Am I late to notice some personnel change?

In Linux, signal handling is probably not popular any more?

But is the mechanisms for dealing with safety issue in signal handling exactly the same as in event handling in some programming languages?

Does event handling provided in programming languages not have to deal with the same safety issue?

Anyone knows about it? or if it is a good question on stack overflow in Russian?

Does no one know about event handling and signal handling?

or can post my question on stack overflow on behalf of me?

Tim, this is not the place to air your concerns about Stack Overflow.

Please don't bring irrelevant topic here

If you know about event handling and signal handling or can post a question on behalf of me, then thank you for doing so.

I don't think it's appropriate to evade a question ban, either.

That sounds like conservative talking point

The mods when Tim is on a roll (╯°□°)╯︵ ┻━┻

By safety issue, I mean reentrance, thread-safety

It is introduced in CS:APP 12.7

My suspicion is that event handling in high level language should face the same issue, but I don't know

Anyone with a happy experience in a higher education institution or a workplace knows about it?

safety third

@JeffSchaller I have the stickers!

there. are. stickers?!?

stickerthepla.net/#001

and t-shirts, and fridge magnets, and ... (but I only have stickers)

nice! I should a S.W.E.A.T. pledge while I'm at it; nice messages

Wait, you're telling me you guys agree with that idea? That we should put "the business of getting the job done" and making money above safety?

@terdon there's wayy too many clicks to suspend people. I just think we could do things a little better around here ;)

@FaheemMitha see that message. It implicitly (in context and after prev. message) echoes a very common pattern under SO questions: "if you can't answer get lost". I see it all the time when askers are told that their question is unclear or otherwise off-topic on SO.

@AndrasDeak OK, please stop this. Discussing other users and pointing out the parts of their behavior you don't like isn't nice.

Fair enough, sorry

thanks

@terdon no, I take it (“safety third”) as a joke, but I take safety seriously (both when programming, and in my “real” life)

@StephenKitt Doesn't seem to be a joke. He does have some valid points, but I certainly don't want a world where we value getting the job done more than getting it done safely.

That said, I am constantly amused by how often the Brits take this idea to extremes.

@terdon yes, I hadn’t seen that video before, and for him it certainly isn’t a joke; but note that when the captain of the crab boat told him that safety was his own problem, he did take it very seriously

@terdon yes, or Americans with all the warnings all over the place

it is said that every warning has a story and a lawsuit behind it

@StephenKitt Oh yes. It was the captain's attitude I found shocking.

It's a dangerous job; don't sign up for it if you're not up for 40-foot waves, I guess?

There's no regulation that will protect you from the ocean

(have life vests and radios? sure)

some safety regulations ironically make things more dangerous

Anyone ever seen a comparison of the quality of Stack Overflow in different human languages?

@JeffSchaller Sure, but if you're the captain of a vessel at sea and you don't consider it your primary job to ensure the safety of your crew but instead see your job as getting them back rich, if not necessarily alive, then I have some serious problems with your attitude.

@terdon I have to assume that the captain wants to make it back as well. and you could assume that "back & rich" implies "back"

I was just talking to a friend who operates cranes the other day about how they have horn regulations. They have to beep their horn once before they start the engine of any piece of equipment. Twice if they are going to drive forward, and three times if they are going to drive backward. The result is a construction site with never ending horn honks. And everyone on the site just completely ignores horns so much that they don't even consciously hear them anymore.

true. Back and well would be better though. I'd rather be well than rich, thank you.

"getting safely back to shore is not the goal of crab fishing" might be another way to put it

@jesse_b yeah, that makes sense. I mean, it's stupid, but I can see how it would happen. I'm not saying you should follow regulations, I'm just saying that you should consider safety first. Sadly, when that's codified it often results in absurd and less safe conditions. No argument there.

@jesse_b seems like it'd be easier if people used their eyes :(

@JeffSchaller yeah. I take issue with that. Which may be one of the reasons I don't crab fish, of course.

Heh, yes I am not a huge fan of regulations. It's the principle of safety first I agree with. Not necessarily, or even often, the rules and regulations imposed in order to achieve it.

