Use code-coverage.cmake from cmake-scripts
#124
Comments
|
Okay I figured out why the What is even weirder is that when generating the code coverage data, it includes |
|
Okay, I don't know how I didn't see this before. It's so simple. To instrument a library target and an executable one, just call Although it also imports |
It should though, e.g. using Makefiles on Mac, the code coverage data is stored in cmake -Htest -Bbuild/test -DENABLE_TEST_COVERAGE=1
cmake --build build/test
./build/test/GreeterTestsTo summarise the data you could simply call > gcovr
------------------------------------------------------------------------------
GCC Code Coverage Report
Directory: .
------------------------------------------------------------------------------
File Lines Exec Cover Missing
------------------------------------------------------------------------------
source/greeter.cpp 8 8 100%
------------------------------------------------------------------------------
TOTAL 8 8 100%
------------------------------------------------------------------------------TBH I don't do much local analysis myself and rely on platforms such as codecov.io for a nice visualisation. But all the data should be there and you should be able to easily generate the html reports locally. I'm not really sure about the advantage of adding additional abstraction for the coverage, as the flags are straightforward to use and give most user control without additional required code. To read more about gcov and how the flags work, see the wikipedia entry. |
|
As far as I see it, the current solution doesn't support |
|
This doesn't seem to work anymore: |
|
@acgetchell try a |
|
@ClausKlein That fixes it, thanks! |
Currently, the template just applies
-fprofile-arcs -ftest-coverageto theGreetertarget fromtest. As far as I could tell, it doesn't really generate the code coverage data locally, and I think it'd be useful to be able to do that.The
code-coverage.cmakescript exposes targets that we can run if we want to generate the data locally too. It also does checks that the current ModernCppStarter solution doesn't.The text was updated successfully, but these errors were encountered: