing this software is designed. It supports simple features i.e. snapping, trimming, splitting, and clips cutting.
That intuitive word again. I don't believe a word of it. So far I haven't found any video software that comes close to being intuitive. About the most interesting thing about this description is its grammar.
Blender, which appears to be a 3D program. Not what I'm looking for.
Cinelerra. No FreeBSD port available.
LiVES, “aimed at the digital video artist who wants to create their own content, the video editor who wants to produce professional looking video, and the VJ who wants to captivate with spectacular images.”
Is that me? In any case, FreeBSD has a port, so I could consider it.
Open Shot, “an award-winning free and open-source video editor for Linux, Mac, and Windows, and is dedicated to delivering high quality video editing and animation solutions to the world.”
Well, I'd be happy if it made my life a little easier. But there's a FreeBSD port for it, so another to note.
Flowblade. No FreeBSD port. And it's clear why. Recall that this page was updated 20 February 2020:
The version 0.12 of flowblade movie under was released for Linux in January 2014 and the version 0.14 is in the pipeline to be released in June 2014.
Vivia: No port.
Lightworks and Davinci Resolve: cost money.
|
More video editor pain
|
Topic: multimedia, technology, opinion | Link here |
So first let's try to get the video software that I have to run. It's excruciating for multiple reasons.
First another look at avidemux. Why is there no executable? Ah, that's for me to know and you to guess. There are a total of five ports: avidemux, avidemux-cli, avidemux-plugins, avidemux-qt4 and avidemux-qt5. What do the pkg-descrs say?
A video editor that allows editing of AVI, OGM, and MPEG videos. The MPEG
support provides the ability to convert to DVD compliat PS streams. It
contains various filters for deinterlacing, cropping, resizing, etc. Allows
for cutting without re-encoding. Has the ability to re-encode and re-sample.
Utilizes ECMAScript (aka. Javascript) to provide batch scripting and
processing support.
WWW: http://avidemux.org/
That's in avidemux. But it lies. This port is not a video editor. It has no executables at all. And none of the other ports have any description whatsoever!
OK, let's guess at avidemux-qt5. Try to install the package.
avidemux: package size conflict
It suggested that I try pkg upgrade -f, which sounded dangerous. Try building from source? Spent a couple of hours tripping over things like being given a configuration file to build a port, then having it not build and having the port that depends on it fail:
+---------------------------- py37-Jinja2-2.10.1 ------------------------------+
# +--------------------------------------------------------------------------+ #
# #+[x] BABEL Enable Babel extension # #
# #+[x] EXAMPLES Build and/or install examples # #
# +--------------------------------------------------------------------------+ #
+------------------------------------------------------------------------------+
# < OK > <Cancel> #
+------------------------------------------------------------------------------+
(continue)
===> py37-sphinx-1.6.5_2,1 depends on package: py37-Jinja2>=2.3 - not found
*** Error code 1
I think that that's because a different version of the port was installed. In each case I had to build and reinstall the port.
Or did it? No, the port was not installed, but it seems that it wanted to be installed manually anyway:
=== root@teevee (/dev/pts/7) /usr/ports/devel/py-Jinja2 50 -> Make deinstall reinstall
===== Tue 10 Mar 2020 12:44:15 AEDT on teevee.lemis.com: Make deinstall reinstall
===> Deinstalling for py37-Jinja2
===> py37-Jinja2 not installed, skipping
===> Deinstalling for py37-Jinja2
===> py37-Jinja2 not installed, skipping
===> Installing for py37-Jinja2-2.10.1
===> Checking if py37-Jinja2 is already installed
actual-package-depends: dependency on py37-MarkupSafe>=0.23 not registered
actual-package-depends: dependency on py37-Babel>=1.3 not registered
===> Registering installation for py37-Jinja2-2.10.1
Installing py37-Jinja2-2.10.1...
Surely there's a simpler way. After much pain I discovered that yes, indeed, my package database was corrupted, and that pkg update -f (silently) fixed the breakage, so I could install the port.
And, of course, on consideration the port decided that it needed to remove Emacs. I'm used to that, something clearly implemented by the Vi warriors. But no, this time I couldn't reinstall it! avidemux wants ImageMagick 7, and Emacs wants version 6. It's hard coded in the Makefile.
OK, that's simple enough. Modify the Makefile:
--- Makefile (revision 528187)
+++ Makefile (working copy)
@@ -176,7 +176,7 @@
M17N_IMPLIES= XFT
M17N_LIB_DEPENDS= libm17n.so:devel/m17n-lib
M17N_CONFIGURE_WITH= m17n-flt
-MAGICK_LIB_DEPENDS= libMagickCore-6.so:graphics/ImageMagick6
+MAGICK_LIB_DEPENDS= libMagickCore-7.so:graphics/ImageMagick7
And that gave me:
=== root@teevee (/dev/pts/7) /usr/ports/editors/emacs 110 -> make all
...
===> Configuring for emacs-26.3_3,3
env: ./configure: No such file or directory
===> Script "configure" failed unexpectedly.
Please report the problem to emacs@FreeBSD.org [maintainer] and attach the
"/eureka/home/src/FreeBSD/svn/ports/editors/emacs/work-full/emacs-26.3/config.log"
including the output of the failure of your make command. Also, it might be
a good idea to provide an overview of all packages installed on your system
(e.g. a /usr/local/sbin/pkg-static info -g -Ea).
Huh? No complaint on the screen? Went off looking for /eureka/home/src/FreeBSD/svn/ports/editors/emacs/work-full/emacs-26.3/config.log, and it wasn't there!
Dammit, it's a silly idea to use ImageMagick in Emacs anyway. Reconfigure without it and rebuild.
|
Rediscovering avidemux
|
Topic: multimedia, technology, opinion | Link here |
So, FINALLY, I had avidemux-qt5 installed, at round 14:00. Let's run it:
The screen seems to be 1920 x 1080 px
[destroy] 03:49:21-533 Destroying preview
[renderDisplayResize] 03:49:21-533 Render to 0x0 zoom=1.0000, old one =0 x 0, zoom=1.0000, renderer=0x0
[UI_resize] 03:49:21-534 Resizing the main window to 152x230 px
[RDR] Resizing to 0 x 0
[initGUI] 03:49:21-535 OpenGL enabled at built time, checking if we should run it..
What are all those empty lines? Ah, they're not empty, just tastefully set to displaying white foreground on your choice of background. You can mark them, for example, which also makes them visible. Background? You have chosen my background, haven't you?
WHY do people do that? In fact, it wasn't quite as bad as that, since my xterm background is BlanchedAlmond, not white, so what I saw looked something like:
The screen seems to be 1920 x 1080 px
[destroy] 03:49:21-533 Destroying preview
[renderDisplayResize] 03:49:21-533 Render to 0x0 zoom=1.0000, old one =0 x 0, zoom=1.0000, renderer=0x0
[UI_resize] 03:49:21-534 Resizing the main window to 152x230 px
[RDR] Resizing to 0 x 0
[initGUI] 03:49:21-535 OpenGL enabled at built time, checking if we should run it..
[initGUI] 03:49:21-535 OpenGL not activated, not initialized
Probing for VDPAU...
VDPAU not available
[ADM_ad_plugin] Scanning directory /usr/local/lib/ADM_plugins6/audioDecoder
[ADM_ad_plugin] Cannot parse plugin
[ADM_av_loadPlugins] 03:49:21-627 [ADM_av_plugin] Scanning directory /usr/local/lib/ADM_plugins6/audioDevices
[ADM_av_loadPlugins] 03:49:21-627 [ADM_av_plugin] Cannot parse plugin
[ADM_ae_plugin] Scanning directory /usr/local/lib/ADM_plugins6/audioEncoders
[ADM_ae_plugin] Cannot parse plugin
[ADM_dm_plugin] Scanning directory /usr/local/lib/ADM_plugins6/demuxers
[ADM_av_plugin] Cannot parse plugin
[ADM_mx_plugin] Scanning directory /usr/local/lib/ADM_plugins6/muxers
[ADM_av_plugin] Cannot parse plugin
[ADM_vd6_plugin] Scanning directory /usr/local/lib/ADM_plugins6/videoDecoders
[ADM_vd6_plugin] Cannot parse plugin
[ADM_ve6_plugin] Scanning directory /usr/local/lib/ADM_plugins6/videoEncoders
[ADM_ve6_plugin] Cannot parse plugin
[ADM_ve6_plugin] Scanning directory /usr/local/lib/ADM_plugins6/videoEncoders/qt5
[ADM_ve6_plugin] Cannot parse plugin
[ADM_vf_plugin] Scanning directory /usr/local/lib/ADM_plugins6/videoFilters
[ADM_vf_plugin] Cannot parse plugin
[ADM_vf_plugin] Cannot parse plugin
[ADM_audioByName] 03:49:21-628 Device not found :ALSA
Found 1 video encoder(s)
Found 1 audio encoder(s)
Found 0 format(s)
[UI_RunApp] 03:49:21-628 Load default settings if any...
[setContainer] 03:49:21-628 Cannot find muxer for format=MKV
[AVDM_hasVolumeControl] 03:49:21-628 The current audio device doesn't support volume control
[UI_RunApp] 03:49:21-628 [autoUpdate]Current date 1284 , last check = 0
[save] 03:49:21-628 Saving prefs to /home/grog/.avidemux6/config3.tmp
[postInit] 03:49:21-726 myQApplication exec
[postInit] 03:49:21-726 Checking for crash...
ignored: custom
ignored: config3
No crash file (/home/grog/.avidemux6/crash.py)
[downloadFinished] 03:49:22-372 Download finished
[downloadFinished] 03:49:22-372 Error downloading update http://www.avidemux.org/update//update_for_???.html
[downloadFinished] 03:49:22-372 Er=Error transferring http://www.avidemux.org/update//update_for_???.html - server replied: Not Found
On the screen I first had to find the directory. Like all Microsoft space software, avidemux has no concept of working directories, so I had to climb trees. While taking screen shots, my xv window kept disappearing. It proved that avidemux always wants to be in charge, so it just raises its windows at the slightest provocation. Ugh!
Finally I selected a video clip:
|
|
|||||||||
Huh? That's a standard QuickTime format. Surely avidemux can handle that. The documentation implies, rather than says, that it can. OK, try with a track from a DVD. Same thing.
OK, what decoders or muxers can it handle? Installed Yet Another Port, avidemux_cli, which of course was too polite to have a man page. On a hunch ran avidemux --help, which offered me an equally emetic output, over 70 lines of debug output, including the names of all the authors and maintainers (so you know whom to blame?), before coming up with (somewhat truncated:
*** Automated : 19 entries*************
help-->0
Command line possible arguments :
--append, append video (one arg )
--audio-codec, set audio codec (copy|Lame|FDK_AAC|LavAC3|Opus|TwoLame|...) (one arg )
--avisynth-port, set avsproxy port accordingly (one arg )
--help, print this ( no arg )
--info, show information about loaded video and audio streams ( no arg )
--load, load video or workbench (one arg )
--nogui, Run in silent mode ( no arg )
--output-format, set output format (MKV|MP4|ffTS|ffPS|AVI|RAW|...) (one arg )
--video-codec, set video codec (Copy|x264|x265|xvid4|ffMpeg2|ffNvEnc|...) (one arg )
********** Automation ended***********
*********************************
*********************************
End of program..
*********************************
*********************************
That at least suggests that it should handle QuickTime, but --info produced more unintelligible output, including the same surroundings:
Initializing Dithering tables
[UI Toolkit] Running version 02:00
[Script] Scanning directory /usr/local/lib/ADM_plugins6/scriptEngines
[Script] Cannot parse plugin
[Script] Scanning directory /usr/local/lib/ADM_plugins6/scriptEngines/cli
[Script] Cannot parse plugin
[ADM_ad_plugin] Scanning directory /usr/local/lib/ADM_plugins6/audioDecoder
[ADM_ad_plugin] Cannot parse plugin
[ADM_av_loadPlugins] 03:35:12-825 [ADM_av_plugin] Scanning directory /usr/local/lib/ADM_plugins6/audioDevices
[ADM_av_loadPlugins] 03:35:12-825 [ADM_av_plugin] Cannot parse plugin
[ADM_ae_plugin] Scanning directory /usr/local/lib/ADM_plugins6/audioEncoders
[ADM_ae_plugin] Cannot parse plugin
[ADM_dm_plugin] Scanning directory /usr/local/lib/ADM_plugins6/demuxers
[ADM_av_plugin] Cannot parse plugin
[ADM_mx_plugin] Scanning directory /usr/local/lib/ADM_plugins6/muxers
[ADM_av_plugin] Cannot parse plugin
[ADM_vd6_plugin] Scanning directory /usr/local/lib/ADM_plugins6/videoDecoders
[ADM_vd6_plugin] Cannot parse plugin
[ADM_ve6_plugin] Scanning directory /usr/local/lib/ADM_plugins6/videoEncoders
[ADM_ve6_plugin] Cannot parse plugin
[ADM_ve6_plugin] Scanning directory /usr/local/lib/ADM_plugins6/videoEncoders/cli
[ADM_ve6_plugin] Cannot parse plugin
[ADM_vf_plugin] Scanning directory /usr/local/lib/ADM_plugins6/videoFilters
[ADM_vf_plugin] Cannot parse plugin
[ADM_vf_plugin] Cannot parse plugin
[ADM_audioByName] 03:35:12-825 Device not found :ALSA
*** Automated : 19 entries*************
info-->0
Nothing to get infos from
********** Automation ended***********
My guess is that the important part is between “Automated” and “Automation ended”: nothing.
That's enough. Time for a different editor.
|
Learning kdenlive
|
Topic: multimedia, technology, opinion | Link here |
In general, it seems that kdenlive is the most popular editor. It refuses to work over a network, but maybe it would be OK locally. Installed on teevee and tried again.
It runs! And produces about 200 lines of error and warning messages on the xterm, for once without trying to change the colour. Presumably this is typical of developers thinking that they can flush their debug messages down the xterm, and is of no further interest.
But then it brings a typical confusing error message:
|
|
|||||||||
Huh? What project file? And what does “error occurred while parsing element” mean? My guess is it's its inimitable way of saying “file does not exist”. Certainly I neither created nor specified a project file, and I don't want one.
Bad boy! No project file, no play. And clearly it ignores command line parameters, and like avidemux it wants its window to be on the top all the time. And even worse than avidemux, it seems to forget directory paths as soon as absolutely possible.