@JeffSchaller yeah but there is always a real risk of being run over by heavy equipment. The operator usually has large blind spots and people on the ground are often wearing hearing protection and not looking in the direction of the machinery. He also lives in Canada so they often work in extreme temperatures like -20-40C so people are often wearing hoods and scarves that further impair their vision

Especially since we have a natural tendency to take safety regulations or features into account when handling things — so for example drivers pay less attention in a car which can stop on its own to avoid a crash.

@StephenKitt Do you happen to know safety issue in event handling and signal handling?

The other thing I don’t like is when a safety feature projects problems onto others; for example my car beeps when I’m reversing (because it does so silently, it’s a hybrid), and I find that reassuring as a driver, but also annoying because the message to others (pedestrians in particular) is that they should watch out for me instead of me watching out for them.

@StephenKitt I remember hearing about that somewhere; how safety margins (in e.g. cars and roads) led to increased speeds, causing us to take up the slack created by those safety margins

@Tim yes but this isn’t the place to discuss that; you’ve asked plenty of questions about that on the main site, feel free to re-read the answers you’ve been given there.

@StephenKitt I saw you rode a bike the other day

@Tim yes, I ride a bike many days, but that was my first fast road ride of a decent length in a while

@StephenKitt I agree with that though. I think the responsibility falls on both driver and pedestrian. Even though laws almost always side with pedestrian it's much easier for a pedestrian to avoid a car than the opposite. I really hate when people get this mindset that "cars have to yield to me in a crosswalk so I can just dart right across the street and they have to stop." That is true but the law wont make them not dead if that driver isn't paying attention.

@StephenKitt My question here is different from previous though

@jesse_b yes, I’m not saying that pedestrians shouldn’t be responsible too; it’s just the message that’s put across by beeping vehicles.

@jesse_b for some reason pedestrians here know to wait until cars stop at a crossing ;)

it might or might not have to do with terrible road culture

In some cases the configuration makes it unavoidable, e.g. on trucks with no rear visibility, but perhaps truck manufacturers could work on improving that instead of shifting the problem onto others.

@Tim so wait for your question ban to expire.

The unfair ban?

@StephenKitt Yeah, but the British thing where you have a loudspeaker on the thing yelling "Caution, this vehicle is turning right. Caution, this vehicle is turning right. Caution, this vehicle is turning right" drives me up the wall.

On the whole road thing, I find it fascinating to watch old films and see how we moved from a road owned by pedestrians and kids and slower vehicles (bicycles, horses and carts etc.), where cars were tolerated, to completely the opposite...

Just beep, if you must (and you must, especially for blind people) but don't shout full sentences at me.

@terdon yeah that’s just crazy

like car alarms that shout “step away from this vehicle” as well, which I’ve only ever seen in the UK

mind you in the US you get similar things, but where people are employed to shout warnings over and over again

@StephenKitt that would be a ... K.I.T.T. car...

@JeffSchaller LOL

@AndrasDeak Yeah it was mostly a thing I noticed in NYC/north NJ. A LOT of people there basically have the mindset that they are daring cars to hit them

As soon as the crosswalk light turns green for pedestrians (and often times even if it's red) they will dart right into the road almost trying to get hit by cars that run the red light (which happens a lot and is not right either)

@jesse_b when I visited NYC, I remarked to myself that there's not much in the way of traffic laws, so much as the law of physics. "Can the car make it through there?" "Can the bicyclist get through there?" "Can the pedestrian make it across?" ...

You can imagine my surprise when I went to Uppsala (Sweden) for a conference, and I just looked at a ped. crossing and cars stopped. It was like another planet.

Yeah driving in NYC is insane. People will cut across 5 lanes of traffic without there being an opening.

@JeffSchaller same in the Netherlands, but at slower speeds, and with more eye contact, and it feels far less dangerous as a result

traffic flow videos from the Netherlands are amazing

if only humans could learn from each other in some way, we'd be golden

there should be a site where people could ask questions and learn the best way to do things

an ... exchange of ideas, if you will

Would it be stackable?

ideally

if it used jQuery, that'd be the icing on the cake

If only privatizing cars were banned, in the same way many harmless accounts on SO are

Many people are very angry about migrants into their towns. One of the reasons is that they can't find parking spots.

