@MichaelHomer Ok, thank you. Those are the respective standards, I assume?

15 messages deleted

I deleted the conversation/monologue about bowel movements. This is really not the place for it.

@FaheemMitha Yes, or ISO8601 is, and then RFC3339 has some extensions/applications for network protocols. Your string is fully ISO8601-compliant

@MichaelHomer Ok, thank you.

I wonder if a post/question about how shell=True works and what it does, exactly, would be of interest. My specific example is the Python 3 subprocess library, but other languages probably have similar functionalities.

@FaheemMitha there’s stackoverflow.com/q/3172470/4500798 on SO already

What this keyword does has always been a bit of a mystery to be, but I've never really investigated it.

@StephenKitt Yes, thank you for that link. I was looking for a more general perspective though. Like for example, what's actually going on under the hood.

@FaheemMitha that’s addressed in the documentation

@StephenKitt Thanks, that's helpful. But does Python actually call a shell child process in the case shell=True?

@FaheemMitha what do you expect to happen with Popen(['/bin/sh', '-c', args[0], args[1], ...])?

@StephenKitt I don't know. My knowledge of such things is fairly meagre.

@FaheemMitha the shell is specified as the command to run, so yes, Python actually calls a shell child process

@StephenKitt Ok, thank you for the confirmation.

I don't even really know what Popen does. Perhaps I did once.

Some stuff is easy to forget.

@StephenKitt Thank you taking the time to explain.

I suppose in the shell=False case Python invokes the program with the arguments directly. I see that would be workable if the program is a C program. But I'm sketchy how that would work if it wasn't.

@FaheemMitha same as running a program from a shell prompt, with the exception of the last-resort shell fallback

Hmm, I guess the #! could be invoked by the kernel to invoke the appropriate interpreter.

Is #! controlled by the kernel directly?

@FaheemMitha yes

@StephenKitt Ok, that's what I thought.

I suppose without a shebang, Python might have problems. I could experiment.

@StephenKitt Oh I see you wrote the answer for that one.

@FaheemMitha yes, and it even explains what happens without a shebang ;-)

@StephenKitt Yes, I see it does. I was just reading it now.

That's certainly a very through explanation. Thanks.

Have an upvote, though you don't need it.

@FaheemMitha thanks :-)

So the execve system call (I suppose it's a system call) handles both the binary case and the hashbang case. And if those both don't work, it passes it back to the shell, assuming there is a shell to pass it back to.

So presumably Python using execve too.

@FaheemMitha not quite; if none work, it returns an error to the caller, whatever that is (the kernel doesn’t care whether it’s a shell).

@StephenKitt Oh, so the shelf gets an error from execve it then tries running the program in the shell?

Is this execve thing standard around Unix-like systems?

@FaheemMitha yes, see the six-point scenario at the end of my answer

@StephenKitt Ok, I see.

@FaheemMitha yes, see the description of the exec family in POSIX, and execlp and execvp

@StephenKitt Assuming the system is POS)X conformant. Is that a given?

@FaheemMitha nowadays it’s the only context in which it makes sense to try to answer questions such as

@StephenKitt Ok.

remember that POSIX is largely a consensual description of existing behaviour in Unix-style systems

Presumably not counting systems like Xenix, though.

I bet there are people out there using Xenix, too.

@StephenKitt Yes, I know that it's a standard, and therefore to some extent aspirational, like all standards.

@FaheemMitha no, it’s not aspirational, that’s my point

@StephenKitt But all systems can't possibly have identical behavior. There must be variations.

Unless POSIX exactly codifies the intersection of all, which sounds improbable.

@FaheemMitha yes, of course, but behaviour described in POSIX isn’t described as a target to reach, it’s described as the common behaviour of existing systems (at the time when it’s documented)

Random trivia: Stallman gave POSIX its name. That's mentioned on the Wikipedia page, but I don't know how well known it is.

@StephenKitt So it really it an intersection, then? That would presumably leave some variation undocumented.

@FaheemMitha as far as possible, yes, excluding systems whose behaviour is too different; differences are either documented (if that’s possible) or defined as implementation-specific

@StephenKitt Ok. Sounds like a lot of work. But I suppose these days systems are more conformant.

@FaheemMitha yes, but few are actually certified as such

And there are less of them, too.

Linux effectively killed off a lot of the proprietary UNIXes. They were still around in the 1990s, but one hears less and less of them.

I’m not sure about that; from a POSIX perspective, each Linux distribution is a different system

Some still exist in hobbyist form. Like OpenSolaris, perhaps.

@StephenKitt It is?

All the software is exactly the same.

@FaheemMitha no, it isn’t

take /bin/sh for a start

It's mostly in how it's put together, and the packaging system.

I spend my time hopping between different Linux systems, and there are many variations which would probably affect the results of running the POSIX conformance tests.

@StephenKitt Oh. I thought they were all pretty much the same.

It certainly seems like that from my perspective. Though it's been a while since I used anything but Debian. It's been an isolated existence for a while now.

Is anybody looking at the suggested edits for the tag wiki excerpts? ;-)

@Tsundoku yes, and thanks for taking care of them

it would be nice if you could mention when tags shouldn’t be used

for example in your agrep description, go into a bit more detail about what makes questions agrep-specific

Hmm, I thought I could reopen that tag wiki excerpt again...

@Tsundoku you need to wait for your edit to be approved or rejected before you can change it again

Why do we have an adobe-acrobat tag?

unix.stackexchange.com/questions/tagged/adobe-acrobat

The tool doesn't run on *nix, so I can't see any point in having a tag for it.

@Tsundoku you just edited its tag excerpt, do you think it's a useful tag?

linuxconfig.org/…

@terdon I think that tag exist because a version for Linux used to exist.

"The latest native Linux version [from] 2013."... and then instructions for installing it on a 2020 distribution :|

@Tsundoku A version of acrobat reader existed, but never of acrobat as far as I know.

@jesse_b reader, not acrobat

And still. Do we really need a tag for a tool almost nobody uses these days?

adobe reader used to be called "Adobe Acrobat Reader"

I haven't seen the Linux version of reader in more than a decade

I assumed the tag would apply to any product in that product family.

@jesse_b Yes, but there is "Adobe Acrobat" a powerful tool for creating PDF files, and "Adobe Acrobat Reader" a pdf viewer. The two are not the same.

It's actually still called adobe acrobat reader on adobe's website

The actual application is just adobe reader but the title of the download page is adobe acrobat reader

Yeah, the reader is. But not Acrobat itself, right?

There is no adobe-reader tag, so getting that tag in would require retagging existing questions.

So if we were to have a tag, that should be adobe-acrobat-reader anyway, and I just don't see any point in having the tag at all. We currently have 9 questions tagged with it, and of those, only two seem to actually asking anything to do with Acrobat. The rest seem better served with the generic pdf tag.

IIRC they used to be the same application but you just don't have access the edit functionality with the free version and then I guess they split reader into it's own application at some point, I still refer to it as acrobat though

Oh, wait, there's acrobat-reader...

@jesse_b Ah, maybe. I always knew of them as separate products, but never really used them.

@Tsundoku with a whopping 3 questions...

I'm leaning more and more towards burminating both of them.

This question is about reading a password-protected file created using Adobe Acrobat on Windows. This is one way in which the tag adobe-acrobat appears to have been made relevant to this site. It's that question's only tag. If we want to get rid of it, should it be replaced just with pdf?

@Tsundoku that would make sense, yes

Yeah, that's the only one that I found actually relevant to the specific tool, but that could be covered by the pdf tag. I don't see any reason to keep a tag around for so few questions, most of which were simply tagged so because the poster probably knew acrobat.

I've sometimes wondered why Adobe was (and is) so uninterested in Linux-based systems. And presumably equally uninterested in BSD systems. It can't be so hard to port it. After all, it presumably works on OS X, which is just another Unix.

just do it :)

@FaheemMitha they did have a Linux version for a while. But what's the point? The users who want a professional PDF creating tool are not really using Linux, so why would they put any resources into that?

@terdon Is there a specific downside to having that tag? Or just extra clutter?

@terdon Just the Reader part. And why wouldn't Unix people use Acrobat if available? By the same people who created Postscript and PDF?

It's actually quite hard to get Acrobat type editing functionality with free software. Not sure why. And the other tools available for Linux aren't that great. I've experimented with some of them.

@FaheemMitha extra clutter, worse organization, and when the tag exists, new users will find it and apply it.

@FaheemMitha Well, they did have a reader for many years.

For annotations, I've been using LaTeX for a while now. It's not perfect, the usual TeX overhead, but it makes the competition look pretty silly.

But these days, there are other alternatives that work perfectly well so it was probably not worth maintaining.

@terdon They did, yes. But you couldn't do much with it.

@terdon Such as?

Like I said, I use TeX because the alternatives are not great. And some are proprietary too.

