You are not logged in.
I want my parents to upgrade their Arch installation every week. I can get that done with PackageKit which is working absolutely perfect.
However, it turns out that it's not doing a 'full upgrade'. For stuff like 'Replace libusbx with core/libusb? [Y/n]' I still have to start a terminal, become root, remount partitions rw and run the pacman manually.
I already checked /etc/Packagekit (what's with the capitals?) and did not find a knob to make the gpk-update-viewer perform a genuine 'sysupgrade' as stated in Pacman.
For example, Packagekit happily mentions to update nfoview (which I intentionally downgraded). But mentions nothing about libusb replacing libusbx.
[root@delta gebruiker]# pacman -Suu
:: Starting full system upgrade...
:: Replace libusbx with core/libusb? [Y/n] n
oplossen van afhankelijkheden...
controle van interne conflicten...Packages (1): nfoview-1.13-1
Total Download Size: 0,07 MiB
Total Installed Size: 0,34 MiB
Net Upgrade Size: -0,20 MiB:: Proceed with installation? [Y/n]
Anyone got any ideas?
fs/super.c : "Self-destruct in 5 seconds. Have a nice day...\n",
Offline
I've never tried packagekit but if you are saying you wish to automatically accept the upgrades without first reading pacman's output that is likely to cause a problem sooner or later. Also, if it it really is only doing partial upgrades that will be a problem.
All men have stood for freedom...
For freedom is the man that will turn the world upside down.
Gerrard Winstanley.
Offline
I've never tried packagekit but if you are saying you wish to automatically accept the upgrades without first reading pacman's output that is likely to cause a problem sooner or later. Also, if it it really is only doing partial upgrades that will be a problem.
So it's a problem either way? What's your point? ![]()
I would be interested in this as well; I understand the alpm-packagekit writer in not wanting to force upgrades when pacman explicitly wants user-interaction; but an option for auto/unmaintained upgrades would be nice to have
Offline
I've never tried packagekit but if you are saying you wish to automatically accept the upgrades without first reading pacman's output that is likely to cause a problem sooner or later. Also, if it it really is only doing partial upgrades that will be a problem.
loafer wrote:I've never tried packagekit but if you are saying you wish to automatically accept the upgrades without first reading pacman's output that is likely to cause a problem sooner or later. Also, if it it really is only doing partial upgrades that will be a problem.
So it's a problem either way? What's your point?
I would be interested in this as well; I understand the alpm-packagekit writer in not wanting to force upgrades when pacman explicitly wants user-interaction; but an option for auto/unmaintained upgrades would be nice to have
Yes: loafer is explaining the trade-off. Why would you? -> Fully automatic updating. Why would you not? -> Stuff breaks. But stuff might break anyway while doing partial upgrades. So no hard feelings about that (otherwise I should have chosen Ubuntu).
The fact that Arch is treating this seperately is quite 'interesting'. Why the artificial seperation? AFAIC, Pacman does not even have a switch to update like PackaKit is doing right now. What makes replacing a package more daunting with a valid replacement than upgrading another package's version. One could even argue that replacing a package with another is considered an upgrade consisting of a complete rewrite for all I care.
fs/super.c : "Self-destruct in 5 seconds. Have a nice day...\n",
Offline
So it's a problem either way? What's your point?
His point is that you shouldn't automate the upgrade process. Arch just isn't built for that.
Offline
you could try creating a feature report asking for a new parameter, something like this :
--alwaysconfirm
Answer Yes to all “Are you sure?” messages. Be aware that if this breaks your system, you can't blame it on Allan.Although i think using this paramater would cause much more problems then it solves.
Last edited by Lone_Wolf (2014-03-02 13:50:22)
Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.
clean chroot building not flexible enough ?
Try clean chroot manager by graysky
Offline
You could also complement the weekly upgrade with a cron.hourly script that calls
yes|pacman -SyuLast edited by Spider.007 (2014-03-02 13:59:20)
Offline
No, that's an awful idea. Use pacman's --noconfirm switch if you want to auto-dismiss prompts, it'll use the default (recommended) response, instead of just blindly agreeing to it.
Sakura:-
Mobo: MSI MAG X570S TORPEDO MAX // Processor: AMD Ryzen 9 5950X @4.9GHz // GFX: AMD Radeon RX 5700 XT // RAM: 32GB (4x 8GB) Corsair DDR4 (@ 3000MHz) // Storage: 1x 3TB HDD, 6x 1TB SSD, 2x 120GB SSD, 1x 275GB M2 SSD
Making lemonade from lemons since 2015.
Offline
I see many arguments as to why simply ignoring these messages is bad.
But I hear no arguments as to why a package replacement requires user intervention and a package upgrade does not. Once again, some updates which are complete rewrites of pieces of software could be considered replacements for the older package.
The difference between an update and a 'replacing update' is blurred.
Furthermore, I'm not expecting my parents to parse terminal output and make informed decisions. They just want to read their mail and view websites. Who cares if their libusbx is replaced by libusb?
fs/super.c : "Self-destruct in 5 seconds. Have a nice day...\n",
Offline
Might I suggest that Arch might not be the best choice of distribution for your parents? Maybe they would be happier with something like Linux Mint?
Who cares if their libusbx is replaced by libusb?
Who knows? Who cares if their firefox is replaced by midori? Or their VLC is replaced by mpv? Pacman doesn't know what package replacements are important to you, and which ones you don't care about. All it sees is that package-A replaces package-B, and asks you, the user, how you want to proceed.
Sakura:-
Mobo: MSI MAG X570S TORPEDO MAX // Processor: AMD Ryzen 9 5950X @4.9GHz // GFX: AMD Radeon RX 5700 XT // RAM: 32GB (4x 8GB) Corsair DDR4 (@ 3000MHz) // Storage: 1x 3TB HDD, 6x 1TB SSD, 2x 120GB SSD, 1x 275GB M2 SSD
Making lemonade from lemons since 2015.
Offline
But I hear no arguments as to why a package replacement requires user intervention and a package upgrade does not.
Unfortunately I cannot answer this. Maybe the devs can provide a reason.
The difference between an update and a 'replacing update' is blurred.
The way I see it is an update simply means a newer version, e.g. v1.0 --> v1.1, where as a "replacing update" might mean the program name was changed, e.g. ethereal --> wireshark, or in the case of libusb, two projects were merged: libusb and libusbx.
Furthermore, I'm not expecting my parents to parse terminal output and make informed decisions. They just want to read their mail and view websites.
Yet you have them running Arch Linux?
Last edited by anonymous_user (2014-03-02 18:15:07)
Offline
Who cares if their libusbx is replaced by libusb?
Who knows? Who cares if their firefox is replaced by midori? Or their VLC is replaced by mpv? Pacman doesn't know what package replacements are important to you, and which ones you don't care about. All it sees is that package-A replaces package-B, and asks you, the user, how you want to proceed.
Excellent point.
@ Rexilion, it's not just about libusb and libusbx as you currently see it. For example, a few months back, procps-ng replaced sysvinit-tools, and sysvinit-tools tools provided killall5 which was needed for me to shut down my OpenRC based system. So, I declined the pacman prompt, waited for the AUR sysvinit package to include killall5, then proceeded with the install.
Last edited by x33a (2014-03-02 18:30:58)
Offline