The biggest reason I left NJ was the lack of parking but I never once thought about migrants and this is probably the first time in my life I've ever used the word "migrants"

If you don't have a car to drive, then you don't have the worry.

If you don't have a car to drive you don't have to hate/resent people that do

@JeffSchaller I hear OSHA isn't that effective. Though I don't know what this is apropos of.

@AndrasDeak It's generally good practice to create a repos one level down.

I'll try creating a repos with a name. I hope your script still works.

Apropos of really nothing at all, there is a mildly entertaining dance film called "Fast Forward".

@FaheemMitha yup

I actually did git init tmp_repo but I wanted to omit the cd in the shell script :P

@AndrasDeak Looked?

So git help branch says:

> The command’s second form creates a new branch head named <branchname> which points to the current HEAD, or <start-point> if given.

Am I supposed to be able to figure out what "the command's second form" corresponds to?

@FaheemMitha second variant in the synopsis

that comes up quite a lot in manpages

If I literally just count occurrences of git branch, I get to:

> git branch [--track | --no-track] [-f] <branchname> [<start-point>]

So maybe that's the one.

@FaheemMitha yes, that's the one

@StephenKitt I don't recall.

I normally do git checkout -b branch_name to create a new branch and check it out. I just used git branch for the example because I wanted to go on with master

So current HEAD if start-point is omitted.

@FaheemMitha yeah, just looked at it, hinting at a desire to cross

@AndrasDeak You mean you were just standing on the pavement near the sign, and hadn't pressed a button or anything?

I assume there is a button. I remember using them. Not in India, of course.

Probably the UK.

@FaheemMitha no button, no lamp, just a crossing marked on the road and with a sign, yeah

I was walking along the pavement

@AndrasDeak Huh. That's surprising.

In many countries in Europe, if a pedestrian gives a suggestion that they intend to cross, drivers are supposed to stop for them. How that actually translates in practice depends on the country; but in the North of Europe, anything that could possibly be interpreted as indicating a wish to cross will result in drivers stopping.

I assume they are going quite fast. Do they just slam on the brakes?

In the East of Europe you stand facing the crossing, with practically one leg raised, and wait a few minutes before the street runs out of cars before you cross ;)

@FaheemMitha no, they have to stop in a safe manner...

@AndrasDeak yes, France is like that too!

@AndrasDeak In Bombay, the cars will sometimes try to run you over. Depends on their mood. The motorbikes will definitely try to run you over. If they see you in the way, they'll speed up.

You can even stand on the crossing and have to wait for cars to finish passing you.

Actually, sometimes the cars will speed up too. The private cars are generally worse, I think.

So

git branch stuck_on_foo

is now attached to HEAD? And if HEAD changes, does it follow? Or does it say with the same commit?

@FaheemMitha stuck_on_foo points at the commit that’s currently the HEAD, and stays there unless you switch to the branch and move.

say you’re on branch bar; git branch foo will create foo pointing at the current HEAD

if you commit then, bar will move to the new HEAD, but foo won’t

@StephenKitt Ok. So it doesn't move when HEAD changes. Got it.

the only branch that moves with HEAD is the current branch

Mercurial has different branch concepts, but nothing I've used behaves like that. Possibly bookmark does.

@StephenKitt master?

I don't see how something that points to a commit could be interpreted as a branch, though.

@FaheemMitha the current branch, whatever that is

@StephenKitt Isn't it master by default?

@FaheemMitha that depends on the repo settings

@StephenKitt Ok.

@FaheemMitha if you're interested in learning git I very strongly suggest first reading the git parable and then playing this fun tutorial game.

@FaheemMitha a branch is something that points at a commit

but it moves with HEAD when it’s current

unlike tags

The parable gives a high-level idea of the philosophy, and the game makes it very obvious how commits and operations on them behave.

