Re: [CMake] Possiblity to speedup the build by compiling multiple files in one compiler invocation

Is this of use to you? http://www.gmixer.com/archives/46/comment-page-1#comment-320 Regards, Arjen From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Martin Nielsen Sent: dinsdag 26 april 2011 15:11 To: cmake@cmake.org Subject: [CMake] Possiblity to speedup the build

Re: [CMake] Replacing compiler flags for certain project subdirectories

Tim, I don't think cmake can facilitate this out of the box (to my knowledge). The problem is that some compilers and fortran code just don't mix with +O3 (blas, lapack). We have worked around it by reimplementing a mechanism from our old build system, where optimization for fortran c and cxx

Re: [CMake] Specify an alternate C++ compiler for some source files

Since our default compilers are not gcc (icc, pgcc, cc, xlc_r etc) we've introduced FLEXLM_C_COMPILER and friends to be able to compile this from a single source tree. But perhaps it would be cleaner to off-load it with ExternalProject_Add(). Seriously, how often do you want to compile LAPACK,

Re: [CMake] Unify entire tree into a single static library?

Doug, There is no convenient way to do this in Cmake you have you program around this issue with a GLOBAL list of source files. Michael Wild has offered a nice example at http://www.mail-archive.com/cmake@cmake.org/msg27780.html Hopefully this is useful to you as well. Regards, Arjen From:

Re: [CMake] Expected boost path structure.

Just to make sure: set BOOST_ROOT=X:\32bit\VC.90\Libraries\boost-1.45.0 Won't work? Or setting it to this value in the Computer Properties Environment variables section? I thought cmake worked fine with the path separator that is native to the OS. Would be strange if it didn't. Arjen

Re: [CMake] feed parsed information from source files into compile process

Thanks Tyler, -Original Message- From: Tyler Roscoe [mailto:ty...@cryptio.net] You can't use add_custom_command() to drive this because the results of any custom commands will only be available at build time. You need stuff to happen at CMake configure time. I noticed.

[CMake] feed parsed information from source files into compile process

Hi, I'm looking for a way to parse information from source files and feed the outcome as input to the build process. In more detail, I have Fortran files that contain information about what optimization flags should be fed to the compiler. Currently there is a python script that does

Re: [CMake] cmake 2.8.3 forces -rdynamic for pf90 in pure fortran projects

Hi Brad, Please try the patch below. It gets rid of -rdynamic for pure F90. I will try in a C/C++/F77/F90 project soonish (hopefully today). I didn't check if other flags get cleared as well. Thanks, Arjen diff --git a/Modules/Platform/Linux-PGI.cmake b/Modules/Platform/Linux- PGI.cmake index

[CMake] cmake 2.8.3 forces -rdynamic for pf90 in pure fortran projects

Hi, -rdynamic is used for linking with pgf90 even though it is not supported. The flag is imported from Linux-GNU.cmake and I guess it is included because cmake (2.8.3) checks for C and C++ compilers by default. http://www.vtk.org/Bug/view.php?id=9985 talks about this bug, but I'm unsure what

Re: [CMake] Converting a large C++-Project to CMake

Benjamin, If you haven't already, you might want to take a lot at JOM, which is essentially parallel nmake: http://qt.gitorious.org/qt-labs/jom. JOM is supported by CMake. Thank you Pedro. One of our developers already tried to plug jom into our current build system, but some trouble with

[CMake] cmake 2.8.2 doesn't build on suse 9.3

Hi, For some reason bootstrap on a plain SuSE 9.3 fails to find sys/wait.h. Manually defining HAVE_SYS_WAIT_H to 1 in config.h allows cmake and friends to be built. Regards, Arjen -rw-r--r-- 1 root root 6620 2005-06-11 00:58 /usr/include/sys/wait.h -- Looking for include files

Re: [CMake] Merge two static libraries

Hi, -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Marcel Loose The downside to this solution is that you have duplicates of the object files that are part of common, but that's the price you'll have to pay if you want to have just one

