Re: [boost] [Release 1.34] Supported Compilers - another view

From: David Abrahams (dave_at_[hidden])
Date: 2006-02-01 09:09:04

"Gennadiy Rozental" <gennadiy.rozental_at_[hidden]> writes:

>>> Actually, maybe my concerns can be addressed by going back
>>> into boost test and using the "minimal" option.
>>
>> That might help, but the author is opposed to adding the facilities
>> needed to make that viable for me.  I can't turn off the crash
>> handlers on Windows, for example, so debugging a problem in a "minimal
>> test" application is often prohibitive.
>
> Ok. Let's say I do this. How would your test behave during regression tests 
> run? Hung the system? Crash? Show dialog message? Remember: no CLA.

What do you mean, "no CLA?"  It _is_ possible to specify command-line
arguments in a Jamfile.  

Regardless, an environment variable would be a reasonable approach.

>>> Hmmm - I don't we want to invoke the debugger in the
>>> regression tests.
>>
>> What happens on the testers machine if it runs this program?
>>
>>     #include <cassert>
>>     int main() { assert(0); }
>>
>> If it throws up a dialog and if we have no test monitor to kill it, I
>> agree that we ought to have something in place to make sure no dialog
>> comes up.  But I was sure the test script *did* start a monitor that
>> could kill off any hung applications (?)
>
> What if somebody else running regression tests that do not have such 
> monitor?

If we don't have a portable monitor that everyone can use, I wouild
think we'd want some crash protection from a library.  On the other
hand, I have lots of tests that just use good-old-<cassert>, and those
have never, to my knowledge, caused a problem for testers.  So that
tells me the only need for a monitor is for killing hung processes.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com