Re: [gentoo-project] Call for agenda items - Council meeting 2014-04-08

[Patrick Lauer <patrick@gentoo.org>]

On Monday 07 April 2014 14:54:43 Samuli Suominen wrote:
> On 07/04/14 11:26, Patrick Lauer wrote:
> > On Monday 07 April 2014 11:02:44 Samuli Suominen wrote:
> >>>> at least as safe kludge to avoid disrespect like this:
> >>> I wonder if you noticed that I've been a dev since 2004 ...
> >>> 
> >>> and disrespect, well, you actively work towards breaking things.
> >>> 
> >>> Even when told, multiple times, that you're breaking things,
> >>> 
> >>> And that we will revert breakage when detected ...
> >> 
> >> What you told was you will impose your personal opinions as policies,
> >> and you will use the
> >> QA team membership as a cover to enforce it upon others.
> > 
> > I don't even know what words I could use to change your mind.
> 
> You can get me to change mind by writing up a policy that says
> dynamic deps can't be relied upon, and getting rest of the QA
> team, perhaps council, on board with it.
> 
> #gentoo-qa, yesterday:
> 
> 17:50 < floppym> Hmm... documenting when to revbump in relationship to
> RDEPEND is not a bad idea. However, I think the rules would
> be a little more subtle than "always". We would also need some way of
> dealing with eclasses which generate
> RDEPEND, and also with revbumping stable ebuilds.
> 

Soo, here's the usecase:

/usr/portage empty except profiles/ (needed for various reasons)

Everything else supplied by binpkgs


... note how there's no way at all ever for your dynamic deps idea to affect 
the visible packages and updates.

So the only way that has a chance of working, wuarrghl, is to run emerge -e 
world on every friggin update always. So about a CPU-day per build env, or 
about two CPU-weeks for more comprehensive binpkg collections. 


I'd suggest not making life more difficult :)