@overexchange you might help your case by giving a reason why you're creating a brand new docker container with an EOL OS; see also the article I linked in with my comment. Also, why did you tag the Question with Ubuntu when it seems to be Debian-based?
@overexchange I'm basing that off all the "jessie" instances in your question as well as derobert's answer "your bigger issue is that you're using jessie-backports. That was EOL'd along with Jessie"
@overexchange I'm no docker expert, but I agree with derobert's general guidance: stop creating new things on deprecated (3 year old!) technology. The jenkins page on docker says "This image has been deprecated in favor of the jenkins/jenkins:lts image provided and maintained by Jenkins Community as part of project's release process. The images found here will receive no further updates after LTS 2.60.x. Please adjust your usage accordingly."
Genuine question, apropos of nothing in particular: if someone persistently spams this room with low-quality irrelevant questions betraying not even the barest modicum of effort or competence or care or improvement over time, how long would they need to do that for before one of that majority of regular attendees of the room with the ability to do so intervened to stop that from happening?
@MichaelHomer I'm rather new to the idea of moderating chat, so if people have ideas around better guidelines for what's acceptable here, let's hear them.
I know we recently had someone in here that was semi-coherent; if it's disruptive, then we should let them know, ask them to stop, and then escalate to something like a chat suspension.
More recently, we've had someone posting questions (either from SO or U&L) here; given that the room's topic is "General discussion for unix..." it seems to me not very far off-topic. I can't put my fingers on it at the moment, but I have a memory of there being advice to use the chat room to raise awareness of an (unanswered) question.
We're here to learn, and maybe to teach; I'd encourage us to be a little careful about what we want this room's topic to be, given that it's the chat link from the main site.
My ownpersonal two cents around Questions in general is to show your effort; as to posting them in chat, my own request would be that you (1) consider your audience, and (2) explain why you're posting in the chat room -- do you think someone there could answer it? Do you want help rephrasing it? Do you think it'll solve a problem they're having?
searching for "https://" in the chat room's search box shows a range of people posting both Questions and Answers here, so I'm curious if there's a line that we need to draw about what's appropriate.
@MichaelHomer I have a hard time with it because occasionally I'll post a question in here, usually for other reasons, but the situation you are referring to does just seem different to me. At this point it does feel sort of like an annoyance.
But I know I am certainly not one to be the judge of what should or should not be in this room. I'm probably barely allowed here as it is :p
apparently my personal inclination is the positive version of "don't 'question 'dump'", which I just learned is "pasting a question link into a chat room without context'
I think in an ideal world it should be allowed but the user should also be able to recognize when it's becoming too much and self regulate
well maybe not without context, even if the context is just "Hey I'm really in a bind with this issue, I would appreciate it if anyone in here is able to help"
IMO what is annoying about the particular user is that it seems they post every single one of their questions here, as if that is just part of the process of asking a question on SO. It doesn't seem like they even give the community a chance to answer before assuming it won't happen.
I like to think that this room enjoys technical discussions (keyboards, awk, RAID arrays, computers vs tablets, etc); but they're all multi-person discussions.
For myself, I'd love to help a person write a good question (hint, hint)
If you asked me, I'd say this room is not a great place to answer questions -- people go to the site(s) to answer questions; and they can do that by looking for questions they want to answer
@Jesse_b get a chainsaw
;)
What I meant was, someone coming in here saying "I've got a problem (and I want to ask it on the site) but I can't describe it very well"
> The value of status may be 0, EXIT_SUCCESS, EXIT_FAILURE, [CX] [Option Start] or any other value, t
But I haven't found a clear standard that states that an exit status of 0 means success.
I answered a question over on AU asked by a C programmer who was having their program exit with 1 for success, just as you would for a function returning true, and I wanted to point the to the official docs that show why exit status should be non-0 for failures.
> On POSIX systems, the value of this macro is 0. On other systems, the value might be some other (possibly non-constant) integer expression.
but still not an official posix document
@terdon: I think this is it: https://pubs.opengroup.org/onlinepubs/009695399/functions/exit.html
> A value of zero (or EXIT_SUCCESS, which is required to be zero) for the argument status conventionally indicates successful termination. This corresponds to the specification for exit() in the ISO C standard. The convention is followed by utilities such as make and various shells, which interpret a zero status from a child process as success. For this reason, applications should not call exit(0) or _exit(0) when they terminate unsuccessfully; for example, in signal-catching functions.
