[Gluster-devel] Update: GlusterFS code coverage

Amar Tumballi Suryanarayan atumball at redhat.com
Wed Jun 5 07:00:16 UTC 2019


All,

I just wanted to update everyone about one of the initiatives we have
undertaken, ie, increasing the overall code coverage of GlusterFS above 70%.
You can have a look at current code coverage here:
https://build.gluster.org/job/line-coverage/lastCompletedBuild/Line_20Coverage_20Report/
(This shows the latest all the time)

The daily job, and its details are captured @
https://build.gluster.org/job/line-coverage/

When we started focus on code coverage 3 months back, our code coverage was
around 60% overall. We kept the ambitious goal of increasing the code
coverage by 10% before glusterfs-7.0 release, and I am happy to announce
that we met this goal, before the branching.

Before talking about next goals, I want to thank and call out few
developers who made this happen.

* Xavier Hernandez - Made EC cross 90% from < 70%.
* Glusterd Team (Sanju, Rishub, Mohit, Atin) - Increased CLI/glusterd
coverage
* Geo-Rep Team (Kotresh, Sunny, Shwetha, Aravinda).
* Sheetal (help to increase glfs-api test cases, which indirectly helped
cover more code across).

Also note that, Some components like AFR/replicate was already at 80%+
before we started the efforts.

Now, our next goal is to make sure we have above 80% functions coverage in
all of the top level components shown. Once that is done, we will focus on
75% code coverage across all components. (ie, no 'Red' in top level page).

While it was possible to meet our goal of increasing the overall code
coverage from 60% - 70%, increasing it above 70% is not going to be easy,
mainly because it involves adding more tests for negative test cases, and
adding tests with different options (currently >300 of them across). We
also need to look at details from code coverage tests, and reverse engineer
to see how to write a test to hit the particular line in the code.

I personally invite everyone who is interested to contribute to gluster
project to get involved in this effort. Help us write test cases, suggest
how to improve it. Help by assigning interns write them for us (if your
team has some of them). This is a good way to understand glusterfs code
too. We are happy to organize sessions on how to walk through the code etc
if required.

Happy to hear feedback and see more contribution in this area.

Regards,
Amar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-devel/attachments/20190605/47742cae/attachment-0001.html>


More information about the Gluster-devel mailing list