This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Best Practice for using contrib/dg-cmp-results
- From: Joel Sherrill <joel dot sherrill at oarcorp dot com>
- To: <gcc at gcc dot gnu dot org>
- Date: Thu, 18 Dec 2008 13:39:47 -0600
- Subject: Best Practice for using contrib/dg-cmp-results
Hi,
I have made a lot of progress running the GCC Test Suite
on RTEMS targets on the CFARM and reporting results.
Now I would like to begin to run the tests on a regular
basis and use dg-cmp-results to point out variances from
one run to the next. I think this is critical because
there are ~12 RTEMS target and when multiplied by the
SVN trunk and branches, this is potential information
overload.
Do people save their gcc.log files in a directory structure
arranged by branch, target, and date? And then compare gcc.log
against a saved one?
What is the accepted best practice in this situation?
Thanks.
--joel