I’m gonna be real. I dont think home directory files should handled by something named tmpfiles. I think something named tmpfiles should only handled volatile data as it is colloquially known. I get there’s a lot that can be considered that in home directories. But user data should be handled by something that is made for that and users inherently know by the name that this will mess your home directory. This only applies to monoliths and monoliths to be like systemd. Maybe you’re doing too much with a single program if your naming becomes problematic.
I’m gonna be real. I dont think home directory files should handled by something named tmpfiles.
But… but… it was in the documentation! /s
What killed me about the whole thing was how defensive the dev was about the whole thing, basically calling the reporter a moron for running a command without extensive knowledge of the entire system. I don’t care how good the documentation is, if open file proceeds to format your hard drive in some circumstances, you done goofed as a dev.
I agree his answer sucks, but perhaps the fault still lies with the distribution developers (who should know better), not the authors of systemd. In that context I can understand the resentment expressed by the dev. It’s not directed toward the end user but toward the distro developers, who have implemented systemd in a broken and dangerous way.
That first response was awful, like fuck you man, maybe I should read the doc but maybe your software should be better designed if it handles my data like that. Jesus Christ…
Tbf, this is something that only some distros do. Those distros should be reprimanded for handling home directories with the tmpfiles system, not systemd.
I dunno, I don’t have a camera feed into your life. But considering that is the first thing you respond to a clarification it most certainly wouldn’t surprise me if you did.
I’m gonna be real. I dont think home directory files should handled by something named tmpfiles. I think something named tmpfiles should only handled volatile data as it is colloquially known. I get there’s a lot that can be considered that in home directories. But user data should be handled by something that is made for that and users inherently know by the name that this will mess your home directory. This only applies to monoliths and monoliths to be like systemd. Maybe you’re doing too much with a single program if your naming becomes problematic.
But… but… it was in the documentation! /s
What killed me about the whole thing was how defensive the dev was about the whole thing, basically calling the reporter a moron for running a command without extensive knowledge of the entire system. I don’t care how good the documentation is, if
open file
proceeds to format your hard drive in some circumstances, you done goofed as a dev.I agree his answer sucks, but perhaps the fault still lies with the distribution developers (who should know better), not the authors of systemd. In that context I can understand the resentment expressed by the dev. It’s not directed toward the end user but toward the distro developers, who have implemented systemd in a broken and dangerous way.
That first response was awful, like fuck you man, maybe I should read the doc but maybe your software should be better designed if it handles my data like that. Jesus Christ…
Especially when “tmpfiles” is an existing term of art with a very specific meaning
Tbf, this is something that only some distros do. Those distros should be reprimanded for handling home directories with the tmpfiles system, not systemd.
What distros are those?
The only reason its still called tmpfiles is because of backwards compatibility
Well, the back is my hand is compatible with the dev’s face. You don’t see me slapping people all willy-nilly.
I dunno, I don’t have a camera feed into your life. But considering that is the first thing you respond to a clarification it most certainly wouldn’t surprise me if you did.