On 3/19/2013 5:51 PM, Niranjan M wrote:
--debug-trycompile
...
-- Detecting C compiler ABI info
Unable to find the executable at any of:
DIR/mysql-connector-c-6.0.2/CMakeFiles/CMakeTmp/cmTryCompileExec
DIR/mysql-connector-c-6.0.2/CMakeFiles/CMakeTmp/Debug/cmTryCompileExec
On 04/19/2013 07:08 PM, Robert Dailey wrote:
C:\work\buildlibrary
Instead of:
C:\work\build\library
The slash between build and library is missing. Any reason for
this? It doesn't do it if the source binary directories are on the
same drive letter.
Almost certainly this is:
On Sun, Apr 21, 2013 at 12:32 AM, Paul Smith p...@mad-scientist.net wrote:
I looked up this message and found another report of this error which
was purported to be fixed in cmake by this patch:
commit 2dc17f88dd2de900154f153f521b803ec9b7c377
Author: Brad King brad.k...@kitware.com
Date
On 04/24/2013 09:11 AM, Paul Smith wrote:
On Wed, 2013-04-24 at 09:02 -0400, John Drescher wrote:
I installed it and it ran one time and I didn't see the error, but it's
intermittent so that's not definitive. Unfortunately one of my cmake
files failed (later on; this is a cmake file from a
Stephen,
I am able to reproduce this with the simple test case shown below
in any multi-config generator (VS IDE and Xcode).
On 04/24/2013 11:27 AM, Brad King wrote:
On 04/24/2013 09:11 AM, Paul Smith wrote:
The package I'm trying to build is here, FWIW:
https://github.com/nuodb/nuodb-php
On 04/25/2013 03:34 AM, Stephen Kelly wrote:
I haven't had time to investigate fully, but this patch should 'fix' the
problem:
[snip]
I'll investigate later to see if it's the right fix and why.
Great! I've turned that patch into this commit:
On 04/29/2013 10:53 AM, Raffi Enficiaud wrote:
I read the content of the change in the link you provided, and I have a
newbie question:
Why not using one timestamp (different filename) per library/binary/project?
The current approach evolved historically. A per-target approach would
On 04/29/2013 04:37 PM, Raffi Enficiaud wrote:
- process 1 is accessing generate.stamp for reading
Nothing ever reads the file. Only its existence and modification
time matter.
- process 2 is moving some tmpfile to generate.stamp - race
The only race was for multiple processes simultaneously
Hi Jon,
Thanks for looking at this.
On 05/09/2013 09:42 AM, Jonathan Anderson wrote:
I have created a patch that seems to provide the functionality
that I want (attached).
The patch looks interesting. We'll need it to include documentation
and tests for the feature before acceptance.
First,
On 05/09/2013 11:12 AM, Jonathan Anderson wrote:
I've attached a CMakeLists.txt file that demonstrates what
I want to do (and what I can do with my patch).
Thanks, that makes sense.
Does add_library(... SHARED OBJECT ...) work? I think you
mean to set the POSITION_INDEPENDENT_CODE target
On 05/10/2013 10:31 AM, Paul Smith wrote:
I've separated my project statement from:
project(MyProject C CXX)
into:
project(MyProject NONE)
enable_language(C)
enable_language(CXX)
on Windows (VS 10 Win64 generator):
On 05/10/2013 11:23 AM, Paul Smith wrote:
Is there any way to work around this? I've tried setting
CMAKE_CONFIGURATION_TYPES before project()
Make sure you set it as a cache entry:
set(CMAKE_CONFIGURATION_TYPES
Debug;Release;MinSizeRel;RelWithDebInfo
CACHE STRING Supported configs)
On 05/10/2013 11:14 AM, Brad King wrote:
OTOH the Xcode generator just initializes all four configurations
in EnableLanguage up front:
http://cmake.org/gitweb?p=cmake.git;a=blob;f=Source/cmGlobalXCodeGenerator.cxx;hb=v2.8.10.2#l182
I've now taught the VS generator to do the same:
http
On 05/14/2013 10:28 AM, Paul Smith wrote:
On Tue, 2013-05-14 at 09:51 -0400, Brad King wrote:
On 05/10/2013 11:14 AM, Brad King wrote:
OTOH the Xcode generator just initializes all four configurations
in EnableLanguage up front:
http://cmake.org/gitweb?p=cmake.git;a=blob;f=Source
On 05/13/2013 02:08 PM, Pritchett-Sheats, Lori wrote:
IMPLICIT_DEPENDS Fortran ${F_file}
The custom command is running a C preprocessor, not a Fortran compiler,
so use IMPLICIT_DEPENDS C.
list(APPEND PGSLibIface_SOURCE_FILES ${new_file_f90})
Do you mean to append to
On 05/15/2013 04:10 PM, Mark Cianciosa wrote:
I have a fortran code that I'm compiling twice to generate two libraries.