Re: [CMake] cmake 2.8.2 doesn't build on suse 9.3

: dinsdag 5 oktober 2010 18:59 To: Verweij, Arjen Cc: cmake@cmake.org Subject: Re: [CMake] cmake 2.8.2 doesn't build on suse 9.3 On 10/05/2010 08:34 AM, Verweij, Arjen wrote: For some reason bootstrap on a plain SuSE 9.3 fails to find sys/wait.h. Manually defining HAVE_SYS_WAIT_H to 1 in config.h allows

Re: [CMake] How to invoke 64 bit cmd.exe from cmake script onWindows 7 (x64).

I would be interested to learn how. Because of similar issues we got rid of a shell script on Win64 because UnxUtil zsh.exe doesn't run from the 64bit cmd.exe. Also, when in a 64bit VS shell, I now need to run the 32bit nmake.exe because otherwise my 32bit flex.exe will crash. I have tried

Re: [CMake] Setting CMAKE_CXX_FLAGS..., CMAKE_C_FLAGS... and CMAKE_C_STANDARD_LIBRARIES_INIT and friends on a per target basis

I can't say I share this experience. We use this mechanism to substitute standard flags with less or more optimization and/or other options. Usually -O3 instead of -O2, and I don't get them both on the commandline. Arjen From: Brian Davis [mailto:bitmi...@gmail.com]

Re: [CMake] Setting CMAKE_CXX_FLAGS..., CMAKE_C_FLAGS... and CMAKE_C_STANDARD_LIBRARIES_INIT and friends on a per target basis

Hi Brian, We override the standard CMAKE_LANG_FLAGS (C CXX Fortran) and use it as a baseline. Then, for individual source files it is possible to override them in a separate file using: set_source_files_properties(${CMAKE_BINARY_DIR}/path/to/file.f PROPERTIES GENERATED TRUE COMPILE_FLAGS

Re: [CMake] 64 bit windows: rsp file not employed in build.make

Hi, The objects1.rsp file is generated by CMake during configuration. The start/end temp file syntax tells NMake to use response files at build time. Why is this scheme only used for shared libs and not static libs? For static libs it works automagically somehow. The link rule uses

Re: [CMake] ifort 9.1 unable to compile a simple test program

Hi, Interesting. It works for ifort 11.1. So does ifort /incremental:yes myfile.f ICE for any file with ifort 9.1? Hrm, no not really :| Not too sure why this is breaking. Is this technically an ICE? I see ifort reporting it, but it occurs when linking and both ifort and xilink call

[CMake] 64 bit windows: rsp file not employed in build.make

Hi, I'm investigating a problem where on 64-bit Windows the generated build.make (NMake or NMake JOM) does not make use of response files. I have tried transplanting the settings file for the 32 bit platform and minimizing differences, but still I cannot get it to use rsp files. This is a

[CMake] comments wrong in Windows-ifort.cmake

Patch attached :P Arjen Windows-ifort.cmake.patch Description: Windows-ifort.cmake.patch ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and

Re: [CMake] ifort 9.1 unable to compile a simple test program

Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: maandag 9 augustus 2010 16:15 To: Brad King Cc: cmake@cmake.org Subject: Re: [CMake] ifort 9.1 unable to compile a simple test program Hi, Interesting. It works for ifort 11.1. So does

Re: [CMake] 64 bit windows: rsp file not employed in build.make

Hi Bill, I'm also unsure why in one case the linking is done through link.exe and by lib in the other. One is using visual studio, and one is using the intel compiler. They have different files in the Platform directory. That is why they can be different. See Windows-icl.cmake. Thanks,

Re: [CMake] ifort 9.1 unable to compile a simple test program

\CMakeTmpcmTryCompileExec.exe Hello Can I just turn this off in Windows-ifort.cmake or is this symptomatic of another problem? Thanks, Arjen -Original Message- From: Brad King [mailto:brad.k...@kitware.com] Sent: vrijdag 6 augustus 2010 22:28 To: Verweij, Arjen Cc: cmake@cmake.org Subject: Re: [CMake] ifort

