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

na12 · 2010-01-22 18:25:48

Yaourt skype upgrade without problem.

Last edited by na12 (2010-01-22 18:26:49)

kjslag · 2010-01-23 01:23:45

is installing/building/updating packages from core/extra/community from the source code a planned feature?

drcouzelis · 2010-01-23 02:42:41

kjslag wrote:

is installing/building/updating packages from core/extra/community from the source code a planned feature?

I don't think so, since you can do those things by using pacman and abs.

Is there a special feature that can't be done using those two programs?

doorknob60 · 2010-01-23 06:02:03

This is slightly unrelated, but in zshrc is there a way to make it so when I type "y -R" then it runs "sudo pacman-color -R" Aliases seem to only work without spaces, but I'm too used to y -R (y is my old alias for yaourt, I've now replaces y with packer, but now it bugs me sometimes ).

bruenig · 2010-01-23 20:08:49

na12 wrote:

Yaourt skype upgrade without problem.

The dependency listed on the skype PKGBUILD is

' lib32-pcre'

This dependency does not exist.

The dependency that does exist is:

'lib32-pcre'

Not sure if it makes sense to get in the business of trying to guess and fix PKGBUILD typos. That is what editing the PKGBUILD is for.

Last edited by bruenig (2010-01-23 20:15:17)

bruenig · 2010-01-23 20:14:05

MajorTom wrote:

Well, you've got a point there. Somehow I envisioned using packer exclusively for all my searches and installs and using pacman for querying, removing and everything else.
Let's stay away from Bloatsville then.

I went ahead and add -Syyu -Suyy -Syuy on the list of legitimate update commands. Given that all I do is pass those arguments to pacman unparsed, it does not slow the program down at all, so it is not a big deal.

Last edited by bruenig (2010-01-23 20:56:49)

bruenig · 2010-01-23 20:31:20

Added support for people who utilize the PKGDEST feature of makepkg. For those unfamiliar, PKGDEST is a feature of makepkg that moves packages to a certain directory once they are built (so you have a sort of local cache of all built packages). This created problems because packer was assuming that the built package was going to be in the current directory. Now it sees if PKGDEST is set and if so uses that path to find where the built .tar.gz is. If not, assumes it is in the current directory.

Because makepkg.conf was already being sourced, it only introduces 1 more conditional test into the code for aur installation which is a negligible speed hit especially for that part of the code where speed is not very important anyways.

MajorTom · 2010-01-23 22:18:14

bruenig wrote:

add -Syyu -Suyy -Syuy on the list of legitimate update commands

packer FTW!
Thank you.

Any chance you could make it recognize and pass pacman's "--ignore" as well?

Last edited by MajorTom (2010-01-23 23:29:15)

bruenig · 2010-01-24 01:10:28

MajorTom wrote:

bruenig wrote:
add -Syyu -Suyy -Syuy on the list of legitimate update commands
packer FTW!
Thank you.
Any chance you could make it recognize and pass pacman's "--ignore" as well?

I have attempted to add it. It definitely works for aur packages.

For pacman, it passes the --ignore argument during update and it passes the --ignore argument during install. So, like pacman --ignore firefox -Syu and pacman --ignore firefox -S foo. When it is fetching pacman dependencies for aur packages, it does not currently pass the argument, but it does always prompt you for pacman dependencies for aur packages, so you already have an opportunity to say no. I don't intend on adding that extra layer of prompting because it would be obnoxious and kind of difficult even.

Last edited by bruenig (2010-01-24 01:25:16)

MajorTom · 2010-01-24 03:04:20

This is awesome. Just tested '-Syu --ignore' and it worked perfectly.

I'm glad you've got the time to make all this happen. This is certainly appreciated. You've had my AUR vote right from the start and I hope many others will do the same soon, too.

At this point all I need is the "quiet" search results option I mentioned before (-Ssq). BTW: I've just realized that yaourt couldn't do that!

kalpik · 2010-01-24 04:40:28

bruenig wrote:

Not sure if it makes sense to get in the business of trying to guess and fix PKGBUILD typos. That is what editing the PKGBUILD is for.

I agree! But it does not even let me edit the PKGBUILD. It just fails saying dependency does not exist.

bruenig · 2010-01-24 04:59:24

kalpik wrote:

bruenig wrote:
Not sure if it makes sense to get in the business of trying to guess and fix PKGBUILD typos. That is what editing the PKGBUILD is for.
I agree! But it does not even let me edit the PKGBUILD. It just fails saying dependency does not exist.

Ah. Hmmm. Well. Should probably fix the PKGBUILD. I think I will implement -G to allow people to do this kind of shit when PKGBUILDs are so dysfunctional.

kalpik · 2010-01-24 05:01:16

Heh..

bruenig · 2010-01-24 05:14:36

Added -G option because PKGBUILD creators are often dumb.

kalpik · 2010-01-24 05:17:11

Thanks! One suggestion, why not allow editing PKGBUILD even when dependency is not found? Right now, it exits.. Perhaps make it say that dependency does not exist, but even then allow editing PKGBUILD?

MajorTom · 2010-01-24 05:32:50

bruenig wrote:

Added -G option because SOME PKGBUILD creators are often dumb.

^fixed

Why the choice of -G and not, say, -d/-w (pacman's download only)? Just curious... With this, packer is getting to behave more like aurget.

Last edited by MajorTom (2010-01-24 05:36:08)

res · 2010-01-24 05:45:27

brueg, are you going to bother with mktmp or am I forced to fork for reals?

