Twitter: @ewenmcneill -- July 2022

Thu Jun 30 20:40:01 +0000 2022 (#)

RT @kawaiiconNZ: Signal feed established on Twitter and YouTube: https://youtu.be/4hBLf2vQc8k

T-1 and counting šŸš€


Fri Jul 01 00:52:56 +0000 2022 (#)

Very much appreciating the live stream #kawaiicon (and itā€™s been a great selection of talks so far)

āœØšŸ’–āœØ https://twitter.com/kawaiiconNZ/status/1542600442739556352


Fri Jul 01 02:20:39 +0000 2022 (#)

RT @Yash25571056: According to a systematic review on the COVID-19 survivors, "Seventy-nine percent of COVID-19 survivors at 1 months andā€¦


Fri Jul 01 02:23:25 +0000 2022 (#)

RT @RealSexyCyborg: I mean if the smartest people you know get caught up in the whole social/economic pressure of being unmasked at times-ā€¦


Fri Jul 01 07:00:29 +0000 2022 (#)

RT @drjulie_b: So when asked for the scientific basis of their decisions around RAT test advice, @minhealthnz replies with mostly citationsā€¦


Fri Jul 01 08:52:05 +0000 2022 (#)

RT @Zeptobars: Which microchip protects the most property and human life? Automotive? Healthcare? I think this one could be the winner. DW0ā€¦


Fri Jul 01 21:57:49 +0000 2022 (#)

RT @kawaiiconNZ: Signal feed re-established for day 2: https://youtu.be/IyeDSyvYvZs


Sat Jul 02 00:30:47 +0000 2022 (#)

Fascinating (and encouraging) talk at #kawaiiconNZ by Patrick Hindmarsh about the rapid development of the NZ Covid Tracer app in 2020 (~2 weeks total from ā€œdecision to buildā€ to ā€œin production across NZā€).

Thanks for including the talk #kawaiicon2!

#kawaiicon


Sat Jul 02 01:34:45 +0000 2022 (#)

RT @andrewtychen: One of the unsung heroes of our response, @phindmarsh: the architect behind NZ COVID Tracer. His team had to make a LOT oā€¦


Sat Jul 02 01:37:49 +0000 2022 (#)

Replying to @andrewtychen

FTR for those that may not know the context: Patrickā€™s talk was submitted a year ago, for a conference that was going to be in late 2021ā€¦ which got delayed by Covid. (As best I can tell from the timeline, Patrick was at MoH for ~2 years ā€” ie a pretty typical IT job length.)


Sat Jul 02 01:51:55 +0000 2022 (#)

RT @andrewtychen: So if Kawaiicon can get 1500+ people through vaccine pass and ID checks (including having up-to-date passes) amongst a coā€¦


Sat Jul 02 01:54:41 +0000 2022 (#)

RT @jpdanner: We used to think we were underground hackers! @Metlstorm

#kawaiicon #kawaiiconnz @kawaiiconNZ


Mon Jul 04 01:44:44 +0000 2022 (#)

Replying to @minxdragon

Yes, this ā¬†ļø.

Iā€™m very ā€œIā€™m not even sure gender is a useful categorisation category, why is everyone so obsessed with itā€ these days.

(Unfortunately marketing people imagine gender is more meaningful to them than, eg, star sign, so I donā€™t see the obsession stopping soon šŸ˜¢)


Mon Jul 04 01:44:52 +0000 2022 (#)

RT @minxdragon: But, on the other hand, I think our society is entirely too fixated on gender and needs to stop. Everything is about genderā€¦


Mon Jul 04 02:13:15 +0000 2022 (#)

Replying to @aurynn

Covid-19 spread at the conference?

So far Iā€™ve only heard 1 positive case by lunchtime Friday, and ā€œa fewā€ by lunchtime Saturday. Which is (a) too soon to be conference spread, and (b) below NZ population case percentages (ie ~5 in 1500 is about .33% versus NZ at about 0.5-1%).


Mon Jul 04 06:26:11 +0000 2022 (#)

Replying to @BR3NDA, @__fincham, @plambrechtsen, @pepperraccoon, @kawaiiconNZ and @puck_

I too am curious if there was spread at the conference itself, as the conference itself seems to have been extremely cautiously run.