OK, climb the Microsoft directory tree to /Photos/yvonne/20200306 and try to open something. Of course it found nothing. Somehow managed to select something and was rewarded with even more brightly coloured vomit:
|
|||||||||||||||||||||||
What's that? Round here I lost track of what was going on. At one point ended up with streams of this vomit (at least, I think it's this attempt that produced it):
org.kde.solid.udisks2: Failed enumerating UDisks2 objects: "org.freedesktop.DBus.Error.ServiceUnknown"
"The name org.freedesktop.UDisks2 was not provided by any .service files"
org.kde.solid.udisks2: Failed enumerating UDisks2 objects: "org.freedesktop.DBus.Error.ServiceUnknown"
"The name org.freedesktop.UDisks2 was not provided by any .service files"
virtual QStringList Solid::Backends::UPower::UPowerManager::allDevices() error: "org.freedesktop.DBus.Error.ServiceUnknown"
/////////// createClipFromFile "/Photos/yvonne/20041006/dscn3802.jpg" "988" "/Photos/yvonne/20041006/dscn3802.jpg"
/////////// final xml "<producer type=\"5\" in=\"0\" length=\"125\">
<property name=\"resource\">/Photos/yvonne/20041006/dscn3802.jpg</property>
</producer>
"
/////////// requestAddBinClip "988"
/////////// found id "1014"
/////////// constructed
/////////// added true
org.kde.solid.udisks2: Failed enumerating UDisks2 objects: "org.freedesktop.DBus.Error.ServiceUnknown"
"The name org.freedesktop.UDisks2 was not provided by any .service files"
org.kde.solid.udisks2: Failed enumerating UDisks2 objects: "org.freedesktop.DBus.Error.ServiceUnknown"
"The name org.freedesktop.UDisks2 was not provided by any .service files"
virtual QStringList Solid::Backends::UPower::UPowerManager::allDevices() error: "org.freedesktop.DBus.Error.ServiceUnknown"
/////////// createClipFromFile "/Photos/yvonne/20041006/dscn3803.jpg" "988" "/Photos/yvonne/20041006/dscn3803.jpg"
/////////// final xml "<producer type=\"5\" in=\"0\" length=\"125\">
<property name=\"resource\">/Photos/yvonne/20041006/dscn3803.jpg</property>
</producer>
"
/////////// requestAddBinClip "988"
/////////// found id "1015"
/////////// constructed
/////////// added true
org.kde.solid.udisks2: Failed enumerating UDisks2 objects: "org.freedesktop.DBus.Error.ServiceUnknown"
"The name org.freedesktop.UDisks2 was not provided by any .service files"
org.kde.solid.udisks2: Failed enumerating UDisks2 objects: "org.freedesktop.DBus.Error.ServiceUnknown"
"The name org.freedesktop.UDisks2 was not provided by any .service files"
virtual QStringList Solid::Backends::UPower::UPowerManager::allDevices() error: "org.freedesktop.DBus.Error.ServiceUnknown"
That went on and on, and I couldn't find a way of stopping it short of shooting down the process. It's worth looking at those file names; they're JPEG files from over 15 years ago, and presumably it had intended to go through all 200,000 or so images. Why? I certainly tried to limit it to a directory with only 3 files.
Later I was able to establish that there's no easy way to open a file; first you need to create a project file, and then click on the button below &Project Bin, which produces this menu:
|
|
|||||||||
Clearly “Clip or Folder” is their jargon for “file or directory”.
After doing something there, ended up with a display that really showed my 3 files:
|
|
|||||||||
Now isn't that helpful? It truncates all three file names so that they're the same. Managed to trim the selected one (youtube-63061408.MOV), which seemed to work, but drawing the end marker forward was glacially slow and used 100% of all four cores of the machine!
So it seems that I need to drag the clip to the “timeline”, another intuitive term. Did that and ended up with:
|
|||||||||||||||||||||||
11 seconds! I probably ended up using one of the areas that I should have trimmed off. Clearly there's nothing “intuitive” about this program (or any other, it seems): I'm going to have to RTFM.
|
Cheap holiday cruises
|
Topic: general, health, opinion | Link here |
I usually ignore advertising on the web, but while trying to find my way through the video recorder documentation I was accosted by advertising for holiday cruises for extremely low prices, like 2 days for $208. Princess? Aren't they the people who run the Diamond Princess (currently in quarantine in Yokohama) and the Grand Princess (currently in quarantine in Oakland)? Yes.
They must be hurting. I wonder how much these fares are discounted. Of course, they're not on the level with their pricing either. This was sent to us in Australia, and you have to read carefully to discover that the pricing is in US dollars (currently round 50% more). Also (typical USA) no alcoholic drinks, so you can bet your bottom dollar that they'll be expensive.
| Wednesday, 11 March 2020 | Dereel | Images for 11 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Video editing: avidemux success!
|
Topic: multimedia, technology, opinion | Link here |
Spent most of today writing up my diary entries for yesterday. It paid off, though: while reading what I had collected, I noticed that there were five ports with names starting in avidemux: avidemux, avidemux-cli, avidemux-plugins, avidemux-qt4 and avidemux-qt5. I had already established that I needed avidemux and avidemux-qt5, and had installed avidemux-cli to admire the screen vomit. That leaves avidemux-plugins and avidemux-qt4. Clearly avidemux-qt4 is an older version. And avidemux-plugins? No description, of course, and at first I had thought of it as extras, but what if it's part of the base port? After all, what use is the avidemux port with no executables? So I installed it. It works! FINALLY, after two days that seemed like an eternity, I can run avidemux3_qt5! And it's not as glacially slow as kdenlive.
How could such a mess occur? What I see is:
Reading the change logs, it seems that the Qt 4 version should have been removed a year ago, so really there should only be one package. Can't they make it like that again? The whole collection is only a couple of megabytes.
|
NBN outages again
|
Topic: technology, general, opinion | Link here |
We've had 22 days with no threat of an National Broadband Network outage! What joy!
Of course, it was too good to last. The next one is on its way:
- Start date and time: Fri 20th March 2020 00:00 AEDT
- End date and time: Fri 20th March 2020 06:00 AEDT
- Window: 6.0 hours
You may experience the following interruptions during the maintenance
- 15 min
Of course, a 15 minute outage (maybe!) in the middle of the night is almost acceptable. I just hope that they won't build up courage and start their day-long outages again.
|
Richmond Hokkien mee
|
Topic: food and drink, opinion | Link here |
I've found that the best Penang Laksa paste that I can find comes from a company called “Richmond”. Well, that's what the emblem says; there's no further reference to the name on the package. It appears to come from a company in Penang called Melexport Sdn. Bhd. So I've bought a couple of other pastes from them to try out. Their curry laksa proved to be a disappointment, but I had also bought some paste for Hokkien noodles from them:
|
|||||||||||||||||||||||
Finally today I tried it out. The curry laksa may have been a disappointment, but this reached new lows. Yvonne didn't want to eat it, and I can't blame her. I'll have to spread the rest over a number of breakfasts.
On investigating the various pastes, I suspect that each comes from a different maker. This particular one didn't even have a maker's name on it, just “Product of Malaysia”. Looking at the Wikipedia article, though, it seems that it could be Penang Hokkien Mee, or at least an attempt at it.
| Thursday, 12 March 2020 | Dereel → Napoleons → Dereel | Images for 12 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
avidemux for lagoon
|
Topic: multimedia, technology, opinion | Link here |
So finally I have avidemux working on teevee. Now to put it on lagoon for Yvonne.
=== root@lagoon (/dev/pts/0) ~ 108 -> pkg install avidemux-qt5 avidemux-cli avidemux-plugins
pkg: No packages available to install matching 'avidemux-qt5' have been found in the repositories=== root@lagoon (/dev/pts/0) ~ 109 -> pkg search avidemux
avidemux-2.6.11_15 Simple GUI based video editor
avidemux-cli-2.6.11_9 Simple GUI based video editor (cli)
avidemux-plugins-2.6.11_18 Simple GUI based video editor (Plugins)
Huh? That's exactly what I installed on teevee yesterday:
=== root@teevee (/dev/pts/7) ~ 115 -> pkg search avidemux
avidemux-2.7.4_1 Simple GUI based video editor
avidemux-cli-2.7.4 Simple GUI based video editor (cli)
avidemux-plugins-2.7.4 Simple GUI based video editor (Plugins)
avidemux-qt5-2.7.4 Simple GUI based video editor (QT5 UI)
Dammit, what's wrong with this system? Does pkg update -f help here too? No.
OK, start building from source.
===> avidemux-qt5-2.7.4 depends on file: /usr/local/lib/qt5/bin/moc - not found
*** Error code 1
===> avidemux-qt5-2.7.4 depends on file: /usr/local/lib/qt5/bin/linguist - not found
*** Error code 1
===> avidemux-qt5-2.7.4 depends on file: /usr/local/lib/qt5/bin/qmake - not found
*** Error code 1
===> avidemux-qt5-2.7.4 depends on shared library: libADM6avutil.so.56 - not found
*** Error code 1
Each of those were a broken build. The first ones were relatively simple to install as a package, but what package is the last one? Look on teevee:
=== root@teevee (/dev/pts/7) ~ 118 -> locate libADM6avutil.so.56
/usr/local/lib/libADM6avutil.so.56=== root@teevee (/dev/pts/7) ~ 119 -> pkg which /usr/local/lib/libADM6avutil.so.56
/usr/local/lib/libADM6avutil.so.56 was installed by package avidemux-2.7.4_1
Oh. Now I see that the versions offered on lagoon are older than on teevee. That's what I get for not updating the ports for two whole months.
And why don't I update more frequently? Because I no longer trust the packages system. I've had so many problems over the last four months that I try to do it as seldom as possible. Clearly I'm going to have to do so now, but first I'll make a copy of the entire system in /destdir so that I can just boot back to the old system if something goes wrong.
Once this stuff used to work.
|
Trimming trees
|
Topic: general | Link here |
The people from Powercor are in the area again, trimming trees that dare to grow anywhere near power lines, like this one:
|
||||||||||||||||||||||||
The power line is visible in the large version of the image; it goes off to the left (seen from my position) of the truck. Why do they have to cut the trees on the right?
That's not so bad, though. This used to be a tree until this morning:
|
||||||||||||||||||||||||
|
New flash unit
|
Topic: photography, opinion | Link here |
My planned new flash unit arrived today, a Godox “Mini Pioneer” 300DI for the kitchen space. What does the model number mean? The 300 means 300 J (or “WS” or similar, as photographers love to misquote). The old one is a “Smart” 300SDI, clearly the same power. What other differences? The 300DI is heavier, though the same size. It's silver grey instead of brassy. It has an even less precise power output setting than the 300SDI (marked only “¼”, “½” and “1”), while the 300SDI has about 6 LED stripes with no obvious meaning; clearly they're both nowhere near as accurate as the DE400s that I have in the office, which have a strange numeric scale reading from 5.0 to 7.0—what? It seems that the number is a base 4 logarithm of the (relatively accurate) power output: 7.0 is 4 times the power of 6.0, 6.0 is 4 times the power of 5.0. Photographic suppliers really confuse me.
For the use in the kitchen, the power adjustments aren't of interest: I won't use them. Probably I'll leave them in the “full power” position, but it's convenient to have some kind of adjustment if I change my mind. All these units have remote and local triggering, model light (which I certainly won't use here). But the new one has two distinct advantages over the 300SDI: it's about 25% cheaper, and the color makes it less obtrusive (so I used it to replace the 300SDI, which I moved to a less visible location):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
So now I have the flashes set as I planned nearly a year ago. On that occasion I wanted to run the second unit at 200 J, but clearly I'll try again.
|
Another grid power fluctuation
|
Topic: Stones Road house, general | Link here |
Another grid power failure this evening at 22:53:23, again only one second.
| Friday, 13 March 2020 | Dereel | Images for 13 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Genaealogy information
|
Topic: general, opinion | Link here |
I've signed up for free accounts on a couple of genaealogy sites, include geni.com. Recently they've taken to sending me teasers related (maybe) to items I have stored. Today it was my mother's death, pointing me to the death announcement that the Herberts put in the newspaper, and wanting US $14 odd to read it.
Anything else? Yes, an inconsistency:
Audrey Eileen Lehey has a burial event listed in the future.Tip: Go to the profile of Audrey Eileen Lehey and review her burial date (3/5/2019).
Well, yes, that date is definitely wrong. She was buried on 5 March 2019, not 3 May. But that's this polite US American site changing things around so that US Americans can understand it, and everybody else can be confused.
But clearly they understood the date as 5 March. So why should it be listed in the future (especially now, where it's over a year in the past?). My guess is that I entered the information on the same day, and for Geni, in the always-yesterday USA, it looked like it was in the future.
How to fix it? Press “review”. Done!
|
Colour balance pain
|
Topic: photography, opinion | Link here |
A while back I tried a Microsoft solution for Ashampoo photo optimizer's colour balance problems: deinstall and reinstall. And it seemed to have worked. So when today I wanted to fix the photo of the new kitchen flash unit, I ran it through the optimizer (run the cursor over an image to compare it with its neighbour):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And then there's the old flash unit, taken a little over a month ago:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Those colours are completely different! Even the new “optimized” one is too green, but more to the point, the balance is different between the two. My guess is that the colour of the flash unit confused the optimizer. What I really need to do is find a profile (sorry, “preset”) for DxO PhotoLab that does the same transformations that Ashampoo tries to make.
|
The daily grid power issue
|
Topic: Stones Road house, general | Link here |
Yet another grid power failure (or at least problem) today at 13:36:53.
| Saturday, 14 March 2020 | Dereel | Images for 14 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
ALDI special buys
|
Topic: general, health, opinion | Link here |
Today ALDI had a Nokia 5.1 mobile phone on special. These things tend not to stay around very long, so off early to town to pick one up.
Timing: I estimated 25 minutes, and wanted to be on time. They open at 8:30. Estimate 25 minutes' drive, give 5 minutes grace, so leave (pass the gate) at 8:00. How long do I need to actually get in the car and out the gate? I guessed 4 minutes, which was close enough.
In fact, I arrived at 8:20, so that's only 20 minutes driving time. Things didn't look good. The car park looked as full as it ever is during opening hours, and by the time they opened the doors there must have been 70 people waiting:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Was the phone that interesting? I couldn't recall anything else in their weekly offerings that would be likely to go away that quickly. But when we went in, I headed straight for the cash registers (that's where they keep small, relatively valuable items), got the first of an estimated 20 phones, and was the first through the registers as I paid for it (separately from the other items, presumably for similar considerations).
So what were the other people looking for? Toilet paper! They had a palette of it in the middle of the aisle when I arrived, but only a few minutes later it was as empty as the shelves, even though the only allowed one package per person:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
The last photo shows the purchases of four different people on the conveyor belt, all with their regulation package of toilet paper. Why? When this nonsense started 10 days ago, it appears to have been a panic reaction. But I would have thought that by now people would have realized that it's unnecessary.
In passing, I was approached by a manager after taking the last photo: sure, take photos of the shop, but not of the personnel, please. OK, request noted. But why? That's an employee in the first photo, of course, but he/she isn't recognizable.
|
avidemux on lagoon, next attempt
|
Topic: technology, multimedia, opinion | Link here |
Yvonne off to Ballarat today for Pilates, so took the opportunity of updating her ports. After (re-)copying the entire root file system to the /destdir partition, ran pkg upgrade. And it just worked! It's nice to see something working as advertised.
Then back to install avidemux-qt5. Not found. OK, what do we have?
=== root@lagoon (/dev/pts/0) ~ 188 -> pkg search avidemux
avidemux-2.6.11_15 Simple GUI based video editor
avidemux-cli-2.6.11_9 Simple GUI based video editor (cli)
avidemux-plugins-2.6.11_18 Simple GUI based video editor (Plugins)
Dammit, that's still the old version! Here I have two up-to-date machines running FreeBSD 12-STABLE, with up-to-date packages, and they offer me different versions of the software. Somehow this whole system is too complicated, and thus fragile.
OK, nothing for it: build from source. Once again I'm reminded me of what I wrote in Porting UNIX Software over 25 years ago:
Even the complete X11R6 windowing system takes only about 4 hours on a 66 MHz Intel 486 PC.
Ah, the good old days. Our modern computers are over 100 times as fast, and it still takes days... But what I wrote next still applies:
The real time-consumers are the bugs you might encounter on the way: if you're unlucky, you can run into big trouble, and you may find yourself getting to know the package you're porting much more intimately than you wish, or even having to find and fix bugs.
The first problem was:
===> avidemux-qt5-2.7.4 depends on file: /usr/local/lib/qt5/bin/qmake - found
===> avidemux-qt5-2.7.4 depends on shared library: libADM6avutil.so.56 - not found
===> avidemux-qt5-2.7.4 depends on shared library: libADM6avutil.so.56 - not found
*** Error code 1
Where's that? Here started a sequence of comparisons with what was on teevee:
=== root@teevee (/dev/pts/7) ~ 141 -> locate libfdk-aac
/usr/local/lib/libfdk-aac.a
/usr/local/lib/libfdk-aac.so
/usr/local/lib/libfdk-aac.so.2
/usr/local/lib/libfdk-aac.so.2.0.1=== root@teevee (/dev/pts/7) ~ 142 -> pkg which /usr/local/lib/libfdk-aac.so
/usr/local/lib/libfdk-aac.so was installed by package fdk-aac-2.0.1
=== root@lagoon (/dev/pts/0) /usr/ports/multimedia/avidemux 202 -> pkg install fdk-aac
...
