Re: [boost] [optional] New RC_1_34_0 regression

From: Vladimir Prus (ghost_at_[hidden])
Date: 2007-04-04 11:11:57

Michael Goldshteyn wrote:

> 
> "Vladimir Prus" <ghost_at_[hidden]> wrote in message
> news:E1HZ27f-0007ib-DI_at_zigzag.lvk.cs.msu.su...
>> I disagree. There's no way we can release anything this
>> century if we keep on checking in things are branch is
>> frozen. The time window for commits is over at this point.
>> If this patch breaks one of release compilers, we'll spend
>> another week just to get back to zero regressions.
>>
>> The final decision is up to Thomas, of course, but I hope
>> that no code changes will be allowed, no matter what.
>>
>> - Volodya
> 
> I just want to add my 1.5 cents to this. I want to see this build get
> released as much as anyone. But, let us not confuse the regression test's
> output of no regressions with the fact that there is in fact a regression.

Indeed! Because there's no "output of no regressions". The value
of zero reported currently is the value of *unexpected* failures. 
If a failure is marked as expected, it's not reported, even if that
failures was not present in Boost 1.33.1. I'm sure that there are regressions
that are marked expected at the moment. 

Previously, we decided on this list that on a certain date, all remaining
failures will be marked as expected. Now, you're pointing out that
we have a regression, for which we don't even have a test.
In light of that decision, we probably can add a test for that problem,
and immediately mark all failures of said test as expected. But that
would not be very helpful. I don't think we should give this problem 
any bonus points just because it's not discovered by the tests yet. 

- Volodya