So far all cases Iā€™ve heard about could have arrived with Covid-19, then tested positive once RAT could pick it up (1-2 days after symptoms).


Mon Jul 04 06:29:07 +0000 2022 (#)

Replying to @BR3NDA, @__fincham, @plambrechtsen, @pepperraccoon, @kawaiiconNZ and @puck_

Iā€™d also expect Bluetooth pings from the conference could just be delayed uploads from early (eg Friday morning) ā€œarrived at conference with unknown Covid-19 then found out at conferenceā€ cases.

Iā€™d call it a win if conf case percentage is less than/same as NZ case percentage šŸ¤”


Mon Jul 04 06:37:39 +0000 2022 (#)

Replying to @Dram_Of_Sleep

FWIW there were announcements (Fri/Sat) that a few people at Kawaiicon had tested positive (likely arrived not knowing they had it).

Monday seems extremely likely to be ā€œMoH gave them Bluetooth data upload codeā€ day.

Hopefully masks/ventilation/positive cases leaving worked.


Mon Jul 04 06:42:39 +0000 2022 (#)

Replying to @__fincham, @BR3NDA, @plambrechtsen and @puck_

Interestingly the NZNOG conference seems to have gotten away with it, at least based on not hearing of any Covid-19 cases linked to/after it. Despiteā€¦ inconsistent mask use (and no enforcement), and inadequate ventilation. (Much smaller though: ~150 people.)


Mon Jul 04 07:09:08 +0000 2022 (#)

Replying to @puck_

Yeah, I ended up only going to half a tutorial of NZNOG in person, then deciding the risk was unnecessarily high and watching the remainder of the talks via live stream. (Iā€™d expected more precautions based on original NZNOG announcements than happened in practice.)


Mon Jul 04 07:27:09 +0000 2022 (#)

Replying to @tveastman

The right most box (of 5) is Orientgene; the ā€œ20 packā€ is Healgen. Despite their labelling (I have both in my house.)

AFAICT Healgen / Orientgene are the same except for the length of the swab stick/disposal bag, and the instructions. (Seem to be grouped together in studies.)


Mon Jul 04 07:29:43 +0000 2022 (#)

Replying to @tveastman

Also FWIW the ā€œ20 packā€ is ā€œwe ship 20 singles together to you at onceā€.

And 4 * $29.99 is $119.96. So 3 cents less than $199.99. Feels like a ā€œmarketing roundingā€ only difference to meā€¦ šŸ¤”


Mon Jul 04 09:54:35 +0000 2022 (#)

Replying to @__fincham

Iā€™d be interested to know what you think of the Dentec masks when they arrive. They seem to be highly recommended, but I couldnā€™t easily find a source shipping to NZ and the style seemed ā€œvisually bulkyā€. (So I ordered the GVS one first, which Iā€™ve ended up not using regularly.)


Tue Jul 05 00:26:26 +0000 2022 (#)

RT @kawaiiconNZ: Mōrena, we've been made aware of a few attendees testing +ve for COVID. Please ensure you are being responsible, test if yā€¦


Thu Jul 07 01:21:33 +0000 2022 (#)

Replying to @__fincham

I havenā€™t done much with my precursor device yet beyond updating its firmware a few times (theyā€™re close to a 1.0, but the factory shipped version was a 0.9x).

(Sorry I donā€™t think I want to sell mine. IIRC theyā€™re expecting more parts late 2022 for another production run.)


Thu Jul 07 01:22:43 +0000 2022 (#)

Replying to @__fincham

Beware the screen on the precursor is optimized for ā€œcan review electronicsā€ and is not particularly readable. (And later firmware changed to a tiny font which makes readability even harder.)


Thu Jul 07 01:29:20 +0000 2022 (#)

Replying to @__fincham

Yes itā€™s a LCD, and I think a memory LCD (for display hold while the rest is powered down).

Thereā€™s a backlight you can turn on, which helps a bit, but itā€™s still very ā€œ1990s LCDā€ like in readability. (Chosen for simplicity, not ā€œcompete with iPhone displayā€)


Thu Jul 07 01:33:45 +0000 2022 (#)

Replying to @__fincham

Iā€™m fairly sure Iā€™ve heard which LCD it is. But Iā€™m not finding a written document, so maybe it was part of one of the early 2020s talks about the precursor hardware? (LCA2020 talk as mostly software, CCC 2020 was mostly hardware.)


