Re: [Repost] What is "Error: Protected multilib versions"

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, 2012-04-19 at 17:14 +0200, Kevin Kofler wrote:
> Adam Williamson wrote:
> > AFAICS the real problem here is that an update got unpushed. It seems
> > like Richard got the 64-bit version of libvirt -3 installed, then the -3
> > update got unpushed, then something wanted to install the 32-bit version
> > of libvirt. Obviously, since the update had been unpushed, it was
> > impossible to find the matching 32-bit version.
> > 
> > As long as updates can be unpushed, that one can pop up.
> 
> Oh, so this is yet another example of fallout from the braindead decision to 
> enable updates-testing by default for Branched.

Um...no? The case listed *below* would be, but updates-testing doesn't
seem particularly relevant to the above case. If updates-testing didn't
exist at all, you'd still have the potential of this problem happening
if updates could be unpushed.

> > The other classic case where we get a lot of this (and similar errors)
> > is when we push the fedora-release update which disables
> > updates-testing; people have the 64-bit version of something installed
> > from updates-testing, then they need to have the 32-bit version
> > installed, but now updates-testing is disabled...
> 
> And that's the other reason why that decision is broken.
> 
> updates-testing should NEVER be on by default. There is no expectation of 
> upgrade path in updates-testing (whereas there is one even in Rawhide!), so 
> enabling it by default is very broken.

Yes and no. It's 'very broken' in the sense that we know it can cause a
bit of this kind of pain for people who install pre-releases and update
them to final releases. However, we're perfectly aware of the general
sorts of issues the process can cause and consider them to be an
acceptable trade-off for the considerable _benefits_ of having
updates-testing enabled by default (lots more testing of the packages).

We really mean the whole thing about pre-releases eating babies, and
this is just one instance of that. If you're not prepared to do a bit of
yum handholding to work around issues like this now and again, you have
no business installing a pre-release of Fedora. I'm not sure it's
practically possible to make that *not* the case.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net

-- 
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel



[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Fedora Announce]     [Fedora Kernel]     [Fedora Testing]     [Fedora Formulas]     [Fedora PHP Devel]     [Kernel Development]     [Fedora Legacy]     [Fedora Maintainers]     [Fedora Desktop]     [PAM]     [Red Hat Development]     [Gimp]     [Yosemite News]
  Powered by Linux