[1/1] Installing fdk-aac-2.0.1...
Why didn't the dependencies work? It took me a while to realize that the checks for installation don't take the system name. On installing a port, the work subdirectory contains a number of dot files:
-rw-r--r-- 1 root wheel 0 10 Mar 12:39 .build_done.fdk-aac._usr_local
-rw-r--r-- 1 root wheel 0 10 Mar 12:37 .configure_done.fdk-aac._usr_local
-rw-r--r-- 1 root wheel 0 10 Mar 12:37 .extract_done.fdk-aac._usr_local
-rw-r--r-- 1 root wheel 0 10 Mar 12:39 .install_done.fdk-aac._usr_local
-rw-r--r-- 1 root wheel 0 10 Mar 12:37 .license_done.fdk-aac._usr_local
-rw-r--r-- 1 root wheel 219 10 Mar 12:37 .license-catalog.mk
-rw-r--r-- 1 root wheel 127 10 Mar 12:37 .license-report
-rw-r--r-- 1 root wheel 0 10 Mar 12:37 .patch_done.fdk-aac._usr_local
-rw-r--r-- 1 root wheel 0 10 Mar 12:39 .stage_done.fdk-aac._usr_local
I had the ports tree mounted via NFS, so both lagoon and teevee used the same hierarchy. I had thought that the files included the name of the system, but it's only a stylized install path. So the build system sees the .install_done file (which in fact relates to teevee) and doesn't try to rebuild.
What a pain! That means that I can't use the same tree for different systems. I had to check out the entire ports tree again:
=== grog@lagoon (/dev/pts/16) ~ 3 -> svn checkout https://svn.FreeBSD.org/ports/head /usr/ports
After that was finished, continued with better results. But there were still surprised: for some reason, there was no package for lame. Looking at the Makefile confirms this, but doesn't explain it:
LICENSE_FILE_PATENTS= ${FILESDIR}/lame_patents.txt
LICENSE_PERMS_PATENTS= dist-mirror dist-sell no-pkg-mirror no-pkg-sell auto-accept
# The reason for "no-pkg" was discussed with core:
# Core feels that the risk versus any potential benefit is not in the
# project's best interest and we should continue to not package
# audio/lame.
What risk? The benefit is that one might come slightly closer to having a working packages system. OK, build and install it and continue.
Next, vapoursynth failed:
[3/5] Installing ImageMagick6-6.9.10.57,1...
pkg: ImageMagick6-6.9.10.57,1 conflicts with ImageMagick7-7.0.8.57_1 (installs files into the same place). Problematic file: /usr/local/bin/Magick++-config
I've seen this before. If I do that, I need to rebuild Emacs. No, let's configure vapoursynth (whatever that is) to build without ImageMagick.
And finally avidemux built! OK, let's install:
===> Installing for avidemux-2.7.4_1
===> Checking if avidemux is already installed
===> Registering installation for avidemux-2.7.4_1
pkg-static: Unable to access file /usr/ports/multimedia/avidemux/work/stage/usr/local/lib/libADM_coreVDPAU6.so:No such file or directory
pkg-static: Unable to access file /usr/ports/multimedia/avidemux/work/stage/usr/local/include/avidemux/2.7/ADM_coreVdpau/ADM_coreVdpau.h:No such file or directory
pkg-static: Unable to access file /usr/ports/multimedia/avidemux/work/stage/usr/local/include/avidemux/2.7/ADM_coreVdpau/ADM_coreVdpauInternal.h:No such file or directory
*** Error code 74
What's that? Clearly it's related to VDPAU, an old, worn-out magic rendering algorithm. Reconfigure, make clean, all and install, and it worked!
===== Sat 14 Mar 2020 18:38:32 AEDT on lagoon.lemis.com: Make install
Installing avidemux-2.7.4_1...
2.12 real 0.69 user 0.53 sys
One down, three to go. But by then it was evening.
|
New mobile phone
|
Topic: technology, general, opinion | Link here |
So now I have a new mobile phone, a Nokia 5.1 Plus. Why? I hate mobile phones. But maybe the issue was that my Nokia 3 was just too gutless—only about 100 times the processing power of the Intel 486 that I mentioned in Porting UNIX Software above. That would suggest that it could build the complete X system in 15 seconds.
The advantage of the 5.1 Plus is that it's sold by ALDI, so I have two months to play with it, and if I don't like it, I can just return it. Fortunately it doesn't look like this image from Wikipedia, the stuff of my nightmares (keyboards with missing keys):
OK, set up. How about that, a USB-C connector, and the battery was already 75% charged, both a significant difference from the Nokia 3. How do I get it set up? For once, something easy: a program that copies data from the old phone to the new, using Bluetooth. And, of course, the obligatory software update.
I was busy with lagoon, so I didn't have much time to pay attention to the details. At the end I discovered that most of the stuff had been copied, but not my ring tone, and not the all-important WiFi File Transfer, which I need, amongst other things, to load the ring tone.
Why wasn't it installed? “This app may not be completely compatible with your phone”. Oh. Installed it anyway and discovered that it didn't offer http connectivity, only ftp. Is that a problem? I need to think of the implications. But surely, surely, there is software available for that most basic of network functionality, copying data from one system to another?
I had previously used AirDroid, but now it seems to require an account! What kind of brain damage is that? Did a little further searching and found sendanywhere, which claimed to be able to do that. But when I fired it up, it just asked me to “Enter key”.
Why have people confused a simple operation? Somewhere there must be documentation, but pressing “Help” produces even more confusion:
|
|
|||||||||
OK, potentially “Sending files directly t o a device” is what I want. But then I get:
|
|
|||||||||
And that's where things become clear: I want to select from a sane interface, which means from the other end. That makes ftp look much more attractive. Mañana.
|
The 70th birthday party
|
Topic: general, opinion | Link here |
Garry Mariott and Diane from next door celebrated their joint 70th birthday this evening, and of course we were invited. Are celebrations of that nature appropriate in this time of COVID 19, especially since many of the guests will have come from all around Australia? I did some serious thinking, but ultimately we decided that we didn't want to upset Garry and Diane, so off to the Dereel Hall at 17:00:
|
||||||||||||||||||||||||
I had never noticed it before, but the hall is noisy! Diane in particular introduced us to a number of people, but I could barely hear what she was saying. Shake hands? Not in the current health situation. But it seems that I was the only person who thought so, and though some made fun of the “touch elbow” greeting, I fear they were left with the feeling that I was even stranger than I am. When I half inhaled a walnut and burst out in a fit of coughing, nobody appeared to be concerned.
We weren't really looking forward to the affair—somehow big gatherings are not our kind of thing. But the noise really made it unbearable, and after an hour Yvonne developed a headache, and we had to leave, even before dinner. I'm sorry for Garry and Diane's sake that we had to leave. Clearly we're even less sociable than I thought.
|
New Hibiscus
|
Topic: gardening, opinion | Link here |
To get over my fit of coughing at the Hall, outside for a while, where I found this bush:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
What is it? It looks like some kind of Hibiscus.
| Sunday, 15 March 2020 | Dereel | Images for 15 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The daily ports pain
|
Topic: technology, multimedia, opinion | Link here |
By yesterday evening I had managed to install the base avidemux package. Only three more to go. Next I tried avidemux-qt5. It went for a while and then:
cd /usr/ports/multimedia/avidemux-qt5/work/.build/commonQt4/ADM_render && /usr/local/bin/g++9 -DADM_UI_TYPE_BUILD=ADM_UI_QT4 -DADM_render6_QT5_EXPORTS -DQT_CORE_LIB -DQT_GUI_LIB -DQT_NETWORK_LIB -DQT_NO_DEBUG -DQT_WIDGETS_LIB -D_FILE_OFFSET_BITS=64 -D_LARGE_FILES -I/usr/ports/multimedia/avidemux-qt5/work/.build/config/qt5 -I/usr/local/include/avidemux/2.7 -I/usr/local/include/avidemux/2.7/ADM_core -I/usr/local/include/avidemux/2.7/ADM_coreAudioFilter -I/usr/local/include/avidemux/2.7/ADM_coreAudioParser -I/usr/local/include/avidemux/2.7/ADM_coreAudio -I/usr/local/include/avidemux/2.7/ADM_coreAudioDevice -I/usr/local/include/avidemux/2.7/ADM_coreAudioEncoder -I/usr/local/include/avidemux/2.7/ADM_coreDemuxer -I/usr/local/include/avidemux/2.7/ADM_coreDemuxer/unix -I/usr/local/include/avidemux/2.7/ADM_coreDemuxerMpeg -I/usr/local/include/avidemux/2.7/ADM_coreImage -I/usr/local/include/avidemux/2.7/ADM_coreMuxer -I/usr/local/include/avidemux/2.7/ADM_coreScript -I/usr/local/include/avidemux/2.7/ADM_coreUI -I/usr/local/include/avidemux/2.7/ADM_coreUtils -I/usr/local/include/avidemux/2.7/ADM_coreVideoEncoder -I/usr/local/include/avidemux/2.7/ADM_coreVideoFilter -I/usr/local/include/avidemux/2.7/ADM_coreVideoCodec -I/usr/local/include/avidemux/2.7/ADM_coreImageLoader -I/usr/local/include/avidemux/2.7/ADM_coreJobs -I/usr/local/include/avidemux/2.7/ADM_coreSocket -I/usr/local/include/avidemux/2.7/ADM_coreAudioCodec -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/ADM_muxerGate/include -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/../common -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/../common/ADM_audioFilter/include -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/../common/ADM_commonUI -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/../common/ADM_videoFilter2/include -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/../common/ADM_videoEncoder/include -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/../common/ADM_editor/include -I/usr/local/include/avidemux/2.7/ffmpeg -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/../common/ADM_audioCodec/include -I/usr/local/include/QtWidgets -I/usr/local/include/qt5/QtNetwork -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/ADM_UIs/include -I/usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/ADM_openGL/include -I/avidemux/qt4/ADM_openGL/include -isystem /usr/local/include/qt5 -isystem /usr/local/include/qt5/QtCore -isystem /usr/local/lib/qt5/mkspecs/freebsd-clang -isystem /usr/local/include/qt5/QtWidgets -isystem /usr/local/include/qt5/QtGui -O2 -pipe -fno-omit-frame-pointer -DLIBICONV_PLUG -fstack-protector-strong -Wl,-rpath=/usr/local/lib/gcc9 -I/usr/local/include -DLIBICONV_PLUG -Wl,-rpath=/usr/local/lib/gcc9 -fmessage-length=0 -fvisibility=hidden -fvisibility-inlines-hidden -O2 -pipe -fno-omit-frame-pointer -DLIBICONV_PLUG -fstack-protector-strong -Wl,-rpath=/usr/local/lib/gcc9 -I/usr/local/include -DLIBICONV_PLUG -Wl,-rpath=/usr/local/lib/gcc9 -fPIC -std=c++11 -Wattributes -fPIC -o CMakeFiles/ADM_render6_QT5.dir/GUI_xvRender.cpp.o -c /usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/common/ADM_render/GUI_xvRender.cpp
In file included from /usr/ports/multimedia/avidemux-qt5/work/avidemux_2.7.4/avidemux/qt4/common/ADM_render/GUI_xvRender.cpp:26:
/usr/local/include/X11/Xlib.h:44:10: fatal error: X11/X.h: No such file or directory
44 | #include <X11/X.h>
| ^~~~~~~~~
compilation terminated.
gmake[4]: *** [commonQt4/ADM_render/CMakeFiles/ADM_render6_QT5.dir/build.make:105: commonQt4/ADM_render/CMakeFiles/ADM_render6_QT5.dir/GUI_xvRender.cpp.o] Error 1
gmake[4]: Leaving directory '/usr/ports/multimedia/avidemux-qt5/work/.build'
gmake[3]: *** [CMakeFiles/Makefile2:804: commonQt4/ADM_render/CMakeFiles/ADM_render6_QT5.dir/all] Error 2
gmake[3]: Leaving directory '/usr/ports/multimedia/avidemux-qt5/work/.build'
gmake[2]: *** [Makefile:133: all] Error 2
gmake[2]: Leaving directory '/usr/ports/multimedia/avidemux-qt5/work/.build'
*** Error code 1
Huh? X.h is in /usr/local/include/X11/, not quite coincidentally the same directory as the calling header file. Isn't it? Checked. Yes, it's there:
=== root@lagoon (/dev/pts/0) /usr/ports/multimedia/avidemux-qt5 293 -> locate X11/X.h
...
/usr/local/include/X11/X.h
And surely /usr/local/include is in the include list for the compiler invocation. Isn't it? Isn't it fun to read through invocations like the one that started this compilation? But yes, there are not one, but two -I/usr/local/include in that 3344 character long invocation.
Why is this so complicated? I have the ability (but most definitely not the desire) to find out what went wrong here, but how many people do? This should just work.
Again, I didn't have much time to look at the problem. It's been nearly a week! And this is a package that should just install with a simple, single pkg install avidemux.
|
Another grid outage
|
Topic: Stones Road house, general | Link here |
Another grid power failure today at 10:24:52, this time long enough (I'll call it 3 seconds) to count as a real outage. I even got a complaint from my office UPS, suggesting that the inverter isn't able to completely smooth out these issues.
|
Tagine aux abricots et amandes
|
Topic: food and drink, opinion | Link here |
It's been a long time since we last ate a tagine (or is that tajine? I still can't decide). Today, rather in a hurry, I made one, changing recipes in mid-cooking. Started with my fake tagine before I realized that it didn't include the almonds that we had specially bought for the purpose. They're in the tagine d'agneau aux abricots et amandes. But the recipes are similar enough, so I changed after starting to cook the meat.
This was the first time we have cooked a tagine since installing the induction cooker. Our tagine is made of aluminium, so it doesn't work on the cooker. And I still haven't bought a heating pad. But surely I have something ferromagnetic to put between tagine and cooktop? Yes, of course, an oven pan:
|
|||||||||||||||||||||||
To my surprise, it was completely flat and lay nicely on the cooktop—until it got hot. But it still heated enough to simmer the contents nicely, something that's very difficult on the gas cooker. But afterwards there was a distinct mark on the cooktop:
|
|||||||||||||||||||||||
Why? It proves to be very difficult to clean the cooktop anyway, but this is significantly more than any previous marking. Still, the warping means that it's not a general solution for heating non-induction cookware on the induction cooker.
Here are the results, after we had served the first servings:
|
|||||||||||||||||||||||
|
Advantages of unmown lawn
|
Topic: animals, gardening | Link here |
We have many more birds around the house lately. Murray, a person I met at the party last night, tells me that a lot is due to habitat loss in Gippsland as the result of the recent bushfires. That would probably explain the galahs, but not these crimson rosellas:
|
|||||||||||||||||||||||
|
|
|||||||||
My guess is that their presence is due to the unseasonally green grass.
|
Image framing errors
|
Topic: photography, opinion | Link here |
A couple of photos I took recently have had bad framing:
|
||||||||||||||||||||||||
|
|||||||||||||||||||||||
Why? In each case I had plenty of time to compose. My guess is that the data displayed at the bottom of the viewfinder confused my judgement. Time to turn it off by default; it still shows exposure parameters when I half-press the shutter button.
And no, the white balance is correct in that second image. It just happens to be all rather orange.
| Monday, 16 March 2020 | Dereel | Images for 16 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
avidemux debugging
|
Topic: technology, multimedia, opinion | Link here |
So why did my avidemux build fail yesterday? I really need to find out. I had established that it claimed not to be able to find /usr/local/include/X11/X.h, but locate told me that it was there.
locate gives a historical view. Was the file really there?
=== root@lagoon (/dev/pts/0) /usr/ports/multimedia/avidemux-qt5 293 -> locate X11/X.h
...
/usr/local/include/X11/X.h=== root@lagoon (/dev/pts/0) /usr/ports/multimedia/avidemux-qt5 294 -> l /usr/local/include/X11/X.h
ls: /usr/local/include/X11/X.h: No such file or directory
Gone! Clearly after the locate database was built. But when, exactly?
=== root@lagoon (/dev/pts/0) /usr/ports/multimedia/avidemux-qt5 295 -> l -dT /usr/local/include/X11/
drwxr-xr-x 12 root wheel 2048 14 Mar 12:27:38 2020 /usr/local/include/X11
That looks like the time I was upgrading the ports. What else happened round that time?
=== root@lagoon (/dev/pts/0) /usr/ports/multimedia/avidemux-qt5 297 -> find /usr/local/ -mtime -3 -mtime +2 | xargs ls -ldtr | less
drwxr-xr-x 2 root wheel 1024 14 Mar 12:27:35 2020 /usr/local/include/X11/fonts
drwxr-xr-x 2 root wheel 512 14 Mar 12:27:35 2020 /usr/local/share/licenses/libxml2-2.9.10
drwxr-xr-x 2 root wheel 3072 14 Mar 12:27:35 2020 /usr/local/include/libxml2/libxml
drwxr-xr-x 2 root wheel 512 14 Mar 12:27:35 2020 /usr/local/lib/cmake/libxml2
drwxr-xr-x 2 root wheel 512 14 Mar 12:27:36 2020 /usr/local/share/licenses/readline-8.0.1
drwxr-xr-x 2 root wheel 512 14 Mar 12:27:36 2020 /usr/local/include/readline
drwxr-xr-x 2 root wheel 1024 14 Mar 12:27:36 2020 /usr/local/share/readline
drwxr-xr-x 2 root wheel 512 14 Mar 12:27:36 2020 /usr/local/share/doc/readline
drwxr-xr-x 2 root wheel 512 14 Mar 12:27:38 2020 /usr/local/share/licenses/libX11-1.6.9,1
drwxr-xr-x 12 root wheel 2048 14 Mar 12:27:38 2020 /usr/local/include/X11
drwxr-xr-x 12 root wheel 512 14 Mar 12:27:38 2020 /usr/local/lib/X11
drwxr-xr-x 2 root wheel 512 14 Mar 12:27:38 2020 /usr/local/lib/X11/locale/C
drwxr-xr-x 2 root wheel 512 14 Mar 12:27:38 2020 /usr/local/lib/X11/locale/am_ET.UTF-8
That was the very start of the list, the first files modified on that day. How did they compare with the output of pkg upgrade?
Installed packages to be UPGRADED:
...
xorgproto: 2019.1 -> 2019.2
Proceed with this action? [y/N]:
Installing xorgproto-2019.2...
Extracting xorgproto-2019.2: .......... done
[1/103] Deinstalling xorgproto-2019.1...
[1/103] Deleting files for xorgproto-2019.1: .......... done
Am I correct in assuming that these files were in xorgproto? On teevee:
=== grog@teevee (/dev/pts/8) ~ 309 -> pkg which /usr/local/include/X11/X.h
/usr/local/include/X11/X.h was installed by package xorgproto-2019.2
Yes. And was it really in this specific package on lagoon?
=== grog@lagoon (/dev/pts/16) ~ 9 -> tar tvf /var/cache/pkg/xorgproto-2019.2.txz
-rw-r--r-- 0 root wheel 505 1 Jan 1970 +COMPACT_MANIFEST
-rw-r--r-- 0 root wheel 19382 1 Jan 1970 +MANIFEST
...
-rw-r--r-- 0 root wheel 2815 3 Mar 12:10 /usr/local/include/X11/DECkeysym.h
-rw-r--r-- 0 root wheel 6046 3 Mar 12:10 /usr/local/include/X11/HPkeysym.h
-rw-r--r-- 0 root wheel 4022 3 Mar 12:10 /usr/local/include/X11/Sunkeysym.h
-rw-r--r-- 0 root wheel 20137 3 Mar 12:10 /usr/local/include/X11/X.h
-rw-r--r-- 0 root wheel 13548 3 Mar 12:10 /usr/local/include/X11/XF86keysym.h
-
Yes. And the other files mentioned there were also missing. It seems that the subsequent removal of the files from xorgproto-2019.1 removed files from xorgproto-2019.2. Does the lack of a sequence number for xorgproto-2019.2 have any relationship? But clearly this is serious breakage.
So what do I do now? Reinstall the package? There's no reinstall subcommand. How about removing it and reinstalling?
=== root@lagoon (/dev/pts/0) /var/db/pkg 310 -> pkg delete xorgproto
Updating database digests format: 100%
Checking integrity... done (0 conflicting)
Deinstallation has been requested for the following 265 packages (of 0 packages in the universe):
Installed packages to be REMOVED:
...
The operation will free 2 GiB.
Proceed with deinstalling packages? [y/N]:
Clearly that way madness lies.
But this package is a tarball, and it contains sane file names. Why not just extract it? That's what I did, and it worked.
Next, build avidemux-qt5. More problems! Dammit, that's another day where I've used all my spare time trying to build something that should Just Install. I give up.
But I have a package on teevee, and it's the same version of FreeBSD as lagoon. Why not just install it?
=== root@lagoon (/dev/pts/6) /eureka/home/grog 1 -> pkg install /src/FreeBSD/svn/ports/multimedia/avidemux-qt5/work/pkg/avidemux-qt5-2.7.4.txz
Cannot solve problem using SAT solver, trying another plan
Checking integrity... done (2 conflicting)
- ImageMagick6-6.9.10.57,1 conflicts with ImageMagick7-7.0.8.57_1 on /usr/local/bin/Magick++-config
- ImageMagick6-6.9.10.57,1 conflicts with ImageMagick7-7.0.8.57_1 on /usr/local/bin/Magick++-config
Cannot solve problem using SAT solver, trying another plan
Checking integrity... done (0 conflicting)
The following 1 package(s) will be affected (of 0 checked):
Installed packages to be REMOVED:
emacs-26.3_1,3
Number of packages to be removed: 1
The operation will free 139 MiB.
Proceed with this action? [y/N]:
OK, we just need to remove Emacs. You didn't need that, did you? What, really?
OK, let's build a version on teevee without ImageMagick support. That ran fine. But where's the package? It should have been in /usr/ports/multimedia/åvidemux-qt5/, but there's nothing there. It turns out that somebody has thought it better to put it in /usr/ports/multimedia/åvidemux-qt5/work/pkg. That makes sense, but it's yet another instance of changes that make life harder for people who once knew their way around the system.
OK, install on lagoon:
=== root@lagoon (/dev/pts/6) / 3 -> tar xvf /src/FreeBSD/svn/ports/multimedia/avidemux-qt5/work/pkg/avidemux-qt5-2.7.4.txz /usr/local
It works, sort of. But where's the /usr/local/lib/libADM_coreVDPAU6.so?
=== grog@lagoon (/dev/pts/10) ~ 3 -> avidemux3_qt5
ld-elf.so.1: Shared object "libADM_coreVDPAU6.so" not found, required by "avidemux3_qt5"
Oops, built without that. OK, build a new avidemux_qt5 without Vdpau support, install. Where's the /usr/local/lib/libADM_coreVDPAU6.so? It seems that even when you disable Vdpau, you end up with an executable that refers to it.
Were there any others?
=== grog@lagoon (/dev/pts/10) ~ 5 -> ldd /usr/local/bin/avidemux3_qt5 | grep "not found"
libADM_coreVDPAU6.so => not found (0)
libADM_coreVDPAU6.so => not found (0)
libADM_coreVDPAU6.so => not found (0)
That's something, I suppose.
Dammit, we have a /usr/local/lib/libADM_coreVDPAU6.so on teevee. Move it across to lagoon and try again. It runs! Well, sort of:
|
|
|||||||||
Grrr! But it seems that it was, in fact, OK. Cancel the message and I was able to run the thing. That, of course, is strange enough. But at least I got some output.
I ventured into this particular ports pain a week ago, and it's still not really finished. At the very least I have to document it. How many problem reports do I need to write? I'd guess about 10.
|
The first chili of summer^Wautumn
|
Topic: gardening | Link here |
Things have been growing in the garden, though slowly. We still have harvested almost no tomatoes. But my tiny Thai chili plant is full of (unripe) chilies:
|
|||||||||||||||||||||||
Looking more carefully, though, there are two ripe ones hidden in the foliage:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I fear that they could be in the “lift skull cap” category.
| Tuesday, 17 March 2020 | Dereel | Images for 17 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Installing a gate opener
|
Topic: Stones Road house, general | Link here |
CJ Ellis along this morning to take a look at our front gate:
|
|||||||||||||||||||||||
To go through it with a car requires stopping, getting out, opening the gate, returning to the car, driving 10 metres through the gate, stopping, getting out, closing the gate, returning to the car and driving on: a pain. When we were rich we installed a remote controlled electric gate in Wantadilla, and we can probably afford one again. Things have changed since then; now there are photovoltaic powered openers for conventional gates, and they're even quite cheap, as I investigated last month. But we need to install it, and that's what CJ came to look at. We're planning a second personal access gate on the right, requiring reworking this part:
|
|||||||||||||||||||||||
Nothing seems particularly complicated. Clearly we'll have to get the parts, but that shouldn't be a big issue.
|
Using avidemux
|
Topic: technology, multimedia, opinion | Link here |
So now, after only a week, I have a slightly castrated version of avidemux running on lagoon. Tried to show Yvonne how to use it, but failed. Although I had used it before, and though it's relatively sane, I forgot what the symbols mean. On a second attempt I was more successful. But then she tried by herself, trimmed video clips, save them and... they weren't there!
After some investigation, discovered that avidemux is just as bad as all this GUI software that has no concept of directory trees. Or maybe almost as bad. You can specify at least one file on the command line, which produces many and varied illegible texts:
|
|||||||||||||||||||||||
And the first time round it happily provided the correct directory for saving the output file (/Photos/yvonne/20200317, the same one in which the input file was located). But then she processed some files in /Photos/yvonne/20200216. Where did avidemux save the output? /Photos/yvonne/20200317, of course. That's what you did last time.
To be fair, I can't single out avidemux here. All GUI software is like this, or worse. I think that if you want to tell a GUI software developer how to get anywhere in a building, you first take him to the main entrance and then give directions from there, one step at a time. In this case, I already have a workaround: give the package its own directory, never to be changed, and use scripts to move things in and out to maintain your sanity.
OK, look for documentation. Found this tutorial, not easy to follow, but which at least showed me that there should be a volume slider on the window. Not there on mine. In fact, no sound at all. Bug in the port?
The other current issue with avidemux is the display size. It seems that you can only display the clip in one of 4 sizes, selected with the keys 1 to 4. No way to use the space available. Maybe it's there, but hidden in the apparently non-existent documentation.
So: do I try to fix it, or do I look for a better editor? There's small choice in rotten apples.
| Wednesday, 18 March 2020 | Dereel | Images for 18 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Cockatoo storm
|
Topic: animals | Link here |
Another mass arrival of cockatoos today, accompanied by a few galahs and what appear to be long-billed corellas:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's not the first time, but I'm always surprised by how the different kinds of birds flock together.
|
Taming chilis
|
Topic: food and drink, opinion | Link here |
Monday's chili was as explosive as I had feared. Yesterday I tried a couple of thin slices off the pointy end (less hot), but it was too much for me. Today I tried frying the rest, invoking the ire of Yvonne, who couldn't breathe the air. But what was left, though hot, was perfectly edible.
|
Life in COVID-19 lockdown
|
Topic: health, general, opinion | Link here |
So the government has finally come to the decision to “lock down” the population, restricting movement and gatherings of more than 500 (I think) people. The recommendations: stay at home, don't meet with people, work from home if possible.
I've been doing that for decades! And indeed, nothing seems to have changed, at least not directly. But today Yvonne went into town shopping. If my experience on Saturday suggested extreme panic reactions, today was much worse. It seems that the shops were out of food as well, not just toilet paper. Well, at least durable food like spaghetti (which I had wanted), rice, flour and beans. Even the kitchenware that I had wanted to try had sold out within hours; usually they would have them for at least a week. And I suspect that that's what Marrickville Coles Totally Out Of Typewriters And Vintage Suitcases, As Panic Buying Continues is all about. I wonder how long this will last. At some point, everybody will be saturated with toilet paper and flour.
|
Groggy's ports mistakes
|
Topic: technology, opinion | Link here |
I made a mistake on IRC this morning:
* groggyhimself grumbles about ports.
Ah, but that's my fault. I'm not just using ports, I'm mixing them with packages. Naughty boy!
Is that wrong? I claim not, but others point to configuration differences between my system and the system on which the packages were built. Yes, the dependencies could be different. But that's exactly the point of the dependency database. I'm pretty sure that I'm doing it by the book, with some explicit exceptions, like extracting the package directly without passing "go" and registering with the database.
Then people suggested that part of my problem was that I had an old installation. Well, sorry, that's not a reason: the whole point of the package system is to keep an installation up to date. If an old installation suffers, it's because the package system isn't doing its job. But the suggestion is also irrelevant: I reinstalled both systems only a couple of months ago (and have been grumbling ever since).
Still, how about an audit? There's a command pkg audit. Oh, that just checks for security issues. How about pkg check? Yes, that looks better, and with some redirection (the man page for pkg just has a one-liner; you need the man page for a non-existent pkg-check, which doesn't explain all options in its list), it seems that pkg check -as will do the check.
Not just that: it tries to fix what it finds. In fact, it's just the ImageMagick conflict that has been giving me such pain. And then there's pkg check -ad, which found myriads of conflicts in the maze of twisty little python ports (why?), but none in the ones I had been looking at.
About the only useful information that my RTFM brought me was that pkg reinstall is spelt pkg upgrade -f. And that works.
But these misconceptions about ports worry me. They tend to confirm my suspicions that people don't really use the packages system.
|
Yet another video editor
|
Topic: technology, multimedia, opinion | Link here |
Finally I have avidemux working, sort of. But it has some irritating habits, and it's clear that it's relatively limited. What else is there? I had already tried kdenlive, running into problems with both performance and interface. Clearly interface is something that I need to get used to, but performance is an issue that might remain.
So today Callum Gibson suggested openshot, something that I had been thinking of anyway. To emphasize the suggestion, he pointed to this page, which interestingly enough listed the four best packages (in order) as openshot, kdenlive, shotcut and avidemux. So another reason to try it.
Install with pkg? Almost:
Installed packages to be REMOVED:
emacs-26.3_1,3
Yes, that's the ImageMagick issue again. Dammit, OK, remove Emacs, see if I care. Built Emacs again from ports without ImageMagick support, and all was well.
Lots of fun trying to get openshot to work. Of course, the package doesn't tell you how to run it:
=== grog@lagoon (/dev/pts/15) ~ 23 -> openshot
bash: openshot: command not found
Fool! It's openshot-qt. Callum Gibson tells me that this would show up automatically on a GNOME or KDE desktop, but I have my doubts; integration doesn't seem to be the ports collection's strong suit. And even if it is the case, that's not an excuse. Ports shouldn't be tied to “desktops”.
To its credit, the (apparently obligatory) vomit on the xterm was almost all uncoloured, and the interface was almost bearable. But why the 5 tracks by default, sorted the wrong way round, of which (even in the tutorials) only 4 and 5 are visible on the default screen?
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
To display the initial tracks, you need to scroll down. What kind of mental confusion can have given rise to that?
One of the things that occurred to me and which avidemux can't do is picture-in-picture. Callum doubted that openshot could do it, while I was sure that it could. And sure enough, it can, as this video shows.
But it only half worked for me. And the program hung after a while. After considering various options, including the network, noticed the CPU time usage:
PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
30131 grog 48 41 0 4953M 2694M umtxpi 1 5:11 397.43% python3.7
It was maxing out all four cores! I went to lagoon and got it to work better, but even when doing nothing at all, it seems to use 25% of a CPU:
PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
60936 grog 25 20 0 1478M 227M select 3 2:09 26.78% python3.7
Why does this happen? It can't just be processing, because the program shouldn't be doing anything at all when it's actually using 25%.