Thu Jul 07 04:52:10 +0000 2022 (#)

Random #Covid19NZ observation: supermarkets seem to once again be out/almost out of the ā€œgoodā€ paracetamol (ie well recognised brands). Just like Feb/Mar 2022, when NZ decided to let Covid-19 run free.

(Stocking was low but there the last couple of months, in between.)


Thu Jul 07 05:27:36 +0000 2022 (#)

Replying to @kawaiiconNZ

Do yā€™all have a rough idea of how many attendees have subsequently tested positive?

Iā€™m directly aware of 3 attendees testing positive Mon/Tue/Wed which seem linked. (Plus the ā€œoneā€ and ā€œa fewā€ during Con, who presumably caught it immediately pre-con and found out at Con.)


Thu Jul 07 06:52:09 +0000 2022 (#)

Replying to @tveastman and @kawaiiconNZ

Yes, Iā€™m expecting that most of the cases will correlate with ā€œdinner/drinksā€ type social events.

Given NZā€™s current ā€œreporting positiveā€ rates Iā€™m guessing NZ is at ā€œabout 1% of country per weekā€. So anything under 1% (~15 out of 1500) seems like ā€œcon practices reduced riskā€.


Thu Jul 07 10:01:34 +0000 2022 (#)

RT @RealSexyCyborg: All metal version of the Buterin-Wu Upper Room UVGI Fixture. This is an air sanitizing device designed for local manufaā€¦


Fri Jul 08 02:30:58 +0000 2022 (#)

RT @NewsroomNZ: Experts are concerned that the Ministry of Health's internal and public health advice isn't aligned with the latest scienceā€¦


Fri Jul 08 02:36:50 +0000 2022 (#)

TIL that the NZ MoH were concerned people would become complacent about #NZCovid19 if encouraged to wear good masks (P2, N95, etc).

Fortunately that didnā€™t happen.

Instead people became complacent (and 1500+ died) because of the MoH/government ā€œsheā€™ll be right, mateā€ attitude. https://twitter.com/NewsroomNZ/status/1545194510153654274


Sun Jul 10 21:31:05 +0000 2022 (#)

TIL thereā€™s a fairly direct design influence from the Datapoint 2200 (very early desktop computer) to the Intel 8008, and then on into the Intel 8080 / Zilog Z80, and then on into the Intel 8086 / 8088. And then into 40 years of PC Intel/AMD CPUs šŸ˜®

https://bread80.com/2022/07/09/comparing-datapoint-2200-8008-8080-and-z80-instruction-sets/


Sun Jul 10 21:35:25 +0000 2022 (#)

Replying to @ewenmcneill

The Datapoint 2200 was an early 1970s design. So weā€™re talking a 50 year lineage of CPU design influence here. Give or take some byte code ordering changes (8080 to 8086), itā€™s basically 50 years of ā€œextensionsā€ being added, repeatedly, to the same base. https://en.m.wikipedia.org/wiki/Datapoint_2200


Mon Jul 11 07:09:24 +0000 2022 (#)

RT @farmgeek: If we are serious about adapting to life with Covid - and not just by pretending it doesn't exist - changing our approach toā€¦


Mon Jul 11 07:16:25 +0000 2022 (#)

RT @TRyanGregory: A reminder:


Mon Jul 11 07:19:23 +0000 2022 (#)

Replying to @hroethgar

Men with beards cause ā€˜indecent thoughtsā€™ in other men too, because they donā€™t look like women.

Possibly thatā€™s better somehow? šŸ¤”


Mon Jul 11 21:36:40 +0000 2022 (#)

Replying to @__fincham

The problem in 2Q 2020 was the Government was making up new restrictions by the day, including dramatic reductions on who could visit anything in person.

So a pause in ā€œroutine operationsā€ for Level 4 made sense.

ā€œUntil after the pandemicā€ not so much šŸ˜ƒ


Mon Jul 11 21:37:30 +0000 2022 (#)

Replying to @nzgizmoguy and @__fincham

I mean yā€™all wouldnā€™t be the first to declare the pandemic overā€¦ :-)


Mon Jul 11 21:42:47 +0000 2022 (#)

Replying to @__fincham

If itā€™s hardware RAID, thatā€™s the RAID cardā€™s problem.

