clover reporting with aggregated/group packages.

[niv]

Hi,

Is it possible to create report with aggregated package level coverage. For example: I have 100 packages under source dir and wanted to generate report with aggregation/group of selected packages(grouping of releated packages, so that I will get component level report which tells % of coverage at component level)

src: com.xyz.a .....

com.xyz.z

grouping: 1.com.xyz.a to com.xyz.f

2.com.xyz.g to com.xyz.m ...

report:

com.xyz.a to com.xyz.f(af component) 40%

com.xyz.g to com.xyz.m(gm component) 20%

Thanks

[niv]

thanks a lot! It solved the problem :) Now I am seeing same stats.

I have one more question here..

some new java file(interface) was added newly to src from last build. This new interface(.java) file instrumented(but I did not see any clover specific lables added to this file) and it has only one method def

ex: public interface Abc{

void xyzAbc();

}

It is missing in clover coverage report.

Is this something expected?.

Interfaces(like above ex) won't show up in the clover report even after instrumentation?.

[Marek Parfianowicz]

Can you confirm that you new interface has no code in a method body?

I'm asking because, by default, Clover does not show classes or interfaces having no executable code - you can change this behaviour by setting the showEmpty to true.

<clover-report>
  <current>
    <format type="html" showEmpty="true"/>
   </current>
</clover-report>

See: clover-report / current / format

[Marek Parfianowicz]

You could generate separate reports for every component by defining proper inclusion/exclusion filters. See <clover-setup> / < fileset> for instance.

[niv]

Hi,

Thanks for the reply. Instead of clover-setup can I exclude them in <clover-report> <fileset <exclude ...

Is there any difference in coverage if I exclude files/packages in clover-report instead of clover-setup?.

thanks

[niv]

Hi,

Thanks for the reply. Instead of clover-setup can I exclude them in <clover-report> <fileset <exclude ...

Is there any difference in coverage if I exclude files/packages in clover-report instead of clover-setup?.

thanks

[Marek Parfianowicz]

If you exclude files in <clover-setup> then they will not be instrumented at all. Which means that they will not be registered in the clover.db. As a consequence they will not be included in the report as well.

If you exclude files in <clover-report> then such files will be instrumented (in <clover-setup>) and will be registered in a clover.db, but they will not be included in the final report.

So as you can see in both cases you will achieve the same.

What's the difference then?

1) There is a difference in case you want to generate several reports with a different content using the same database.

In such case you shall instrument all files (no exclusions in <clover-setup>) and exclude files in <clover-report> calls. Thanks to this, you will not have to recompile the code in order to get different content of reports.

2) There may be also a performance difference if you use exclusions in <clover-setup> - files which are not instrumented will run faster.

[niv]

Hi,

thanks for information. It is really helpfull.

I have created history point and run history report in the same day.

I am seeing different project stats for both "clover code coverage report" and "Historical coverage report". why these stats are different?. What really history stasts means?. Please let me know?.

ex: Clover Coverage Report statistics for project clover db

stmts:600,100 LOC :2151267 Methods:89940

classes: 11103 NLOC:1212077 Files: 9310

packages:919 TOTAL Coverage :41.4%

Historical coverage Report Project stats: LOC: 2265934 Methods: 94883

NCLOC: 1280188 Classes: 12092

Files: 10033 Packages: 1130

TOTAL: 40.4%

Thanks,

[Marek Parfianowicz]

I think this is caused by a fact that a historical report is based on data found in recorded history points. It means that you have to configure proper exclusion patterns for history points as well. See:

https://confluence.atlassian.com/display/CLOVER/clover-historypoint

the <fileset> nested element and the 'filter' attribute (if you use context filters)

[niv]

I did not find any option so far to group selected packages for coverage.