I came across one potential issue. Seen on lagoon's log:
+NVRM: API mismatch: the client has the version 440.31, but
+NVRM: this kernel module has the version 390.87. Please
+NVRM: make sure that this kernel module and all NVIDIA driver
+NVRM: components have the same version.
Another result of a bad package update. pkg updates the nvidia driver to the wrong version:
=== root@lagoon (/dev/pts/0) /usr/ports/multimedia/avidemux-qt5 337 -> pkg info|grep nvidia
nvidia-driver-440.31 NVidia graphics card binary drivers for hardware OpenGL rendering
nvidia-settings-440.26 Display Control Panel for X NVidia driver
nvidia-xconfig-440.26 Tool to manipulate X configuration files for the NVidia driver=== root@lagoon (/dev/pts/0) /usr/ports/multimedia/avidemux-qt5 338 -> locate nvidia-driver
nothing useful
I fixed the problem on teevee a while back, but forgot to do so on lagoon. I'll have to do that next time Yvonne isn't there. But it can't be the cause of the performance problems, because it happens on teevee too.
|
Rubbish collection in Australia
|
Topic: general | Link here |
I've been used to plastic rubbish bins for over 50 years, but it's only in Australia that I've seen things like this:
|
|||||||||||||||||||||||
Why? The system originated in Germany, but when you look for photos of German rubbish collection, you end up with something like this (taken without permission from Wochenblatt and Volksfreund):
The rubbish bins look pretty much the same, modulo colour. But there are people there handling them (and taking them to and from the machine at the back of the truck). In Australia, it's a one-man operation (the driver), and the bin is handled automatically. And this automation, it seems, is also designed to return the bin as shown in my photo.
I wonder why they don't do this in Germany. I haven't found anything online that suggests that it does.
|
Android: file access and ringtones
|
Topic: technology, music, opinion | Link here |
For some reason, my new Nokia 5.1 Plus is only partially compatible with WiFi File Transfer: it doesn't support HTTP connections. On the old Nokia 3 I could connect with a web browser like this:
|
|||||||||||||||||||||||
But on the new one (called flachmann now that the old tablet of that name is no longer usable), I need to use FTP:
|
|
|||||||||
Is that a problem? After using it, I don't think so. In fact, it can make things easier. How do I set my custom ring tone? The first obvious thing to do is to upload it to the phone. Where is it? A quick search showed that it's /home/grog/public_html/Day/20190321/O-Freunde.mp3, and the list of the top level directories on the phone shows an empty directory /Ringtones. OK,
=== grog@eureka (/dev/pts/28) ~ 29 -> ftp ftp://flachmann.lemis.com:2121/Ringtones/
Connected to flachmann.lemis.com.
220 SwiFTP 3.0 ready
331 Send password
230 Access granted
Remote system type is UNIX.
Using binary mode to transfer files.
200 Binary type set
250 CWD successful
ftp> put /home/grog/public_html/Day/20190321/O-Freunde.mp3
local: /home/grog/public_html/Day/20190321/O-Freunde.mp3 remote: /home/grog/public_html/Day/20190321/O-Freunde.mp3
227 Entering Passive Mode (192,109,197,247,152,110).
451 Couldn't open file "/home/grog/public_html/Day/20190321/O-Freunde.mp3" aka "/storage/emulated/0/home/grog/public_html/Day/20190321/O-Freunde.mp3" for writing
Oops, ftp is too stupid to remove the directory part of the path name. OK, try again.
=== grog@eureka (/dev/pts/28) ~ 30 -> cd ~/public_html/Day/20190321/=== grog@eureka (/dev/pts/28) ~/public_html/Day/20190321 31 -> l
total 1
-rw-r--r-- 1 grog lemis 795,036 31 Mar 2019 O-Freunde.mp3=== grog@eureka (/dev/pts/28) ~/public_html/Day/20190321 33 -> ftp ftp://flachmann.lemis.com:2121/Ringtones/
Connected to flachmann.lemis.com.
...
ftp> put O-Freunde.mp3
local: O-Freunde.mp3 remote: O-Freunde.mp3
227 Entering Passive Mode (192,109,197,247,136,247).
150 Data socket ready
100% |************************************************************************************************************************************************| 776 KiB 1.37 GiB/s 00:00 ETA
226 Transmission complete
795036 bytes sent in 00:01 (709.39 KiB/s)
That was painless enough, though it's interesting to look at the transfer speeds that ftp claims.
Next, to the Settings/Sound menu on the phone. Select ring tone. Yes, it's there, and I can set it. That was easier than I had feared, and I think I'm making friends with the ftp interface.
Now to find the version of Wachet auf that I have in my head. Somehow it's not the same setting as BWV 140. But clearly it's a more appropriate ring tone.
| Thursday, 19 March 2020 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another NBN outage notification
|
Topic: technology, general, opinion | Link here |
Recently I had over two weeks with no looming National Broadband Network outages, but last week another came—“only” 15 minutes, and in the middle of the night. I can live with that.
And then today another one came:
NBNCo has let us know that they are planning network maintenance in your area, and that your service at <strong>29 STONES RD, DEREEL VIC</strong> will be affected.<br>
The details are:
- Start date and time: Fri 20th March 2020 00:00 AEDT
- End date and time: Fri 20th March 2020 06:00 AEDT
- Window: 6.0 hours
You may experience the following interruptions during the maintenance
- 90 min
Not quite as acceptable, but in the middle of the night, so again, I can live with it. But one thing seemed strange, so I checked. Yes, the same window: Friday morning, 0:00 to 6:00. Does this mean two outages, one of 15 minutes and one of 90 minutes? It's tomorrow, so we'll find out soon.
|
Emacs strangeness
|
Topic: technology, opinion | Link here |
Seen in my calendar this morning:
Mar 19 Greg installs BSD/386 0.3.1, 1992
28 years! But that's nothing compared to an editor that I installed some time in 1980, MINCE, a clone of Emacs for the Z-80 microprocessor. And I've been using some form of Emacs ever since (GNU Emacs since 1989).
But lately I've been having surprises. I sometimes get areas of the screen marked with a yellow background. Why? What are they? I've done some searching, and most answers on the web seem to reduce to RTFM. Clearly I've pressed some inappropriate key combination, but which? There's view-lossage to help, but so far it has evaded me.
But now I got one that I haven't seen before. I typed /2 to enter, well, /2. But I got ½. And later I wanted to enter "under", but the first two keystrokes produced ü.
Did I accidentally restart Emacs? I almost never do that, and I hadn't now:
USER PID %CPU %MEM VSZ RSS TT STAT STARTED TIME COMMAND
grog 5045 0.0 0.4 1465144 149188 v0 S 11Sep19 124:48.46 /usr/local/bin/emacs -display eureka:0.1 -geometry 100x50-53+0 -font 6x13 (emacs-24.5)/
OK, what's special about /? I noted that after pressing it I got a vestigial underline following it. Press space and the underline went away, and the space wasn't inserted. OK, c-h k /. It said:
/ (translated from / /) runs the command sgml-slash, which is an
interactive compiled Lisp function in `sgml-mode.el'.
It is bound to /.
OK, so somehow I seem to have set sgml-mode. How? Too much lossage. Discard the buffer and start again, but somehow I feel I have missed something important.
| Friday, 20 March 2020 | Dereel | Images for 20 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
NBN breaks promise
|
Topic: technology, general | Link here |
Between 0:00 and 6:00 this morning the National Broadband Network had promised not one, but two outages. And there were none! That makes seven weeks with no planned outages, and also none in the immediate future. That's something of a record.
|
Moths
|
Topic: animals | Link here |
Seen on the fly screen outside the lounge room this morning:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly a kind of moth, but what? There were about 5 of them on the screen.
|
Garden flowers in early autumn
|
Topic: gardening, opinion | Link here |
It's the March equinox already! Time for the monthly garden flower photos. And summer barely seems to have started.
On the other hand, we've had plenty of rain, and in particular the lawn shows it—exactly when we had sold our ride-on mower and entrusted the lawn mowing to our gardener Mick Solly, who then got sick and whose own mower then failed. So we have the unusual situation of lots of long grass at the beginning of autumn. That's most obvious round our Box elder:
|
|||||||||||||||||||||||
In fact, the shoot is doing relatively well; at least it doesn't think it's autumn yet:
|
|||||||||||||||||||||||
Other trees in the front garden are also doing better than in previous years. The Paulownia kawakamii is not only flourishing, it has also produced a sucker:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
And the Schinus molle (“pepper tree”) that we transplanted a couple of months ago is looking better all the time. Here three months ago and today:
|
||||||||||||||||||||||||
|
|||||||||||||||||||||||
The Corymbia ficifolia (red-flowering gum) is also still flowering, though not very much:
|
|||||||||||||||||||||||
Normally they flower profusely in the summer, and that's all. Assuming that it survives the winter, that's what I'd expect of this one too.
Other plants that are doing better than previously are the Clematis. Here the Edo Murasaki:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's not spectacular for a clematis, but it's the best I've had in the past 2½ years. And the General Sikorski is flowering:
|
|||||||||||||||||||||||
But apart from that flower, there's a complete new shoot to the left that has grown this year, and it has (barely recognizable) buds on it too:
|
|||||||||||||||||||||||
And the Fuchsia triphylla in the shade area, which had looked quite sad before, is now a lot better:
|
|||||||||||||||||||||||
The Hibiscus rosa-sinensis “Uncle Max” that I planted in the north bed is now looking a lot better:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That part of the garden is particularly crowded now, at least partially because of the tomatoes (which are only now starting to give any reasonable quantity of fruit) and the cannas. I hope that the Hibiscus will survive this winter better, and then we'll try to give it a little more space.
The parent tree is still recovering from its rather drastic repotting a couple of months ago:
|
||||||||||||||||||||||||
It's showing signs of recovery, though one of the branches has clearly given up the ghost. And the curry tree that I repotted is looking much better, as are the cuttings that I made:
|
||||||||||||||||||||||||
|
|||||||||||||||||||||||
On the other hand, as I feared, the last Buddleja × weyeriana has died. It really happened quickly, and I don't know why. Here photos from two months ago (first two), a month ago and today (last two):
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Why did that happen? It's all the stranger because the cuttings that I took last month seem to be doing well:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Black cockatoo storm
|
Topic: animals | Link here |
While taking the garden photos, hundreds of yellow-tailed black cockatoos flew over the house:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
They're nothing like as common as the sulphur-crested cockatoos that we see almost every day, and of course I didn't have the right lens with me.
| Saturday, 21 March 2020 | Dereel | Images for 21 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Peace!
|
Topic: general, multimedia, technology, opinion | Link here |
Somehow the last week has been particularly stressful, and it had its overhang into today: I didn't get yesterday's diary entry committed until 16:54. But that's about all I had to do today. Yes, Yvonne managed to find an interesting bug in avidemux: when saving a clip, if you forget and use the standard Emacs c-x c-s, avidemux does save a clip. But it's not the one you selected: it's the one to the left of the A mark (in other words, the part cut off at the beginning).
“Well don't do that, then”.
|
More black cockatoos
|
Topic: animals | Link here |
While walking the dogs, still with the wrong lens, found more yellow-tailed black cockatoos:
|
|||||||||||||||||||||||
I suppose it's time to carry a long telephoto with me.
| Sunday, 22 March 2020 | Dereel → Ballarat → Dereel | Images for 22 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Finally mowing the lawn
|
Topic: gardening, opinion | Link here |
Mick Solly along this morning to finally mow the lawn. He has been sick, so has his lawn mower, and he arrived with a not completely healed rib fracture. Normally it wouldn't have been necessary to mow the lawn at this time of year. Here a comparison of this time last year (normal situation) and yesterday:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
At least things are looking a little better now.
|
OED access in times of COVID-19
|
Topic: general, technology, opinion | Link here |
I access the Oxford English Dictionary several times a week via the State Library of Victoria. It's convenient and free. Somewhere I have a DVD of the Second Edition, but it's easier via the web.
But today I was unable to access the service. “You could not be signed in. This account has no valid subscription for this site”. Damn. What has gone wrong there? The only credentials I need are the ones I need to access SLV, and they were OK.
OK, report a bug? No provision for that on the SLV web site, just “ask a question”. OK, “ask a question”, specifying a reason. What, bugs? That's no reason. At least not in their list. Still, somebody will see it tomorrow, and hopefully do something about it.
I sent my “question” off at 9:28:43 on a Sunday morning. What's a reasonable time to expect a response? Certainly not what what happened: 6 minutes later, at 9:34:45, I got a response from David Flegg:
Normally we would contact the vendor to identify and resolve the problem however as of 3pm 19 March staff have been sent home until at least Wednesday 25 March and so are not in a position to assist in this manner.
Some staff are working from home (like me). I will forward the issue to the relevant staff and hopefully they can do something about contacting the vendor. If not we hope to be back on-site on Wednesday 25 March. If we are we will deal with the issue at that time.
That's interesting for a couple of reasons:
|
Shopping in Ballarat
|
Topic: general, animals, gardening, opinion | Link here |
So we're supposed to be idle at home and not go anywhere. But then Yvonne saw a horse statue for sale in Ballarat:
|
|||||||||||||||||||||||
Of course she had to have it, and that meant going to town to pick it up. While there we dropped in at ALDI. Things aren't looking quite as bad as last week, though they have converted the toilet paper aisle to other purposes:
|
|||||||||||||||||||||||
And now they're sold out of disinfectants and things:
|
|||||||||||||||||||||||
At least that I can understand.
From there off to Bunnings to look for various things: a connector for irrigation hose fittings, some plants, and some strip lights for the cooking area in the kitchen.
We didn't find the plants we were looking for, which was just as well. We were expecting to pay $5 to $8, but the cheapest were over $20. We just couldn't find the fitting for the hose, but the closest fittings would have cost over $10, and for $23 we bought a new hose, fittings at both ends and a reel. Somehow the prices are ridiculous.
And then the strip lights. I had expected to pay $10 to $15, but of course there are no incandescent lights any more. After much comparison (I must finally get used to measuring light output in Lumens rather than Watts) bought two 39 cm 600 Lumen LED strip lights for the princely sum of $40 each. It wasn't until we got home that we discovered that there way no sane way of attaching them to a surface.
And the horse? We bought it, of course. That was a foregone conclusion.
|
More GPS navigation fun
|
Topic: technology, general, opinion | Link here |
My GPS navigator works even worse than before. It reached the end of time a year or so ago, and last month I left it out in the sun (wouldn't you think a navigator could deal with that?):
|
|||||||||||||||||||||||
Since then it doesn't work too well, and I'v ebeen thinking of a replacement for a while. Today's trip to Ballarat was worth another attempt. Off to the toyshop to see what was on offer.
I discovered an app already installed: “Australia Topo Maps”. Where did that come from? I had to search the toyshop to find that it's here. Like most of these things, there's no obvious documentation, and it doesn't seem to be related to Australia. While playing around with it, came up with this unbelievably bad map:
|
|
|||||||||
|
|
|||||||||
Where is it? To judge by the barely legible Cyrillic script, somewhere in Russia? No, Kelantan, as the text “Kampong Durian Daun”—old spelling—confirms. But what use is it? So I tried some new apps.
First, Maps Directions & GPS navigation, which got good ratings. Not from me: I couldn't get it to work. Every time I tried to do anything, I ended up with another advertisement trying to get me to install stuff that was sometimes only marginally related, and I had to fight my way past it, a different way each time. After the third time I gave up.
Why do people do this? Yes, advertisements are a fact of life, but if you annoy people so much that they don't even try the product, you're missing the point.
Next, Offline Maps & Navigation, which was similar, but not quite as bad. I got it installed and set the address in Ballarat. It gave me a clearly optimum route:
|
|||||||||||||||||||||||
I've seen this with other navigators too: they hate to make you turn around. But in this case I wasn't pointing anywhere, and in any case, going round the loop is a 3 km detour.