[CMake] ifort 9.1 unable to compile a simple test program

Hi, Ifort 9.1 is unable to compile a simple test program on WinXP 64 bit with cmake 2.8.2. I can execute the linkline manually and it seems to link fine. How can I debug what's going on? Thanks, Arjen E:\buildcmake Z:\f90example -G NMake Makefiles --debug-trycompile debug trycompile on --

Re: [CMake] ifort 9.1 unable to compile a simple test program

testCXXCompiler.cxx testFortranCompiler.f E:\build\CMakeFiles\CMakeTmp -Original Message- From: Brad King [mailto:brad.k...@kitware.com] Sent: vrijdag 6 augustus 2010 17:49 To: Verweij, Arjen Cc: cmake@cmake.org Subject: Re: [CMake] ifort 9.1 unable to compile a simple test program On 08/06/2010 11

Re: [CMake] ifort 9.1 unable to compile a simple test program

I did it is below the cmake error in the previous email. No error just a prompt And no noticeable difference to me. Arjen -Original Message- From: Brad King [mailto:brad.k...@kitware.com] Sent: vrijdag 6 augustus 2010 21:39 To: Verweij, Arjen Cc: cmake@cmake.org Subject: Re: [CMake

[CMake] linking large lib on windows 64 bit

Hi, I'm trying to link my project on Windows. Part of the task is constructing a big archive. Relevant compilers: Intel C/C++, Intel Fortran, Microsoft VS 2005 SE SP1. Apparently the link line is too long (expanded command line lib bla too long). I have attempted to circumvent it with: #

Re: [CMake] eliminating -blibpath on AIX

SKIP_BUILD_RPATH ? -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Dave Semeraro Sent: woensdag 4 augustus 2010 15:58 To: cmake@cmake.org Subject: [CMake] eliminating -blibpath on AIX Hello, I am building projects on an AIX system and I

Re: [CMake] eliminating -blibpath on AIX

Patch Modules/Platform/AIX.cmake ? -Original Message- From: Dave Semeraro [mailto:semer...@ncsa.illinois.edu] Sent: woensdag 4 augustus 2010 16:42 To: Verweij, Arjen Subject: Re: [CMake] eliminating -blibpath on AIX Arjen and Denis, Neither suggestion worked. There seems to be no way

Re: [CMake] [cmake-developers] Bug fix requests for the *next* release of CMake...

I'd like to see those fixed: http://public.kitware.com/Bug/view.php?id=11030 That one is closed. ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages

[CMake] libraryname decoration

Olaf, Why? I'm still waiting for someone to post a reason of why a decorated name is a problem for them. Also waiting on an answer to the code duplication issue. Perhaps you should write a proposal that takes care of details, like how you would like to see this decorated. Then write a patch

Re: [CMake] Debug/Release configurations for codeblocks ?

Hi, The default is for CMake to generate a Release configuration for Makefile based projects. ie, if you do NOT specify or otherwise set CMAKE_BUILD_TYPE then you will be a Release configuration. I think this statement is false for Windows, or I am misunderstanding Windows-cl.cmake -- I had an

Re: [CMake] Debug/Release configurations for codeblocks ?

Hi, Isn't build_type = Release sufficient? Not for the way I'm mistreating cmake ;) set (CMAKE_BUILD_TYPE RELEASE FORCE) SET (CMAKE_EXE_LINKER_FLAGS_INIT FORCE) set (CMAKE_C_FLAGS_RELEASE_INIT ) set (CMAKE_CXX_FLAGS_RELEASE_INIT ) set (CMAKE_Fortran_FLAGS_RELEASE_INIT ) Also I have to add

Re: [CMake] [VS gen] Multiple configurations code

David, I hear what you are saying, but a lot of legacy Makefile based system support make and make debug alongside each other. I'm still looking for the cmake way to easily mark a small part of a project to be built with debug flags. For instance, if you have functionality in a static archive,

Re: [CMake] Debug/Release configurations for codeblocks ?

