@StephenKitt i used to file bugs against packages for that, long ago but I gave up. lots of packages have descs like "dev package for libfoo", "shared lib package for libfoo", "binaries package for foo", and i'd think "that's obvious, but WTF is foo and why might i want to install it? those descriptions are nothing more than a re-statement of the package names"
@FaheemMitha i submitted a bug report for pdsh in September with the new home page for pdsh (which is little more than a link to the github page now...the old LLNL page used to have docs and usage examples as well as the source - I got tired of not being able to link to it when I posted an answer using pdsh, so I hunted for the new homepage). The package hasn't been updated yet.
I'm facing an issue here where I can't use apt-get, any arguments will fail such as apt-get install subversion or apt-get update.
I understand this problem may arise from the proxy configuration. However I believe I've set it up properly from the settings window and other internet functionalit...
@Tim I agree with Wildcard’s characterisation. There are two other big considerations in my mind: some languages don’t prescribe any way of handling errors, and others allow multiple different approaches.
Thus in C for example there are patterns of error transmission (not handling), as discussed earlier in the chat; in some cases the way to figure out if an error occurred is to set errno to 0 and check its value afterwards, because some functions have no way of indicating that an error occurred.
In Java, you get exceptions, checked and unchecked. What happens in practice is that most projects end up handling checked exceptions (which the compiler forces you to deal with) by re-throwing them wrapped in unchecked exceptions, so you either track everything manually or you ignore most errors.
What’s nice about Go is that error handling is explicit and expected, i.e. it’s defined technically and also part of the idiom, or even philosophy of the language. That results in well-designed Go-idiomatic libraries (or rather, frameworks) taking the existence of errors into account. For example, the Kubernetes controller paradigm defines how errors should be handled, which means developers know what to do.
Handling errors is really the same as backup though: backup is not much use if you don’t check restores, and error-handling isn’t much use if you don’t know what to do with the errors. The best Java project error-handling I know is in a set of projects which are explicitly designed to feed into an Elasticsearch infrastructure which is connected to the monitoring system (common enough) but also the issue tracker.
That means that errors which appear in the logs can easily be turned into bug reports, so they are, and problems are fixed.
That also goes with a development culture in the company where logging useful errors is the norm, and you’re in more trouble if you omitted logs which would have helped in a production event, than if you log too much — thanks to the good logging infrastructure which means that human involvement is minimised and logs are actually usable.
And that too has to be baked into the language from the beginning, otherwise you end up with things like Java’s streams and futures which will get you some of the way and then leave you high and dry, so you write a mixture of nice functional-style code and then a bunch of explicit error-handling (or just sweep it under the rug).
@JeffSchaller Good. I've never used AskUbuntu nor their chat, so I don't know the culture around there, but that would probably be the correct place for bringing attention to unloved AU questions (their Meta site could possibly also be useful).
@JeffSchaller Firstly, I have no idea... the purpose of retrieving public key. Am novice in apt package management. I have no idea, why there are so many commands: apt-get, apt-update.... So, I don't understand your comment on fetching keys fromm key server
@overexchange I could have been clearer in my comment, but I posted it as a suggestion to follow the Docker documentation for installation. It was only a comment, and may not be what you need to do.
well, two suggestions. First, notice that the error is from gpg and it says "no data" -- this isn't "package management" yet, just "apt-key" failing because of a gpg failure because of something else. If you suspect networking, then it'd be helpful in the question if you gave a quick overview of the network setup and whether other operations work or not. Are there firewalls between you and the keyservers? A proxy? -- all good information for the question
It works.. if I run apt-key adv --keyserver hkp://p80.pool.sks-keyservers.net:80 --recv-keys 58118E89F3A912897C070ADBF76221572C52609D on my ubuntu machine
$ apt-key adv --keyserver hkp://p80.pool.sks-keyservers.net:80 --recv-keys 58118E89F3A912897C070ADBF76221572C52609D
Executing: /tmp/apt-key-gpghome.lHZ4HKJQps/gpg.1.sh --keyserver hkp://p80.pool.sks-keyservers.net:80 --recv-keys 58118E89F3A912897C070ADBF76221572C52609D
gpg: key F76221572C52609D: 7 signatures not checked due to missing keys
gpg: key F76221572C52609D: public key "Docker Release Tool (releasedocker) <[email protected]>" imported
gpg: Total number processed: 1
gpg: imported: 1
so this is not working when I run amidst building image
https://forums.docker.com/t/gpg-key-for-docker-repo-fail-to-fetch-from-key-server/24253/2 says: "Problem was solved via: Changing the network settings inside VirtualBox from: NAT to Bridged adapter!
