A co-worker sees it on his Windows box (VS 7.1), but I don't on mine (VS 8). So I don't know how to trigger it on Windows.
But I see it consistent for me on Linux, both with that simple example and ParaView. Of course, I don't see the msvcrt(d) link warnings. Clint On Tuesday 11 December 2007 1:53:36 pm David Cole wrote: > It's something more than just this simple example... I've tried it with > CMake 2.4.5, 2.4.7 and today's CVS CMake and it does not happen on my > machine. (The Debug C always properly links to Ad only, not also to A...) > > I am using VS 2005 standard edition. > > The problem was occurring on the ParaView3 dashboards (until I masked it > out by mucking with CXX_FLAGS last Friday) on morva and minastirith > dashboards. Here's an example: > http://www.paraview.org/ParaView3/Testing/Sites/morva.kitware/Win32-vs8/200 >71207-0100-Nightly/BuildWarning.html > > morva is using CMake 2.4.5 and VS8 Professional edition, minastirith is > using CMake 2.4.7 and VS8 Standard edition. > > Any other clues about what additional steps / settings might be required to > trigger this error condition? > > On 12/10/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > > Can this be fixed for 2.4.8? It looks like it was already fixed for 2.6, > > but > > I couldn't find a bug report for it. > > > > ================================================= > > ADD_LIBRARY(A a.c) > > ADD_LIBRARY(Ad a.c) > > > > ADD_LIBRARY(B b.c) > > TARGET_LINK_LIBRARIES(B debug Ad optimized A) > > # if building shared libs, cmake correctly links B with -lAd OR -lA > > > > ADD_EXECUTABLE(C c.c) > > TARGET_LINK_LIBRARIES(C B) > > # cmake incorrectly links C with -lB -lAd -lA if build type is Debug > > =================================================== > > > > Clint > > > > On Wednesday 05 December 2007 3:13:39 pm Bill Hoffman wrote: > > > I have a beta release for 2.4.8 ready for cmake. This will be the last > > > release of the 2.4.X branch. The next release will be 2.6.0. So, > > > please make sure you test it if you are interested in a 2.4.8. Send > > > any issues to me or the cmake list. Thanks. > > > > > > The files can be found here: > > > > > > http://www.cmake.org/files/v2.4/*RC-4* > > > > > > > > > The changes from 2.4.7 are as follows: > > > > > > Changes in CMake 2.4.8 RC 4 > > > * fix for cpack and messing up PATH with NSIS > > > Changes in CMake 2.4.8 RC 3 > > > * fix for bug 5363: GET_TARGET_PROPERTY(... DEBUG_LOCATION) > > > returns value containing $(OutDir) > > > * Better error from ctest if nightly time not set > > > * Fix for exception handling flags in VS 2003 and up > > > * Avoid relinking exclude-from-all directory targets before install > > > Changes in CMake 2.4.8 RC 2 > > > * fix for bug 5590 relative paths in windows not working across drives > > > * fix warning/error with TAR_VERBOSE flag > > > Changes in CMake 2.4.8 RC 1 > > > * Fix for kde4-config location > > > * Fix for self extracting .sh files on solaris > > > * Remove KDE3_ENABLE_FINAL (did not work) > > > * KDE3 fix for 64 bit location of plugins > > > * mark PYTHON_EXECUTABLE as advanced > > > * Fix for version numbers on NetBSD > > > * Add more search directories (install prefix and cmake location) > > > * include WindowsPaths in Windows.cmake not just Windows-cl.cmake > > > * documentation fix for file, find_package, try_run > > > * add IS_ABSOLUTE to if > > > * INSTALL() everything which doesn't have a COMPONENT set, is assigned > > > to the COMPONENT "Unspecified" > > > * make #cmakedefine output match autoconf when undefined > > > * document cmake remove -f > > > * document order of -D and -P > > > * add support for DragonFly and GNU hurd > > > * fix for fortran depends doing too many scans > > > _______________________________________________ > > > CMake mailing list > > > CMake@cmake.org > > > http://www.cmake.org/mailman/listinfo/cmake > > > > _______________________________________________ > > CMake mailing list > > CMake@cmake.org > > http://www.cmake.org/mailman/listinfo/cmake _______________________________________________ CMake mailing list CMake@cmake.org http://www.cmake.org/mailman/listinfo/cmake