Hi, Which is why I stated Makefile based generators. Not nmake or msvc. Alright, I didn't know you implied Unix Makefiles :) Arjen ___ Powered by www.kitware.com Visit other Kitware open-source projects at

Re: [CMake] cmake generators on windows: nmake versus jom

Hi Alex, Will this do? http://www.cmake.org/Wiki/CMake_Generator_Specific_Information#Makefile_generators Regards, Arjen -Original Message- From: Alexander Neundorf [mailto:a.neundorf-w...@gmx.net] Sent: zondag 25 juli 2010 19:09 To: cmake@cmake.org Cc: Verweij, Arjen; Brad King

Re: [CMake] how to set ar/ranlib flags per target

() - myadd_static_library() ? - to indicate if you want to use a 32-bit or 64-bit archiver. Regards, Arjen -Original Message- From: Hickel, Kelly [mailto:kelly_hic...@bmc.com] Sent: donderdag 22 juli 2010 13:46 To: Hickel, Kelly; Verweij, Arjen; cmake@cmake.org Subject: RE: how to set ar/ranlib flags

Re: [CMake] cmake generators on windows: nmake versus jom

Brad, I found another repository with jom binaries. It works fine with 0.9.4. Apparently the original link I used to download JOM was ancient. Thanks, Arjen -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: woensdag 21

[CMake] cmake generators on windows: nmake versus jom

Hi, I'm having trouble generating the Makefiles I want on Windows. I can't generate Unix Makefiles because cmake fails to compile a simple program. But also using generators that should be very much alike, NMake Makefiles and NMake Makefiles JOM, the first will finish and the second will not.

Re: [CMake] incremental build slow because of F90 (cmake_copy_f90_mod)

Brad, Is this normal behaviour after all? Or at least intended? Thanks, Arjen -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: dinsdag 13 juli 2010 13:27 To: cmake@cmake.org Subject: Re: [CMake] incremental build slow

Re: [CMake] cmake generators on windows: nmake versus jom

2010 17:00 To: Verweij, Arjen; cmake@cmake.org Subject: RE: [CMake] cmake generators on windows: nmake versus jom Hi Arjen, This may be totally the wrong direction, but we have seen similar errors that appear to be caused by resident programs that crawl the local filesystem. Desktop search

Re: [CMake] how to set ar/ranlib flags per target

Kelly, I suffer from the same problem and solved it like this: SET(CMAKE_C_ARCHIVE_CREATE CMAKE_AR -X64 cr TARGET LINK_FLAGS OBJECTS) SET(CMAKE_C_ARCHIVE_APPEND CMAKE_AR -X64 r TARGET LINK_FLAGS OBJECTS) SET(CMAKE_C_ARCHIVE_FINISH CMAKE_RANLIB -X64 TARGET) SET(CMAKE_CXX_ARCHIVE_CREATE

Re: [CMake] cmake generators on windows: nmake versus jom

Hi, The generators are testing and inspecting the toolchain. Something is not building as expected so it reports the problem up front rather than waiting for some subtle error when building the main project. OK. NMake Makefiles JOM [snip] -- FIXME: altered Windows.cmake for debugging

Re: [CMake] how to set ar/ranlib flags per target

-Original Message- From: Hickel, Kelly [mailto:kelly_hic...@bmc.com] Sent: woensdag 21 juli 2010 22:46 To: Verweij, Arjen; cmake@cmake.org Subject: RE: how to set ar/ranlib flags per target -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf

Re: [CMake] cmake 2.8.2 win32: invalid escape sequence when finding fortran compiler

Spot on Bill: C:\Users\madymo\build\CMakeFiles\CMakeTmpcat CMakeLists.txt cmake_minimum_required(VERSION 2.8) PROJECT(CMAKE_TRY_COMPILE Fortran) SET(CMAKE_VERBOSE_MAKEFILE 1) SET(CMAKE_Fortran_FLAGS /w /I:C:\Program Files\VNI\CTT5.0\include\IA32 /fpe:3 /nologo ${COMPILE_DEFINITIONS})

Re: [CMake] cmake 2.8.2 win32: invalid escape sequence when finding fortran compiler

Hi Bill, I don't think we ended up with those flags after installing the intel compiler for the dashboard we run at Kitware. Are you sure that came from the installer? Not really. I wasn't around when that system was set up. I tried from cmd.exe, MKS ksh and cygwin shell to get something with

[CMake] cmake 2.8.2 win32: invalid escape sequence when finding fortran compiler

Hi, I'm trying a little F90 example on Windows with cmake. Apparently something is amiss, but I can't find much on Google about it; at least not about this particular case. All I do is cmake ..\f90example and the following appears in my terminal: S:\mnt\usr3\people\verweija\buildcmake

Re: [CMake] CMAKE_LANG_FLAGS added to link rule

Hello Michael, Apparently, there may be - at least with GCC and shared libraries - a need to specify flags on the link line if they have been specified on the compile line before. Though, I don't know an example of a damaged binary due to different sets of flags for compiling and linking.

Re: [CMake] CMAKE_LANG_FLAGS added to link rule

Hi, AFAIK there is nothing that will be broken on AIX by removing FLAGS from the link rule for Fortran executables. OK, I will just redefine the Fortran linker to keep the cmake generator compatible then start scrubbing link flags based on compiler documentation, because: What flags are

Re: [CMake] incremental build slow because of F90 (cmake_copy_f90_mod)

...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: maandag 12 juli 2010 20:31 To: Brad King Cc: cmake@cmake.org Subject: Re: [CMake] incremental build slow because of F90 (cmake_copy_f90_mod) Hi Brad, It's not *supposed* to do that. Can you produce a minimal project

Re: [CMake] incremental build slow because of F90 (cmake_copy_f90_mod)

Wrong log.txt :P -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: dinsdag 13 juli 2010 13:27 To: cmake@cmake.org Subject: Re: [CMake] incremental build slow because of F90 (cmake_copy_f90_mod) Sent Brad source code

[CMake] incremental build slow because of F90 (cmake_copy_f90_mod)

Hi, We're finding incremental builds to be slow, because the F90 source seems to be privy to a processing step known as cmake_copy_f90_mod. While cmake and the compiler race through C/CXX/Fortran(77) source, this is not the case for F90. What does cmake_copy_f90_mod do? I can hardly find any

[CMake] CMAKE_LANG_FLAGS added to link rule

Hi, It seems that in our project the contents of CMAKE_LANG_FLAGS is added to the link line. This causes problems for some of our platforms. Is there a mechanism to separate language specific compile and link flags out? It would be nice if you could define CMAKE_LANG_COMPILER_FLAGS and

Re: [CMake] CMAKE_LANG_FLAGS added to link rule

of the environment variable, e.g. CMAKE_CXX_FLAGS and CXXFLAGS from the shell, which means it doesn't make sense to include it when linking. Am I overlooking something here :) Regards, Arjen From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: maandag 12 juli 2010 18

Re: [CMake] incremental build slow because of F90 (cmake_copy_f90_mod)

Hi Brad, Actually the copy-f90-mod step is a *feature* to make incremental rebuilds faster. It is an optimization for the case that the .f90 source file that provides a module gets recompiled but the *interface* to the module does not change. Many compilers update the timestamp on the .mod file

Re: [CMake] CMAKE_LANG_FLAGS added to link rule

To: Verweij, Arjen Cc: cmake@cmake.org Subject: Re: [CMake] CMAKE_LANG_FLAGS added to link rule On 07/12/2010 01:12 PM, Verweij, Arjen wrote: As a follow-up: redefining CMAKE_Fortran_LINK_EXECUTABLE to not include FLAGS gets rid of the CMAKE_Fortran_FLAGS.. but perhaps other (useful) stuff as well

