This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
gcno file question
- From: Joel Sherrill <joel dot sherrill at oarcorp dot com>
- To: "gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>
- Date: Thu, 24 Mar 2011 07:53:42 -0500
- Subject: gcno file question
Hi,
RTEMS has been using simulators and some
programs we wrote for coverage analysis
for a while now.
I am looking into writing a converter which
takes coverage data from simulators and produces
.gcno files. The coverage data is often just
a bitmap of which addresses were executed. There
is no frequency, just yes/no. We can already
map that information back to file/line.
+ Is this enough to produce a .gcno file from?
+ What records need to be generated as a minimum?
As a technical sanity question, the RTEMS code
is in a library and we are merging coverage
data from multiple executables to get unified
coverage data. We abstract away physical
address into offsets into methods and file/line.
Does generating a .gcno from this merged data
sound feasible?
Thoughts, insights, comments appreciated.
Thanks.
--
Joel Sherrill, Ph.D. Director of Research& Development
joel.sherrill@OARcorp.com On-Line Applications Research
Ask me about RTEMS: a free RTOS Huntsville AL 35805
Support Available (256) 722-9985