/dev/chat

« first day (3244 days earlier) ← previous day next day → last day (1718 days later) »

Kusalananda

10:54 AM

I'm doing all the fancy triple letter things today! Both VNC and VPN.

Jeff Schaller

11:12 AM

@Kusalananda consider casting a VTC from the VPN'd VNC window :)

Kusalananda

@JeffSchaller "Voting" in that context is a bit of an odd word to use when you're a moderator. It just closes the question... :-/

Jeff Schaller

@Kusalananda I couldn't come up with another three-letter acronym containing V & N, so I settled on VTC

Kusalananda

Reminds me about the voting in Ankh-Morepork (Discworld novels): [something like] It's one man, one vote, and the Patrician was the man, and he had the vote.

Not much point in running VNC though, as I'm always use just a terminal anyway, so SSH over the VPN would be quite enough.

Right, off to the bus.

Faheem Mitha

11:44 AM

@Kusalananda "They were the salt of the earth. Square, hard, and bad for the health"

From Mort, I think. I've mostly only read the early Discworld novels.

The actual quote is:

> They were the sort of people generally called salt of the earth. In other words they were hard, square and bad for your health...

But I did get the novel right.

terdon

Different quote though. I remember the one @Kusalananda means, and it was about the Patrician.

Also from Mort, but:

> “Ankh-Morpork had dallied with many forms of government and had ended up with that form of democracy known as One Man, One Vote. The Patrician was the Man; he had the Vote.”

Faheem Mitha

12:09 PM

@terdon Yes, of course it's a different quote. But similar, in a pull-the-rug-out-from-under-feet kind of way.

The notion of considering the file containing function code as arg 0 of that function. Does that originally derive from C, or Unix, or both?

Kusalananda

@terdon That's the one. I scraped it up from the "15 years ago" sediments of my brain.

@FaheemMitha That has to do with how some of the exec() family of library functions work.

Faheem Mitha

12:26 PM

@Kusalananda C library functions?

Kusalananda

So, I would say it derives from C and/or Unix (it's difficult to distinguish them that far back).

Yes, C library functions.

Faheem Mitha

@Kusalananda Outside C/Unix, is it a thing?

TeX uses it to alter behavior depending on the name of the executable, but I find that unexpected behavior.

Kusalananda

@FaheemMitha I bet many languages inherits it.

Faheem Mitha

I know there are other programs do that too.

Faheem Mitha

12:58 PM

Apparently I've forgotten the plot of Mort. Sic transit Gloria.

Are pipes in Python conceptually similar (or the same) as Unix pipes?

I'm currently looking at the subprocess library/module.

And are pipes themselves implemented as processes?

For example, this answer:

A: Get exit code and stderr from subprocess call

The accepted solution covers the case in which you are ok mixing stdout and stderr, but in cases in which the child process (for whatever reason) decides to use stderr IN ADDITION to stdout for a non failed output (i.e. to output a non-critical warning), then the given solution may not desirable....

Kusalananda

I just realised one of my colleagues was born one year after I started university. Life is strange.

Faheem Mitha

I'm not clear what the communicate function is doing here, other than providing an "end" for the pipes. But perhaps that is its function?

@Kusalananda Why is that strange?

Now, for example, if your colleague started getting younger, that would be strange.

Kusalananda

@FaheemMitha Pipes are not implemented as processes. What that Python code is doing is opening a pipe to a process.

The communicate function is a placeholder for "talk to the process and listen to what it has to say".

Faheem Mitha

@Kusalananda You mean Popen?

Kusalananda

Yes

Faheem Mitha

1:12 PM

@Kusalananda Yes, so just creating a pipe isn't enough? It needs an end? So to speak.

Kusalananda

Or, the script as a whole.

Oops, sorry.

@FaheemMitha Well, you use the pipe until you're done using the pipe, and thon the code checks the status of the process to see whether all went well.

Faheem Mitha

I guess I don't quite see why the subprocess thingy can't just pass the std output and error from the subprocess/child process to the parent by itself. Why does it need further handholding? Perhaps it needs a name?

I mean, something to collect what emerges from the pipe, so to speak.

Kusalananda

You should look up examples of C code using the pipe() call. There's a bit of faffing abot taht you'd have to do. Doing similar things in the shell is comparably easy because the shell handles the "plumbing".

Faheem Mitha

@Kusalananda Suppose you don't want to check anything, and just want output/error to be sent to the parent unfiltered?

The common thing seems to be to collect and inspect stuff. It would be nice if this could be handled automatically.

If there is stuff sent to standard error from the child, the parent should terminate, for example.

@Kusalananda In shell I think that sort of stuff is relatively easy.

Kusalananda

@FaheemMitha Not everything on standard error is an error. The shell, for example, prompts on stderr.

"Standard error" is a misnomer. It should really be called "standard other".

Faheem Mitha

1:20 PM

@Kusalananda It does? Isn't that naughty?

Kusalananda

No, it's not naughty. It's even required by POSIX.

Faheem Mitha

@Kusalananda I think if Unix wants a standard other, it should get a standard other. I like standard error.

2 hours later…

terdon

3:07 PM

Q: Do progress reports/logging information belong on stderr or stdout?

Is there an official POSIX, GNU, or other guideline on where progress reports and logging information (things like "Doing foo; foo done") should be printed? Personally, I tend to write them to stderr so I can redirect stdout and get only the program's actual output. I was recently told that this ...

Faheem Mitha

I think "standard error" is an unfortunate name.

Kusalananda

3:50 PM

Related issue in the Github repo of the development version of the ksh93 shell: github.com/att/ast/issues/1380

1 hour later…

Kusalananda

5:13 PM

@StephenKitt I sorted the "dot-file" thing as I agree with you. But I really can't remember where I saw "dot-script" being mentioned. I thought it actually was in POSIX, but searching I can't find it. It's also not from the bash manual or from zsh. Did I just invent a new word?

Ah, the ksh93 manual has it.

« first day (3244 days earlier) ← previous day next day → last day (1718 days later) »

all rooms