I’m beyond irritated by the presence of an output breaking easteregg in a GNU utility under standard usage. I should be able to depend on my outputs being repeatable.
Honestly, the fact that this got into the codebase makes me question the code review process of GNU and what else is potentially hiding in there.
Like the OP and maintainer said, no one ran into this for 6 years—that's a darn long time in the lifetime of a program. And, as the maintainer said, "it was never meant to affect non-error cases." He pushed a change to fix that [1] and this won't break any builds in the future.
Personally, I really appreciate that a little whimsy was added into a terminal command that's literally built for reading manuals for command-line programs. IMO, this was a net-positive for humanity.
I understand what you are saying, but keep in mind, you are also bitching about software you didn't pay for, probably didn't spend time coding, yet get to use for free that probably made your career and livelihood, AND you get the source code for.
Maybe if you took a step back, you wouldn't feel so irritated. You could be stuck working on an AS400.
Not that I think it has much relevance as an argument, but my employer pays for Ubuntu. Which in turn seems to be this developers employer. Presumably this is part of his job there, since the commit was made during the day.
As I said, not that I think it matters. But "no one pays for open source" frequently just isn't correct these days.
>But "no one pays for open source" frequently just isn't correct these days.
I understand it's paid for one way or the other, but to use it, it is free. I'm comparing it to before, if you wanted to try a Microsoft product or something, you had to buy it and once you broke the seal, you owned it* (or the little plastic disk and box).
Also, if the OP was a maintainer, I can see where he would get frustrated. I assume he was just some overzealous user.
Anyway, the fact that open source even exists is probably the best thing to happen to computing since it's invention (and I'm not even much of an open source user).
Is it not part of a standard set of utilities with standard, expected behavior? You can't really complain if you create some public API and people use it in a way you did not entirely anticipate.
> The string "gimme gimme gimme" can be found in RHEL, OpenSUSE, Fedora, Debian and probably more, so it's not really distro specific. You can grep your man binary to verify.
Technically the behavior is standard, for some definition of standard.
I know it can be tedious, but if you don't trust a piece of free software you have the opportunity to analyze its source code and decide for yourself if it's ok for your use case.
Also, it's debatable whether `man -w` without further arguments is standard usage, since it's clearly an incomplete command (that should return nonzero, but that's another story).
Being "beyond irritated" by a harmless easter egg in a developer facing tool which hasn't even been found in 6 years, is some pretty serious pedantry. Maybe the kind of pedantry you see in people who work with safety related embedded software like aircrafts or automobile software.
Unless you're someone who write that kind of embedded software, and depends on the stderr output of man to control some critical parts of that system (which would be insane) your level of pedantry is unwarranted.
Man is an UI for people. Complaining about its output is like complaining about the GUI of a PDF reader. Sure, you could take snapshots of it, parse it in your system, and be slightly annoyed once by a small easter egg, but it would be kind of stupid because that's not what it was build for.
I don't think it is that irritating as such. But it's the principle. It much easier to convince people that they should be prudent than at which moments jokes are appropriate. I recently found out that some mplayer developer thought it was funny to include "Core dumped ;)" as the last line when you dump a stream. I'm sure that it was funny to them, but less so when you are just trying to figure out if it works after battling deceptive copyright protection, and learning all the settings of different tools, for a couple of hours. Now try and convince the developers that this, or other jokes, isn't funny.
> Man is an UI for people.
Even the man page includes a couple of examples where it isn't.
You're welcome to write your own version of coreutils, then, if it's that a big deal for you (you may want to take the opportunity to raise the size of buffers everywhere to fit your machines resources realistically).
Honestly, the fact that this got into the codebase makes me question the code review process of GNU and what else is potentially hiding in there.