Boost logo

Boost :

Subject: Re: [boost] Report # 29 (aka 1.45 blockers) is too narrow
From: Jim Bell (Jim_at_[hidden])
Date: 2010-10-25 09:41:50


On 1:59 PM, Vladimir Prus wrote:
> vicente.botet wrote:
>
>> ----- Original Message -----
>> From: "Vladimir Prus" <vladimir_at_[hidden]>
>> To: <boost_at_[hidden]>
>> Sent: Sunday, October 24, 2010 7:40 PM
>> Subject: Re: [boost] Report # 29 (aka 1.45 blockers) is too narrow
>>
>>
>>> Jim Bell wrote:
>>>
>>>> I think that boost report # 29
>>>> (https://svn.boost.org/trac/boost/report/29) is too narrow a query. If
>>>> you remove the "Milestone" filter you get nearly fifty tickets marked
>>>> Showstopper, going back 22 months.
>>>>
>>>> A ticket's originator may not know what milestone to fill in, and
>>>> his/her ticket shouldn't be lost because of that.
>>> Or the ticket's originator may have used "showstopper" just because
>>> he overestimated the importance of the bug. Therefore, "showstopper"
>>> severity does not necessary mean anything.

I appreciate not wanting to be at the mercy of any yahoo with a barrage
of new tickets. (Like me!)

But if we're all in this together (and we are), then existing tickets
need to be reviewed more carefully than they are at present.

>> It is up to the maintainer or someone else to change the severity if not correct.

But it's up to the one making a boost-1.45 release to make sure all
showstopper and regression tickets are resolved, holding up the release
as necessary.

>>>> I think they all need review.
>>> And -- who exactly do you suggest review them -- given that per above,
>>> you actually have to review every single open bug, not just bugs with
>>> specific severity.
>> We need to review all of them of course and by respect to the reporter, we should start by more
>> critical. For the more critical we can start by the newests.
> Ok, would you please review them, in the suggested order?
> I'm merely pointing out that Stowstopper+1.45 is a list of issues that were explicitly
> designated as potential showstoppers for 1.45, and is meant to make it less likely
> to have 1.45 released with such showstoppers unresolved.
>
> While it's good to review other issues from time to time, that goal is entirely separate
> from the goal of not forgetting to resolve the issues that we think should be really
> resolved for 1.45.

So all agree that the 1.45 blockers report needs to be changed, at least
to remove the milesone.

What about adding 'regression' to it as well? That field of dandelions
on the test matrix reduces confidence in all of boost.

And previously filtered showstoppers need to be at least re-categorized,
if not addressed.


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk