Re: [boost] Implicit failures in MetaComm regression tests?

From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2004-06-30 01:18:53

Doug Gregor writes:
> The Boost.Function test case "lambda_test" for intel-win32-7.1-vc6
> (and its STLport variant) come out as unexpected passes. I'd like to
> move these over to explicit passes, but they don't seem to be marked
> as expected failures in status/explicit-failures-markup.xml.

Yep, they are not. They are marked as unexpected passes because they
were failing in the "reference release" -- currently 1.30.2 [1].

> How can I
> make these expected passes, to get that nice cushy all-green feeling?

You can simply remove the corresponding failures from
"boost\libs\expected_results.xml" (just checked in the CVS).

[1] In the early days, in order to produce a meaningful report in
    presence of a huge amount of unmarked failures, we decided to
    simply take the "last known good" release (1.30.0), run the
    regressions on it, and save the results for use as a reference
    point for future releases. This is the key idea that allows us to
    produce the next release that is at least as good as the previous
    one. 

    For instance, 1.30.2 was released with no regressions from 1.30
    and quite a few bug fixes -- http://tinyurl.com/2jyvq.
    Consequently, it became the next "last known good" release, and,
    at this point, the last one. 1.31.0 was released with a
    significant number of regressions; see http://tinyurl.com/2jyvq.

-- 
Aleksey Gurtovoy
MetaCommunications Engineering