It’s definitely a problem with the network configuration."
since you're creating a docker container, then maybe the problem is with your docker networking? Also good to add the above and the docker n/w info to your Q
As am not good technically in this area... I would better delete the question
@JeffSchaller am I suppose to raise query in U&L for below issue:
The following packages have unmet dependencies:
docker-engine : Depends: libsystemd-journal0 (>= 201) but it is not installable
Recommends: aufs-tools but it is not going to be installed
Recommends: cgroupfs-mount but it is not going to be installed or
cgroup-lite but it is not installable
Recommends: apparmor but it is not going to be installed
Recommends: yubico-piv-tool (>= 1.1.0~) but it is not going to be installed
@overexchange certainly it'd be on-topic, but I'd strongly recommend you explain your environment -- including OS version and whether (as I suspect) this is a Docker setup; if so, then the dockerfile
simply searching for one of those messages leads me to github.com/moby/moby/issues/20698 which has several suggestions, so your question could also include any relevant information along those lines
funnily enough, another result is from Ask Ubuntu, whose accepted answer is a link-only answer pointing to a Digial Ocean page :(
I'm not an expert here, but when I follow my nose in Docker land from "FROM jenkins:2.60.3", I go through openjdk:8-jdk-stretch eventually to "debian:stretch" -- not at an Ubuntu base. But again, maybe I've misstepped somewhere.
@Jesse_b Oh, she's doing just fine. Going through a bit of a tired patch at the moment, but the other day she was cooking food and making a chocolate cake. And she's been rearranging things (not furniture) in the apartment. All from the comfort of her Permobil.
@MichaelHomer wow, so just make a second profile, build it up to enough rep to cast close votes (but not enough to get gold tag badges), and then you can cast your close votes instead of directly closing questions...seems kinda convoluted. :)
@MichaelHomer Yeah, that's allowed as long as there is no sock-puppet voting happening between the accounts. I don't really like it though, it a bit schizophrenic, and somewhat confusing.
Well I don't think Stéphane would do it and those accounts are both named the same with the same picture to make it easy enough to see but other people with malicious intent would likely make it more difficult
I guess there really couldn't be anything in place to check though except for the extreme case where they upvote so many questions the system automatically overturns it. They could just use separate IP addresses, and upvote multiple users, including themselves, doing such conservatively enough to not get flagged
@Jesse_b We can see voting patterns to some degree, but we can't see individual votes. Yes, there are automatic checks for catching some things. I'm not sure it's based on IP addresses.
You basically have to put more effort into cheating than to answer/ask questions to successfully fool the system.
Some people are in fact murderers, and other people wouldn't consider themselves murderers but would easily hack into critical systems that could cause life or injury but would never directly harm a person. They get a detachment by so called "cyber" crime
If murderers learned that you can easily manipulate self driving cars to kill people they would likely do that
and therefore you can't just ignore bugs based on the fact that people are good
Kids drop bricks on the highway killing people all the time
@Jesse_b I honestly doubt it. Most murders are...there's some specific word for it which I forget, but basically most murders are committed as spur of the moment actions in anger.
The 2017 Interstate 75 Rock-Throwing Incidents were a series of Rock and sandbag throwing incidents from two highway overpasses by teenagers along Interstate 75 caused two deaths in separate incidents in 2017: one in Michigan and one in Ohio.
== Michigan rock throwing death ==
Kenneth White, 32, was killed on October 18, 2017 when a 6-pound (2.7 kg) rock thrown by a group of five teenage vandals among which was Mikadyn Payne, crashed through the windshield of the van he was riding in on Interstate 75 in Michigan, in Vienna Township, 80 miles (130 km) north of Detroit. Other cars were also damaged...
It's much easier to go to an empty road in the middle of the night and completely detach yourself from what you're actually doing while you're simply painting a line in the road
When something goes wrong automatically, it goes wrong ALL OVER.
@Jesse_b that's true.
@Jesse_b yes, but my stance (not the comic's point) is that human drivers are MUCH more likely to drive safely in spite of misdrawn lines than self-driving cars are.
One time in Mexico City I looked out the window from the second floor and saw something wild. There was a relatively short one-way street that was about 8 lanes wide, between two two-way streets (i.e. between two "T" intersections).
I think in order for self driving cars to at least be semi safe they have to use more than just radar. All cars will have to emit wireless signals that broadcast as much information as possible including rate and direction of travel, announcement that brakes will be applied, announcements that steering will be adjusted, alarms that there are obstacles ahead or behind, etc
@Jesse_b last I read, AI is incapable of crossing a busy four-way stop sign intersection.
Because nobody actually follows the right of way laws about that. There is communication and agreement between people, and we make it work. Practically using telepathy.
@Jesse_b bingo, exactly. Which requires human intelligence and judgment.
