That would make sense if you were checking for memory problems / bad code as part of unit testing or final assembly testing. Two approaches are possible:
- writing a test tool that will use the valgrind API through its library, largely creating a user interface that replaces the valgrind executable (it seems like a headache)
- Valgrind can also output XML for memcheck (option -xml = yes)
- Use Expect http://en.wikipedia.org/wiki/Expect to automate the valgrind command and check the output
- Write shell scripts that run valgrind and grep output
- Perhaps add this functionality to something like DejaGNU that uses expect and tcl
Perhaps these tools can automatically create error elements in your tracker (although they can fill up quickly).
, , valgrind, . - , , !
http://www.redhat.com/f/summitfiles/presentation/June2/Developer%20Tools/Cox,Malcom_Automated%20Testing.pdf
, , DejaGNU Valgrind,
!