Nicholas Clark wrote:

On Thu, Jun 12, 2003 at 09:35:18AM +0200, Leopold Toetsch wrote:

You have to used global labels for subroutines, then all is fine.

But being able to craft input that causes IMCC to segfault (actually any form of segfault) is bad on principle, isn't it?


For sure. But this is slightly a different problem. The error happens at runtime and can easily reproduced by this totally syntactically valid program:

        bsr -1
        end

The error is caught at runtime only with the -b or -t switch. The same holds of course for branching e.g. inside of an instruction or past the end.


Nicholas Clark

leo




Reply via email to