packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

bruenig · 2010-01-05 06:07:38

There are some people who wish to keep aur and pacman separate. For those people there is aurget and slurpy.

There are others who wish to integrate them for certain tasks like myself. For those people there really is only yaourt which is excruciatingly slow and in many ways bloated. For instance, why would you need to use yaourt -R package ever? All that happens when you do that is it calls pacman. What a joke.

I contend that an aur+pacman wrapper only need to perform 4 major functions.

1) Search pacman and aur at the same time with one query (-Ss)
2) Get info from packages in the repository (-Si)
3) Install packages from either (especially be able to handle the dependencies of aur packages which can sometimes have other dependencies on aur which themselves might have dependencies on aur, etc.) (-S)
4) Update packages at the same time (again with an interest of handling aur packages which might have new dependencies in newer versions and such) (-Su)

If you agree with this or just are curious and want to test my program, please download the script from my github and leave feedback. This is a pretty early release (just began working on it 3 days ago), but all four of those functionalities are implemented. There will be cosmetic changes for sure to come and please comment on what you think those should be or any other bugs or problems you might have had.

github link
aur link
packer wiki

EDIT:
I have forked pacdiffviewer from yaourt and modified it to work as a stand alone program due to many people saying they coud not switch from yaourt because it has pacdiffviewer. To avoid path conflicts, I have named it pacdiff.

github link
aur link
forum post about pacdiff

EDIT: Title edited due to gratuitous number of complaints.

Last edited by bruenig (2010-04-22 02:04:13)

.:B:. · 2010-01-05 18:40:23

Kudos on the original project name!

Navi · 2010-01-05 18:43:04

This is so awesome. bruenig is fast and does whatever I tell it to!

Thanks!

belak51 · 2010-01-06 13:50:08

Sweet. Yaourt was the reason I left arch linux. Looks like I might be willing to dual boot again... looking forward to trying this out...

frigaut · 2010-01-06 14:32:34

belak51 wrote:

Sweet. Yaourt was the reason I left arch linux.

What do you mean? No one obliges you to use yaourt... Note that I am in no way affiliated to the yaourt dev team. yaourt may be bloated and somewhat slow, but it does the job. and again, no one makes you use it. There are alternatives.

belak51 · 2010-01-06 14:47:45

I don't like how it asks about 5 times whether you install the package... if I say I want to install it, I want to install it. Like you said, it's also slow and bloated. I love the idea of the AUR but not that there are no official decent tools that work with it.

Stythys · 2010-01-06 16:43:03

there *never* will be official tools to work with it. Leaving arch because of yaourt is ridiculous. other distros don't even *have* the AUR, leaving *because* of the aur makes no sense.

second, use --noconfirm. it won't ask you five times whether you want to install it or not. try reading man pages for once.

and third, there are other decent tools that are being developed.

Last edited by Stythys (2010-01-06 16:43:44)

belak51 · 2010-01-06 20:05:57

I simply expressed interest in this new script because it was simple and (I hope) it works. I agree that I never really gave Arch a chance, but that's no reason to bash me. "Other distros", like gentoo have similar things. Gentoo's overlays, and it's simple to add new repos in other distros. My interest in arch comes because I think I've finally advanced past my "Ubuntu days" enough to read the man pages.

Just a hint, please don't be mean to people who are looking at joining the community... it has a tendency to put them off. Don't get me wrong, I see where you're comint from, I just think it was a tiny bit harsh. I kept using yaourt because everyone said it was the best thing out there for using the AUR. I never bothered with all the configuring. I know enough now to try again. Please be patient with me.

frigaut · 2010-01-06 20:22:22

Yeah, Stythys was a bit harsh with you, granted. But that's because your original bashing of yaourt was a bit on the troll side :-)
I mean, yaourt has its drawbacks, but you will understand that you can upset people (and the yaourt devs in particular) when you say this software "was the reason you left arch linux".
That said, if you feel you'd like to give arch another try, please do. The arch community is particularly helpfull (but don't be a troll !). I was also a ubuntu user and was ready for an upgrade...which I found in arch (after much distro hoping). Been using it for 6 months now and really like it a lot.

