[boost] Re: [build/regression testing] runtime failures

From: Hubert Holin (Hubert.Holin_at_[hidden])
Date: 2005-03-24 04:30:21

Somewhere in the E.U., le 24/03/2005
  
   Bonjour

In article <d1smb2$tip$1_at_[hidden]>,
 "Gennadiy Rozental" <gennadiy.rozental_at_[hidden]> wrote:

> Is there any way to see a run command in case test fails at runtime? The 
> same as we could see compiler command line if it fails at compile time
> 
> Gennadiy

      I do not know the answer to your question, but I would like to add 
a request also dealing with runtime failures.

      Currently either a test passes or it fails. I'd like finer 
granularity.

      For instance, at least for my libraries, it may happen that 
"enough" functionality passes that I would like them to show as "pass", 
while at the same time not *everything* passes, so I would like to warn 
prospective users. For instance, without template templates, some 
specializations of some special functions will not be defined for some 
types of variables, but they are only infrequently likely to be used, or 
due to deficiencies in the implementation of standard library special 
functions, accuracy of my library's special functions may suffer for 
some types (such as "long double"). Currently I use messages output by 
the test suite to inform about this state of things, but this is not 
reflected in the compiler status chart.

   Merci

         Hubert Holin