|
From: Josef W. <Jos...@gm...> - 2005-07-20 08:44:42
|
On Wednesday 20 July 2005 04:41, sv...@va... wrote:
> Author: njn
> Date: 2005-07-20 03:41:31 +0100 (Wed, 20 Jul 2005)
> New Revision: 4205
>
> Log:
> Reinstate stack trace printing on assertion failures. It's terrible
> for the module dependency graph, but it's very useful.
I suppose this can be solved by using callbacks ("interface") without
disturbing the module dependency graph: m_libcassert.c would signal via the
interface that the action of "pp_sched_status" is needed at some point in
time, but would not have the implementation, and thus no dependency.
The user of m_libcassert can register the correct handler when it also uses
m_stacktrace.
Josef
|