On Thu, Sep 1, 2011 at 6:05 PM, Sergej Pupykin wrote:

we should not recompile anything. But it is usefull package with 24 votes which has many dependencies in AUR. It is just a way to make things easier for guys who want to use tcp_wrappers.

I'm sure that whoever wants to use tcp_wrappers, and therefore is willing to recompile all the services that support it on every update, won't mind building one more package. I'd say we keep it out of the binary repos. :)

On 09/01/2011 07:52 PM, Lukas Fleischer wrote:

On Thu, Sep 01, 2011 at 07:05:01PM +0400, Sergej Pupykin wrote:

...

Dave,

we should not recompile anything. But it is usefull package with 24 votes which has many dependencies in AUR. It is just a way to make things easier for guys who want to use tcp_wrappers.

I assume all developers build packages in chroots and it should not break anything even if some developer occasionaly install tcp_wrappers.

There are packages that we don't want in the binary repos, even if they have a lot of votes. awesome [1] is another good example. It has almost 400 votes but uses cairo-xcb which is broken.

dude wtf, don't give him ideas!

We decided to remove tcp_wrappers. -1 to moving it back to [community].

-- Ionuț

Le Vendredi 2 Septembre 2011 12:56:28, Sergej Pupykin a écrit :

At Thu, 1 Sep 2011 18:52:42 +0200,

Lukas Fleischer wrote:

...

There are packages that we don't want in the binary repos, even if they have a lot of votes. awesome [1] is another good example. It has almost 400 votes but uses cairo-xcb which is broken.

Btw, cairo-xcb is _broken_. I did not search for tcp_wrappers issues globally, but I did not understand from http://www.archlinux.org/news/dropping-tcp_wrappers-support/ what is exact issue with tcp_wrappers.

See https://bugs.archlinux.org/task/23929

On Fri, Sep 02, 2011 at 06:04:04PM +0400, Sergej Pupykin wrote:

At Fri, 02 Sep 2011 11:22:02 +0200, Laurent Carlier wrote:

...

See https://bugs.archlinux.org/task/23929

It is not issue. This is rather controversial philosophy/architecture questions and care about home users (If they forget that libwrap does not cover all network applications).

As I understand tcp_wrappers still works fine despite of 14+ years old release.

No, that's exactly the issue. We decided, after bringing up the idea for the third time [1] that tcp_wrappers needs to go. There was plenty of support and _zero_ opposition to this decision, so we went ahead with it. d [1] http://mailman.archlinux.org/pipermail/arch-dev-public/2011-July/020931.html

On 1 September 2011 17:05, Sergej Pupykin wrote:

Dave,

we should not recompile anything. But it is usefull package with 24 votes which has many dependencies in AUR. It is just a way to make things easier for guys who want to use tcp_wrappers.

I assume all developers build packages in chroots and it should not break anything even if some developer occasionaly install tcp_wrappers. And tomorrow we'll get some Feature Request with the title "Build X with tcp_wrappers support" - Just add it to depends array.

-1. Bring it to AUR! -- Andrea