On 09/10/2015 04:03 PM, Stephen Kelly wrote: >> One possible fix to the problems is to avoid allowing empty >> values to be appended at all as is done for non-buildsystem >> properties. > > I'm sure that's possible. Is it acceptable to fix it?
Yes. The inconsistency with non-buildsystem properties is a worthwhile fix anyway. Empty entries in buildsystem properties have no semantic meaning (except perhaps the property value being exactly the empty string as against unset), so there should be no compatibility problem. Please look at fixing that and reference in the commit message when it first changed behavior. This will of course involve updating the expected output in the test cases I added. Then the new tests for directory properties I attached at the start of this thread can be added too. > Another possible fix is to store a vector of sentinal positions for each > buildsystem property. This could be done in addition to the above fixes, but may not be necessary unless needed to support empty property values. Thanks, -Brad P.S. Please refrain from further refactoring until post-3.4 development opens. I'd like to get existing regressions fixed without risking new ones in the weeks leading up to the freeze. -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
