[eresi-dev] Is etrace able to trace (GCC)c++ simbols ?

[Groleo Marius]

Hi,
it was too late in the night to perform the tests yesterday so I
don'em quickly this morning.

The tests were simple. I modified  eresi/testsuite/tracing/Makefile
to have -O0 and generate
the assembly code too. I ran tracer and etrace on the obtained executables.
I did a vimdiff between host_ok.s and host_fail.s and AFAIS the only
differences we're in stack pointer substraction.

The tracer ok/fail links contains the source code that was compiled,
the make command and tracer output.
etrace ok/fail links contains the log of { traces add testcrypt;
tracerun } commands.

tracer ok :  http://pastebin.ca/1576203
etrace ok:  http://pastebin.ca/1576207
host_ok.s: http://pastebin.ca/1576208

tracer fail:  http://pastebin.ca/1576204
etrace fail: http://pastebin.ca/1576206
host_fail.s: http://pastebin.ca/1576209


HTH


On Tue, Sep 22, 2009 at 11:44 PM, Julien Vanegue
<julien.vanegue at gmail.com> wrote:
>
> Yes, this is why in some cases I guess that optimizations only happens on buffer size >= 8, or something. The test is worthwhile and easy to do.
>
> Julien
>
>
>
> _______________________________________________
> eresi-dev mailing list
> eresi-dev at lists.einherjar.de
> http://lists.einherjar.de/mailman/listinfo/eresi-dev
>



--
Regards, Groleo!