Quite a lot of software would have to be removed if this was required... And obviously software is downloaded from somewhere, and that’s tracked in debian/copyright (and debian/watch), but sometimes the download location is now Debian’s infrastructure!
A more Debian-centric view would be that package descriptions should be improved so that you can decide whether to install them without referring to their homepage...
@FaheemMitha if it’s orphaned upstream, the source is only in Debian, so I’m not sure it’s very useful to list Debian as the homepage; that’s not a desirable condition, we don’t necessarily want to be the long-term caretakers.
@Jesse_b It depends on the semantics of the call. Some functions returns a positive integer or zero for success, but -1 or the EOF constant for error (e.g. fscanf()), others return some other specific constant on error.
Yes, you really need to check the docs, there’s no rule. C functions which return some other value will often return negative values on error. C functions which return a boolean to indicate success or failure usually return 0 for success (see close).
This leads to odd code such as if (close()) { // something went wrong }
I'm currently on a train. I like travelling by train, but right now we're just standing still on the track in the middle of a forest... which is less than ideal as I'm about to miss a connection.
@Kusalananda When I used to take the train to work every morning it got delayed once for several hours within walking distance to my station but they wouldn't let us out. I got very angry about it but it turns out a train in front of us had hit and killed a pedestrian
I felt bad about being angry after I found that out
Faheem, I wanted to ping you - are you a regular at the TeX.SE? I'm wondering about bringing some of my Wildcard-inspired challenges there, and wondered if you had any guidance for asking Questions at TeX
(if you can compare/contrast to the U&L crowd, that's the one I'm familiar with)
@JeffSchaller Depends on how you define regular. You can of course see my activity on the site. It's not high. I do hang out in chat a fair amount there, though not as regularly in hear. Since around 2014 or so, I suppose.
TeX SE is reasonably helpful and friendly, particularly towards beginners. Comparable to here, probably. And certainly much better than places than SO. However, just a tip. They like MWEs over there. Like, really like them. As in, they're unhappy if you don't.
@FaheemMitha I haven't learned how to do that yet; I suspect I'm getting "good" errors -- it's just that I don't know how to approach doing this-or-that
Hmmm; well, my existing code is a rats-nest of "tried this, didn't work" -- my real questions would be more along the lines of "I want this stuff in a numbered list here, with lines over there"
I mean, I could show all the non-working versions :)
Of course, my perception may be skewed by the high rep users I talk to in chat. Who tend to be members of the LaTeX team. Like David, Ulrike, Joseph, egreg.
They're definitely quite fussy. There are many other people in the site, and they're probably less fussy. Though for some reason it's also the case that the high rep users virtually always answer my questions. As you can check.
But TeX SE also seems unusually skewed in terms of who answers the questions. It's disproportionately the high rep users. Unlike here, for example. Or probably even SE. Though I haven't seen any statistics.
@JeffSchaller Try to isolate the problem. You can also try bouncing things off people in chat. They won't bite.
fair enough; thank you for the insights! I spent enough time last night banging my head against the wall that I realized I should probably just ask some smart people how to do it
They're kind to beginners. I've been around enough that they don't bother to be particularly nice to me. It's like: Faheem again, trying to take a shortcut and refusing to show us the code...
It's easy enough to install other stuff. If you start using LuaTeX you might want to try to stay more current. But LuaTeX is more if you want to do backend stuff. It's not super common. Though I've started doing it a fair amount for my purposes.
@StephenKitt Right. Yes, that should be ok most of the time. But Debian packages, at least in stable, do date, as you know. But backporting TeX Live is fairly easy, except I had problems with biber a couple of times.
I mean, sometimes you need to stay a bit more current. And LuaTeX and the overall system is still changing quite fast. For example, they're in the process of merging Harfbuzz with LuaTeX, which is quite major.
That will probably knock out XeTeX as a competitor. At least David Carlisle thinks so.
