On 17. Sep, 2010, at 16:49 , David Aldrich wrote:
> Hi Michael
>
>> Ok, to clear things up:
>
>> I hope you get the idea.
>
> Thanks very much, I think I get it now! I have implemented an out-of-source
> build tree for our project, as you suggested, and it is building fine.
Good.
>
> I'm
On 17. Sep, 2010, at 16:50 , Eric Noulard wrote:
> 2010/9/17 SC :
>> David, Michael,
>>
>> Thank you for your help. [David, sorry for this double mailing to you]
>>
No, this is intentional, I suspect to avoid conflicts. There are people
doing stuff like
add_library(foo foo.c foo.
Hi Brad,
> This is why I was asking all those questions. I bet the library
> has not been built with a proper SONAME field.
You are completely right, the library I was using to import has no SONAME
field. I know exactly what I am doing wrong. I did build a proper library with
the SONAME set wi
Thanks,
but could it be that FindGit.cmake isn't complete?
Because i tried Git_WC_INFO and i get an Unknown CMake command error. Or is the
command for git different?
Best Regards
Am 17.09.2010 um 17:50 schrieb Alexander Neundorf :
> On Friday 17 September 2010, norulez wrote:
>> Hello,
>>
>>
I have tried many variations to the HKEY_LOCAL_MACHINE setting and am unable to
get a run through of Cpack.
I know it must be something simple I am missing.
I need help with this.
I am on a windows 7 / 64bit machine. NSIS is unicode and installed in the 32
bit
program files directory.
Thanks,
I also get:
2> nvcc Hâ∞(Ç=y░: nvcc cannot find a supported cl version. Only MSVC 8.0
and MSVC 9.0 are supported
2> CMake Error at
CMakeFiles/gpuReconPlugin_generated_gpuPMatrixRecon.cu.obj.cmake:194
(message):
2>Error generating
2>
C:/Users/bdavis5/Documents/QS/QS-NIH/source/branches/trunk/
On Friday 17 September 2010, Brad King wrote:
> On 09/17/2010 03:35 AM, Molsen, Hannes wrote:
> > In CMakeTestCCompiler.cmake I found a flag called
> > "CMAKE_C_COMPILER_FORCED".
>
> This is meant to be set by cross compiling toolchain files that explicitly
> set the compiler ABI information. It i
On 09/17/2010 01:06 PM, Brad King wrote:
> In the case of an imported target it checks
> for the IMPORTED_NO_SONAME property:
[snip]
> Try setting that for your imported target (to value "1"). It looks
> like this property is missing from the documentation though! I'll
> fix that.
Documentation
I recently had to upgrade to vs 2010 to use NSight debugger for NVIDIA GPU
degging and was using 2.8.0. This version would load all but my CUDA
related projects the cpuReconPlugin and the gpuReconPlugin. These two libs
are defined in the same directory and share common source (which has been
prob
On 09/17/2010 12:31 PM, Pere Mato Vila wrote:
> Thanks for your interest. In fact your questions made me think
> in the direction that the problem must be in the way the
> library libGaudiKernel.so was built or installed. In fact I
> remembered that I have a soft link in the installation
> path. T
Hi Brad,
Thanks for your interest. In fact your questions made me think in the
direction that the problem must be in the way the library libGaudiKernel.so was
built or installed. In fact I remembered that I have a soft link in the
installation path. The directory /build/mato/GAUDI/GAUDI_v21
On Friday 17 September 2010, norulez wrote:
> Hello,
>
> is there a way in cmake to determine if it is a svn or git working copy to
> use the individual commands?
>
> Such as:
>
> If svn then FindPackage(Svn).
> Else if git then FindPackage(Git)...
something like the following should work:
if(
Hi Pere,
What platform are you using (uname -a)? Does it use ELF binaries?
On 09/17/2010 11:24 AM, Pere Mato Vila wrote:
> $ ldd libLHCbMathLib.so
>
> /build/mato/GAUDI/GAUDI_v21r10p1/InstallArea/x86_64-slc5-gcc43-opt/lib/libGaudiKernel.so
> (0x2afdec07d000)
>...
> $ ldd
> /
Hello,
is there a way in cmake to determine if it is a svn or git working copy to use
the individual commands?
Such as:
If svn then FindPackage(Svn).
Else if git then FindPackage(Git)...
Thanks in advance
Best Regards
NoRulez
___
Powered by www.