There is a USE statement inside a preprocessor conditional. It looks like
cmake is not picking up the dependency because of it. The code in question is
SUBROUTINE
On 05/04/2013 08:12 AM, Leif Walsh wrote:
I think the colors and carriage returns (without line feeds) are lost
because the build tool sees its controlling terminal is cmake, not a
real terminal program capable of showing color or redrawing, so it
doesn't output them. Cmake may be logging this
On 05/23/2013 12:59 PM, Bogdan Cristea wrote:
I have noticed that the latest cmake release, 2.8.11, does not detect sgemm
using check_function_exists macro from mkl_rt.lib library. The error message
says;
Cannot find C:\Program.obj
It seems to me that the paths with spaces are no longer
On 05/23/2013 02:58 PM, Bogdan Cristea wrote:
Attached you shall find the test file
Thanks, Bogdan.
Stephen, I can reproduce this on Linux using code of the form
include(CheckFunctionExists)
set(CMAKE_REQUIRED_LIBRARIES /path/with space/libfoo.a)
check_function_exists(foo HAVE_FOO)
The
On 05/24/2013 04:24 PM, Brad King wrote:
It does not individually re-quote the library names so when the
generated CMakeLists.txt file is parsed it separates on spaces.
Fixed:
http://cmake.org/gitweb?p=cmake.git;a=commitdiff;h=e65ef08b
-Brad
--
Powered by www.kitware.com
Visit other
On 05/28/2013 11:38 AM, Zamir Khan wrote:
I should probably clarify the situation (I'm not sure I provided enough info
up front). The larger dependencies like ITK are pre-built (I guess one would
assume this; we don't rebuild ITK or VTK with every build of our project). So
I thought it
On 05/31/2013 05:47 AM, Leek, Jim wrote:
Oh, if there any cmake way to get the fortran link flags
I need to pass to g++? I know -lgfortran is all I need in
this case, but if there was a generic way to get cmake to
figure it out for me, that would be nice.
CMake is supposed to figure this out
On 05/31/2013 01:35 PM, Alain Leblanc wrote:
What happens when enable_language(Fortran) is invoked and CMake
can not find the gfortran shared libraries?
CMake does not actually search for the libraries. It runs the Fortran
compiler front-end to build a test executable and passes options to
ask
On 05/31/2013 04:49 PM, Leek, Jim wrote:
CmakeCXXCompiler.cmake:
set(CMAKE_CXX_IMPLICIT_LINK_LIBRARIES
stdc++;mingw32;moldname;mingwex;msvcrt;advapi32;shell32;user32;kernel32;mingw32;moldname;mingwex;msvcrt)
CmakeFortranCompiler.cmake:
set(CMAKE_Fortran_IMPLICIT_LINK_LIBRARIES
On 6/13/2013 4:36 PM, Malfettone, Kris wrote:
Just wanted to add some extra information. I found a way to work around
the problem by first generating with:
cmake –G Visual Studio 11 path_to_src
Then reconfiguring directly after with:
cmake –G Visual Studio 11 -DCMAKE_GENERATOR_TOOLSET=v100
Patrick,
On 08/01/2013 01:41 PM, David Cole wrote:
He's probably just looking for the generator itself in the cmake
--help output. They are listed in the dropdown gui if you use cmake-gui...
The Win64 generators are not listed explicitly in the command line
--help output anymore, although
On 08/15/2013 07:40 AM, David Cole wrote:
http://cmake.org/gitweb?p=cmake.git;a=commitdiff;h=b2ac3463b5ef5dfd3fc662c12cdd177e5dc1
git says it's only in 'next' and 'nightly' so far:
Actually that one was rewritten and is in master:
On 08/15/2013 09:09 AM, Brad King wrote:
Fixed and test added:
http://cmake.org/gitweb?p=cmake.git;a=commitdiff;h=9040ec91
Actually another side effect of the previous behavior that is
good to fix is in a case like this:
message(STATUS a # oops, forgot close paren
message(STATUS b
On 08/22/2013 04:39 PM, Alexander Neundorf wrote:
Beside that, I played around with it a bit.
