On 12/12/2016 10:18 AM, Jakub Cajka wrote:
----- Original Message -----
> From: "Stephen Gallagher" <sgallagh(a)redhat.com> To:
> devel(a)lists.fedoraproject.org Sent: Monday, December 12, 2016 4:09:41 PM
> Subject: Re: F26 System Wide Change: Golang 1.8
>
> On 12/12/2016 10:03 AM, Jakub Cajka wrote:
>>>
>>> The mass-rebuild is currently scheduled to begin on February 1st. Will
>>> there be a release-candidate available for Golang 1.8 by that time? If
>>> it is not, how will the contingency plan work?
>>
>> There should be, if there would be serious issues(which I believe are
>> highly unlikely as even with current beta release I haven't hit any
>> considerate issue while rebuilding Fedora locally), I assume we will
>> stick to the previous version of Go as stated in contingency plan. Should
>> I expand that section in any direction?
>>
>
> Mostly I'm just concerned that any issue discovered during the
> mass-rebuild that would require a downgrade to Go 1.7 would require another
> mass-rebuild to accomplish it. Which really means that once this process
> starts, we're committed to finishing it (and fixing whatever comes up).
> Thus, it seems like there isn't a realistic contingency strategy (and we
> should make that clear in the proposal for FESCo to judge the risk).
I believe it is the same as in other issues found at the mass-rebuild time,
backing the change and rebuilding affected packages. Or am I missing
something?
As the Betas and RC of golang are coming out I do my local "scratch rebuilds"
and I'm confident even at current moment with Beta1 mass re-buld would be
successful(from Go side).
To be clear, I'm not saying this situation is a problem. I'm mostly saying that
the "contingency plan" is likely unrealistic. Meaning if FESCo gives the
go-ahead, it means this is happening, period. I figured it makes sense to adjust
the contingency plan to make this clear.