Well, PDFTeX, to be precise. Because it has a PDF writer built in.

@FaheemMitha What do you mean? It's a PDF reader, what do you want to do with it?

@FaheemMitha I use evince

@terdon I mean, you can't do much with just a reader. I want to edit.

@terdon That can do some sort of annotation, last I checked.

@FaheemMitha Ah, sorry. I misunderstood. I thought you were referring to the reader only.

@terdon My fault for not being more explicit.

I might have read this before, but I don't recall doing so.

If I read that 100 times, I doubt I would recall doing so :)

Way over my head.

The corresponding paper looks very fancy. Well, the first one.

@terdon The problem is simple. Just sphere packing.

I didn't realise it had practical applications, though.

The solution is definitely not simple.

@FaheemMitha Sure. "Simple" as in it took one genius in the 17th century to come up with a solution in 3 dimensions (call them 4, if you like) and then another 3 centuries for someone to prove that. And now, since that was obviously too simple, they're playing with higher dimensions.

Simple, I say. Simple as anything! :P

@terdon I meant the problem, not the solution. In the sense that it's easy to understand.

Most math results aren't easy to understand. Even the relatively simple ones.

@FaheemMitha I can't understand what a sphere is in higher dimensions, let alone how to pack a bunch of them.

E.g. classical complex analysis is a simple area, as these things go. But one still cannot explain the results in elementary terms.

@terdon Sure you do. All points within distance r of a fixed point.

Linux doesn't have many options for creating accessible PDF files outside LibreOffice. And it doesn't have any tools for checking the accessibility of PDF files, or none that I am aware of.

The basic sphere packing problem is indeed simple, when we're talking oranges and a grocer. But the simplicity is inversely proportional to the number of dimensions you bring to bear.

For some constant r.

@terdon Dimenstions make no difference in terms of the comprehensibility of the problem.

@Tsundoku probably not. Linux is more geared towards servers and science neither of which does all that much with PDFs.

It's harder to visualize, perhaps.

@FaheemMitha I beg to differ.

@terdon Ok.

@terdon Linux has a broad choice of desktops (too many, according to some), and scientific papers are usually made available in PDF format.

I was just looking at the paper. That sort of hard numerical analysis, especially in an area without established methods, is hair raisingly difficult. She must really have given it her all.

And there are distributions geared at people with disabilities.

Bear in mind that I am very deeply ignorant of mathematics. All I know about higher dimensions is that you get the 2nd dimension by raising the first to the second power, the 3rd by raising the second to the third power and (this is where it breaks down for me) "so on". I cannot grok what raising the cube to the 4th dimension means. Let alone fancy things like the 24th!

@Tsundoku Well yes, but we make them in LaTeX :P Or submit a doc and the journal makes them into PDFs.

I'm just saying that the vast majority of people willing and able to pay for a PDF creation tool will not be Linux users. Which gives very little incentive to Adobe to spend their effort developing for us.

@terdon It's just Cartesian geometry. An n tuple correspond to a point in n dimensional space.

It was one of the great historical advances in mathematics, back in its day, because it connected algebra and geometry together for the first time.

@terdon I guess I just don't understand why so much effort was required.

It works for OS X, I suppose. And like I said, that's just another Unix.

@FaheemMitha Remind me to explain some complex concept involving DNA, negative strands, positive selection and speciation and tell you how it's "just biology" :P

@FaheemMitha it wasn't back when adobe reader was a thing for Linux.

And macOS has all sorts of special sauce for GUI stuff which Linux presumably lacks.

@terdon I never could get along with biology. Did I mention I was in a biology research group once? It did not go well.

Yep

And I never could get along with mathematics. I find math beautiful, often interesting, but completely opaque to me.

@terdon Hmm. GUI stuff. Yes, I guess that could be a complicating factor. Though one would hope their code would be sufficiently modular to cope with that.

Which is why I let the computer handle the math for me and collaborate whith mathematicians when it gets too hard!

Anyway, Adobe doesn't care, so I guess the whole discussion is moot.

I wonder if I'm the only person on the planet to use TeX for PDF annotation. Probably not.

@FaheemMitha the UI framework is like nothing available on Linux (yes, even with GNUstep), so porting graphical applications is non-trivial

oops, which is basically what terdon said above

@StephenKitt If you say so. I don't know what's involved.

@FaheemMitha I don't think I've ever needed to annotate a PDF file.

They can't abstract stuff out?

@terdon I do it all the time. They love their forms here.