@AndrasDeak I'm not really. And reading a Git tutorial would probably put me to sleep. Plus I have a short attention span. Also, I'm lazy.

It's not a tutorial, it's a game

@AndrasDeak Oh.

A puzzle game, to be precise

Sometimes I wonder how I ever managed to learn anything.

Of course it's a tutorial as well :) But that's a side-effect ;)

The only thing the tutorial game doesn't cover is how to resolve merge conflicts, because it abstracts away commits (and you need to go down into a commit's contents to see merge conflicts)

Thanks, Andras! Now I've gone from "gosh, I'd like to learn more about git" to "here's some excellent resources; now, when are you going to do it?"

I hereby waive all responsibility for any mental damage caused by learning git :P

Full disclosure: I love git, the way it works, but the command-line interface is a mess

hehe

tar --help is a valid tar command!

Wouldn't that depend on if it's gnu tar or not though

yes

the safe option is tar cf archive.tar *

no -?

Have I spent so many hyphens in vain?

@AndrasDeak it's tar, there is no rhyme or reason

@StephenKitt touch file && tar cf archive.tar *; just case the directory's empty :)

@AndrasDeak there are two standards to choose from -- hyphens or no hyphens :)

But I keep using tar -cvzf and tar -cvJf etc. with success... am I using You Can't Believe It's Not tar?

@JeffSchaller phew

@AndrasDeak no, hyphens are OK now, but some older variants don’t support them

I'll get off your lawn

Old man yells at cloud

I’m always amazed at what people will do when faced with flaky networks or hardware

I mean yael runs Hadoop, I wonder how that can possibly work correctly on a network which can’t transfer a 4.6G ISO

I'm looking at git-scm.com/book/en/v2/Git-Branching-Branches-in-a-Nutshell

or randomly corrupt them

I wonder how he's transferring it. Any robust network file sharing method should be fine but using windows copy via samba share or something probably is the main issue

But I'm unclear when this pointer to a commit will move with the commit, and when it won't. Assuming both options are possible.

@FaheemMitha if you're checked out on a branch (your HEAD is attached) the commit will move with the commit. If you checked out a commit or a tag then you have a detached HEAD and the commit won't move. Assuming I understand your question.

must be using RFC 2549 without the ECC algorithm

@FaheemMitha play the tutorial game, it illustrates that quite well

Using windows explorer to copy large files over network shares is often problematic even with perfectly functional networking

@FaheemMitha You can try my example repo, look at the git log to find a commit hash, git checkout that commit, read the warning about the detached HEAD, then commit something to see what happens

@AndrasDeak Yes, you understood the question. And checked out just means that the HEAD pointer is pointing to the same commit as say the pointer "foo" is pointing to?

Then we say that "foo" is "checked out"?

@FaheemMitha to be honest I'm shaky on the terminology and dynamics, I mostly have an intuitive understanding due to the tutorial :)

@AndrasDeak There doesn't seem to be all that much to understand. I guess I could experiment.

"pointing to the same commit" is insufficient, especially since a single commit can have multiple branches on it. You have to actually check out a branch.

But yeah, experimentation works best

@AndrasDeak I understand. But once checked out, HEAD and the branch "pointer" will be pointing to the same commit, correct?

$ gittree
* 65861db (HEAD -> stuck_on_foo, master) bar
* 31282bc foo
$ git checkout 65861db
Note: switching to '65861db'.

You are in 'detached HEAD' state. You can look around, make experimental
[snip]

HEAD is now at 65861db bar
$ touch baz
$ git add baz
$ git commit -m 'baz with detached HEAD'
[detached HEAD 89f48a2] baz with detached HEAD
 1 file changed, 0 insertions(+), 0 deletions(-)
 create mode 100644 baz
$ gittree
* 89f48a2 (HEAD) baz with detached HEAD
* 65861db (stuck_on_foo, master) bar

@AndrasDeak Sometimes I read documentation too. But like I said, I have a short attention span.

@AndrasDeak HEAD is detached from... bar? Or something else?

From any branch, presumably. As I said I'm not good with terminology. And I never commit with a detached head.

the others here can probably be more useful in this regard

I only know how to use git :)

@AndrasDeak So detached from master too?

