Re: can we hope a stable version in the near future?

[Dor Laor <dlaor@redhat.com>]

Farkas Levente wrote:
> Avi Kivity wrote:
>   
>> Farkas Levente wrote:
>>     
>>>> There is the maint/ series on git.kernel.org.  It doesn't have formal
>>>> releases though.
>>>>     
>>>>         
>>> do you plan any formal release? and it'd be nice to see the relationship
>>> between the current devel tree and the stable tree to eg. last stable
>>> 0.5 current devel 0.78.
>>>   
>>>       
>> The key to a formal release is a formal test suite.  We've been building
>> one (for a long while) but it isn't in production yet.
>>
>> The plan is for it to be open so people can add their favorite guests,
>> to ensure they will not regress.
>>     
>
> the question is not when but what happened with those bugs which cause
> test fail? the problem currently not that we don't know problems, but
> there are many known bugs just the reason and the solution not known. so
>  test suite can't help too much here (may be find more bugs).
>
>   
Test suite will help since it's job is to run regression tests each 
night or even each commit.
Once a new regression is introduced it will immediately revert it.

Now, when we have only very poor, old regression suite, it does not 
happen so regressions
are detected by users, weeks after being committed.

We'll publish the test suite (based on autotest) next week. The more 
users will use it the better.
Anyway our maintainer will run it each night.
>>> on the other hand the real question are you plan to somehow stabilize
>>> any of the following release in the near future? in the last 1.5 years
>>> we wait for this. or you currently not recommend and not plan to use kvm
>>> in production? it's also an option but would be useful to know. in this
>>> case we (and probably many others) switch to xen, virtualbox, vmware or
>>> anything else as a virtualization platform.
>>>   
>>>       
>> kvm is used in production on several products.  Just not the kvm-nn
>> releases I make.  The production versions of kvm are backed by testing,
>> which makes all the difference.  Slapping a 'stable' label over a
>> release doesn't make it so.
>>     
>
> there are many open source project which has stable and devel
> versions:-) actually almost all projects have a stable release along
> with the devel version. but kvm has not any in the last few years,
> that's why i think it's high time to stabilize 'a' version ie. frozen
> feature list and fix all known bugs.
>   
You're right about the need for stable release, that's the idea of the 
'maint' branches.
maint/2.6.26 for both kernel and userspace is stable (using userspace 
irqchip).
Now we'll stabilize another user/kernel pair based on 2.6.28

Thanks, Dor