Directories Included in Coverage Reports

[Joel Sherrill]

Now I can reply to this and there will be context. :)

On Thu, Mar 18, 2021 at 5:14 PM Joel Sherrill <joel at rtems.org> wrote:

> Hi
>
> Alex has been working hard on getting covoar into shapt and generating
> reports as he goes. Eventually, this work will all be merged and I will
> work to generate these reports on an automated basis. But we need to have
> some discussions on what is and is not covered. Copying over part of an
> email from Gedare from users@
>
> > Asking stupidly, are there tests now? We didn't catch that libtrace was
> missing in the new coverage reports:
> >
> >
> https://ftp.rtems.org/pub/rtems/people/joel/coverage/coverage-2021-02-28/
> >
> > I can add it to the queue that libtrace is added. If more subsystems
> > are missing that we want to track coverage on, please point them out.
> >
>
> In cpukit the following are not covered. I have no strong opinion, but
> unless there is justification NOT to report on it, we should? Several
> of these don't have any tests, so then the question is whether they
> should have tests/examples written that use them.
>
> libdebugger
> libdl
> libdrvmgr
> libfs
> libgnat
> libi2c
> libmisc/capture
> libmisc/fb
> libmisc/monitor
> libmisc/mouse
> libmisc/redirector
> libmisc/rtems-fdt
> libmisc/serdbg
> libmisc/shell
> libmisc/utf8proc
> libmisc/uuid
> libmisc/xz
> libpci
> librtemscxx
> libtest
> libtrace
> mghttpd
> telnetd
> zlib
>
> It may be good just to know that our testsuite covers 0% of some of
> these directories, so that users can know... use at your own peril :o
>

Some of those are third party so we probably need a policy on that.

Some of those have tests (libdl, uuid and utf8proc should). Alex can
add these if we identify them.

Some of those are hardware frameworks that probably could be tested
with the right stub BUT that might conflict with the BSP so these are
probably off limits.

So we need a master list of all, rationale, and note why each is in or out.

There is a cost per set analyzed but it isn't a big deal if this is done
in batch mode at night. And hopefully we can parallelize this soon.

--joel



> --joel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20210318/0c875901/attachment.html>