@Jesse_b The reason you have different kinds of return values for errors in C functions is that there's no single "undefined" or "false" value you could return from any function. You have to return something that fits the type, but is not something that could be the result of a successful call.
So a function that returns a pointer to an object might use NULL (basically zero) to signify an error (think malloc()), but a function that returns the number of bytes processed could return a zero as part of normal operation, so it has to use something like -1 instead (read()).
while interpreted languages the typing often isn't that strict and they have those special "undefined" values (undef in Perl, None in Python, nil in Lua). And C also doesn't have exceptions which you'd commonly use in Python and C++...
@StephenKitt we just got bitten by this bug at work. Do you have any insider knowledge? It's a kernel issue, so it affects all distros, do you have any way of knowing if anyone is working on this that isn't available to me?
Missed my train connection in the end, by literally 30 seconds. Next train was only 20 minutes later though, so I'm OK. Spent the whole day talking so I'm dead tired. Work retreat.
In the European Union, banks are now forced to move away from SMS. Most are using their own banking apps on smartphones, but there are other solutions too.
@FaheemMitha A secret from who though? Everyone? I'm no security expert but I feel if your phone is compromised you are probably completely compromised already
@FaheemMitha it’s better if it’s not public knowledge at least, IMO. If searching “Faheem Mitha” on a search engine leads me to your phone number then that’s not good.
@StephenKitt I've not seen banks here offering alternatives. But India isn't exactly on the bleeding edge of security and technology. Or current, even.
@Jesse_b Supposedly people can get the provider to get a new SIM, using social engineering or some such. Basically hijack your number.
@Jesse_b Actually, I have two numbers already. But they're both public. Quite common in India, because the networks aren't that good. So people often use two networks.
@StephenKitt Yeah true since the topic is specifically about SMS. I also was considering things like google authenticator or duo for 2fa
Then you also have to consider that most of those apps can be "recovered" via SMS so if they compromise your number they can probably access them as well
@Jesse_b yeah the latter are safe enough, although in many circumstances they don’t qualify as 2FAs (e.g. you shouldn’t authenticate with a phone-based 2FA on an app on the phone, or a web site accessed on the phone).
@StephenKitt True, I've never used them for that. I use duo for various websites that I suppose I could technically authenticate through my phone but I would still consider it 2fa because I'm using a password and the phone
@terdon in any case the fix has to come from the Ubuntu kernel team; since 4.15 isn’t an upstream LTS kernel it won’t get a backport from the kernel people (or from Red Hat)
TMobile currently has a 30 day free trial of hotspot, and allow me to keep the device regardless of whether I will continue. Anyone interested can give it a try. https://www.t-mobile.com/offers/free-trial#
But the device is locked, while Freedom Pop uses AT&T network (unlocked Tmobile hotspot might work). How sad
@Tim I think there are two fundamentally incompatible philosophies about error handling. One is that the programmer shouldn't have to think about error handling all the time; it should be hidden away and work magically how it's supposed to. The other philosophy is that error handling should be explicit and entirely non-magical.
@AaronHall Yes. Sorry, I guess that comment was a bit cryptic.
@Wildcard When it comes to error handling, it's helpful if it's clear where the error is happening, to start with. It's distressingly common for that not be the case.
@Wildcard I now remember that PFPL says using option type to represent both normal values and errors. Is that one of two philosophies that you mentioned?
There is also exception handling. Is that one of the two philosophies?
when to use option type and when to use exception handling?
Actually, the whole concept of an exception, dropping out of the program, is rather ugly to start with. It's much better if the program can pause execution. But few environments do that. And it's so rare now, that if a program does do it, a lot of people wouldn't understand what it's doing.
TeX does do that, as it happens. For example. But it's a very old program.
And until recently, I didn't understand what it was doing, either.
The Lisp family does do it, but it's very much the exception these days.
Below command:
RUN apt-get update -y && \
apt-get install apt-transport-https curl python-dev python-setuptools gcc make libssl-dev -y && \
easy_install pip
is giving error:
Step 9/14 : RUN apt-get update -y && apt-get install apt-transport-https curl python-dev python-setuptools ...