Last edited by frigaut (2010-01-06 20:22:38)

bruenig · 2010-01-06 22:52:16

B is such a troll.

Gen2ly · 2010-01-06 23:17:41

frigaut wrote:

Yeah, Stythys was a bit harsh with you, granted. But that's because your original bashing of yaourt was a bit on the troll side :-)
I mean, yaourt has its drawbacks, but you will understand that you can upset people (and the yaourt devs in particular) when you say this software "was the reason you left arch linux".
That said, if you feel you'd like to give arch another try, please do. The arch community is particularly helpfull (but don't be a troll !). I was also a ubuntu user and was ready for an upgrade...which I found in arch (after much distro hoping). Been using it for 6 months now and really like it a lot.

Very nice.

Still use Yaourt myself... gets the job done, glad for the work that they do.

O btw, this is no troll to bruenig ... program looks pretty cool.

Last edited by Gen2ly (2010-01-06 23:19:33)

bruenig · 2010-01-07 01:39:01

Still use Yaourt myself... gets the job done, glad for the work that they do.

I don't understand this. With some benchmarks, I found yaourt to be about 29 times slower than my program (for searching especially) or other programs like slurpy. Why would you use such a things when there are legitimate alternatives?

Last edited by bruenig (2010-01-07 01:40:07)

frigaut · 2010-01-07 02:01:58

bruenig wrote:

Still use Yaourt myself... gets the job done, glad for the work that they do.
I don't understand this. With some benchmarks, I found yaourt to be about 29 times slower than my program (for searching especially) or other programs like slurpy. Why would you use such a things when there are legitimate alternatives?

Well, Bruenig, you uploaded your "packer" yesterday!!! Give some time to your audience :-)

keenerd · 2010-01-07 02:55:02

This is fast. And fails even faster if you don't have sudo set up. The one nice thing about yaourt is that if will check if you can sudo pacman, otherwise it asks for your root password. I like this behavior.

drcouzelis · 2010-01-07 02:55:15

Wow! Thank you for making "packer"! It is way fast. I have made it my new Arch package manager via aliases. I can't even think of anything it is missing.

pacman-color + packer + pacman-cage = winning combination for me.

Also, I totally added it to the wiki: http://wiki.archlinux.org/index.php/AUR_Helpers Please update it as you see fit!

some-guy94 · 2010-01-07 03:07:52

I don't know where else to post this, so here's a patch that remove the need for the tmpfile, and stores everything in variables
http://pastie.org/769777

bruenig · 2010-01-07 08:37:12

some-guy94 wrote:

I don't know where else to post this, so here's a patch that remove the need for the tmpfile, and stores everything in variables
http://pastie.org/769777

Myself and a few others spent about 2 hours in ##off-archlinux on freenode going through a whole series of different combinations of external apps (awk, sed, grep, cut), tmpfiles, variables, etc. And you may be surprised, but it was found to be significantly faster to use tmpfiles than variables on searches. It is a pain to use them, but I'd rather deal with that than slower speeds. Wait until you see the mess of a tmpfile clusterfuck on the new -Ss function now (which includes support for multiple search terms).

keenerd wrote:

This is fast. And fails even faster if you don't have sudo set up. The one nice thing about yaourt is that if will check if you can sudo pacman, otherwise it asks for your root password. I like this behavior.

I will make it do this instead. But I don't fully understand what this means. By, checks if you can do sudo pacman, do you mean checks if you can execute sudo, and then if not uses su -c? I could look through that maze of code, but your input would be better than that anyways. How would you like it to deal with permissions. Right now, you can either run packer or root, or if you are not doing so, it tries to use sudo. I am sure there is a better way, but how?

Last edited by bruenig (2010-01-07 08:40:18)

bruenig · 2010-01-07 09:15:31

Updated version which is on the aur right now fixes the broken PKGBUILD and fixes the update function. You should probaly packer -S packer manually to get this version because if you are waiting on it to update, it will fail. So yeah. It is on the aur. Go and get it. (lots of changes added to it by the way).

