> Kishore, Jonnalagadda (IE10) wrote: > >> KJI> This would have worked just fine except that INCLUDE_DIRECTORIES > >> returns > >> KJI> semicolon separated list of directory names not a space separated > >> list > >> > >> Seems to be true for any list oft values as I experienced also. Is it > >> true content or only the formatting of the MESSAGE() command? > > > > Hmmm... You are right. It seems to be the formatting of the message > > command. My problem was with including the definitions part. Generally > > speaking, it seems to be a problem with cmake introducing backslashes in > > the command. In my case, the command failed with including the > > definitions. I do not have a clear idea of the pattern in which the > > backslashes are introduced. > > > > Warm regards, > > Kishore > > You must be doing something wrong. I have been using this macro for ages: > MACRO(QT_ADD_MOC output) > GET_DIRECTORY_PROPERTY(_flags DEFINITIONS) > SEPARATE_ARGUMENTS(_flags) > > GET_DIRECTORY_PROPERTY(_includes INCLUDE_DIRECTORIES) > FOREACH(_dir ${_includes}) > LIST(APPEND _flags ${CMAKE_INCLUDE_FLAG_CXX}${_dir}) > ENDFOREACH() > > FOREACH(_header_file ${ARGN}) > GET_FILENAME_COMPONENT(_header_file ${_header_file} ABSOLUTE) > GET_FILENAME_COMPONENT(_basename ${_header_file} NAME_WE) > SET(_moc_file ${CMAKE_CURRENT_BINARY_DIR}/moc_${_basename}.cpp) > ADD_CUSTOM_COMMAND( > OUTPUT ${_moc_file} > COMMAND ${QT_MOC_EXECUTABLE} ${_flags} -o ${_moc_file} > ${_header_file} > DEPENDS ${_header_file} > ) > LIST(APPEND ${output} ${_moc_file}) > ENDFOREACH() > ENDMACRO()
I was doing the same thing except that I was not using the command SEPARATE_ARGUMENTS. I added that and now the command does not fail. I checked the command execution with VERBOSE=1 and notice that backslashes still exist but not in a way to cause the command to fail. SEPARATE_ARGUMENTS in this contect does not make much sense to me. The description says it replaces the space between arguments with semi-colons. I don't want semicolons or the command would consider it the start of a new command. Yet it works as expected! :-) Warm regards, Kishore _______________________________________________ CMake mailing list CMake@cmake.org http://www.cmake.org/mailman/listinfo/cmake