Reminder that compilation from source is absolutely futile to support because of the arbitrary nature of libs and deps on the machine at time of build and replication of the environment for repro. KThx
Yeah. Most chaotic packages are pre built binaries. Saves you time, and usually not problematic, but has the potential to screw you in a big way. Kinda like opening anything in Outlook in Windows.
I would argue that it’s their own fault then. Laziness is not a valid excuse to put yourself so much at risk. If you start doing it consistently, it becomes a habit and won’t take much effort. Of course, the familiarity with PKBUILD syntax has a learning curve
But a peer-reviewing system would be a better approach in AUR. Weird that it’s not been implemented yet.
I guess it can be assumed that a good number of people read the PKGBUILDs, so at some point malware would be found. A peer-reviewing system would give people a false sense of security, since the AUR is a user repository, where breakage should be expected (compared to the official repos).
How would peer reviewing in a user repo be more a sense of false security compared to official repos? I don’t know any of the arch maintainers, so for me it’s also pure trust they don’t do shady stuff.
Peer reviewing would not be failproof for sure, but at least it would give more security than not reviewing the pkbuilds, and especially to those that aren’t too familiar with them
I think the argument is pretty solid as an alternative to writing PKGBUILDs yourself. Sure it doesn’t hold up for people unfamiliar, but Arch is build on the idea of getting yourself familiar with it.
Agreed. People should learn to read PKGBUILDs, but given how popular Arch(-based) distributions are, I do think many people won’t bother. Afterall, many people download random things all the time.
Add comment