If 2.8.12 is required, it errors out, as documented.
If I use 2.8.12 and require 2.8.11:
[snip]
So this seems to behave differently than in 2.8.11 even if 2.8.11 is required
(and so the policy
On 08/31/2013 10:50 AM, Marcin Wojdyr wrote:
What's the best practice for linking Fortran program with C++ library?
The library is external, built using autotools, either dynamic or static.
The problem is that runtime c++ library is not linked when the library
is static.
I though that the
On 09/03/2013 01:57 PM, Marcin Wojdyr wrote:
On 3 September 2013 15:46, Brad King brad.k...@kitware.com wrote:
One way to do this is to tell CMake that the library uses C++
using the IMPORTED_LINK_INTERFACE_LANGUAGES property on an
That's what I was looking for, but for me instead of adding
Brad King wrote:
Alexander Neundorf wrote:
CMakeLists.txt: ADD_CUSTOM_COMMAND(OUTPUT
${CMAKE_CURRENT_BINARY_DIR}/generated.h COMMAND sh ARGS -c echo
//generated.h DEPENDS blah.h #DEPENDS
${CMAKE_CURRENT_SOURCE_DIR}/blah.h )
[snip]
if I put only DEPENDS blah.h in the custom command
Alexander Neundorf wrote:
Hi,
I tried to check for the function res_init()
with
CHECK_FUNCTION_EXISTS(res_init HAVE_RES_INIT)
This always fails, although it shouldn't.
The problem is this line in resolv.h:
/usr/include$ grep res_init *
resolv.h:#define res_init
Alexander Neundorf wrote:
I want to check for a function which isn't part of libc, but of libutil.
I tried this:
INCLUDE(CheckFunctionExists)
SET(CMAKE_REQUIRED_LIBRARIES util)
CHECK_FUNCTION_EXISTS(openpty HAVE_OPENPTY)
#SET(CMAKE_REQUIRED_LIBRARIES) # should I reset it here again ?
xavier delannoy wrote:
- How can I add a ressource file (.rc file) in a VS7 project (Win32
Console project) ?
Just include the .rc file in the list of source files.
- How to set specific args to mc.exe (Microsoft Message Compiler) ?
When is this tool being invoked in your setup?
-Brad
[EMAIL PROTECTED] wrote:
Hi
I'm using a copy of CMake from cvs and have found that the
CMAKE_SHARED_LINKER_FLAGS_build_type flags are ignored by the VC
generators.
I have made a fix to the localVC7 generator -see below - but vc6 and
perhaps vc8 still need attention.
Please create a bug
Mike Talbot wrote:
Hello,
I am trying to use cmake and ctest to run unit tests using cppunit. The
unit tests are built as shared libraries and then passed to the cppunit
test harness (DllPlugInTesterd_dll.exe) to run them. My CMakeLists.txt
looks like this:
ADD_LIBRARY(MyUnitTest SHARED
Alexander Neundorf wrote:
We've used
IF(${CMAKE_C_COMPILER} MATCHES ^cl$)
and this should work for you for now, but...
Sure ? This one seems to be set to:
D:/Programme/Microsoft Visual Studio 8/VC/bin/cl.exe
So I'd go with
IF(CMAKE_C_COMPILER MATCHES cl\\.exe)
Actually I think the
Alexander Neundorf wrote:
Von: Brad King [EMAIL PROTECTED]
Of course, that is a simple solution. Each of those files should set a
variable to indicate it has been loaded. I'll make this change. The
:-)
How about platform-id's everywhere ? E.g. NETBSD, OPENBSD, IRIX etc. ?
Yes
Christian Ehrlicher wrote:
Sorry, what's the difference between WIN32 and MINGW? And why do you want
WIN32 instead of WINDOWS?
The name WIN32 instead of WINDOWS is historical, just like _WIN32 is
defined by MS compilers even on 64-bit windows. I've made a note to add
WINDOWS.
When
Tao Zhang wrote:
I have an executable depends on some nonexisting files.I use
ADD_CUSTOM_COMMAND to produce these files.
http://www.cmake.org/Wiki/CMake_FAQ#How_can_I_generate_a_source_file_during_the_build.3F
-Brad
___
CMake mailing list
Jorgen Bodde wrote:
Well should it be enabled by default? Because I am doing nothing else
but generating my project (I am not changing GUI settings) so by default
it is not included right ?
