Re: [boost] Boost regression notification (2005-09-15 [RC_1_33_0])

From: Douglas Gregor (doug.gregor_at_[hidden])
Date: 2005-09-17 17:13:13

On Sep 17, 2005, at 5:32 AM, Thorsten Ottosen wrote:

> Douglas Gregor <dgregor <at> cs.indiana.edu> writes:
>
>
>>   range (3)
>>
>
> Is there any way to deduce if these errors persist after one  
> updates the test
> code`?
>
> For example, the last change I made took days to show up in the  
> regression-test.
> That makes it sooo hard to make even small fixes.

The only way to solve this is with more testing resources. Many of  
our testers can turn around results nightly, but some cannot and  
those will take a few days.

> If only there was som easy way to know, per compiler, that the  
> latest change
> had removed/added X failures.

Buildbot has some smarts to do this kind of thing (e.g., by  
determining what changes in CVS happened from one run to the next),  
but we don't have this capability yet.

     Doug