Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-08-02 Thread Zac Medico
On 08/02/2011 11:29 AM, Ciaran McCreesh wrote: > On Tue, 2 Aug 2011 20:18:17 +0200 > Michał Górny wrote: >>> I think I prefer the second option (copying from Exherbo). A better >>> integration with the package manager than USE flags should result >>> in a better user experience. >> >> Are you will

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-08-02 Thread Paweł Hajdan, Jr.
On 8/2/11 11:18 AM, Michał Górny wrote: >> I think I prefer the second option (copying from Exherbo). A better >> integration with the package manager than USE flags should result in a >> better user experience. > > Are you willing to update and EAPI-bump all the eclasses? May I remind > you that

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-08-02 Thread Alex Alexander
On Tue, Aug 2, 2011 at 21:18, Michał Górny wrote: > On Tue, 02 Aug 2011 10:58:56 -0700 > ""Paweł Hajdan, Jr."" wrote: > >> On 8/2/11 12:35 AM, Michał Górny wrote: >> > On Sun, 31 Jul 2011 10:19:03 -0700 >> > ""Paweł Hajdan, Jr."" wrote: >> >> I'm interested in some sort of suggested/recommend de

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-08-02 Thread Ciaran McCreesh
On Tue, 2 Aug 2011 20:18:17 +0200 Michał Górny wrote: > > I think I prefer the second option (copying from Exherbo). A better > > integration with the package manager than USE flags should result > > in a better user experience. > > Are you willing to update and EAPI-bump all the eclasses? May I

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-08-02 Thread Michał Górny
On Tue, 02 Aug 2011 10:58:56 -0700 ""Paweł Hajdan, Jr."" wrote: > On 8/2/11 12:35 AM, Michał Górny wrote: > > On Sun, 31 Jul 2011 10:19:03 -0700 > > ""Paweł Hajdan, Jr."" wrote: > >> I'm interested in some sort of suggested/recommend deps for > >> www-client/chromium, but I'm not sure if eclass

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-08-02 Thread Paweł Hajdan, Jr.
On 8/2/11 12:35 AM, Michał Górny wrote: > On Sun, 31 Jul 2011 10:19:03 -0700 > ""Paweł Hajdan, Jr."" wrote: >> I'm interested in some sort of suggested/recommend deps for >> www-client/chromium, but I'm not sure if eclass is the right >> implementation. > I don't think we can start drafting until

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-08-02 Thread Ciaran McCreesh
On Tue, 2 Aug 2011 09:35:05 +0200 Michał Górny wrote: > I don't think we can start drafting until we agree on one solution. > AFAICS there are two major ideas: > 1) using special USEflags for that (which I can draft if you like), > 2) copying DEPENDENCIES syntax from exherbo. I guess there's nothi

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-08-02 Thread Michał Górny
On Sun, 31 Jul 2011 10:19:03 -0700 ""Paweł Hajdan, Jr."" wrote: > I'm interested in some sort of suggested/recommend deps for > www-client/chromium, but I'm not sure if eclass is the right > implementation. > > I think I agree with Ciaran that this should be implemented as a PMS > update. Let me

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-07-31 Thread Paweł Hajdan, Jr.
On 7/31/11 8:27 AM, Michał Górny wrote: > The last discussion on new solutions optional runtime depends lead to no > agreement. Thus, I'd like to propose a solution extending the usability > of current methods of handling them. I'm interested in some sort of suggested/recommend deps for www-client

Re: [gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-07-31 Thread Ciaran McCreesh
On Sun, 31 Jul 2011 17:27:21 +0200 Michał Górny wrote: > My idea is to create an eclass which would take a optional dependency > list (e.g. through some kind of SDEPEND variable or so [being a bash > array!]) and print it out to user in pkg_postinst(). Please don't. This should be worked out and

[gentoo-dev] RFC: an eclass to handle optional runtime depends

2011-07-31 Thread Michał Górny
Hello, all. The last discussion on new solutions optional runtime depends lead to no agreement. Thus, I'd like to propose a solution extending the usability of current methods of handling them. My idea is to create an eclass which would take a optional dependency list (e.g. through some kind of S