Read the other recent discussion on Debian moving to mount /tmp as tmpfs in upcoming stable release. It’s just confusing at the least if not a bad idea outright that systemd-tmpfiles goes out to delete files in /tmp regardless of how and by whom files were created. /tmp is a shared folder.
Guys, look. You know when I say systemd is evil and will screw you over, and never like any of my examples?
HERE. RIGHT HERE. This is the evil and/or incompetence I don't want interacting with my machines. I, and the rest of the community, frankly, is being held hostage.
After only reading the headline, I expected Poettering to be quoted saying, "Functioning as designed. Go pound sand." The response was as expected, but apparently Poettering's attitude has become contagious within the systemd devs:
> Initially the bug report was shot down by systemd developer Luca Boccassi of Microsoft with:
> So an option that is literally documented as saying "all files and directories created by a tmpfiles.d/ entry will be deleted", that you knew nothing about, sounded like a "good idea"? Did you even go and look what tmpfiles.d entries you had beforehand?
> Maybe don't just run random commands that you know nothing about, while ignoring what the documentation tells you? Just a thought eh
I'm mystified at the attitude. Why not a "er. I see how that's a footgun, although it is the documented behavior. Lets figure out how to make this safe, just remove it because it's not very useful, or at least put big warnings on it."
Who cares if the user ran random commands? Users are going to do that sometimes, even highly experienced users sometimes have a bad day. Commands ought not erase the home directory unless it's fairly obvious that they're going to do so.
The task of the programmer is to make the computer work well even with imperfect use... if we imagine a spherical perfect user in simple harmonic motion we'd need only equip them with a repl prompt and they could just directly write the code for the computer to do exactly what they want.
Given that we're authoring code for third parties at all we must be assuming that the user is bounded in terms of time, knowledge, and potentially capability and attention... and in every case any fault is a share consequence of the author and the user. If nothing else at the extreme it's the programmers fault for failing to anticipate a sufficiently advanced idiot or the users fault for choosing to run some bozo's software. :)
And here the user could simply have privately cursed the software and went on without reporting it, leaving it to some other hapless victim to blow their foot off... so a bit of explicit gratitude would have been justified in both directions, for reporting the bad behavior and for looking into the report.