My question wasn't whewre it went, it seems to be missing from the
generated project as VS.NET
Alexander Neundorf wrote:
maybe I don't find it, but I always have to struggle to find the cmake
cvs web interface:
http://public.kitware.com/cgi-bin/viewcvs.cgi/?root=CMake
How about adding a link to it on
http://www.cmake.org/HTML/Download.html ?
Done.
-Brad
Alexander Neundorf wrote:
While you're at it, how about adding the link to the FAQ on this page
http://www.cmake.org/HTML/Documentation.html
below the link to the CMake Wiki page ?
I know the FAQ link is on the bar on the left side, but I always at first
click the Documentation link
Filipe Sousa wrote:
cl : Command line warning D9036 : use 'RTC1' instead of 'GZ'
Fixed in CVS.
-Brad
___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake
Asmodehn Shade wrote:
Hi all,
Just to confirm I have the same problem here... CMake 2.2.3
Fresh install of VS8Express + PSDK (option environment variable checked
ON). Was able to build the win32 template (from the command line) and
run it.
I have checked my paths, I can access cl from the
Filipe Sousa wrote:
Brad King wrote:
There is no default, if the build type is not set then _neither_ will be
linked. The reason there is no default is a bit complicated (I once
Are you sure? I did this test with the fallowing CMakeLists.txt
[snip]
[EMAIL PROTECTED] ~/tmp/default/build
Mike Jackson wrote:
The -install_name option is _already_ being specified by cmake
somewhere so when I tried your suggestion, I get an error during link
phase that the -install_name option is being specified twice.
Is there a master config file somewhere for cmake that I can just edit?
Mike Jackson wrote:
The -install_name option is _already_ being specified by cmake
somewhere so when I tried your suggestion, I get an error during
link phase that the -install_name option is being specified twice.
Is there a master config file somewhere for cmake that I can just
edit?
Filipe Sousa wrote:
pos = lib.find_last_not_of( \t\r\n);
if(pos != lib.npos)
{
lib = lib.substr(0, pos);
//** this code always eats the fist char **
}
Actually it always eats the last character but thanks for tracking it
down. We've fixed it. The fix also
Filipe Sousa wrote:
Now that cmake has EXECUTE_PROCESS command, should we use this instead
of EXEC_PROGRAM?
If you want. It is much more powerful. I'll add this to the documentation.
I notice that both commands captures output in a different way:
[snip]
EXECUTE_PROCESS does not remove
Martin Baumann wrote:
I use the MinGW-version available to download. It comes with gcc 3.4.2.
Hmmm...
I had similar problems with other Utilities, just as you said. With your
advice I could make it!
Now there are errors linking 'Filtering':
Linking CXX shared library
Xavier Delannoy wrote:
my CMakeLists.txt need an environement variable.
I try this, with no success.
/tmp cat CMakeLists.txt
IF(DEFINED ENV{TEST})
MESSAGE(test is defined)
ELSE(DEFINED ENV{TEST})
MESSAGE(test is not defined)
ENDIF(DEFINED ENV{TEST})
I've implemented support
Wheeler, Frederick W (GE, Research) wrote:
The old tclsh Dart had the following return values. Does ctest use the same
values? I don't see anything about ctest return values on the man page or the
cmake wiki.
# Dart status for continuous builds:
# 0: source changed, submitted ok
# 1:
James Bigler wrote:
So I'm using SWIG to generate an interface file to python.
Can I get the interface file (myfile.py) to be built in the library path?
I need two things to run the interface. The myfile.py file and the
_myfile.so library with all the c code. myfile.py gets placed in a
Linton, Tom wrote:
But what I want to do is to source a file that sets up various
paths and environment variables in addition to the compiler
paths. Yes, I could manually do this
within CMakeLists.txt but I'd rather source the actual file that
is installed in the compiler distribution tree, to
Sylvain PRAT wrote:
In fact, I meant duplicate copies for *release* versions :
relwithdebufinfo, minsizerel and release. That is, is there a
CMAKE_MINSIZE_POSTFIX, CMAKE_RELWITHDEBUG_POSTFIX or something like that?
This was added a few weeks in the development version. It will be
included
Christian Ehrlicher wrote:
can sombody tell me why ${OUTPUT} is empty?