Re: [CMake] incremental build slow because of F90 (cmake_copy_f90_mod)

Hi Brad, It's not *supposed* to do that. Can you produce a minimal project that reproduces that behavior? This is reassuring. I'm probably not able to reproduce it with something small, but I could try. I would imagine that if it's not supposed to do it, it likely won't :) Apparently, the

Re: [CMake] CMAKE_LANG_FLAGS added to link rule

Brad, That sounds fine, I'm already in the habit of specifying link flags per language anyway. In fact I will take a looksie and recompile cmake with your suggested fix. Regards, Arjen Looking at the implementation I see it gets architecture flags on the Mac (-arch and -isysroot), which is

Re: [CMake] CMAKE_LANG_FLAGS added to link rule

Brad, That bit of code is #IFDEF'ed for __APPLE__. I doubt it is causing the problem on AIX. Regards, Arjen -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: maandag 12 juli 2010 20:47 To: Brad King Cc: cmake@cmake.org

Re: [CMake] problems with creating static library on aix

Hi, Apparently it had to do with the setting for ncargs. Increasing it to 256 (blocks of 4K) solved it. Regards, Arjen -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: dinsdag 6 juli 2010 18:55 To: cmake@cmake.org

[CMake] problems with creating static library on aix

Hi, We're experiencing two problems with creating static libraries through cmake on our AIX system. The first one is that a cmake based build environment will come up with 64-bit XCOFF, but ar and friends won't play ball unless you add -X64. I suppose this issue has just been overlooked and

Re: [CMake] pgf90 error

Hi, Sorry for reviving this, but I've read it and I didn't understand. Will some version of 2.8.x clear this flag on Linux when the compiler is pgf90 or is this now a documented feature for which you need to clear the flags manually? For the record, when using pgf90 you also need to clear

[CMake] pgf90: influencing link flags for the binary

Hi, After clearing a lot of FLAGS, the final link rule for my executable has been reduced to something that resembles what is being used to link it with the current make system. But I still want to get rid of some other stuff like -lpgc and -lnspgc. Is there a possibility to influence this as

Re: [CMake] pgf90: influencing link flags for the binary

Never mind, it's CMAKE_XXX_IMPLICIT_LINK_LIBRARIES. Regards, Arjen -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Verweij, Arjen Sent: dinsdag 30 maart 2010 16:02 To: cmake@cmake.org Subject: [CMake] pgf90: influencing link flags

Re: [CMake] How to run shell command in CMakeLists.txt

Hi, Please guide me how to create one static lib from multiple static libs. Can I use ar -x libname and then again make ar from the *.o files? Is there any other way? -- Thanks Regards, Kiran Patil Don't even try. Create the single static lib in the first place. There has been a long

Re: [CMake] static library from several subdirectories

Hi Markus, -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Markus Raab I have exactly the same problem, please share if you get a good solution. Any luck yet? I think we got it nailed so far, but now it needs to refactoring and reusage of

Re: [CMake] Nice handling of code coverage option?

Hi Brad, -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Brad Hards Any comments or suggestions on this? We use gcov and lcov for our C++ code. I think you should take care to make sure you have the same version gcov as gcc/g++. Also I

Re: [CMake] Nice handling of code coverage option?

Hi Brad, -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Brad Hards Do you have something else? Not really, but I don't build debug either (no -g). I don't know about an iron clad way to prevent cmake from optimizing anything, so I just

Re: [CMake] scope of set_source_files_properties?

Hi, -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Simmons, Aaron All the sources are going into a single *library*. I could make a bunch of static libraries corresponding to each folder, but I'm not sure how to subsequently link all them

Re: [CMake] scope of set_source_files_properties?

