How the coverage percentage is getting increased in merge

[Sivaraj]

Team, I have basic and sanity test cases both are running aganist same code line. If I run the basic and sanity in single shot I am getting 27% coverage. But, I am running the basic alone and sanity alone and finally merged both clover.db. Now I am getting 29% coverage. So, Which one is correct.

Eg:        Basic + Sanity = 27 %

             Basic db + merge + Sanity db = 29%.

Please help me .. thanks in advance.

[Marek Parfianowicz · Accepted solution]

Hi Sivaraj,

Did you check all steps listed by me in: 

• https://answers.atlassian.com/questions/32525559 and 
• https://answers.atlassian.com/questions/32520636 ?

As I wrote before, the most probably the way you set up Clover in your build is incorrect. And this is a primary reason why you see "weird" coverage numbers. 

In case you can't find the error, feel free to raise a ticket at support.atlassian.com - please attach entire build log (with debug logging) + build file(s) + clover database(s) generated.

Cheers
Marek 

[Sivaraj]

Thanks for your response. Everytime we are cleaning the clover db and cloversrc. But still we are getting same . So the conclusion is "Even though we are merging the clover db we will get identical coverage percentage." Can you please confirm on same.

[Marek Parfianowicz]

In case you do run your tests as follows: 1) Create single clover.db and run basic and sanity tests in one run, generate HTML report. and 2) Create clover.db for basic tests, run them; next create clover.db for sanity tests, run them; next merge two clover databases into one; next generate HTML report from a merged database. Then it's quite probable that you have the same source files registered in BOTH Clover databases. For this reason, during a merge, only ONE version of the source file will be picked. Clover does not perform any merge of the same source file from different databases; a reason is simple - in each database you could have a different version of the file, i.e. with different content. As a consequence, it may happen that code coverage for such file in a report generated from a merged database will be lower than in case when single database was created. So your conclusion "Even though we are merging the clover db we will get identical coverage percentage." is not correct.