I don't see any reason why it should not work with msvc. Can sombody
give me a hint where to look for in the sourcecode?
This is a problem with VCExpress 2005 only because the tool itself does
not produce output. CMake is
Brad King wrote:
Christian Ehrlicher wrote:
can sombody tell me why ${OUTPUT} is empty?
I don't see any reason why it should not work with msvc. Can sombody
give me a hint where to look for in the sourcecode?
This is a problem with VCExpress 2005 only because the tool itself does
Zachary Pincus wrote:
I'm having trouble figuring out how to turn CMAKE_SKIP_RPATH on for a
single subdirectory (or for that matter, just one library).
I've tried all possible permutations of
SET(CMAKE_SKIP_RPATH ON [CACHE BOOL [FORCE]])
at the top of the CMakeLists.txt file in my
Jan Palach wrote:
Hi! Can i use the command SUBDIRS with PREORDER flag?
As of CMake 2.2 the PREORDER flag is not useful or needed. The make
system that is generated does not recurse through the directory
structure. Instead it recurses on a target-level basis and always
builds things in
Brandon J. Van Every wrote:
John Biddiscombe wrote:
I've used boost a lot, but never spent much time looking at the
configuration or build. Has anyone given much though to how hard it'd
be to CMakeify Boost?
(And even if the boost maintainers aren't interested in using CMake,
we could put
Brandon J. Van Every wrote:
William A. Hoffman wrote:
Actually, I just tried it your way and it worked???
PROJECT(foo)
ADD_LIBRARY(libchicken foo.cxx)
SET_TARGET_PROPERTIES(libchicken PROPERTIES PREFIX )
On my version, I now realize I called SET_TARGET_PROPERTIES before
calling
Brandon J. Van Every wrote:
Since you
only implemented per-target flags recently, your perception is probably
that it's uncommon. Instead you're probably thinking that everyone
makes *.h files and subdirectories.
In CMake's earliest days it was part of a project that had one library
per
Brandon J. Van Every wrote:
1.) The error message is worded poorly.
2.) The error is actually referring to the .dll, not the .lib and
is due to a bug with trailing slashes in the install destination.
3.) The .lib is mentioned even though it is not actually installed.
I've fixed #1 and #2 in
Alexander Neundorf wrote:
Hi,
CMAKE_CXX_FLAGS_RELEASE and CMAKE_CXX_FLAGS_DEBUG are appended to
CMAKE_CXX_FLAGS if the buildtype is release or debug.
Now let's say I'd like to have the following flags:
default: -Wall -fno-rtti -O2 -g
release: -Wall -fno-rtti -O2
debug: -Wall
Zachary Pincus wrote:
Thanks everyone for your suggestions about how to deal with rpaths (aka
install names for OS X dylibs) and installed shared libraries.
Option A is to disable the rpaths and make sure that (1) the libraries
are installed in a standard location, or (2) [DY]
[EMAIL PROTECTED] wrote:
Hi Brad
1) works well - thanks
the MESSAGE cmd won't print it which made me doubt it was working
2) doesn't work since the ; seps are removed before the regular
expression is applied
STRING(REGEX REPLACE ; \n CONTENTS ${CONTENTS})
-Brad
Prakash Punnoor wrote:
Actually I don't know. :-) Or perhaps I confused its behaviour while building
a project as such. Here at least it is no problem if you have some libtool
libs declared within the porject and link them either static or shared with
executables or libs within the same
Brandon J. Van Every wrote:
Rodrigo Madera wrote:
I have a project that build a lot of different libraries (like Boost
does, but it's not Boost) with configuration markings and build
options.
I want to build:
MyLib.lib: the release library
MyLib-D.libdebug library
MyLib-A.lib non
Philippe Poilbarbe wrote:
Here is a suggestion for cmake in command line mode.
Cmake is only able to remove files; it would be useful to be able
to remove directories (with possibility to recursively remove their
content).
Please submit a feature request here:
http://www.cmake.org/Bug
-Brad
[EMAIL PROTECTED] wrote:
I wonder if it's possible to create subfolders with the command SOURCE_GROUP
in Visual Studio. For example, I want to group some header files in the
folder include, some in the folder include\graph and some in the folder
include\image.