Hi, -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Hendrik Sattler That's a convenience library approach, and it's not really supported (partly because there is no way to implement it on some toolchains such as MSVC). The problem

Re: [CMake] static library from several subdirectories

Hi, From: Michael Wild [mailto:them...@gmail.com] You have to call ADD_SOURCES for all your sources BEFORE you do the ADD_LIBRARY call. I am already doing this, but since the files I was adding to the library didn't exist, the call failed. I somehow expected cmake to start executing the

Re: [CMake] static library from several subdirectories

Hi, -Original Message- From: Michael Wild [mailto:them...@gmail.com] Problem is the following: Dependencies between stuff in different directories only work for top- level targets (add_library, add_executable and add_custom_target). So, you would need to add a ADD_CUSTOM_TARGET call

Re: [CMake] static library from several subdirectories

Hi, -Original Message- From: Michael Wild [mailto:them...@gmail.com] I'd recommend to change the add_sources function to do all the preprocessing and then only add the products (.f, .f90, .c etc.) files to the list. You can do the preprocessing like this (of course, you'll have to

Re: [CMake] static library from several subdirectories

Hi, This is a bit length, sorry. -Original Message- From: Michael Wild [mailto:them...@gmail.com] What am I missing? I'm trying again with my simple a/liba.c and b/b/libb.c setup, but this time I start out with liba.cr and libb.cr that need to be converted to liba.c and libb.c.

Re: [CMake] static library from several subdirectories

Hi, From: Michael Wild [mailto:them...@gmail.com] I'd recommend to change the add_sources function to do all the preprocessing and then only add the products (.f, .f90, .c etc.) files to the list. You can do the preprocessing like this (of course, you'll have to change things for your setup,

Re: [CMake] static library from several subdirectories

Hi, -Original Message- From: Michael Wild [mailto:them...@gmail.com] I think the problem is that the ${BIG_LIB} target does not exist when you call ADD_DEPENDENCIES. Probably you should also collect all the custom targets into a global property and then after the ADD_LIBRARY(${BIG_LIB}

Re: [CMake] static library from several subdirectories

Hi, Hopefully I'm misunderstanding you, since otherwise the list would contain close to 4000 files. Well, I don't see a real problem with this. If this makes your files too long, just separate the setting of these source-variables into separate files. I would prefer to keep the source listing

Re: [CMake] static library from several subdirectories

to solve it I will report back here. Regards, Arjen -Original Message- From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of Markus Raab Sent: dinsdag 16 maart 2010 17:43 To: cmake@cmake.org Subject: Re: [CMake] static library from several subdirectories Verweij, Arjen

[CMake] static library from several subdirectories

Hi, For a legacy FORTRAN project, I'm looking for a cmake-way to create a static library from several subdirectories. The current solution is using a custom POST_BUILD step with ar and ranlib, but the resulting archive is not known to cmake and can't be used to link to the main. As an

Re: [CMake] static library from several subdirectories

Hi, For a legacy FORTRAN project, I'm looking for a cmake-way to create a static library from several subdirectories. The current solution is using a custom POST_BUILD step with ar and ranlib, but the resulting archive is not known to cmake and can't be used to link to the main. You can just

Re: [CMake] ctest with cmake without compiler

or an installer, install some test tools, and test the binaries that were provided in a clean environment, like a virtual machine. Testing on build machines is undesirable. Setting CMAKE_MAKE_PROGRAM to (n)make.exe does not help. Regards, Arjen | Verweij, Arjen wrote: | Well, obviously the problem

[CMake] ctest with cmake without compiler

Hi, I'm looking for the switch that kills the compiler check in cmake. Is there such a switch? Through Google I found http://www.mail-archive.com/cmake@cmake.org/msg08245.html but the answer is not in there. Regards, Arjen ___ CMake mailing list

Re: [CMake] ctest with cmake without compiler

without compiler On Wednesday 04 February 2009, Verweij, Arjen wrote: Hi, I'm looking for the switch that kills the compiler check in cmake. Is there such a switch? Which problem do you have with the check ? It should work. Alex ___ CMake mailing list