You are not logged in.
Given how often it is used, I propose supporting -A as an alias for --auronly; in fact that is what I did to my clone of packer. Actually 90% of the time I call packer --auronly. To be careful, we can use a more obscure one like "-." or "-/" to avoid possible collision with pacman options in future (if such radical changes like adding a new capitalized option should ever happen with pacman development ;-) ).
Offline
Why not use a shell alias for that?
alias pa="packer -Syu --auronly"
Offline
I'm not sure how clean or efficient my code is in comparison to packer's, but I've been able to implement basic ABS support for packer (installing with -Sb and updating from an array that you define). I also fixed the fact that it didn't use dependencies that are added/removed by customizepkg (at the cost of a little speed). Installing isn't quite as pretty for ABS and there's a few areas that could be improved, but it seems to be almost as fast as it is for AUR packages. It'll require rsync for the ABS support, and I changed the "pacman" 's to "$PACMAN" 's for better integration with pacman-color (but these aren't essential to the ABS patch). I uploaded the patch here in case anyone's interested in using/improving/burning it: http://pastebin.com/qZKSMJN4
Edit: Cleaned up ABS update support.
Last edited by DarkSniper (2012-03-02 23:25:29)
Failure is not an option... It comes bundled with Windows.
Offline
Quick question: why can't packer to -a command, as in "-Syua"? If it did, it could replace yaourt completely.b
Offline
Quick question: why can't packer to -a command, as in "-Syua"? If it did, it could replace yaourt completely.b
I think the philosophy was to be using both pacman and packer (which is why it doesn't wrap things like -Q* and -R* by default). If you didn't want aur updates, wouldn't you be better off with a simple "pacman -Syu"?
Last edited by DarkSniper (2012-02-26 07:23:58)
Failure is not an option... It comes bundled with Windows.
Offline
Since updating packer, on all my systems I get "sudo: sorry, you are not allowed to preserve the environment" when running packer and it cannot use pacman. I've tried adding "Defaults !env_reset" or "Defaults env_keep += "PACMAN" to sudoers and tried setting/exporting PACMAN=pacman, to no avail.
Do others have this problem? Any hints?
Offline
Yes I have this problem but no solution yet.
Online
You need the line
Defaults setenv
in your sudoers file, then it should work
Offline
I found "set_env" somewhere and it wasn't working. setenv works.
Thanks, jakob.
Offline
Thanks jakob
Online
Ignore the post below. I see now that it's the result of a bad PKGBUILD.
packer seems to want to re-install packages that have already been installed/upgraded for me. For example:
$ packer -Si lib32-sqlite3 | grep Version
Version : 3.7.10-1
$ pacman -Qi lib32-sqlite3 | grep Version
Version : 3.7.10-1
$ packer -Si lib32-curl | grep Version
Version : 7.24.0-1
$ pacman -Qi lib32-curl | grep Version
Version : 7.24.0-1
$ packer -Syu
[...]
:: Starting full aur upgrade...
Aur Targets (4): lib32-curl lib32-boost-libs gufw plexmediaserver
Pacman Targets (2): lib32-libxml2 lib32-sqlite3
^c
Last edited by emphire (2012-03-22 17:56:46)
Offline
Packer doesn't seem to handle AUR pkgbuilds that create multiple packages. It just install the first.. Yaourt can handle them.
Offline
@dieelt: I've been able to build linux-ck with it properly which creates two packages: linux-ck and linux-ck-headers, both of them are passed to pacman and installed. Which package specifically isn't installing correctly for you?
Last edited by DarkSniper (2012-03-28 18:19:21)
Failure is not an option... It comes bundled with Windows.
Offline
I'm having some problems upgrading packages with packer -Syu. Everything seems to work fine, but after the package is created the following error occurs,
==> Finished making: python-keybinder 0.2.2-5 (Wed Mar 28 20:24:58 CEST 2012)
loading packages...
error: 'python-keybinder-*.pkg.tar': could not find or read package
Upgrading packages individually works fine using packer -S foo-package. If I check the packer tmp folders, I see no difference between running with the -S or -Syu flag.
[edit]
I just checked on another system running Arch and there the problem does not occur.
[/edit]
Last edited by FreeTheBee (2012-03-28 19:06:09)
Offline
@DarkSniper The package that do not install both are https://aur.archlinux.org/packages.php?ID=41737, it just installs appmenu-gtk2 and not appmenu-gtk3.
Offline
I know I'm coming into this post really late, and I don't feel like reading through 26 pages to make sure this hasn't been brought up already, but I think it's a bit insensitive to be flinging mud at yaourt devs. Saying their tool is poorly written in the title of your post? That may be your legitimate opinion, and I support your right to express it, but I think some opinions are better left to one's self. How would you feel if I created a new tool and in the title line I write (replacing the horrid packer code)? If this has already been brought up feel free to ignore this, and don't take it as me saying you don't have the right to your opinion.
Offline
New bug:
if you try to install a package and write the package name with upper case letters, for example:
$ packer -S JUMANJI
Aur Targets (1): JUMANJI
Proceed with installation? [Y/n]
/usr/bin/packer: line 287: cd: JUMANJI: No such file or directory
No PKGBUILD found in directory.
Offline
Well Linux is case sensitive is it not?
Offline
Well Linux is case sensitive is it not?
Yup, but all letters in the pkgname should be lowercase, so you can add an extra step of converting all letters the user typed to lowercase.