@StephenKitt ah, but terdon was making it up as he went along. You used some fancy sounding jargon, makes the whole thing far more respectable!

@terdon but I left out “leverage” and “utilize”

@FaheemMitha Oooh, forms. Yes. I think last time I found some online free tool to do it. But I may also have just used evince or libre office.

And I prefer to do it on a computer, because I can easily make small changes, and have the entire thing saved for future reference. Also, there's no question what is written.

@StephenKitt It's OK. You'll learn :P

@FaheemMitha forms work fine in Evince and Okular, including saving them, if they’re correctly-defined PDF forms

Also GNUStep is dead, I think.

most French administrative documents are now available as PDFs-with-forms, I’ve been using that for years

@FaheemMitha still twitching, it had a release in April this year

@terdon @StephenKitt To be clear, I'm not talking about PDF forms. Just arbitrary PDF files.

@StephenKitt Glad to hear it.

@FaheemMitha yes, I get that (annotations), but I’m just saying that well-formed PDF files work fine in free software tools

of course, many people manage to create “forms” in PDF files which aren’t :-(

/me sometimes edits PDFs in Emacs

The expectation is that people will fill in the printed out version by hand. But lunatics like me write on the PDF file.

@StephenKitt Ok. PDF forms would be much easier to handle. They don't really happen here.

@FaheemMitha No, lunatics print out a digital format, fill out the form with a pen and then scan it. What you do is reasonable!

@terdon Actually, the former method would be much faster. At least if it was only done once.

@StephenKitt Do you know the PDF language, then?

Apropos of nothing in particular, I was just looking at some old doctor notes of mine, and could make no sense of them. Why do doctors have such horrible handwriting?

@StephenKitt Whaaa? how?

Sometimes they type it, but unfortunately that's quite rare.

Mine opens them read-only in DocView mode. Can you actually edit them too?

@terdon I'm thinking he writes the annotations in PDF.

Though perhaps not.

@terdon there’s a tool (can’t find it right now) which simulates printing and scanning a PDF, and adds signatures

@FaheemMitha enough to wreak havoc

@terdon PDFs are text files, and if you make sure they’re not compressed they’re plain text (ignoring images and embedded fonts)

@FaheemMitha no, this is when I need to change text or move things around or change font sizes etc.

@StephenKitt Ok.

@StephenKitt Yes, OK, written in postscript. I get how you can edit the source, but how do you tell emacs to open as text and not render the pdf?

@StephenKitt So all that binary gibberish is misleading? Is that just a compression artifact?

@terdon C-c C-c

Magic!

@FaheemMitha not necessarily, PDFs can also be encrypted

and embedded images show up as gibberish

I work with PDF a lot, because TeX has now standardized on PDFTeX. So the default output of TeX is now PDF files. It used to be DVI, which you would convert to PS using dvi2ps. Of course, that path still exists, but I think it's used less and less.

Ah, no. I guess the file I'm looking at is compressed. I can't make heads or tails of this and searching for a string I know exists in the file fails.

@StephenKitt Ok.

So, it is possible to annotate a PDF file by entering PDF code by hand? Or too difficult to be worth the trouble?

And is it possible to uncompress a compressed PDF file?

@FaheemMitha technically yes, but it’s too difficult

of course, PDF annotations are also a thing, but I don’t think that’s really what you’re talking about here

@StephenKitt That's what I thought.

@StephenKitt I wasn't talking about them as a separate entity.

@FaheemMitha yes, I realised that after the fact

it’s possible to add annotations to PDFs, they show up as post-it notes

@FaheemMitha yes, I don’t have the incantation to hand though

@StephenKitt Interesting. If you find it, post it.

ah, found the printing-signing-scanning simulator: gitlab.com/edouardklein/falsisign

@StephenKitt TeX could do such things too.

At least with the new LuaTeX souped up engine.

Probably overkill for this use case, though.

The gap at the top of the original pong game that the paddles couldn't reach was a bug that the creator decided to make a feature

Anyone using Firefox may confirm this bug with other video provider. bugzilla.mozilla.org/show_bug.cgi?id=1657009

Hey @Braiam. How goes it?

As usual, this isn't on topic, but it's unexpectedly entertaining, so I thought I would share it - youtube.com/watch?v=m1W-oXZ_31U

Debate at the Oxford Union. Video by Al Jazeera.

The presenter is British. Does a pretty good job. From 2015.

@FaheemMitha It's going. Finally the administration recognized that remote work is a way to work