>> The problem I have is that when have already built and
>> installed B I can not move anymore the location of the A. This
>> is because the libraries created in B contains the absolute
>> path to exported libraries in A.
>
> The binaries in B would at most contain RPATH entries to point at the
2010/9/17 SC :
> David, Michael,
>
> Thank you for your help. [David, sorry for this double mailing to you]
>
>> > No, this is intentional, I suspect to avoid conflicts. There are people
>> > doing stuff like
>> > add_library(foo foo.c foo.cpp foo.f)
>> > Why is this a problem for you?
> Right, I u
Hi Michael
> Ok, to clear things up:
> I hope you get the idea.
Thanks very much, I think I get it now! I have implemented an out-of-source
build tree for our project, as you suggested, and it is building fine.
I'm wondering what my co-developers will think of it. It is usual for us to
work
David, Michael,
Thank you for your help. [David, sorry for this double mailing to you]
> > No, this is intentional, I suspect to avoid conflicts. There are people
> > doing stuff like
> > add_library(foo foo.c foo.cpp foo.f)
> > Why is this a problem for you?
Right, I understand the point of view
On 09/17/2010 09:31 AM, Pere Mato Vila wrote:
> The problem I have is that when have already built and
> installed B I can not move anymore the location of the A. This
> is because the libraries created in B contains the absolute
> path to exported libraries in A.
The binaries in B would at most c
On Fri, Sep 17, 2010 at 7:43 AM, Brad King wrote:
> On 09/17/2010 07:19 AM, Pere Mato Vila wrote:
> > I am using the nice feature of exporting targets from one CMake
> > project to another. This works really nicely since it avoids
> > explicitly linking my executables and shared libraries with th
Dear Brad,
Thanks for the prompt reply, but still I do not understand. I agree that if I
have a project A that is exporting targets of the install tree and a project B
that is using project A, I can move the install tree of A and I can still
continue to build project B because the properties
Ok, to clear things up:
Source tree: here are all the source files (*.h, *.cpp, CMakeLists.txt, etc).
It's the stuff you check into your version control system.
Build tree (a.k.a binary tree): here go all the products (i.e. generated files,
like CMakeCache.txt, cmake_install.cmake, *.o, librari
Hi Ryan
Thanks for your reply.
CMakeLists.txt always lives in the source directory I assume?
If it does live with the source, is there anyway of moving the clutter of
cmake_install.cmake, CMakeFiles etc. in the source directory?
David
From: Ryan Pavlik [mailto:rpav...@iastate.edu]
Sent: 17 Sep
On 09/17/2010 07:19 AM, Pere Mato Vila wrote:
> I am using the nice feature of exporting targets from one CMake
> project to another. This works really nicely since it avoids
> explicitly linking my executables and shared libraries with the
> dependent libraries of the imported library target. The
> http://public.kitware.com/Bug/view.php?id=3796 )
Oh thank you, that should be the next topic (long awaited) what I wanted
to test. Now it has to wait. I didn't realize this correlation.
>
>
> In the meantime, to get rid of the problem for now, turn the new
> feature off. In the top le
On 09/17/2010 03:35 AM, Molsen, Hannes wrote:
> In CMakeTestCCompiler.cmake I found a flag called "CMAKE_C_COMPILER_FORCED".
This is meant to be set by cross compiling toolchain files that explicitly
set the compiler ABI information. It is not set by --debug-trycompile.
> So I still want to get
Hi,
That didnt make any difference. I am still getting the same error.
There seems to be some permission problems. Ryan mentioned that it looked
like either Cmake or VC++ had ran as administrators but no longer do. Im not
sure how to check that. But i did as he said and deleted the build directory
On Fri, Sep 17, 2010 at 6:51 AM, David Cole wrote:
>
>
> On Fri, Sep 17, 2010 at 1:01 AM, Micha Renner wrote:
>
>> I get the message "CMakePredefinedTarget not available" and then VS
>> message: "Cannot load CMakePredefinedTarget project".
>>
>> What does this mean for the work with CMake and VS2
Your build system would be independent where the output files are: the user
can choose whatever build directory they want, and not be limited to
_gnuDebug and _gnuRelease. (If they so chose, they could create each of
those and configure a build into them, but they are just as likely to not do
so.
Hi Chris
> No, you shouldn't have to, unless you're using in-source builds
> which is very strongly deprecated. Once you've gotten used to
> out-of-source builds you'll never want to go back.