If itā€™s, eg, Linux MD, then RAID 1 with one of the more recent MD RAID set types that puts the RAID metadata at the end of the partition seems to work. (EFI partition is always separate due to file system required.)


Mon Jul 11 21:58:38 +0000 2022 (#)

Replying to @__fincham

Now I check it turns out the first two MD RAID metadata formats are at the end of the partition; the newest ones are at/near the start.

But if you pick the right one (1.0 from a quick glance) the rest of the partition looks normalā€ to anything else. https://raid.wiki.kernel.org/index.php/RAID_superblock_formats


Tue Jul 12 00:11:55 +0000 2022 (#)

Replying to @__fincham

I think the new style MD RAID metadata is better for most use cases (eg can store resync bitmap).

But for the tiny /boot and EFI partitions I use the 1.0 (end of partition metadata) for convenience. Because 1GB doesnā€™t take long to resync anyway :-)


Tue Jul 12 00:37:04 +0000 2022 (#)

šŸ˜¢

What if I told you the air inside the bus was more polluted (in ways that matter to humans) in NZ, than the air outside the buses? šŸ¤”

Prediction: Weā€™re going to have 5+ years of arguing about the need for better indoor air treatment. https://www.rnz.co.nz/news/national/470758/cost-concerns-scupper-fresh-air-plans-for-auckland-buses


Tue Jul 12 00:39:29 +0000 2022 (#)

Replying to @ewenmcneill

And FWIW buses drawing in outside air from up high, not directly at the front, ought to bring in very little vehicle exhaust.

(I guess one bonus of the 20 year switchover to electric vehicles is ā€œoutside air is pollutedā€ will reduce as an excuse not to bring in fresh air.)


Tue Jul 12 06:13:00 +0000 2022 (#)

Replying to @__fincham

Caps lock as Alt? Or even the ā€œwindowsā€ key as a Super or Hyper or similar? šŸ¤”

(I too find switching between macOS and Linux keybindings a bit jarring. Especially ctrl-shift-C/V in terminals.)


Tue Jul 12 22:43:41 +0000 2022 (#)

Well thatā€™s disappointing: @ContactEnergy are now sending out unsolicited marketing emails, with no visible unsubscribe link, to addresses that explicitly opted out of all their junk email (on signup and again later). In breach of their own Privacy Policy, and NZ law.


Tue Jul 12 22:48:21 +0000 2022 (#)

Replying to @ewenmcneill

Apparently (according to the email auto-reply when I complained) Iā€™m supposed to join Facebook (or WhatsApp, owned by Facebook) to communicate with Contact. Yeah, nah.

Maybe they could just keep their unsolicited/unwanted marketing to FaceBook/WhatsApp instead.


Wed Jul 13 02:01:24 +0000 2022 (#)

RT @RealSexyCyborg: "Encouraging" businesses to spend money on ventilation will work about as well as "encouraging" masks has. Health codesā€¦


Thu Jul 14 07:04:03 +0000 2022 (#)

šŸ§µā¬†ļøā¬‡ļø speaking truth to power (in this case Engineers (tm))

Naomi has single handedly done more to advance the practice of practical UV air treatment than anyone else I know of. On a budget scraped together from scraps.

(A+ cultural reference too, in her third language.) https://twitter.com/RealSexyCyborg/status/1547461068301037570


Thu Jul 14 07:04:14 +0000 2022 (#)

RT @RealSexyCyborg: I'm going through the Chinese Far-UVC patents with a Chinese IP lawyer seeing which we have wiggle room on, I'm on WeChā€¦


Fri Jul 15 05:39:14 +0000 2022 (#)

Replying to @IanColdwater

Alternatively there are several ā€œear saverā€ designs that will take ā€œover earā€ mask straps and connect them behind the head, over significant amounts of hair. (Random examples chosen by hair of model shown.) https://www.thingiverse.com/thing:4588608 https://www.thingiverse.com/thing:4266319 https://www.thingiverse.com/thing:4348611


Sat Jul 16 02:14:45 +0000 2022 (#)

ā€œBuilding a haptic input knob from scratchā€

Itā€™s a very clever design, unfortunately currently around a custom motor design (with hollow inner core) that is no longer available. (Theyā€™re trying to find an alternative motor.) https://m.youtube.com/watch?v=Q76dMggUH1M https://github.com/scottbez1/smartknob/


Sat Jul 16 08:00:38 +0000 2022 (#)

Replying to @nickzoic

Possibly. Iā€™m not sure Iā€™ve seen a ā€œpcb motorā€, to know how itā€™s constructed.