Is this possible? I found a
Patrick Cheng wrote:
I have the following scripts, it suppose to check out the ITK code to
the ${TEST_BINARY_DIR} directory, but the result ITK code is actually
being checked out into the cmake binary directory (C:\Program
Files\CMake 2.2\bin):
EXEC_PROGRAM(${CVS_COMMAND}
Alexander Neundorf wrote:
Ok, I tested now the following:
set(CMAKE_BUILD_WITH_INSTALL_RPATH TRUE)
set(CMAKE_SKIP_BUILD_RPATH TRUE)
set(CMAKE_INSTALL_RPATH ${LIBRARY_OUTPUT_PATH}/${CMAKE_CFG_INTDIR}/
${CMAKE_INSTALL_PREFIX}/${LIB_INSTALL_DIR})
Which gives
Alexander Neundorf wrote:
Von: Brad King [EMAIL PROTECTED]
CMake is providing an interface to get whatever RPATHs you want into
the installed binary. It is is up to a project's code to produce the
proper path for its distribution. Only the project authors know how
the binaries will finally
Alexander Neundorf wrote:
some thoughts to the new INSTALL command:
-with the old INSTALL commands the paths were always interpreted relative
to CMAKE_INSTALL_PREFIX. At least I got used to writing /bin, /lib
etc. I.e. that the install dir always starts with a /, so that
James Bigler wrote:
I'm using CMake 2.2.3, swig 1.3.28 on Suse Linux 9.3.
I want to set a couple of different options for a swig interface file
I'm compiling:
# This is all on a single line if my email program munges it.
SET_SOURCE_FILES_PROPERTIES(manta.i PROPERTIES SWIG_FLAGS -Wall
Lars Nordlund wrote:
This simple CMakeLists.txt will (on NetBSD at least) cause the binary
'bar' to be linked with -fPIC. Is this really necessary? I know that it
must be used when producing the shared object 'foo', but I did not know
it was needed for the executable also? I have checked the
Filipe Sousa wrote:
QtCanvas\CMakeFiles\QtCanvas.dir\c:\devel\PndK\QtCanvas\qtcanvas.obj:
QtCanvas\CMakeFiles\QtCanvas.dir\flags.make
QtCanvas\CMakeFiles\QtCanvas.dir\c:\devel\PndK\QtCanvas\qtcanvas.obj:
..\..\devel\PndK\QtCanvas\qtcanvas.cpp
Okay, I've committed a fix to avoid this error in
Philippe Poilbarbe wrote:
I have just installed VTK 5.0 under Linux and this version breaks the
FindVTK.cmake mechanism used to find vtk.
FindVTK.cmake (from CMake 2.2.3) tries to find vtk from the path entries
in directories ../lib/vtk but VTK installs itself in lib/vtk-5.0
Attached there
Bitter, Ingmar (NIH/CC/DRD) [E] wrote:
cmake is complaining about a file it can not find that I think it should
not be searching for in the first place.
How do I tell cmake to echo each CMakeLists file and line that it is
working on so I can trace where the source of this problem comes from?
David Cole wrote:
Are
FILE(GLOB var *.txt)
results guaranteed to be sorted alphabetically? Or is that just
happening by happy accident for me...?
There is nothing in the implementation to my knowledge that guarantees
any kind of order. If the files came from a CVS checkout then they are
Alexander Neundorf wrote:
--- Ursprüngliche Nachricht ---
Von: Brad King [EMAIL PROTECTED]
I get this from your test case in LibsDependencies.cmake:
# this file was generated during the build process
SET(blub_LIB_DEPENDS z;)
Is that not what you expect?
Strange.
With current cvs I
Alexander Neundorf wrote:
Von: Brad King [EMAIL PROTECTED]
...
Then the result is:
SET(blub_LIB_DEPENDS z;)
Hmmm.
And it really works for you ?
Since the problem occurs with either command's APPEND option I wonder
if
there is a problem with opening files in append mode
Brad King wrote:
Alexander Neundorf wrote:
Von: Brad King [EMAIL PROTECTED]
Since the problem occurs with either command's APPEND option I wonder
if there is a problem with opening files in append mode. What
exact compiler did you use to build this CMake?
FILE(WRITE ...) FILE(APPEND
Alexander Neundorf wrote:
FILE(WRITE ...)
FILE(APPEND ...)
EXPORT_...(...)
works now.
EXPORT_...(...)
FILE(APPEND ...)
still doesn't work.
This is because EXPORT_LIBRARY_DEPENDENCIES is a final pass command
that executes after all the rest of the configuration is done. This is
Alexander Neundorf wrote:
Von: Brad King [EMAIL PROTECTED]
The idea of that option was to make *double-quote* includes look the
same for in-source and out-of-source builds. For angle-bracket
includes only the explicitly given include path should be used.