Off anyway, discovering that the display wasn't that bad, though clearly it suffered if the phone was in my pocket, presumably because of marginal reception. It has many of the disadvantages of other navigators, but I can probably live with it.
While doing these comparisons, though, ran into another issue that also bites Microsoft users: which program am I running? With X you almost invariably get a title on the window, though Google Chrome tries to make you share the pain of Microsoft users by trying to be a (second-rate) window manager that omits frames and identification. And that's apparently the only way on mobile phones. Why?
| Monday, 23 March 2020 | Dereel | Images for 23 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Still more flash pain
|
Topic: photography, opinion | Link here |
After buying the horse statue yesterday, the first thing to do was to take a photo.
How much effort should I take? In principle, it doesn't look bad where it is. But that's with a window behind and to one side. Take a photo like that and it will look terrible:
|
|||||||||||||||||||||||
But that's what flash is for, and I have a couple of studio units in the area. But for some reason I decided that the on-camera mecablitz would do a better job:
|
|||||||||||||||||||||||
OK, not bad. But then of course I wanted to compare it with the studio flash. Off with my light meter and measured f/8.7:
|
|||||||||||||||||||||||
What went wrong there? Increased the exposure ultimately to f/5.6, still with no good results:
|
|||||||||||||||||||||||
I've been having problems with flash exposure for over 50 years, but I still haven't solved them.
|
Illuminating the cooking area
|
Topic: food and drink, opinion | Link here |
After some examination, including consideration of kludges, discovered that the LED strip lights that I bought yesterday would indeed fit where I wanted them, as long as I ensured that the power connection (at the end) was at the right end; otherwise the light would shine into the wall. After relatively little work I had (before and after):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly I could do with a third light under the range hood, but I don't know how to attach it there. And I still need to think about the wiring. The lights allow daisy-chaining the power connections, but then I need custom wiring.
|
Hibiscus in flower
|
Topic: gardening, opinion | Link here |
The Hibiscus rosa-sinensis “Uncle Max” is definitely happier now:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But it's not the only hibiscus that is doing well. Clearly the new position suits the one Hibiscus syriacus that I planted up against the wall:
|
|||||||||||||||||||||||
The other one is flowering too, but it's not doing nearly as well.
And then there's the Alyogyne huegelii, once considered a hibiscus. It has grown so much that there won't be any space between it and the “Uncle Max” next year:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Tuesday, 24 March 2020 | Dereel → Ballarat → Dereel | Images for 24 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Vaccinations in the time of COVID-19
|
Topic: health | Link here |
Into Ballarat again this morning. Weren't we supposed to keep at home? Yes, but the doctors think that it would be good for old people like us to at least be vaccinated against influenza. It won't protect against COVID-19, but it will at least lower the likelihood that we'll end up getting both. Normally we would have been due for a revaccination in May, but now they're doing fast vaccination “clinics” in the clinic early in the morning, before others arrive—at least in theory.
While there, Yvonne had a blood test done. For once, the bandage on her arm didn't work, and she had blood streaming down her arm. Here the replacement they put on:
|
|||||||||||||||||||||||
|
Ballarat ghost town
|
Topic: general, health, opinion | Link here |
The effects of the curfews are becoming evident. This is Doveton Street in Ballarat, between Sturt Street and Dana Street looking east, during the morning rush hour:
|
|||||||||||||||||||||||
Normally this is the busiest street in Ballarat.
|
More PV recalibration
|
Topic: Stones Road house, general, opinion | Link here |
The PV inverter had another recalibration fest today. Started at 14:28 and went on all day. I wonder if this is really correct. Interestingly it started with the battery fully charged. Could it be that it wasn't expecting the behaviour at that level of charge?
|
Understanding flash exposure, episode 32768
|
Topic: photography, opinion | Link here |
Why did yesterday's flash photos have such bad exposure? I measured the light with my Sekonic L-308S, which should be accurate. Back to take another look today. There are two ways of measuring: incident light and reflected light. There's a little dome that fits over the sensor for incident light, and which is moved away for reflected light:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
On checking, discovered that the dome wasn't quite in the correct position. It wasn't completely pushed to the left (first photo):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Did that explain things? Only partially. But holding my hand above the meter changed things dramatically, by about 2½ EV. First lesson.
Later I took photos of the kitchen area. The first one was greatly overexposed. Why? Discovered that I wasn't really measuring what I had exposed. To measure it, I went up to the object, held the exposure meter in front of it, and triggered the flash. Then I took the photo from about 3 m away. Exposure should be the same, yes? Well, in principle. But when doing the measurements I was between the flash and the subject, and that was the problem.
Why is this all so hard?
|
File access in the age of Google
|
Topic: technology, opinion | Link here |
Peter Jeremy has given me access to a large collection of files on Google Drive. There must be several terabytes of them. OK, follow the “invitation” and get a list of names. besplex, epsplex and so on. What are they? My understanding is that they're disk images of some kind, but Google Drive tells me that they're “Google Drive folders”. What can I do with them? I have no idea. I'd expect to be able to select them show the component files, but I failed in that attempt.
Once again, the Google interface is so non-intuitive that I can't make head or tail of them. I tried the help, which popped up the obligatory too-small window and offered suggestions like
You can upload files from your computer or create files in Google Drive.
I'm going to have to ask Peter.
|
File access in the age of Android
|
Topic: technology, opinion | Link here |
I've had my new Nokia 5.1 Plus for 10 days now, and so far the experience has been mixed. I've found a number of apps that seem better than what I have used before, but I'm still running into this frustrating now-you-see-me-now-you-don't message stuff. But that might be a Nokia taunt.
One difference is that WiFi File Transfer only provides FTP access, while on the older Nokia 3 it offers HTTP.
I've been thinking for some time about my most basic question, last asked when I got the phone:
But surely, surely, there is software available for that most basic of network functionality, copying data from one system to another?
And then something occurred to me, a concept that even had a name: “ftpfs”, a file system
that runs over FTP. Clearly not a choice for high performance, but in principle FTP offers
the functionality you need to access files remotely. OK, off looking for ftpfs, and sure enough, found a whole lot
of them, even a Wikipedia page, which tells
me that the original implementation was done by no less
than Dennis Ritchie. I also found
a directoryfolder /usr/ports/sysutils/curlftpfs on eureka.
OK, there's a package, conveniently called
=== grog@teevee (/dev/pts/8) ~ 423 -> pkg search ftpfs
fusefs-curlftpfs-0.9.2_5 Mount remote FTP directories
Install that. The usual question: what got installed? man fusefs-curlftpfs? That's an old, worn-out magic word. What's in pkg-plist? Not present. But I had already found that the binary was called (wait for it) curlftpfs.
I later discovered that there's an alternative of the variable PLIST_FILES in the Makefile, saving an inode for small lists, and that pointed me not only to the binary, but also to the man page curlftpfs(1).
OK, how do I run it? Start by starting with inappropriate parameters:
=== grog@teevee (/dev/pts/8) ~ 426 -> curlftpfs
missing host
see `curlftpfs -h' for usage=== grog@teevee (/dev/pts/8) ~ 427 -> curlftpfs -h
usage: curlftpfs <ftphost> <mountpoint>
...
OK, after some trouble specifying the host name, ended up with:
=== root@teevee (/dev/pts/14) /usr/ports/sysutils/fusefs-curlftpfs 17 -> curlftpfs -v ftp://flachmann.lemis.com:2121 /flachmann
* Couldn't find host flachmann.lemis.com in the .netrc file; using defaults
* Trying 192.109.197.247:2121...
* TCP_NODELAY set
* Connected to flachmann.lemis.com (192.109.197.247) port 2121 (#0)
< 220 SwiFTP 3.0 ready
> USER anonymous
< 331 Send password
> PASS ftp@example.com
< 230 Access granted
> PWD
< 257 "/"
* Entry path is '/'
* Request has same path as previous transfer
* ftp_perform ends with SECONDARY: 0
* Remembering we are in dir ""
* Connection #0 to host flachmann.lemis.com left intact
fuse: failed to open fuse device: No such file or directory
Oh, no, you didn't forget to load the fuse module that we didn't tell you about, did you? Guilty as charged. Read fuse: failed to open fuse device, load the module, try again, get exactly the same messages except for the last line. And:
=== root@teevee (/dev/pts/9) /usr/ports/sysutils/fusefs-curlftpfs 74 -> df /flachmann/
Filesystem 1048576-blocks Used Avail Capacity Mounted on
/dev/fuse 976,562 0 976,562 0% /flachmann=== root@teevee (/dev/pts/14) /usr/ports/sysutils/fusefs-curlftpfs 22 -> l /flachmann/
total 1
drwxr-xr-x 1 root wheel 4,096 14 Mar 17:00 AirDroid
drwxr-xr-x 1 root wheel 4,096 19 Sep 2019 Alarms
drwxr-xr-x 1 root wheel 4,096 19 Sep 2019 Android
drwxr-xr-x 1 root wheel 4,096 23 Mar 14:13 DCIM
drwxr-xr-x 1 root wheel 4,096 19 Sep 2019 Download
drwxr-xr-x 1 root wheel 4,096 23 Mar 13:49 mbstph
drwxr-xr-x 1 root wheel 4,096 19 Sep 2019 Movies
drwxr-xr-x 1 root wheel 4,096 19 Sep 2019 Music
drwxr-xr-x 1 root wheel 4,096 19 Sep 2019 Notifications
drwxr-xr-x 1 root wheel 4,096 15 Mar 10:15 Pictures
drwxr-xr-x 1 root wheel 4,096 19 Sep 2019 Podcasts
drwxr-xr-x 1 root wheel 4,096 15 Mar 10:16 Ringtone Cutter
drwxr-xr-x 1 root wheel 4,096 18 Mar 18:24 Ringtones
drwxr-xr-x 1 root wheel 4,096 15 Mar 16:08 SendAnywhere
It's there! I can access all the files on the phone from a real computer! What an amazing relief! Finally I can access data on this phone in a sane manner. By comparison the clearly incorrect values in the df output (which prove to be the same across all phones) are of no interest.
Why was the way here so difficult? For the fun of it, took a few photos with the phone and was able to display them on teevee with no further interaction:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
The last photo (admire that moiré and vignetting) shows some of the file actions.
In passing, it's interesting to note that the camera has a 4:3 aspect ratio. Is that typical for mobile phones?
| Wednesday, 25 March 2020 | Dereel | Images for 25 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Battery calibration without end
|
Topic: Stones Road house, general, technology, opinion | Link here |
When I got into my office this morning, yesterday's PV battery calibration was still going on:
|
|
|||||||||
|
|
|||||||||
Three complete cycles, and it was continuing into the fourth. When it got to the fully charged state of the fourth, I power cycled the inverter, but it's becoming clear that that won't stop it. Time for a stiffer letter to the suppliers, asking for:
As I asked last August, please get me a statement from the manufacturer (not from you) by 10 April answering the questions:
As it is, I can only conclude that the device is defective when measured by the information supplied to me on purchase. Please arrange for it to be repaired or replaced. I am prepared, without obligation, to accept a single calibration cycle per month, but no more.
Tomas, you might like to consider alternate equipment to replace the Ingeteam/BYD combination.
The response was immediate. I sent the message at 12:22, and Fred replied at 12:37. By 13:40 he was on site updating firmware both on the inverter and on the battery. Keeping his distance (he has a cold, potentially a more conventional Coronavirus), he explained that it is possible to completely bypass the recalibration. It seems that it has also happened to customers without grid or generator backup. When the battery reached 0%, they lost all power.
But that's only part of the story. It seems that the request for recalibration comes from the battery (hopefully the firmware update will improve things there), and that it is almost impossible to get technical information from BYD, the maker. Fred can't get the information. I wonder what a report to the authorities will bring. In any case, though it's clear that Fred and Tomas aren't to blame, I can't see any reason not to report the incident if (as I expect) I don't hear back from the makers by 10 April.
|
More lights in the kitchen?
|
Topic: food and drink, photography, opinion | Link here |
While setting up the lights in the kitchen yesterday, found a blob of oil on the cooktop. Clearly it had dropped from the range hood filter, so I put the filter through the dishwasher. But the dripping continued:
|
|||||||||||||||||||||||
That's the extractor fan above the filter. The filter is so useless that oil collects on the fan, then drips down through the filter onto the cooking surface.
The workaround was simple: wipe the oil of the housing of the fan. But while looking at it, it occurred to me that there's a rail between the end of the filter and the wall. Can I screw another strip light to it? Time to take a couple of photos:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
No, sadly. The white oval on the second image is the far-too-large cover for the mounting hole, at 25 mm from the wall. And that's exactly the width of the rail.
But the method shows something interesting. I took this with the Olympus OM-D E-M1 Mark II with the viewfinder screen pointing forward:
|
|||||||||||||||||||||||
|
|
|||||||||
Some people look down on this as a “selfie” aid, and they prefer the tilting screen of the E-M1 Mark I1. But the E-M1 would be useless in this situation. A clear win for the screen of the E-M1 Mark II (and also the Olympus E-30 and the E-M5 Mark III). I had already discovered this with the E-30 years ago when taking panoramas from the north wall of the Kleins Road house.
|
More mobile phone fun
|
Topic: technology, general, opinion | Link here |
So now I have file access to flachmann, my new Nokia 5.1 Plus, via a file system interface. But it's FTP, not exactly state of the art. How about an HTTP access method? I can access the file system on taskumatti (my old Nokia 3) via HTTP, but not via FTP, using the same app, WiFi File Transfer.
OK, off looking. Yes, there's a port fusefs-httpfs, which not only installs, but also produces a message:
===> NOTICE:
The fusefs-httpfs port currently does not have a maintainer. As a result, it is
more likely to have unresolved issues, not be up-to-date, or even be removed in
the future.
And indeed it had unresolved issues. Even armed with my experience with ftpfs, I was unable to get anything useful out of it:
=== root@teevee (/dev/pts/9) /usr/ports/sysutils/fusefs-httpfs 72 -> httpfs2 http://taskumatti:1234 /mnt
file name: taskumatti
host name: taskumatti
port number: 1234
protocol: http
request path: /
auth data: (null)
httpfs2: HEAD: failed with status: 400 Bad Request.
HTTP/1.1 400 Bad Request
Connection: Close
Date: Wed, 25 Mar 2020 01:44:22 GMT
Content-Type: text/html
Content-Length: 63
httpfs2: HEAD: exchange: server error
HTTP/1.1 400 Bad Request
Connection: Close
Date: Wed, 25 Mar 2020 01:44:22 GMT
Content-Type: text/html
Content-Length: 63
Part of the problem, of course, is that HTTP doesn't have a method for showing the contents of a directory, and after some discussion on IRC we concluded that it wasn't intended for that kind of access.
OK, look for a real FTP server. And there are plenty of them. The most obvious name seemed the best first choice: WiFi FTP Server. Installed that, started it, and it worked. What more can you want?
Well, how about access to the SD card? The root was set as “Internal Memory”, but I had the option “/”. Tried that. No content. There's also a “custom” setting, but that didn't work for me either. Later I found, on the web site (you wouldn't expect it on the phone now, would you?):
On Android 5.0 & higher, to access external SD card, in app settings, click mount folder, select "custom" and then select the external SD card in the next screen.
There's a video showing how to do it:
|
|
|||||||||
But that doesn't match my display:
|
|
|||||||||
More head-scratching. Another Android incompatibility, maybe?
| Thursday, 26 March 2020 | Dereel → Napoleons → Dereel | Images for 26 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Nokia 5.1 Plus: yes or no?
|
Topic: technology, general, opinion | Link here |
So: I've had the Nokia 5.1 Plus for 12 days now. Another 48 days before I need to give it back if I don't like it. What advantages does it have? Why was I even interested in a new phone?
The issues I have with the older Nokia 3 are many and varied. The ones that continually irritate me are:
The first problem continues. It's not clear whether it's a general Android issue, specific to Nokia, or just bad luck or lack of intuition on my part, but clearly it's not a reason for buying a new mobile phone.
The second? I haven't had any issues with calls on the new phone, but then I have hardly made any. But making calls is something that any phone should be able to do, including my ancient Samsung GT-I9100T. How about having that set up just for phone calls, and use a more modern phone for the other functions? It's certainly worth a try, especially since my memory tells me that I had fewer problems with the Samsung.
So the third remains. How good are the GPS receivers in modern mobile phones? Tried a couple of comparisons while walking the dogs. They're certainly difficult to keep running, and today the new phone only recorded a couple of waypoints, so I went back to older ones. Here first for 18 March 2020, then for 21 March 2020. On the left the new phone, on the right the old one:
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
Clearly the new phone produces better results. But are they good enough? Not really. On 21 March it included a spurious position 400 m to the east of the house, and the walk down Harrisons Road shows deviations of up to about 30 m.
Can a dedicated GPS receiver do better? By chance I had a package to pick up in Napoleons, so off with three GPS devices to pick it up:
|
||||||||||||||||||||||||
I'll evaluate the differences tomorrow, but the first indications are that the data from taskumatti tags the photo above as being taken from the petrol pumps, and the data from flachmann puts it as taken from inside the building.
|
Another FTP server
|
Topic: technology, opinion | Link here |
Daniel O'Connor pointed me at an alternative FTP server yesterday. FTP? Everything! It's a Servers Ultimate, reminding me that “ultimate” means “last”, and that the German translation „das Letzte“ an extremely negative evaluation is.
And indeed, it had servers.
|
|
|||||||||
5 different FTP servers. What are they all? Of course it's too polite to assume I wouldn't know. Some worked for me, some didn't. But while looking around on their web site, found:
This app will work for 7 days after which you will need to buy the paid version.
BAD Servers Ultimate. They should at least have said that up front. Goodbye.
|
Working from home
|
Topic: general, technology, opinion | Link here |
Due to the COVID-19 pandemic, lots of people are working from home for the first time. And of course lots of other people are trying to sell the solutions to their perceived problems. How-To Geek has many suggestions for The Best Work From Home Apps for iPhone and Android. I only have one: avoid them all. The devices have been optimized for people moving around, to the detriment of use in a single place.
One exception comes to mind, though: The 6 Best Free Video Conferencing Apps. I don't have a camera on eureka, and I don't know of anything that would work with FreeBSD. Doubtless there is something, but my pain with avidemux suggests that I might put up with some pain in the Microsoft space instead. And of course mobile phones have cameras that point at the user. I fear, however, that they will have stupid restrictions like only working over specific links (noticeably mobile telephone links). A couple of examples:
When you work from home for eight hours a day, you can’t just rely on video calls and emails for communication. To stay in touch with your colleagues throughout the day, you need a better system than a WhatsApp group chat.
People, the operative word is work. Video calls are a particularly frustrating way to waste time, and my experience with group chats is also very negative. Limit communication. That way you can do some work without interruption.
If all your friends and family have an iPhone, there’s nothing quite like FaceTime. You can add up to 32 people, and use Animojis or fun effects. It’s one of our favorite apps for video-chatting.
And this is the future? Somehow they're throwing out a lot of the advantages of being by yourself and undisturbed.
| Friday, 27 March 2020 | Dereel | Images for 27 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
OED back online
|
Topic: language | Link here |
The Oxford English Dictionary went offline for me on Sunday. I didn't realize how much I depended on it, and I continued checking a couple of times a day to see if the authentication would finally work. I even considered paying real money (round $100 US) for access.
But finally things are working again, and I also got confirmation from the State Library of Victoria. Thank God for that.
|
Weather station pain
|
Topic: general, technology, opinion | Link here |
My weather station is still having difficulties communicating with the outside unit. That's nothing new: I've had these issues for over 10 years, but for some reason they're getting worse. It's clearly related to the position of the internal unit, and I've been moving it around the office to try to get better communication.
The obvious thing to do is to put it in the lounge room. Like that it almost has a line of sight to the external unit, and the internal measurements are more meaningful. That requires connecting the unit to teevee. But for that I had to work my way through these bizarre MySQL permissions, which I still don't understand. In particular the WITH GRANT OPTION doesn't seem to work, and the host name specifications don't do what I would expect. The first of these commands (executed as root) doesn't grant access, the second does. And neither allow grog to grant anything.
mysql> grant all privileges on *.* to 'grog'@'teevee' with grant option;
mysql> grant all privileges on *.* to 'grog'@'teevee.lemis.com' with grant option;
OK, next run wh1080. Needs different libraries. OK, compile and admire the multicoloured vomit that clang (clearly a modern compiler) produces. Run it:
=== root@teevee (/dev/pts/9) ~ 97 -> while :; do wh1080 Stones-Road grog "" eureka; sleep 5; done
Can't read device: Unknown error or Device busy (16)
Can't read device: Unknown error or Device busy (16)
Can't read device: Unknown error or Device busy (16)
Previous rain 0.0, rain now -2588626550846210933042426723434496.0, difference -2588626550846210933042426723434496.0
INSERT INTO observations
(station_id, date, time, inside_humidity, inside_temp, inside_dewpoint,
outside_humidity, outside_temp, outside_dewpoint, pressure_abs, pressure_msl,
wind_speed, wind_gust, wind_direction, wind_direction_text, rain)
VALUES
("Stones-Road", "2020-03-27", "13:11:31", 52, 25.2, 14.6, 42, 26.0, 12.1, 994.9, 1035.6, 3.9, 6.7, 225.0, "SW", -2588626550846210933042426723434496.00);
Those EBUSY errors are nothing new. I'm sure that they're some bug in the USB stack, but I've never had the nerve to go looking for the problem. But on teevee they seem to be much more frequent.
And what's that rain value? It took me a while to realize that that couldn't possibly be a problem with the device: the rain information is supplied in a 16 bit unsigned value in units of 0.3 mm. Looks like an uninitialized variable. Why does it bite me now? Again, probably a difference between clang and gcc. Clearly something I need to look at, and possibly a reason for changing the kludgy way I treat rainfall.
|
Epazote!
|
Topic: gardening, food and drink, opinion | Link here |
Shortly after moving to Dereel I found a Mexican grocery shop near Melbourne Airport and bought some dried epazote, about 20 g of leaves and twigs. I still have much of it But a year ago I got epazote seeds and planted a very few of them.
That was plenty. It has grown like a weed, probably because it is one. Today Mick Solly came and commented on the size of the bush, and offered to pull it out. That was probably a good idea:
|
|||||||||||||||||||||||
Much of it had gone to seed (thus probably ensuring that I will never get rid of it), but I got enough leaves and small twigs to fill a large plastic bag with 800 g of the stuff. In all likelihood it will last me the rest of my life.
|
GPS comparisons
|
Topic: technology, general, opinion | Link here |
I've been dragging my heels on my GPS receiver comparisons. There's a good reason: it's a pain to access these things. Now that I have file-system like semantics to access the telephones, that's not so bad, but I had forgotten how to access the dedicated unit.
The first step is clear: connect the unit to the computer with a USB cable:
Mar 27 12:14:13 eureka kernel: ugen0.11: <MStar Semiconductor, Inc.> at usbus0
Mar 27 12:14:13 eureka kernel: umass4: <MStar Semiconductor, Inc. Mass Storage Device, class 0/0, rev 2.00/0.00, addr 17> on usbus0
Mar 27 12:14:13 eureka kernel: umass4: SCSI over Bulk-Only; quirks = 0xc001
Mar 27 12:14:13 eureka kernel: umass4:7:4:-1: Attached to scbus7
Mar 27 12:14:13 eureka kernel: da4 at umass-sim4 bus 4 scbus7 target 0 lun 0
Mar 27 12:14:13 eureka kernel: da4: <Mstar > Removable Direct Access SCSI device
Mar 27 12:14:13 eureka kernel: da4: Serial Number 20e41800-0009-0000-0000-000000000001
Mar 27 12:14:13 eureka kernel: da4: 40.000MB/s transfers
Mar 27 12:14:13 eureka kernel: da4: 3550MB (3635200 1024 byte sectors: 255H 63S/T 226C)
Mar 27 12:14:13 eureka kernel: da4: quirks=0x2<NO_6_BYTE>
Mar 27 12:14:13 eureka kernel: da5 at umass-sim4 bus 4 scbus7 target 0 lun 1
Mar 27 12:14:13 eureka kernel: da5: <Mstar > Removable Direct Access SCSI device
Mar 27 12:14:13 eureka kernel: da5: Serial Number 20e41800-0009-0000-0000-000000000001
Mar 27 12:14:13 eureka kernel: da5: 40.000MB/s transfers
Mar 27 12:14:13 eureka kernel: da5: Attempt to query device size failed: NOT READY, Medium not present
Mar 27 12:14:13 eureka kernel: da5: quirks=0x2<NO_6_BYTE>
Presumably da5 is a device for a non-existent SD card. OK, clearly it's a FAT file system, so I'm looking for /dev/da4s1. But there's no partition table! That would be /dev/da4s1:
=== root@eureka (/dev/pts/6) ~ 317 -> ls /dev/da*
/dev/da0 /dev/da1 /dev/da2 /dev/da3 /dev/da4
/dev/da0p1 /dev/da1p1 /dev/da2p1 /dev/da3p1 /dev/da5
OK, hack around in /usr/local/etc/mtools.conf:
drive a: file="/dev/da3s1"
drive b: file="/dev/da4s1"
drive c: file="/dev/da5s1"
drive d: file="/dev/da6s1"
# And for drives that are too polite to use MBR partition tables
drive e: file="/dev/da4"
drive f: file="/dev/da5"
drive g: file="/dev/da6"
But that doesn't work either:
=== root@eureka (/dev/pts/11) ~ 493 -> mdir e:
plain_io: Invalid argument
init E: could not read boot sector
Cannot initialize 'E:'
Clearly mtools are too polite to use naked disks. So I needed an alternative:
=== root@eureka (/dev/pts/6) ~ 318 -> mount -t msdos /dev/da4 /mnt
mount: Using "-t msdosfs", since "-t msdos" is deprecated.=== root@eureka (/dev/pts/6) ~ 319 -> l /mnt
total 1
drwxr-xr-x 1 root wheel 8,192 2 Jan 2008 FLASH
drwxr-xr-x 1 root wheel 8,192 2 Jan 2008 MP3
drwxr-xr-x 1 root wheel 8,192 2 Jan 2008 MP4
drwxr-xr-x 1 root wheel 8,192 2 Jan 2008 Music
drwxr-xr-x 1 root wheel 8,192 2 Jan 2008 PHOTO
drwxr-xr-x 1 root wheel 8,192 2 Jan 2008 Primo 2.0
With a bit of searching, found the track logs in /mnt/Primo2.0/SAVE/PROFILES/01/TRACKS:
-rwxr-xr-x 1 root wheel 20,357 12 Mar 2019 track262.trk
-rwxr-xr-x 1 root wheel 20,477 12 Mar 2019 track263.trk
...
-rwxr-xr-x 1 root wheel 30,357 16 Dec 15:28 track315.trk
-rwxr-xr-x 1 root wheel 30,357 16 Dec 16:02 track316.trk
And track316.trk was the most recent entry!
OK, this device has passed the end of time, and it really thinks that the date is 10 August 2000 (care, US out-of-order date coupled with reverse horror):
|
|||||||||||||||||||||||
But the file name agrees. So where does the date December 2019 come from?
After all that pain, copying the files was trivial. Once again I had a real file system to talk to. But other things during the day kept me from getting any further.
|
More electricity supplier pain
|
Topic: Stones Road house, general, opinion | Link here |
Yesterday I received an electricity bill from Red Energy: $241.47 credit. That's not bad, is it?
Yes, it is. The credit was due to part of what was agreed in January, applying my pensioner concession retroactively. What about the promise of applying the lower rate retroactively? In fact, what about the promise of applying the lower rate? I was being charged the same as before this whole business happened.
OK, time for a complaint, which I had sent off yesterday, in the expectation that they would apologize and apply the correct conditions. But no, today I received a reply:
I have reviewed your account and can confirm we have been waiting for your consent to the quote for the offer to be accepted.
I have checked our records and can see we sent you the consent details by email on the 29/01/2020 but never received a response.
As no response was received the offer was not applied and has now expired.
When you are free, please forward me the email you sent, in response to our consent email dated 29/01/2020, to info@redenergy.com.au for investigation.
I look forward to hearing from you soon and that your day is going well.
Nonsense! Firstly I had given my consent not once, but three times, once on the phone and twice in two separate emails, the second of which produced the automatic ”Your e-mail will be responded to within 1 business day” confirmation and nothing further. The email to which she refers was a reply to a message where I wrote:
Please confirm:
- That you are supplying electricity according to these conditions.
- That you have accepted my "application".
They didn't, of course. And she asked me to forward the message I sent. It was attached to my complaint. I replied accordingly and got another reply (surprisingly, at 18:12), claiming:
Thank you for your reply and I can confirm I did note the details you provided.
And that was the closest she came to addressing the content of my email, which included:
Please also confirm that I have cancelled my direct debit
authorization for this account.
So: it looks like more fun. Is this problem due to incompetence or malice? Bob Hanlon says incompetence, but in this specific case it would be easy to make the case that their behaviour (in particular, not answering emails and then claiming timeouts) is fraudulent.
|
Where's Piccola?
|
Topic: animals | Link here |
Eleven years ago today we went to Shepparton to buy a new pussycat, Piccola. But this evening she didn't come home for dinner. Where was she? Yvonne thought that she was chasing mice, but I was really concerned, and spent some time looking for her, in vain. Cat disappeared on her anniversary?
Then later on a familiar meow!, the kind that only Siamese cats can produce. She had come in through the laundry, finished the food we had put out for her, and took her place on Yvonne's lap, as if nothing had happened. But later she produced the most incredibly obnoxious fart that I have smelt in a long time. Yvonne wanted to blame me or a dog, but not Piccola. After a bit of consideration, I think it smelt of rat...
| Saturday, 28 March 2020 | Dereel | Images for 28 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Buddleja recovery?
|
Topic: gardening, opinion | Link here |
There's no doubt about it, my Buddleja × weyeriana is dead. At least two of the cuttings I made have taken well, and the other two seem to be taking too. But I still wasn't prepared for what I saw today:
|
|||||||||||||||||||||||
That's a flower developing. What a difference from the parent tree outside! And what caused the parent to die?
|
mecablitz flash insights
|
Topic: photography, opinion | Link here |
The last time I used my mecablitz 15 MS-1 ring flash, I was less than satisfied with the results. At the time I guessed that it was simply a disadvantage of TTL metering, but I carried on thinking. Somehow I got the impression that it just wasn't bright enough. It has a guide number of 15 at ISO 100, corresponding to 21 at the standard Olympus ISO 200, so it should have been able to expose this image correctly at up to 2.35 m at f/9:
|
||||||||||||||||||||||||
But that was taken at a distance of about 85 cm. Did I somehow have the thing set to some manual mode? Off to read the instruction manual again, but found no mention of a manual mode. All controlled from the camera.
But the flash tubes have three settings, depending on distance. How does it work? Read further and confirmed that I had had the reflectors set incorrectly, at 20°:
Working distance of approx. 2-10cm: Swivel reflectors by 20° and set the bounce diffuser (see 9.1) in front of reflectors.
Bounce diffuser? What's that? The instructions are not designed for PDFs, and all the illustrations are on a different, fold-out sheet at the back of the book. All I could read was:
The bounce diffuser is required for illuminating the objects in close proximity at distances less than 10 cm.
Position the bounce diffuser asymmetrically over the flash unit so that the reflectors and the AF auxiliary light are covered.
And how about that, I had had the diffuser on the unit since I got it, and didn't realize that it could be removed. Took that off, and all worked well.
|
Tidying house exterior
|
Topic: gardening, general | Link here |
Warrick Pitcher along this afternoon to spread some Kaolin sand that Yvonne wanted in the horse stables, and also to tidy some stuff in the garden. Yvonne handled it all; I was too busy with other things.
|
GPS comparisons, finally
|
Topic: technology, general, opinion | Link here |
So finally I have made maps from the track logs that I created while travelling to Napoleons on Thursday using GPS Visualizer. They're certainly interesting.
