I've remembered that per Satoris' advice the cmake-extras EnableCoverageReport.cmake macro works in a different way that makes it incompatible with the virally-distributed EnableCoverageReport.cmake that we find in mir and unity8, etc.
Specifically, in the existing builds we do
$ cmake ../ -DCMAKE_BUILD_TYPE=coverage
whereas in cmake-extras we do
$ cmake -Dcoverage=1
So I won't remove that file--Martin would we then want a new test for the 'cmake-extras-EnableCoverageReport.cmake' method? Upon consideration that's an unfortunate name-collision with the virally distributed one--or maybe cmake-extras-EnableCoverageReport.cmake should support both enablement mechanisms?
Note that the new method is nowhere in production and is therefore not yet a case that coverage-builder needs to cover.
I've remembered that per Satoris' advice the cmake-extras EnableCoverageR eport.cmake macro works in a different way that makes it incompatible with the virally-distributed EnableCoverageR eport.cmake that we find in mir and unity8, etc.
Specifically, in the existing builds we do
$ cmake ../ -DCMAKE_ BUILD_TYPE= coverage
whereas in cmake-extras we do
$ cmake -Dcoverage=1
So I won't remove that file--Martin would we then want a new test for the 'cmake- extras- EnableCoverageR eport.cmake' method? Upon consideration that's an unfortunate name-collision with the virally distributed one--or maybe cmake-extras- EnableCoverageR eport.cmake should support both enablement mechanisms?
Note that the new method is nowhere in production and is therefore not yet a case that coverage-builder needs to cover.
You can read Jussi's and my correspondence about the new method here: https:/ /code.launchpad .net/~allanlesa ge/libcolumbus/ enable- coverage- option/ +merge/ 218875 .
Meanwhile FindGtest.cmake has indeed made to the archive, will attach that branch to this bug.