I can think of a couple of very real situations were not using mktmp (or /tmp/packaer-$RANOM, whatevs) would break it.

Besides, it should check for $TMPDIR first.

patogen · 2010-01-24 07:44:18

bruenig wrote:

Added -G option because PKGBUILD creators are often dumb.

This was the only thing I was missing. Thanks alot, great effort!

bruenig · 2010-01-24 08:22:46

MajorTom wrote:

bruenig wrote:
Added -G option because SOME PKGBUILD creators are often dumb.
^fixed
Why the choice of -G and not, say, -d/-w (pacman's download only)? Just curious... With this, packer is getting to behave more like aurget.

-Sw would make people think that it was also going to support pacman download only stuff which it wont. Also, someone else said -G was the flag, so I didn't put too much thought into i

Also, -Sw resolves dependencies and this whole thing arose because resolving dependencies was the problem.

Last edited by bruenig (2010-01-24 08:23:06)

bruenig · 2010-01-24 08:25:05

res wrote:

brueg, are you going to bother with mktmp or am I forced to fork for reals?
I can think of a couple of very real situations were not using mktmp (or /tmp/packaer-$RANOM, whatevs) would break it.
Besides, it should check for $TMPDIR first.

mktmp is fail because I cannot access the same directory each time. Also it will bloat up the /tmp dir. I have made it respect TMPDIR on the git.

Last edited by bruenig (2010-01-24 08:28:09)

j2902 · 2010-01-24 20:21:20

I've got a question / feature request.

If packer doesn't recognize the argument pattern, eg, -Rns, could it just call pacman with those arguments?

In an attempt of clarity,

packer -fascfSDfasd package0 package1

would call

pacman -fascfSDfasd package0 package1

(obviously, -fascfSDfasd is not valid input)

Thanks.

Note: I got this to work by changing line 412 to

 -*) runasroot pacman-color $@ ; exit ;;

This probably is a terrible way to implement it, however.

Last edited by j2902 (2010-01-24 22:23:54)

bruenig · 2010-01-25 00:28:45

j2902 wrote:

I've got a question / feature request.
If packer doesn't recognize the argument pattern, eg, -Rns, could it just call pacman with those arguments?
In an attempt of clarity,
packer -fascfSDfasd package0 package1
would call
pacman -fascfSDfasd package0 package1
(obviously, -fascfSDfasd is not valid input)
Thanks.
Note: I got this to work by changing line 412 to
 -*) runasroot pacman-color $@ ; exit ;;
This probably is a terrible way to implement it, however.

I don't understand why you don't just use pacman.

j2902 · 2010-01-25 00:42:48

It's easier / faster to just have one program in bash to call; however, I'm might be missing something. What are the advantages to separating pacman and packer?

Thanks.

edit: Let me clarify.

Specifically, this change allows for closer integration between packer and pacman. The features of packer are useful; however, having to use two separate utilities, packer and pacman, for package management is annoying. This change rectifies this annoyance, and I cannot think of any disadvantages to the change. I cannot imagine how the speed of packer would decrease. However, there could very well be a disadvantage I'm not aware of. If you know of this disadvantage, please tell me. Otherwise, I do not see why this change should not be incorporated into packer.

Last edited by j2902 (2010-01-25 04:09:42)

MajorTom · 2010-01-25 02:30:54

I personally use packer for searches, installs and updates. Otherwise, I'd need to run pacman -Ss and pacman -S/-Su/-Syu/-Syyu (for the official repos) plus aurget -s and aurget -i/-u (for AUR packages).

For everything else involving package management (queries, removals, etc.), I use pacman.

handy · 2010-01-25 04:05:02

I'm glad I stumbled on this thread.

Thanks bruenig for making packer.

I just installed packer & ran packer -Syu, it was certainly quicker than the -Syu --aur I am used to.

When it got to the AUR upgrade the following happened even though I have IgnorePkg = nerolinux in pacman.conf:

:: Starting full aur upgrade...
resolving dependencies...

Aur Targets (1): nerolinux
Pacman Targets (1): rpmextract

Proceed with installation? [Y/n] n

I thought that packer abides by pacman.conf, at least in the IgnorePkg instance. Am I missing something here?

Anyway, even if it doesn't, I'm already in love with packer, it has already replaced yaourt on my system.

Last edited by handy (2010-01-25 04:15:04)

Arch Linux

#151 2010-01-22 18:25:48

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

#152 2010-01-23 01:23:45

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

#153 2010-01-23 02:42:41

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

#154 2010-01-23 06:02:03

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

#155 2010-01-23 20:08:49

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

#156 2010-01-23 20:14:05

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

#157 2010-01-23 20:31:20

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

#158 2010-01-23 22:18:14

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

#159 2010-01-24 01:10:28

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

#160 2010-01-24 03:04:20

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

#161 2010-01-24 04:40:28

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

#162 2010-01-24 04:59:24

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

#163 2010-01-24 05:01:16

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

#164 2010-01-24 05:14:36

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

#165 2010-01-24 05:17:11

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

#166 2010-01-24 05:32:50

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

#167 2010-01-24 05:45:27

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

#168 2010-01-24 07:44:18

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

#169 2010-01-24 08:22:46

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

#170 2010-01-24 08:25:05

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

#171 2010-01-24 20:21:20

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

#172 2010-01-25 00:28:45

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

#173 2010-01-25 00:42:48

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

#174 2010-01-25 02:30:54

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

#175 2010-01-25 04:05:02

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

Board footer