The key requirement here is to have an empty centre of the motor through which to pass cables from bottom to top, so they donā€™t get twisted (without needing a slip ring).


Sat Jul 16 09:13:02 +0000 2022 (#)

Replying to @nickzoic

That definitely looks similar to the motor design they were using. Theirs had a bearing at the centre, and a hole in the permanent magnet rotor as well. But the flat ring of coils seems like the same design idea.

Thanks for the picture, Iā€™d not seen those motors before.


Mon Jul 18 03:40:35 +0000 2022 (#)

TIL that Ansible does its own ā€œsplit on spacesā€ parsing of variables on the ansible-playbook command line, instead of just insisting one variable per ā€œ-eā€ flag.

So you have to protect spaces in variables twice: from the shell and Ansible šŸ¤¦šŸ»ā€ā™‚ļø https://docs.ansible.com/ansible/latest/user_guide/playbooks_variables.html#defining-variables-at-runtime


Mon Jul 18 04:06:49 +0000 2022 (#)

Bonus points to Dell for attempting to PGP sign their Dell Networking OS10 release.

It would however be somewhat more convincing if the key they used hadnā€™t expired 2 days after the signatureā€¦ :-) https://keyserver.ubuntu.com/pks/lookup?search=0x47CB9029&fingerprint=on&op=index


Tue Jul 19 06:54:21 +0000 2022 (#)

RT @RealSexyCyborg: Public health cannot be managed on the basis of individual choice so is by definition a medical dictatorship, yes. Wheā€¦


Wed Jul 20 04:18:16 +0000 2022 (#)

RT @farmgeek: Our Omicron wave in March resulted in more than a million reported Covid cases. The BA4/5 wave we are seeing now - four monthā€¦


Wed Jul 20 04:24:03 +0000 2022 (#)

GoodšŸ§µā¬†ļøā¬‡ļø

The ā€œthree wavesā€ in NZ this year (2022) have all smeared into one 6+ month ongoing wave AFAICT. Weā€™ve never dropped below about 40k-50k recorded #NZCovid19 cases per week (which is ~1% of population per week); now back up to 60k-70k per week. 1/ https://twitter.com/farmgeek/status/1549601385045688322


Wed Jul 20 04:26:10 +0000 2022 (#)

Replying to @ewenmcneill

Since NZ has removed basically all #NZCovid19 precautions which allowed 2020 and 2021 NZ to be safer than a lot of the world, and not implemented any new precautions (eg indoor air treatment), weā€™re now getting new variants in real time, before finishing the old variants šŸ˜¢ 2/2


Wed Jul 20 23:41:38 +0000 2022 (#)

Replying to @ewenmcneill

Additional Ansible Network Device Configuration surprise: at least some of the DEVICE_config modules fetch the entire running config for every single task step. Which isā€¦. slooooowwww.

(They appear to get all of it even when given a parent section šŸ˜¢) https://github.com/ansible-collections/dellemc.os10/blob/master/docs/dellemc.os10.os10_config_module.rst


Wed Jul 20 23:45:41 +0000 2022 (#)

Replying to @ewenmcneill

With some care one can avoid a lot of this ā€œgather facts (ie config) every stepā€ slowness by running DEVICE_facts early on and then passing in:

config: ā€œ{{ansible_net_config}}ā€

(Beware change generation is against stale cached config, so you have to know no relevant changes.)


