RE: Assistance Request

[Dave Korn]

On 12 February 2008 16:35, Joel Sherrill wrote:

> I think so.  I actually got an executable which
> I could run but DejaGNU didn't know how to.
> So that's half the battle. ;)
> 
> ===============================
> PASS: gcc.c-torture/execute/builtins/memset-chk.c compilation,  -O3
> -fomit-frame-pointer -funroll-loops
> core_find_mapping() - access to unmaped address, attach a default map to
> handle this - addr=0x7c9c60 nr_bytes=0x4 processor=0x40017008 cia=0x14980
> 
> FAIL: gcc.c-torture/execute/builtins/memset-chk.c execution,  -O3
> -fomit-frame-pointer -funroll-loops
> ===============================
> 
> But when I run it by hand...
> 
> $ psim-4.9 ./memset-chk.x5
> 
> *** EXIT code 0
> 
> How can I see what is invoked for the simulator run?

  Add lots of "-v" options to RUNTESTFLAGS.

> Is there an easy way to skip to the executable tests?
> I am tired of watching the compile only ones pass. :-D

  Add "execute.exp" to RUNTESTFLAGS.  You can also say e.g. something like
"execute.exp=2001*" to limit which of execute.exp's testcases are invoked.

> I don't know that we will need this for RTEMS.  

  I don't know what RTEMS does or doesn't have, so it might very well be
surplus to your requirements; I threw it into the mix just in case :)


    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....