In my case? yes.

Right.

otherwise you'd just commit to master

$ gittree
* 65861db (HEAD -> stuck_on_foo, master) bar
* 31282bc foo
$ git checkout 65861db
Note: switching to '65861db'.

You are in 'detached HEAD' state. You can look around, make experimental
[snip]

HEAD is now at 65861db bar
$ gittree
* 65861db (HEAD, stuck_on_foo, master) bar
* 31282bc foo

HEAD acts like branch, except if you check out anything else you will move HEAD and you'll lose the reference to your commits you made in the detached state. You can still recover them from the reflog, but that's digging in the garbage.

$ gittree
* 65861db (HEAD, stuck_on_foo, master) bar
* 31282bc foo
$ touch baz
$ git add baz
$ git commit -m 'baz with detached HEAD'
[detached HEAD 64cc08b] baz with detached HEAD
 1 file changed, 0 insertions(+), 0 deletions(-)
 create mode 100644 baz
$ gittree
* 64cc08b (HEAD) baz with detached HEAD
* 65861db (stuck_on_foo, master) bar
* 31282bc foo
$ git checkout master
Warning: you are leaving 1 commit behind, not connected to
any of your branches:

  64cc08b baz with detached HEAD

If you want to keep it by creating a new branch, this may be a good time

Let me know if this is too spammy in this room

@AndrasDeak Yes, I see. That's pretty terrible UI, though. Doesn't this left behind commit show up in the log, though?

Or maybe just terrible design. The UI is terrible too, of course.

Not in the git log, because that shows the ancestors of HEAD I think. It shows up in the reflog.

@StephenKitt there's a video that goes into the history of that; it's a comedy skit but teaches real history. Search for "Adam ruins everything jaywalking" and you should find it. In a nutshell: car manufacturers propagandized the subject.

@FaheemMitha Why is the UI terrible? It uses the best UI ever designed: the command line!

@terdon I meant the Git UI.

@FaheemMitha But the UI is the command line. Unless you're using some sort of graphical tool.

haven't read back all the way but maybe this is one of those cases where people confuse git with github

although I think github has a pretty intuitive ui

@terdon No, the Git UI is all that stuff documented in the Git man pages. Commands starting with git.

OK. Since that's a command line UI, I love it :)

It's not the sanest of command-line UIs. But that's not evident from the above examples.

The only thing I don't like about it is that it suggests commands when you spell something wrong

It's very easy to learn the basic workflow

Nobody has time for that hand holding

@jesse_b I bet you can disable that

I think git does an incredible job in handling something that is mindbogglingly complicated.

And it makes it very hard for you to really screw it up.

It warns you when you try to add something in your gitignore and yells how to disable said warning

@terdon well all VCSs can say that (not you, SVN ;)

Wasn't git the first to do a truly distributed model?

@terdon No. There was Bitkeeper. And I think there were others before Bitkeeper. Or around that time. Including Arch and Monotone.

All three of those definitely predated Mercurial. And they were all distributed.

A good part of Git was based on Monotone, I think. Not the security related stuff, though. Or the db stuff.

I see, thanks. Git is the only one I've had any real experience with. I briefly played with SVN a while back but it didn't take.

Technically, I think Arch is a specification. Tom Lord's Arch, or tla, is/was the main implementation. There were other ones back in the day, but they were really marginal.

tla is still in Debian.

(from help.libreoffice.org/7.0/en-US/text/scalc/guide/cellcopy.html)

@AndrasDeak Spreadsheets are works of the Devil. Best avoided. Along with Word Processors.

Ontologically speaking I don't think spreadsheets are inherently bad. I just haven't seen a satisfying implementation yet.

@AndrasDeak I think the demonic potential of software is greatly underrated.

Good evening/night/morning/afternoon

Is good evening used at the beginning of conversation, and good night at the end?

what would you use for beginning version of "good night"?

In pthreads, which one is used often: semaphore, mutex, condition variable?

@FaheemMitha yes

See also nasal demons