Thu Jul 21 03:15:10 +0000 2022 (#)

TIL that Dell Networking OS10 (Snnnn switches) considers ā€œ,ā€ (ie a comma) to be a special special character in interface descriptions. You have to quote it and escape it with two backslashes on input šŸ˜®

Otherwise CA_STATUS_ERROR or silent truncation šŸ¤¦šŸ»ā€ā™‚ļø https://www.dell.com/support/manuals/en-nz/dell-emc-smartfabric-os10/smartfabric-os-user-guide-10-5-4/description-(interface)?guid=guid-f4caecbb-ca4a-4e54-95fe-932b81c18d5e&lang=en-us


Thu Jul 21 03:18:55 +0000 2022 (#)

Replying to @ewenmcneill

Is the interface description with an embedded comma (,) displayed with double backslashes? No of course it isnā€™t šŸ˜”

I donā€™t think I want to know how broken the config store is that a quoted comma breaks the input parsing, to need escaping twice šŸ˜¬


Thu Jul 21 23:59:15 +0000 2022 (#)

Something tells me that this web page load could be fasterā€¦ šŸ¤”

Weirdly it seems to be around 1 minute wait per page of manual from this Enterprise IT Infrastructure vendor, but every component in the page loads in milliseconds once transfer starts.

Per page šŸ˜¢


Fri Jul 22 00:45:51 +0000 2022 (#)

RT @farmgeek: In case anyone has forgotten: Dr Wiles has literally helped save thousands of lives worldwide through her Covid science commuā€¦


Fri Jul 22 00:48:02 +0000 2022 (#)

Replying to @hroethgar

I just admit Iā€™ve started wondering Iā€™d thereā€™s an IBM 360 frantically scrubbing through magtape to find the relevant page contentā€¦ šŸ˜ƒ


Fri Jul 22 07:26:01 +0000 2022 (#)

Replying to @xssfox

ā€œNormal peopleā€ using computers are more accepting of ā€œI guess it doesnā€™t workā€, and give up much sooner. Itā€™s probably less stressful that way šŸ¤”

(The rest of the thread looks very SNAFU šŸ˜”)


Fri Jul 22 09:32:39 +0000 2022 (#)

RT @slyall: Video of presentation by @DrDreHistorian at Auckland Library earlier this month is now up. Would recommend and checkout his booā€¦


Sun Jul 24 02:49:37 +0000 2022 (#)

RT @atoonk: Remember that giant country-wide outage by @Rogers Canada a few weeks back? Rogers has now released some more insights into whā€¦


Mon Jul 25 06:08:41 +0000 2022 (#)

Oh good, switch model S from vendor D defaults 25GBase-LR interfaces FEC to CL108-RS (good), but switch model N from vendor D defaults 25GBase-LR interfaces FEC to ā€œFEC disabledā€ šŸ˜”

Much testing, very interop.

(Forcing speed, duplex, and FEC on model N brings the link up.)


Mon Jul 25 10:12:12 +0000 2022 (#)

Replying to @ExcitedLeigh

Yeah, itā€™s gestures at the world not great, for lots of people šŸ˜¢

My suggestion, in the short term, would be to try to find a ā€œregular thingā€ per day (evening) you can do with others. Maybe 90%+ online for now. So thereā€™s always ā€œanother thing coming soonā€.


Mon Jul 25 10:15:17 +0000 2022 (#)

Replying to @ExcitedLeigh

Because my past experience is that one thing a month or even one thing a week isnā€™t enough to ā€œpay back the loneliness debtā€ā€¦ and trying to make it be enough just increases the ā€œpost event dropā€.

(Eventually 1 thing per day will probably be too much; thatā€™s a good problem.)


Tue Jul 26 01:00:32 +0000 2022 (#)

Currently debugging why Ansible is returning the running config of the exactly one switch as an err rather than stdout. Thereā€™s some kind of exception; AFAICT never actually logged anywhere.

TIL that JSONRPC has a magic value (-32603) for ā€œInternal Errorā€. https://twitter.com/ewenmcneill/status/1550381638034399232


Tue Jul 26 02:31:55 +0000 2022 (#)

Replying to @ewenmcneill

Some hours later I have my answer for why Ansible ā€œshow running configā€ is failing on one switch:

forward-error-correction enable-cl108

Includes the word ā€œerrorā€ with word boundaries (here ā€œ-ā€œ).

Therefore the command ā€œfailedā€ šŸ¤¦šŸ»ā€ā™‚ļø https://twitter.com/ewenmcneill/status/1551449338026803200


Tue Jul 26 02:34:25 +0000 2022 (#)

Replying to @ewenmcneill

Unfortunately I need the line:

forward-error-correction enable-cl108

in all the N series configs (so the 25G interfaces have matching FEC and come up: the problem I started the day with šŸ˜”), so Ansible treating that config line as a fatal error isā€¦ problematic šŸ˜¢


Tue Jul 26 02:56:51 +0000 2022 (#)

Replying to @ewenmcneill

Bug report filed with vendor Ansible collection. I doubt theyā€™ll fix it (there are year old trivial, needed, PRs that are not merged).