The zigzagging lane lines were a real eye opener for me. Very funny, but also interesting.
@Jesse_b I also think self-driving cars are solving the wrong problem.
Fundamentally, the "API" of our traffic laws was 100% designed around human understanding and judgment, and was never designed to be automatable.
Trying to set up a bunch of machine drivers that could still safely interact with human drivers, regardless of the ratio of human drivers to machine drivers, would require a very different setup.
@Wildcard Well, part of what they do is safely stopping a car when the driver is, for whatever reason, unable to maintain in control of the car (heart-attack, falling asleep, etc.) This I think in a good thing.
Way better than planes OR cars for city-to-city travel.
@Kusalananda to some degree, yes. Buses are fundamentally similar to cars, though. Trains really aren't—they have a track and they have to travel on the track.
Actually, though, "free" local transportation in a downtown area CAN pay for itself in terms of increased foot traffic and therefore patronage of local stores.
the outside of subway trains has little valuable real estate for advertisements because it is rarely ever seen and the internet has greatly reduced the cost of advertisements so I would imagine they could sell every square inch of the train and it would still be less than daily ticket revenue
So taxation doesn't make it free it just means you pay for it once a year, and everyone who doesn't use it also pays for it
Recent "mobs" in NYC subway and in Chile subway were because subways are not free and people can't afford them and get arrested when jumping turnstile or asking for and receiving free swipe
@Tim why? The money, politically, all goes into the same pot. So spending taxpayer dollars on one thing is in direct competition with spending it on something else which is arguably more valuable.
The most direct solution to this is to stop spending taxpayer dollars, and let people direct their funds where they choose.
@Kusalananda the basic contention with government-funded anything is that there are people in favor and people against, but nobody gets to choose not to fund whatever-it-is.
@Wildcard NYC invests a lot of money to install turnstiles and other equipments. Even the ATMs don't work properly. NYC also invest a lot in hiring cops to capture fare evasion. All this money can be used to fund free subway transportation
@Tim you're assuming that the common goal of the people administering the NYC subway is to have a clean, efficient, workable mode of public transportation.
@Wildcard The inverse is that poor people don't have a choice in what they need to spend money on, and so lots of services quickly become unaffordable.
Such as schools and hospitals, and transportation.
Rather than e.g. to get re-elected, or to make money, or to get everybody's agreement that the subway is underfunded and that's why it's running so bad, or just to do as little work as possible.
Governments do not have any incentive to keep the quality high. Once the taxpayers are locked into funding something, there is no actual penalty for doing a bad job at providing it.
@Kusalananda not so. Schools and hospitals are HIGHLY regulated with insane contradictory regulations. This, in combination with the insulation from price on the part of the consumer, leads to skyrocketing prices.
Right. I should not get involved with discussing US government funding, because I know nothing about it. I'm just a Swede brought up with taxed-paid schools and hospitals.
I, for example, pay maximum of about USD 200 per year for medication, regardless of that medication is. I know a guy who has leukaemia, his pills cost something like USD 150 per month (if not more). He still pays a maximum of USD 200 a year for them though. I'm quite happy to let my taxes go to that.
@Tim they may have access to more numeric details than you do. But the conclusions are so very unworkable, that it is not true to say that they do better analysis than you can. I'm sort of complimenting you. ;)
@Tim I never mentioned economists. I mentioned economics.
@Tim and they can get wrong conclusions together, too.
@Kusalananda right. In the U.S., colleges and universities hike up their tuition to astronomical figures far in excess of the actual value of the "education" they deliver. If students had to pay up front they would never do it. Instead, they get "student loans", which can take decades to pay off, and in the frequent cases where the student never does pay off the loan, the taxpayers foot the bill.
@Wildcard sorry for speaking to you without softening my tone
@Kusalananda Congratulations! Do you have to pay the medical expenses for your diabetic cat? Just curious to what extent the medicare for all equivalence in Sweden can cover.
@Tim this is wildly inaccurate and places a greater importance on the use of emotionally charged language than it does on reaching rational, workable conclusions that can be applied to the benefit of any part of the society.
In other words, you are using rhetoric to disguise (or justify) the lack of workability of the economic solutions you are proposing.
@Wildcard I don't do the analysis myself. But I think reducing military expenses and taxing the rich more will fund a lot of things that you can't imagine
@Tim I agree with reducing war expenses. As for "taxing the rich more," it might sound emotionally appealing as a proposal, but when implemented it doesn't have any of the intended effects.
(That point is debatable, of course, since you haven't stated what the intended effects actually are.)
@Tim this is wildly inaccurate and places a greater importance on the use of emotionally charged language than it does on reaching rational, workable conclusions that can be applied to the benefit of any part of the society.