Last edited by bruenig (2010-01-07 09:22:20)

Burt · 2010-01-07 10:39:39

I just tried it and it's just amazing compared to the slowlyness (is that correct?) of yaourt. Great work.

(Too bad it doesn't do -R, could have used it to -R yaourt /o\)

Mikko777 · 2010-01-07 11:00:02

Does it do/ can you add these features:

1: yaourt -Syu --aur --devel --noconfirm (--devel is nice)

2: on first install edit/ show pkgbuild

3: yaourt -C <-- list and edit .pacnew files

4: vote for packages

Those are the ones i like yaourt for.
Performance is like the lowest priority I can think off, who watches updates anyways

frigaut · 2010-01-07 12:53:09

Matthew,

It seems packer does not look at the "provides" flag in the PKGBUILD (or in the pacman database).
Consider the following situation:
X is a package
X2 is a modified X that provides the same functionalities as X, thus is marked as 'provides="X"'
Y is another package that needs "X".
When I install Y, with X2 installed, I want the package manager to recognize this, and not try to pull X.

I have this situation. Below, I am trying to install yorick-yutils (in AUR), which depends on yorick (in AUR). yorick-cvs (in AUR) is marked as providing yorick. Yaourt recognizes that I have X2 installed (yorick-cvs), while packer does not and try to pull yorick:

poliahu:~ $ packer -S yorick-yutils

Aur Targets    (2): yorick yorick-yutils

Proceed with installation? [Y/n] n
poliahu:~ $ yaourt -S yorick-yutils

==> Downloading yorick-yutils PKGBUILD from AUR...
tar: Record size = 6 blocks

    
      First Submitted: Wed, 16 Sep 2009 03:02:13 +0000    
yorick-yutils 1.4.0-1 : General utilies for yorick
( Unsupported package: Potentally dangerous ! )
==>  Edit the PKGBUILD (highly recommended for security reasons) ? [Y/n]("A" to abort)
==>   ----------------------------------------------
==>n


==> yorick-yutils dependencies:
 - yorick (package that provides yorick already installed)   <<<< yaourt has the correct behavior
==>  Continue the building of yorick-yutils ? [Y/n]
==>   ----------------------------------------------
==>

Last edited by frigaut (2010-01-07 13:06:13)

drcouzelis · 2010-01-07 13:02:51

keenerd wrote:

This is fast. And fails even faster if you don't have sudo set up. The one nice thing about yaourt is that if will check if you can sudo pacman, otherwise it asks for your root password. I like this behavior.

I don't understand what you mean. I can run packer as non-root, and when it needs root access it asks me for my password. As far as I can tell, this is the same thing that yaourt does. Are you expecting it to do something else?

Mikko777 wrote:

1: yaourt -Syu --aur --devel --noconfirm (--devel is nice)
2: on first install edit/ show pkgbuild
3: yaourt -C <-- list and edit .pacnew files
4: vote for packages

Those are all great ideas. I will add my opinion as another packer user.

1. I've never used the "--devel" option. Does that mean, it will look for AUR packages that you installed that got the files out of a CVS repository, and download and install the super duper just-updated-at-two-in-the-morning version of it? If that's what it does, then I can imagine that being a nice feature, but it can also be done by doing "packer -S", right? As for "--noconfirm", packer seems pretty unobtrusive to me right now, so I don't really miss that option.

2. I'm confused. Packer asks me if I want to see the PKGBUILD when I install a package from the AUR. You mean, show the PKGBUILD instead of asking? Or, only show PKGBUILD on the first install?

3. I personally prefer to do "sudo updatedb && locate pacnew".

4. I never liked the "vote" feature on yaourt. I think it's confusing. If I just finished installing a package from the AUR, how do I know if I want to vote for it if I haven't used it yet? I prefer to use the program for a while, then use "aurvote".

frigaut · 2010-01-07 13:11:01

drcouzelis wrote:

4. I never liked the "vote" feature on yaourt. I think it's confusing. If I just finished installing a package from the AUR, how do I know if I want to vote for it if I haven't used it yet? I prefer to use the program for a while, then use "aurvote".