But at least I have something to refer to in my workaround documentation. https://github.com/ansible-collections/dellemc.os6/issues/52


Tue Jul 26 03:31:09 +0000 2022 (#)

Replying to @ewenmcneill

Unsurprisingly it turns out this is a common Ansible netcommon issue (it usually causes more obvious task failure).

One can set ā€œansible_terminal_stderr_reā€ with an override set of regexes (no re.I on the ā€œErrorā€ regex works here). https://docs.ansible.com/ansible/devel/network/user_guide/network_debug_troubleshooting.html#task-failure-due-to-mismatched-error-regex-within-command-response-using-ansible-netcommon-network-cli-connection-type https://github.com/ansible-collections/dellemc.os6/issues/52#issuecomment-1194952145


Tue Jul 26 07:02:53 +0000 2022 (#)

Replying to @BR3NDA

1000 characters plus ā€œbottom quoting all messages back to the Big Bangā€? (Thatā€™s my pet peeve with ā€œmodernā€ email.)

TBH I suspect a 1000 character limit would just lead to 1/35 threads for ā€œregularā€ text (not just kernel patches). Or ā€œsee attached .docxā€.


Wed Jul 27 07:15:23 +0000 2022 (#)

Replying to @tveastman and @__fincham

Yes, Iā€™ve seen some really good cosplay style painting onto the Dentec Safety white masks. Looks stunning when thatā€™s done, and like a fashion piece rather than a safety piece per se.


Wed Jul 27 07:20:34 +0000 2022 (#)

Replying to @__fincham

Good news that it fits you well :-)

Whatā€™s the fit like under your chin? Eg how far back under the chin does it go?

Also how do you find ā€œworkingā€ while wearing the mask? Especially sight lines, down below head height? (Eg does it require head tilt down, eye tilt up?)


Wed Jul 27 07:24:23 +0000 2022 (#)

Replying to @plambrechtsen, @__fincham and @aurynn

FWIW there are a few other very new (2022, USA) mask styles optimised especially for talking through (eg large front facing HEPA filter).

General consensus seems to be internal microphones donā€™t survive long in moisture of source controlled mask, FWIW.


Wed Jul 27 07:29:17 +0000 2022 (#)

Replying to @__fincham

Thanks! From those photos it looks like it goes about twice as far under the chin as the GVS Eclipse. Which is good as ā€œbarely under chinā€ was my biggest fit issue with the Eclipse.

(Also good news on the price: cheaper than I expected.)


Thu Jul 28 02:10:49 +0000 2022 (#)

Replying to @__fincham

Most businesses Iā€™ve seen tend to use, eg, corp.PUBLICDOMAIN as their internal DNS name (or their AD domain, which is often like that anyway).

Iā€™m not sure what the use case is for ā€œwell enough run to want internal DNS, not well enough run to own any domain namesā€?šŸ¤”


Thu Jul 28 02:16:40 +0000 2022 (#)

Replying to @__fincham

For businesses multiple uses of the same ā€œcannot be delegatedā€ name tends to cause problems when the business merge (at least as much as the ā€œmodernā€ site local IP reuse).

So TBH Iā€™d suggest businesses own at least one delegated globally unique name anyway. Itā€™s ~$30/year.


Thu Jul 28 02:34:27 +0000 2022 (#)

Replying to @tveastman

The older I get the more Iā€™m astonished that the Y2K issues just got fixed, almost everywhere, in time (so weā€™ll that ā€œnothingā€ happened in 2000).

At the time it seemed obvious this is what society would do even faced with a major problem. Now, not so much.


Thu Jul 28 06:53:34 +0000 2022 (#)

Replying to @tveastman and @nickzoic

šŸ˜¢

(Sadly, having read Collapse by Jared Diamond, I fear you are right.)


Thu Jul 28 09:42:14 +0000 2022 (#)

This talk from VCF East 2022 about reverse engineering the Commodore 128 is an amazing tale of yak shaving, constantly peppered with comments from Bil Herd (in the audience) who designed the Commodore 128.