Note in your example that running
Xavier Delannoy wrote:
On Wed, 22 Mar 2006 09:51:26 -0500
William A. Hoffman [EMAIL PROTECTED] wrote:
ADD_LIBRARY(fooShared SHARED foo.o)
SET_TARGET_PROPERTIES(fooShared PROPERTIES OUTPUT_NAME foo)
ADD_LIBRARY(fooStatic STATIC foo.o)
SET_TARGET_PROPERTIES(fooStatic PROPERTIES OUTPUT_NAME foo)
Gaetan Lehmann wrote:
In our project, with CONFIGURE_FILE() macro, we are configuring the
paths in a script.
How can we set the executable flag on this script, both for file in
build tree and installed one ?
Currently you have to make sure the input file has the proper
permissions to get
Alexander Neundorf wrote:
in KDE we have the problem that e.g. when compiling kdebase the headers
in the source dir should be preferred over the installed ones.
Usually we would do
include_directories( ${KDE4_INCLUDES} ) in the toplevel CMakeLists.txt.
When adding more include dirs in
Eric BOIX wrote:
I'm trying to do the following things:
1/ set up a default CMAKE_BUILD_TYPE for my project (say RelWithDebInfo)
that of course shows up in the GUIs,
To get options in the GUI they need to be cache settings. Use the CACHE
form of the SET command.
IF(NOT
Alexander Neundorf wrote:
cmake cvs supports the COMPILE_FLAGS property for targets, but this isn't
mentioned in the documentation. The attached patch does just that.
I've updated the documentation with slightly more detail.
-Brad
___
CMake mailing
Eric BOIX wrote:
Dear Brad King,
Quoting Brad King [EMAIL PROTECTED]:
To get options in the GUI they need to be cache settings. Use the CACHE
form of the SET command.
IF(NOT CMAKE_BUILD_TYPE)
SET(CMAKE_BUILD_TYPE RelWithDebInfo CACHE STRING
Build Configuration)
ENDIF
Axel Roebel wrote:
On Wednesday 05 April 2006 14:19, Eric BOIX wrote:
Dear Brad King,
Quoting Brad King [EMAIL PROTECTED]:
To get options in the GUI they need to be cache settings. Use the CACHE
form of the SET command.
IF(NOT CMAKE_BUILD_TYPE)
SET(CMAKE_BUILD_TYPE RelWithDebInfo
Axel Roebel wrote:
does anybody know how I may force the use of a c++ compiler
for source files with .c extension?
Before I start renaming nearly 100 files and loose all
the cvs history I would prefer a solution where cmake would
allow me to select a compiler for the part of the files
that
Michael Biebl wrote:
FWIW, this is my solution so far:
MACRO(PROCESS_IIFILES _i_FILES _ii_FILES)
SET(_i_FILES)
FOREACH(_current_FILE ${ARGN})
GET_FILENAME_COMPONENT(_tmp_FILE ${_current_FILE} ABSOLUTE)
GET_FILENAME_COMPONENT(_abs_PATH
Dário Oliveira wrote:
Hi. I've found a topic that solved my las problem but now I'm having
another and I think that hs to do with the fact that my linux is ubuntu
64 bits. The library is in the directory, I can do a ldd command and it
points do the lib-64 directory. Is there any solution for
Gaetan Lehmann wrote:
Hi,
Currently, I'm using the ADD_CUSTOM_TARGET() macro to compile the java
tests in WrapITK. The problem is that those files are rebuild for each
make.
Is there a better way to build java programs ?
Use ADD_CUSTOM_COMMAND, and then add only one ADD_CUSTOM_TARGET
Michael Biebl wrote:
after a successfull make install run I want to display some messages
and execute a command. With automake I used a install-data-hook in the
toplevel Makefile.am for this, so the commands were processed at the
end of the install run.
How can I do this with cmake? I tried to
Michael Biebl wrote:
I have OPTION(INSTALL_LINKS ... ON) in the toplevel CMakeLists.txt
file and want to use that variable in a POST_INSTALL_SCRIPT. Depending
on INSTALL_LINKS I want to create additional links. Unfortunately the
value of INSTALL_LINKS is not propagated to the POST_INSTALL_SCRIPT
601 - 700 of 15210 matches
Mail list logo
