I use Windows btw
Right, but if you don’t already know what those arguments mean, then its not exactly super obvious as to what they do.
For example, I’ve been told that you’re generally not supposed to do pacman -Syyu because it can result in partial upgrades (unless I’m remembering the wrong set/combination of flags, which would just be case in point…) - I tend to remember flags by associating them with words.
For example, the common flags for tar was brought up in one of the sibling comments here, but at least I can remember them by:
- x: eXtract
- z: Use
gzipfor the operation (which originally I remembered as “the file has .gz” in its extension before I knew whatgzipwas) - v: verbose, giving the details about what its doing
- f: file, the file name you’re wanting to work with
- c: create archive
- t: test archive (I use this with the
vflag to see what is in an archive before extracting it)
But with Pacman, even after using it on and off for a couple of years, I can try to estimate what the flag names are, but have no idea if its right without double checking:
- S: Sync with repositories
- s: No idea on this one
- y: Also no idea on this one
- Q: Query (search)
- u: Allow upgrades (?)
Really I think its the fact that some of the flags can be used in different combinations which have different effects - like passing z to tar doesn’t change the effect of the flag whether you use x or c. Yet apparently -Syyu and -Syu are valid but one does a proper supported upgrade, and one does a partial upgrade which is not supported and is generally not recommended unless you know what you’re doing. I also know of no occasion where passing the same flag to tar multiple times mutates its behavior as well.
That makes it feel like an anti-pattern to me, similar to using magic numbers in programming. Maybe there is a valid reason for this decision (such as why the Linux kernel uses magic numbers in syscalls) but the result is still that it feels incredibly foreign to me, despite having 10 odd years of Linux experience under my belt.
I agree with that. My reply was to illustrate how cryptic sometimes the most common pacman commands are, and you have to refer to both –help and man pacman, though I find it easier to go to Arch Wiki or Google it generally. The tar flags are great! Pacman can be a bit confusing.
- s is for search, I believe. When you add the S flag, so it becomes pacman -Ss, it synchronises the package database and search within that database.
- y is refreshing local database or something like that? If you put pacman -Syyu, it will force refresh the master package database so that it can deal with corrupted databases or something like that. Pacman seems advanced to me but I just don’t really remember / know my way around it that well.
- Q is for searching local package database.
- u is for system update, since it is preferable to rebuild the whole package database in your system to synchronise against the database containing the latest packages.
I hope I’m somewhat correct. There are other flags but I don’t remember them. It is elegant if you already are familiar with the commands, but otherwise, other package managers and program commands like tar are much easier since they contain English language imperatives.
As for the magic numbers, that’s the first time I knew of that concept. Quickly looking at Wikipedia seems to suggest that there probably are accepted use cases, though that might be historical and not entirely logical.
Really, sometimes computers can be a mess, we’re just used to that mess.
Those definitions of pacman’s flags sounds correct to me, at least when I correlate them to my (admittedly rough) memory of using them in the past. I generally ended up using yay for handling both Arch and AUR package management, but I did most of my package searching via DDG’s !arch and !aur “bangs” to look at it on the web (or just yay likely-the-package-name) which explains my lapse in the Q/s flags.
As for the magic numbers, that’s the first time I knew of that concept. Quickly looking at Wikipedia seems to suggest that there probably are accepted use cases, though that might be historical and not entirely logical.
I won’t pretend like I’ve seen every possible API out there, but the only time that I personally know the use of magic numbers made sense from a developer standpoint was in Linux’s reboot system call which this question/answer does a good job of explaining the reasoning behind it but the quick rundown is that its a safety feature; A bit-flip could inadvertently trigger the reboot system call to be ran instead of something like say exec() which you definitely wouldn’t want to have done on your PC while you’re working (or even worse, something like a web server) . The magic numbers are required for the reboot system call to eliminate this possibility because its practically impossible to inadvertently pass in the exact correct numbers for the reboot syscall to another one in the same way that its unlikely you’d accidentally unlock the guard to a fire alarm and pull the trigger by just brushing up against it.
(Okay, maybe that rundown wasn’t quicker than the answer on stack-exchange… oops)
I suspect any other use cases for magic numbers are very similar to this (a safety precaution against accidentally calling a destructive API) or some silly (and weak) “security through obscurity” type of principal. Somehow I doubt the “magic arguments” for pacman are for either though 😅
Really, sometimes computers can be a mess, we’re just used to that mess.
Oh yes for sure, computers are quite the double edged sword - vastly handy and important tools but also a royal pain in the ass way too often…
That makes sense about the magic numbers! Though I suspect that that isn’t the case for pacman’s commands. This might be totally false but Arch is designed with simplicity and elegance in mind, and so short single letter commands to manage your entire system is preferred.
Just out of curiosity, are you a developer or sysadmin?