RE: schematics, PCB, PLA, etc. https://www.youtube.com/watch?v=hMMiyUGEygE


Thu Jul 28 09:45:29 +0000 2022 (#)

@unnamed_show Possibly yā€™all would like to talk to Johan Grip who spoke at VCF East 2022 about reverse engineering the Commodore 128 (PCB, schematics, PLA back to terms then verilog, etc)?

(Also highly recommend the talk video if you havenā€™t watched it.) https://twitter.com/ewenmcneill/status/1552590244058648577


Thu Jul 28 22:02:26 +0000 2022 (#)

Replying to @tveastman

Install the missed package updates from the Ubuntu old-releases mirror, then swap back to doing your ā€œlateā€ upgrade?

(FWIW the upgrader ought to tell you to do that, and ā€œno help for youā€ isā€¦ poor UX.) https://old-releases.ubuntu.com/releases/


Thu Jul 28 22:07:33 +0000 2022 (#)

Replying to @tveastman

Yeah old-releases (and the Debian equivalent which was first) is one of those ā€œeverybody knowsā€ (ā€œnobody knowsā€) things šŸ˜¢

TBH the do-release-upgrade ought to just install the updates you missed from the current version for you before doing the restā€¦. šŸ¤”


Thu Jul 28 22:18:57 +0000 2022 (#)

Replying to @tveastman

The deletion off the main mirrors does seem a bit early to me. The ā€œunsupported version, upgrade nowā€ should probably be there for at least a month extra, to make upgrades easy.

IIRC Impish was ā€œ9 monthsā€ from October 2021, which is late June 2022 at earliest.


Thu Jul 28 22:26:15 +0000 2022 (#)

Replying to @tveastman

FTR I was meaning that it should have been on the archives for ā€œ9+1ā€ months (so 10) from release. Which means into early August in practice.

Feels like they pulled it at least a week early. (And do-release-upgrade should have worked around it being off main mirrors anyway.)


Thu Jul 28 22:31:43 +0000 2022 (#)

Replying to @tveastman

Disk space on mirrors is in fairly short supply (if they have more disk they mirror more distros :-) )

But yes, do-release-upgrade should magically paper over old versions being dropped from regular mirrors. ā€œNot found, good luck to youā€ is terrible UX.


Fri Jul 29 01:30:05 +0000 2022 (#)

RT @mipsytipsy: I got my stairs re-carpeted this week! šŸ’œšŸ’™šŸ’ššŸ’›šŸ§”ā¤ļøšŸ’–


Fri Jul 29 01:56:10 +0000 2022 (#)

RT @simonw: This is kind of neat: it turns out GitHub wikis are just Git repositories, which means you can migrate the contents AND historyā€¦


Sat Jul 30 04:26:03 +0000 2022 (#)

MoH frequently adjusting the #NZCovid19 metrics reported, and the frequency of reporting them, just says to me ā€œitā€™s embarrassingly badā€ and that weā€™re now well into ā€œoptimise for the numbers reportedā€ over ā€œoptimise for good health outcomesā€ šŸ˜¢ https://twitter.com/farmgeek/status/1553195694709649408


Sat Jul 30 04:29:51 +0000 2022 (#)

Replying to @ewenmcneill

Unfortunately every country which has allowed a large portion of its population get infected with #NZCovid19, resulting in a ā€œwe didnā€™t expect it to be that highā€ number of deaths, ends upā€¦. massaging the statistics to not look quite as bad.

Seems NZ has joined that club šŸ˜”


Sat Jul 30 05:29:47 +0000 2022 (#)

Replying to @__fincham and @kyhwana

Making hypochlorous acid and putting it in a spray bottle seems like the suggested approach.

QT is a Naomi Wu thread from yesterday; seems like care/testing is required when making as itā€™s easy to accidentally get other byproducts (including bleach). https://twitter.com/RealSexyCyborg/status/1553024174192439298


Sat Jul 30 23:26:18 +0000 2022 (#)

Replying to @eey0re

Yes, I expect ā€œitā€™s a lot of workā€ explains why thereā€™s now no weekend reporting.

But the increases in effort required to figure out whether a death is ā€œa #NZCovid19 deathā€ that counts is the opposite of ā€œdoing less work because the Government doesnā€™t care as muchā€ :-/