Ok, I'm trying to think of how this would work for us.
The source for each of our libraries is in a s
Hi,
I am using the nice feature of exporting targets from one CMake project to
another. This works really nicely since it avoids explicitly linking my
executables and shared libraries with the dependent libraries of the imported
library target. The problem is that imported library is linked
On Fri, Sep 17, 2010 at 4:12 AM, Michael Wild wrote:
> On 17. Sep, 2010, at 9:41 , SC wrote:
>
> > Hi there,
> >
> > I recently upgraded my dev system from Ubuntu 8.04 to Ubuntu 10.4. This
> made me change from Cmake 2.4-patch7 to
> > Cmake 2.8.0.
> >
> > Since then, generating a makefile with "C
On Fri, Sep 17, 2010 at 1:01 AM, Micha Renner wrote:
> I get the message "CMakePredefinedTarget not available" and then VS
> message: "Cannot load CMakePredefinedTarget project".
>
> What does this mean for the work with CMake and VS2010 Express?
>
That means that I made a mistake with this commi
On Fri, Sep 17, 2010 at 3:22 AM, David Aldrich wrote:
> Hi Chris
>
>
>
> Thanks for your reply.
>
>
>
> >The message() and file(REMOVE) commands will be executed when you run
> CMake.
>
> >The command to generate the .cpp file won't be executed until you run
> make.
>
>
>
> I don’t think I have ‘
On 09/17/2010 12:04 PM, Chris Hillery wrote:
> The message() and file(REMOVE) commands will be executed when you run CMake.
> The command to generate the .cpp file won't be executed until you run make.
> That's why it's still around after you're done.
>
> You can't really do exactly what you want
Hi Chris
Thanks for your reply.
>The message() and file(REMOVE) commands will be executed when you run CMake.
>The command to generate the .cpp file won't be executed until you run make.
I don't think I have 'got' CMake yet. I am now definitely confused!
During normal development, i.e. when the
The message() and file(REMOVE) commands will be executed when you run CMake.
The command to generate the .cpp file won't be executed until you run make.
That's why it's still around after you're done.
You can't really do exactly what you want here very easily; you'd need to
have a separate custom
Hi
I want to generate a source file 'SourceFileInfo.cpp', then build a library and
then delete the generated file.
So I wrote:
add_custom_command (
OUTPUT SourceFileInfo.cpp
COMMAND ../VersionInfo/_gnuRelease/versionInfo . KERNEL
DEPENDS ${SRCS}
COMMENT "Generating SourceFileInfo.cpp"
Is the Opal compiler itself written in Opal? I guess it must be, or else you
wouldn't be having this problem.
We have a similar problem with the Zorba XQuery engine. In that case, while
Zorba is coded in C++, some of the C++ files are generated from XQueries,
which leads to a similar chicken/egg p
Hi Michael
> ADD_EXECUTABLE(exe "")
Thanks, that worked nicely.
David
___
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 check the CMake FAQ
On 17. Sep, 2010, at 9:41 , SC wrote:
> Hi there,
>
> I recently upgraded my dev system from Ubuntu 8.04 to Ubuntu 10.4. This made
> me change from Cmake 2.4-patch7 to
> Cmake 2.8.0.
>
> Since then, generating a makefile with "Cmake ." in the Simclist library
> source directory produces the f
Hi there,
I recently upgraded my dev system from Ubuntu 8.04 to Ubuntu 10.4. This made me
change from Cmake 2.4-patch7 to
Cmake 2.8.0.
Since then, generating a makefile with "Cmake ." in the Simclist library source
directory produces the following :
#==
So, you want to embed the "basename" of the output file in the target, right?
In that case you should be aware of the fact that variables/properties like
CMAKE__POSTFIX can change your name and with mult-config generators you
don't know what is at configure-time. AFAIK there is currently no
re
Hi all,
if this is of any help or interest: I added the option --debug-output, which
does not solve the problem, but gives at least some additional information and
a nicer workaround:
=== CMake Output : snip ===
-- Detecting C compiler ABI info
Called from: [2] C:/Program Files/CMake
2.
Hi all,
we are working on a port of our old, handcrafted build system for OPAL
(http://user.cs.tu-berlin.de/~opal/opal-language.html) to cmake.
The Opal compiler produces C-Code, so in theory bootstrapping that beast
on any machine should be easy as long as the C-code is shipped in a
distribution
45 matches
Mail list logo