Potential GSoC Project(s) - Improve Test Coverage

[Joel Sherrill]

  Hi

I know GSoC students are looking for projects so I thought I would post
about a couple that have come up recently and don't have tickets and better
write ups.

This likely can be done by multiple students and all code should be in C.
One must be able to run the RTEMS Test Suite and produce code coverage
reports by subsystem/directory. Example reports are at
https://ftp.rtems.org/pub/rtems/people/joel/coverage/coverage-2021-02-28/.
It is likely that the person doing this could focus on analysis of results
on a single architecture and a secondary for sanity checks and rely on core
developers to generate the reports on all architectures. Results should
improve across all similarly.

This project consists of analysing uncovered code ranges in the report,
determining why they are unexecuted by the test suite, and addressing that.
Most of the time this is done by adding new test cases but there are other
cases where the code can be refactored/reworked to improve testability or
the code is simply unneeded based on the RTEMS configuration (e.g.
profiling or SMP code when not built for that). I think there is a Wiki
page which could be brought over into the Software Engineering Guide with
more detailed guidance.

When identifying uncovered areas to work on, it is generally recommended to
focus on a particular subsystem picking methods which are unexecuted at all
first. Then work on larges uncovered range or the method/area with the most
uncovered code.

There are multiple code subsystems which have uncovered ranges. Work on
this is easy to partition so there are no conflicts or dependencies between
projects if that's how we organize it.

--joel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20210316/f174fd90/attachment.html>