Offline
I know I'm coming into this post really late, and I don't feel like reading through 26 pages to make sure this hasn't been brought up already, but I think it's a bit insensitive to be flinging mud at yaourt devs. Saying their tool is poorly written in the title of your post? That may be your legitimate opinion, and I support your right to express it, but I think some opinions are better left to one's self. How would you feel if I created a new tool and in the title line I write (replacing the horrid packer code)? If this has already been brought up feel free to ignore this, and don't take it as me saying you don't have the right to your opinion.
I agree with this. But the real reason I don't use / switch to packer is:
~ >> date && packer -Syu --auronly && date
Tue Apr 17 14:55:00 PDT 2012
:: Synchronizing aur database...
aur 114 114 [########################################]100%
:: Starting full aur upgrade...
local database is up to date
Tue Apr 17 14:55:10 PDT 2012
~ >> date && yaourt -Su --aur && date
Tue Apr 17 14:55:19 PDT 2012
Foreign packages: - 114 / 114
Tue Apr 17 14:55:20 PDT 2012
~ >>
I would rather use a poorly written yaourt that is 10 times faster at checking for updates (I check for updates a lot). Not sure if it is a configuration error on my end or a wget vs curl thing or whatever but I think I'll wait until packer gets better. There is one thing about packer I do agree with a lot though: Using pacman for pacman tasks and an aur wrapper only for the aur.
Offline
I know I'm coming into this post really late, and I don't feel like reading through 26 pages to make sure this hasn't been brought up already, but I think it's a bit insensitive to be flinging mud at yaourt devs. Saying their tool is poorly written in the title of your post? That may be your legitimate opinion, and I support your right to express it, but I think some opinions are better left to one's self. How would you feel if I created a new tool and in the title line I write (replacing the horrid packer code)? If this has already been brought up feel free to ignore this, and don't take it as me saying you don't have the right to your opinion.
I would rather use a poorly written yaourt that is 10 times faster at checking for updates (I check for updates a lot). Not sure if it is a configuration error on my end or a wget vs curl thing or whatever but I think I'll wait until packer gets better. There is one thing about packer I do agree with a lot though: Using pacman for pacman tasks and an aur wrapper only for the aur.
packer and yaourt have kind of has a neat history. Let me see if I can can summarize it to help you better understand why bruenig wrote what he did.
Before packer, yaourt somehow became the "default" AUR helper. It just ended up being suggested a lot on the Arch wiki and the forums, and many people ended up using it.
At that point in time, development of yaourt was getting pretty stale. bruenig reviewed the source code for yaourt and found:
yaourt was incredibly inefficient and slow.
yaourt was just plain wrong.
(bruenig gave specific examples of these things earlier in this thread)
So, instead of just complaining, bruenig wrote packer, his own pacman and AUR wrapper that only required BASH. It was much faster, simpler, and gave more accurate results than yaourt. Many Arch Linux users found it to be a very nice AUR wrapper.
Now, I can't say that it was because of packer (and bruenig), but after this thread and the creation of packer, yaourt development picked up, the result of which is the very nice yaourt application you use today.
So in summary, yaourt and packer are now both very nice applications, but it wasn't always that way.
Side note: bruenig has a sort of... abbrasive online personality. I find it funny, but I understand that some people wouldn't.
Last edited by drcouzelis (2012-04-18 17:25:53)
Offline
Now, I can't say that it was because of packer (and bruenig), but after this thread and the creation of packer, yaourt development picked up, the result of which is the very nice yaourt application you use today.
So in summary, yaourt and packer are now both very nice applications, but it wasn't always that way.
Interesting. I do not know the history of yaourt vs packer 1st hand because I did not use Arch back then. I have noticed that a lot of people recommend packer over yaourt for some reason. I have heard some argue that it does searches better / more accurately. Other than that is there a great benefit as it stands nowadays to use one over the other?
Offline
I like packer and al but I dislike one thig
download deps and them edit PKGBUILD, not giving the option to change deps and past this edit the PKG
example: you use a git package that build from stable repo, but you want use the git equivalent (for this example are disponibles on AUR), you need download the stables, wayt edit the PKGBUILD and run again the package for download again.
is exaspperant... especialy for not-updated pkgbuild: example...buildables but not updated fopr the news slits packagin of a X package...
Last edited by Jristz (2012-04-18 02:16:18)
Well, I suppose that this is somekind of signature, no?
Offline
I have noticed that a lot of people recommend packer over yaourt for some reason. I have heard some argue that it does searches better / more accurately. Other than that is there a great benefit as it stands nowadays to use one over the other?
I don't know of a great benefit from either. I haven't used yaourt very much. I've used packer quite a bit (I wrote the man page for it) but I wouldn't consider myself a packer expert or anything.
Anyway, I change AUR wrappers every few months to whatever suits my needs / desires at the time and to just try something new.
Offline
I like packer and al but I dislike one thig
download deps and them edit PKGBUILD, not giving the option to change deps and past this edit the PKGis exaspperant... especialy for not-updated pkgbuild: example...buildables but not updated fopr the news slits packagin of a X package...
If I may enter the discussion: this is exactly the reason I consider that packer has a big flaw in its design, and why I developed my very own solution (using other well written tools and ideas that bring many other benefits over packer/yaourt).
Last edited by Spyhawk (2012-04-18 11:48:07)
Offline