I fully agree. let's keep packer un-bloated.

Mikko777 · 2010-01-07 13:30:33

drcouzelis wrote:

1. I've never used the "--devel" option. Does that mean, it will look for AUR packages that you installed that got the files out of a CVS repository, and download and install the super duper just-updated-at-two-in-the-morning version of it? If that's what it does, then I can imagine that being a nice feature, but it can also be done by doing "packer -S", right? As for "--noconfirm", packer seems pretty unobtrusive to me right now, so I don't really miss that option.
2. I'm confused. Packer asks me if I want to see the PKGBUILD when I install a package from the AUR. You mean, show the PKGBUILD instead of asking? Or, only show PKGBUILD on the first install?
3. I personally prefer to do "sudo updatedb && locate pacnew".
4. I never liked the "vote" feature on yaourt. I think it's confusing. If I just finished installing a package from the AUR, how do I know if I want to vote for it if I haven't used it yet? I prefer to use the program for a while, then use "aurvote".

1: yaourt -Syu --aur wont update git/svn/bazaar/hg what ever packages unless the packageversion changes, ie pkgbuild has been modified. So that could be months of updates in devel packages, and reinstalling them all by hand (which i guess would update them) is too much work

2: solved then
3: kinda solved, think it should be external app anyways (it does do automerges and vimdiff which is kinda nice)
4: I like voting simply so that the packager gets a nice "thankyou!" from ppl using the package, plus id be too lazy to vote on website.

keenerd · 2010-01-07 13:47:14

bruenig wrote:

keenerd wrote:
The one nice thing about yaourt is that if will check if you can sudo pacman, otherwise it asks for your root password. I like this behavior.
I will make it do this instead. But I don't fully understand what this means. By, checks if you can do sudo pacman, do you mean checks if you can execute sudo, and then if not uses su -c? I could look through that maze of code, but your input would be better than that anyways. How would you like it to deal with permissions. Right now, you can either run packer or root, or if you are not doing so, it tries to use sudo. I am sure there is a better way, but how?

It sounds like you understand it :-) Here's a use story.

yaourt: As root user, I try to build/install something. It says

==> WARNING: Building unsupported package as root is dangerous.
 Please run yaourt as a non-privileged user.

So I exit root and build. It builds. Time to install.

==>  Continue installing packer? [Y/n] Y
You are not allowed to launch /usr/bin/pacman with sudo
Please enter root password
Password:

packer: As a root user, I try to build something. It goes ahead and does it. Thankfully I Control C out before a malicious AUR script wipes my drive. I try to build as user. It builds. Time to install. It asks for my user password.

Password: 
Sorry, user **** is not allowed to execute '/usr/bin/pacman -U ****' as root on ****.

And then it quits.

In the yaourt script, check out line 572, function launch_with_su().

Arch Linux

#1 2010-01-05 06:07:38

packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#2 2010-01-05 18:40:23

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#3 2010-01-05 18:43:04

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#4 2010-01-06 13:50:08

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#5 2010-01-06 14:32:34

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#6 2010-01-06 14:47:45

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#7 2010-01-06 16:43:03

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#8 2010-01-06 20:05:57

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#9 2010-01-06 20:22:22

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#10 2010-01-06 22:52:16

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#11 2010-01-06 23:17:41

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#12 2010-01-07 01:39:01

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#13 2010-01-07 02:01:58

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#14 2010-01-07 02:55:02

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#15 2010-01-07 02:55:15

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#16 2010-01-07 03:07:52

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#17 2010-01-07 08:37:12

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#18 2010-01-07 09:15:31

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#19 2010-01-07 10:39:39

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#20 2010-01-07 11:00:02

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#21 2010-01-07 12:53:09

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#22 2010-01-07 13:02:51

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#23 2010-01-07 13:11:01

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#24 2010-01-07 13:30:33

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

#25 2010-01-07 13:47:14

Re: packer - New bash aur+pacman wrapper (replaces poorly written yaourt)

Board footer