Re: [patch] debug mode

[François Dumont <francois dot cppdevs at free dot fr>]

On 08/18/2010 03:54 PM, Paolo Carlini wrote:
> ... one final comment ;)
>
> On 08/18/2010 03:11 PM, Paolo Carlini wrote:
>    
> > Also, I see various testcases defining explicitly _GLIBCXX_DEBUG:
> > normally we don't do that: check-debug takes care of testing the
> > debug-mode behavior, uniformly. Please see if you can remove those
> > explicit defines in the testcases.
> >
> >      
> About this: maybe you can remove the explicit -D_GLIBCXX_DEBUG and then
> check in the testcase whether_GLIBCXX_DEBUG is defined or not: if it's
> defined just run the normal test, otherwise call by hand
> __builtin_abort(). In this way you are sure the test will consistently
> xfail at runtime in all cases.
>    
Well ok even if I don't know if you remember but in my first patch 
submition I asked for a way to check that a test fail with an expected 
output. The output with a simple __builtin_abort won't be the same as 
the one obtained when running the test in debug mode. As for the moment 
I cannot make the test as accurate as I would like I will do as you propose.
> Paolo.
>
> PS: don't diff the Changelog additions, when submitting patches. Instead
> attach separately the new Changelog lines. See my patches, for examples.
>    
Sorry, I knew it, just a mistake.