Re: [boost] Regression test now linking to library build failures

From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-07-26 12:55:12

At 10:29 AM 7/26/2002, Peter Dimov wrote:

 >From: "Beman Dawes" <bdawes_at_[hidden]>
 >> There is a fresh version of the Win32 regression tests, still at:
 >>
 >> http://www.esva.net/~beman/jam_regr.html
 >
 >No 'bind' or 'smart_ptr' tests, why?

Because those tests passed for all compilers, and I'm running the daily 
(actually several times a day) tests in a mode that only reports a test if 
it has more that just simple "pass" results.

I've been trying think how we can meet both the need for brevity and the 
need for completeness.

Once possibility is to run two sets of reports.  One that is fairly brief, 
maybe not even showing warnings, but with the advantage it can be viewed 
over the web directly.  Another that is complete, including tests that pass 
all compilers, and the run results even if when the test succeeds.  This 
last would only be made available daily in a .zip format, so would be more 
cumbersome to use.

Would that make sense?  Better ideas?

--Beman