Re: bashdb sig-test.sh failure

[Chet Ramey]

On 10/26/15 1:25 PM, Rocky Bernstein wrote:

> Bash Version: 4.3
> Patch Level: 11
> Release Status: release
> 
> Description:
>         
> Weird trap debug and signal handling problem in bashdb in running test
> test/integration/test-sig..sh. I've tried to narrow to scope of the problem
> by reducing the debugger commands that are invoked as well as the script
> that gets debugged. Attempts to write a small standalone program though
> elude me.  
> 
> Gdb show that bash is attempting to free memory from discard_pipeline()
> inside a restore_pipeline() which is in running run_debug_trap(). See the
> attached gdb backtrace.
> See also https://sourceforge.net/p/bashdb/bugs/36/
> https://sourceforge.net/p/bashdb/bugs/37/ for  where this was first reported. 

Try bash-4.4.  I think this was fixed back in the middle of August.  The
bash-4.3 implementation of save_ and restore_pipeline only provided one
save level, and multiple calls would try to access freed memory.  Depending
on your malloc implementation, this had different effects, all bad.

http://lists.gnu.org/archive/html/bug-bash/2015-08/msg00060.html is the
original bug report for this.

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
                 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRU    chet@case.edu    http://cnswww.cns.cwru.edu/~chet/