First, the whole journey, as logged with taskumatti (my old Nokia 3 phone), flachmann (the new Nokia 5.1 Plus phone) and my old dedicated GPS navigator:
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
I went to Napoleons and returned via the same path, so there should be no loops in the track logs. But both phones show them, in each case going across impassable terrain. I'm assuming that this is due to prolonged reception problems. The GPS navigator shows no such problems, although it was in a worse position for reception.
And how accurate were they the rest of the time? Here's a detail of my passage through Napoleons:
|
|
|||||||||
|
|
|||||||||
|
|
|||||||||
These images are roughly centred in the same place; taskumatti doesn't even register where I turned in to the post office. flachmann does better, but not nearly good enough. The GPS navigator seems to get it right. The “beak” is where, after picking up my item, I pulled aside to check the devices.
So: based on this information, the GPS receivers in both phones are just not adequate. Are others better? I can take the Nokia 5.1 Plus back and get a refund. That's not so easy with others, unless ALDI decides to offer them.
|
Communicating by email
|
Topic: general, technology, opinion | Link here |
My problems with Red Energy yesterday haven't been the only ones where I have problems communicating by email. I've had two others in the last year, first with a company that might be called Resilium, and more recently with Elysian Energy. In all cases, communicating was like pulling teeth.
Why? Part of it is clearly the practice of replying to a message in a separate place from the message itself. This is the way it has always been with conventional mail. But when I wrote lots of conventional mail, over 50 years ago, I might have been replying to a multi-page letter, in which case I would have it on the desk next to me. That way I could go through the letter and reply to the points one at a time.
That's no longer so easy. Modern “desktops” don't make it easy to view two things at once, windows or no windows: the display area is usually too small. The result is that the reply tends to only address one point and ignore the rest.
Of course, that's not the way I do it. Decades ago I wrote guidelines on how to make yourself understood. Sadly, almost nobody outside the Unix space follows them any more.
So: did my message layout confuse people? Possibly, but nobody has said so. They just appear to ignore what they don't understand. But if my layout is confusing, how can I improve it without running into the traps of “first come, only served”?
|
Emacs yellow areas tracked down
|
Topic: technology, opinion | Link here |
While writing yesterday's diary entries, ran into another case of a yellow-stained Emacs buffer. This time I grabbed the lossage:
SPC O K , SPC h a c k SPC a r o u n d SPC i n SPC <switch-frame>
C-x C-f C-g C-x C-s <help-echo> <M-drag-mouse-2> <M-drag-mouse-3>
<switch-frame> <switch-frame> M-x v i e w - l o <tab>
<return>
Clearly this happened as the result of one or more of <help-echo> <M-drag-mouse-2> <M-drag-mouse-3>. But what do they mean? The drag-mouse events are clear, but <help-echo>?
And how did these events happen? That's clearer: mouse vomit.
| Sunday, 29 March 2020 | Dereel | Images for 29 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Piccola: problems?
|
Topic: animals | Link here |
Pretty much the first thing I do when I get up in the morning is open the door to the laundry to let Piccola out. She helps by meowing continuously until I do.
But today it was quiet. Opening the door wasn't encouraging:
|
|||||||||||||||||||||||
Where did that blood come from? I found the answer behind the door:
|
|||||||||||||||||||||||
Yvonne had left the outside door ajar overnight. It's not clear whether the rat came in voluntarily or not. Piccola was unscathed.
|
The effect of battery recalibration
|
Topic: Stones Road house, general | Link here |
Our photovoltaic electricity system has worked well since Fred installed the new firmware. But today, looking back over the week, it's clear what effect the recalibration had:
|
|||||||||||||||||||||||
On 24 March and 25 March, the two days where the system was recalibrating, the system generated 33.89 kWh, as the second bar chart clearly shows. On the other 5 days it generated 168 kWh, an average of 33.6 kWh per day. So over the two days I lost about 33.3 kWh of power. At $0.30 per kWh, that's pretty much exactly $10.
|
Gradation of TV programmes
|
Topic: photography, multimedia, opinion | Link here |
Watching The Doctor Blake Mysteries on TV the other night, I saw this:
|
|||||||||||||||||||||||
|
|
|||||||||
What's interesting about that? I know an Andrew Perry, so I wanted to take a screen shot to show him.
How do you do that? RTFM, but not while watching TV. Today I checked: mpv has no fewer than three options, all selected with some form of s. By itself, s “takes a screenshot”. S (capital) drops the subtitles, maybe. And c-s takes the screen shot as the screen shows it, which seems to be the correct one; the others proved to be smaller in this case.
So this is what I got:
|
|||||||||||||||||||||||
Why is it so dark? That's certainly not what I saw on the screen. With the help of DxO PhotoLab and Ashampoo I got progressively better images (here with the original for comparison):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I'm still puzzled that that was needed.
| Monday, 30 March 2020 | Dereel | Images for 30 March 2020 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Curry tree leaf size
|
Topic: animals, food and drink | Link here |
I've been watching the recovery of my Curry tree for some months now, from when I removed it from the lounge room,
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
to repotting
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
to gradual recovery:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Now it's certainly looking healthier than at any time in the last few years. But the interesting thing is the size of the leaves. I had already noted that the new leaves are smaller than previously, and assumed that they would grow. But in fact they didn't, directly. The new leaves that have grown inside the house are bigger than the ones that grew outside, and even the same frond has leaves of different sizes:
|
|||||||||||||||||||||||
Does it have something to do with the temperature, or the wind? Which are better for cooking?
|
Still more flash issues
|
Topic: photography, opinion | Link here |
I took the photos of the curry tree leaves with the mecablitz 15 MS-1 ring flash. After my insights a couple of days ago I hadn't expected any significant further problems, but I had them anyway. The first photos looked like this:
|
|||||||||||||||||||||||
Oh. How did I set 64/19° ISO? But even so, why was it so underexposed? Tried at the standard 200/24° ISO, but the results weren't much better, though postprocessing helped. Here before and after (run the cursor over an image to compare it with its neighbour):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But why? Tried with my old “Viltrox JY-670 Macro Ring Lite”, and after adjusting the exposure (previously overexposed), got this:
|
|||||||||||||||||||||||
No problems with getting the correct exposure, but the reflections from the leaves are a surprising difference from the mecablitz.
So what's the problem? The Viltrox has a native guide number of 14 at 100/21°, while the mecablitz has 15, close enough to make no difference. I took the Viltrox photo at ¼ power, which corresponds to a guide number of 10 at 200/24°. Looking at the exposure details (f/22, 0.28 m), that corresponds to a guide number of 6.16—but then there's about a 1.5 stop difference for the magnification, giving about 9. That's close enough for me.
And the mecablitz? I really get the feeling that the thing really doesn't have anything like the power it claims. DxO PhotoLab tells me that the mecablitz example was underexposed by 1.5. That's the point after which increasing the “exposure” causes the overexposure indicator to come into play. So this exposure (f/5.6, 0.46 m) corresponds to a guide number of 2.1—at 200/24° ISO! That's 1.4 odd at 100/21°.
OK, how about the first, vastly underexposed image? DxO says 4 EV underexposed. Aperture is f/8, distance 56 cm. That corresponds to GN 2.24 at 64/19°, or 2.8 at 100/21°. That's actually better than the results for the second image. Why? It seems difficult to believe that the flash is really that bad; after all, Metz were once one of the big names in flash.
Once again, how about the reflector position? This was taken in standard position. I've already established that 20° is not very helpful. How about 10°? What does the manual say?
At distances over 10 cm, the reflectors can be swivelled to a greater or lesser extent to achieve a centre weighted average or a balanced illumination.
Well that helps, doesn't it? But the fact that the first image, taken at 56 cm, was less underexposed than the second, taken at 28 cm, suggests that I should maybe try the 10° position next time.
Looking back, it seems that I have never taken any photos with this unit without some problem or another.
|
Planting random plants
|
Topic: gardening | Link here |
A number of things have been hanging around waiting for planting lately. This plant, I think a Syngonium podophyllum, continually produces runners:
|
|||||||||||||||||||||||
Can they grow outside? There's one way to find out, so I cut some off and planted them to the south of the house.
|
|||||||||||||||||||||||
While I was at it, also planted some particularly sickly looking Epazote plants:
|
|||||||||||||||||||||||
They've been in pots for about the last 9 months, and I had left them to die, but I might as well put them in the ground and see what happens. They're supposed to be annuals, but I'll believe that when I see it.
And then there's the garlic that popped up in the same garden last spring, and which I admired as it produced a 2 m high flower stalk:
|
||||||||||||||||||||||||
Mick Solly told me it was garlic, and I didn't believe him. But garlic it is, and one of the biggest cloves I have ever seen:
|
|||||||||||||||||||||||
There were only 6 in total, so I've planted them all. Next year, along with the smaller garlic plants that have sprung up miraculously in the succulent bed, we might even be self-sufficient in garlic.
|
Panic!
|
Topic: technology | Link here |
While watching TV this afternoon, teevee froze. No obvious disk activity, but after a while it rebooted:
Mar 30 16:04:57 teevee savecore[651]: reboot after panic: page fault
Oh. It's been a while since I've seen one of those. Took a quick look at the dump:
=== root@teevee (/dev/pts/8) /var/crash 3 -> kgdb /boot/kernel/kernel vmcore.0
...
Unread portion of the kernel message buffer:
NVRM: Xid (PCI:0000:01:00): 8, Channel 00000005
Fatal trap 12: page fault while in kernel mode
cpuid = 1; apic id = 01
fault virtual address = 0x0
fault code = supervisor read data, page not present
instruction pointer = 0x20:0xffffffff82a3c132
stack pointer = 0x0:0xfffffe00004c28e0
frame pointer = 0x0:0xfffffe002da33d40
code segment = base 0x0, limit 0xfffff, type 0x1b
= DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 12 (swi4: clock (0))
trap number = 12
panic: page fault
cpuid = 1
time = 1585544630
KDB: stack backtrace:
#0 0xffffffff80c0a6d5 at kdb_backtrace+0x65
#1 0xffffffff80bbea3e at vpanic+0x17e
#2 0xffffffff80bbe8b3 at panic+0x43
#3 0xffffffff81090310 at trap_pfault+0
#4 0xffffffff8109035f at trap_pfault+0x4f
#5 0xffffffff8108f9b8 at trap+0x288
#6 0xffffffff81069a1c at calltrap+0x8
Uptime: 24d22h26m17s
Dumping 759 out of 3875 MB:..3%..11%..22%..32%..41%..51%..62%..72%..81%..91%
...
Now isn't that trace useful? It shows the path by which the underlying problem led to the panic. To find out what really happened, I need to go deeper:
(kgdb) bt
#0 0xffffffff81206a6b in cpustop_handler () at /usr/src/sys/x86/x86/mp_x86.c:1408
#1 0xffffffff81206a2d in ipi_nmi_handler () at /usr/src/sys/x86/x86/mp_x86.c:1369
#2 0xffffffff8108f774 in trap (frame=0xffffffff81ed9eb0)
at /usr/src/sys/amd64/amd64/trap.c:203
#3 0xffffffff8106a793 in nmi_calltrap () at /usr/src/sys/amd64/amd64/exception.S:862
#4 0xffffffff811f7b45 in cpu_idle (busy=<value optimized out>)
at /usr/src/sys/x86/x86/cpu_machdep.c:562
Previous frame inner to this frame (corrupt stack?)
Current language: auto; currently minimal
(kgdb) f 4
#4 0xffffffff811f7b45 in cpu_idle (busy=<value optimized out>)
at /usr/src/sys/x86/x86/cpu_machdep.c:562
562 atomic_store_int(state, STATE_MWAIT);
(kgdb)
The “corrupt stack?” comment is incorrect. There's no way to return from cpu_idle(). But how can it generate a page fault exception? A quick look didn't get me much further:
(kgdb) x/20i $rip-20
...
0xffffffff811f7b41 <cpu_idle+225>: sti
0xffffffff811f7b42 <cpu_idle+226>: mwait %rax,%rcx
0xffffffff811f7b45 <cpu_idle+229>: jmp 0xffffffff811f7b48 <cpu_idle+232>
...
The instruction pointer was at 0xffffffff811f7b45, so I went back a while to find the previous instruction; the 20 bytes was to ensure that I got in sync with the instruction stream. So the processor was on an mwait instruction, basically waiting for something to happen. Not the cause of the exception. But the “called” function was nmi_calltrap, which indicates that the error had already happened.
How can that be? My best guess is a hardware problem.
|
Recovering from panic
|
Topic: technology, opinion | Link here |
OK, reboot teevee. Not surprisingly, I had to run fsck on /spool. Should I set journalling? Not now, let's get the machine up and running smoothly first. And there I ran into the problems that I had had earlier this month. I couldn't start X, and when I did it manually I ran into all sorts of problems with key remapping.
I thought I had fixed them. Yes, I had, but the fix only works under certain circumstances. I ran into a situation where xev told me I had the correct mappings (Ctrl to the left of the A key, for example), but on the xterm I was using, the same key functioned as CapsLock. How can that happen?
It seems that there's some timing issue with starting X and remapping the keys. Originally I had this in my .bashrc (abridged):
if [ "`tty`" = "/dev/ttyv0" ]; then
DISPLAY=:0 startx -- -listen tcp > startx.log 2>&1 &
...
But for some reason that didn't even start X, and startx.log was empty. So I had to start it manually (giving me the opportunity to miss the all-important -listen tcp).
In .xinitrc I had:
xmodmap ~/.xmodmap-$me
me is a variable that evaluates to the name of the system; I use the same .xinitrc for all my systems. But that, too, ran into trouble. I had thought that I had fixed that problem earlier this month, but it's still with me. It seems that if I invoke xmodmap from .xinitrc, I run into the problems that I had today. If I comment it out and run identically the same invocation from an xterm, it works. This has nothing to do with the contents of .xmodmap.teevee, which have been unchanged since 7 December 2002, when the file was called .xremap.
That's the current status. But clearly I need something better. I have a number of issues since upgrading to the latest and gratest:
| Tuesday, 31 March 2020 | Dereel | Images for 31 March 2020 |
| Top of page | ||
| previous day |
|
A covered riding arena?
|
Topic: Stones Road house, general, photography, opinion | Link here |
Chris Bahlo has a large covered riding arena:
|
|||||||||||||||||||||||
It's considerably bigger than her house:
|
|||||||||||||||||||||||
The riding area (at the back, not visible here) is 12×18 m in size, and that seems more than enough. Now Yvonne wants one too.
Damn, building the house was enough pain. I don't want anything to do with it. Let her arrange all the details. But of course we do need to discuss it. How many walls? Where? Where do we put the thing? Did a bit of walking around and discovered that her current arena would be the best place, and a 12×18 m structure would easily fit onto it. Walls to the wind? West and south? That would also fit well with the position of the arena, and leave some extra space to the east and north. But what about light? I've had some really terrible photo problems with arenas that were open on all sides. That wouldn't be the case here, but what about the light from the north? When the sun shines in the winter, up to half the floor would be in sunshine and the rest in the shade. Maybe an lean-to in front, like Chris has?
|
Firefox' answer to PDF
|
Topic: technology, opinion | Link here |
To my surprise, Yvonne doesn't display PDF documents. But now she has received some—via Facebook, of course—and she asked me how to look at them. OK, click on the link:
|
|
|||||||||
Now isn't that clever?
Use Chromium to display
the PDF. Why not just use Chromium? Well, I have a number of reasons, but wouldn't it be
more obvious to use a dedicated PDF viewer like xpdf? OK, can do that too. Just find the pathname of the executable by
climbing up from the directoryfolder from which X was started and then
climb down again into /usr/local/bin (assuming you know where it
is). SCREAM! How could such an approach have become “modern”?
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 |
