[CMake] Missing Folders in CPack generated TGZ

2010-04-09 Thread David Graf
Hello

I have a strange cpack problem: I use cpack to generate a Mac install package 
(PackageMaker) and to generate a tar.gz. Logically, the tar.gz and the 
installer should contain/install the same files. Unfortunately, this is not the 
case. Some more or less random directories are missing in the tar.gz. But 
strangely, these files are not missing in the helper folder 
_CPack_Packages/TGZ. _CPack_Packages/TGZ contains the correct directories. 

Is this a known problem? Can I debug this somehow? Can someone tell me the 
command that is used to pack _CPack_Packages/TGZ?

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 at: 
http://www.cmake.org/Wiki/CMake_FAQ

Follow this link to subscribe/unsubscribe:
http://www.cmake.org/mailman/listinfo/cmake


Re: [CMake] Batch Files in CTest (instead executables)

2010-04-01 Thread David Graf
Hello David

The problem I have is not located in the batch file. The EXIT is executed. But 
if the /B option is on, ctest doesn't detect failures anymore. The batch script 
must be killed by executing EXIT without the '/B' options. But then, the batch 
script cannot be used anymore from the command line because it kills the 
command line all the time.

I didn't find a clean solution in the meantime. Now, I am using different 
scripts. One to execute executables form ctest (without /B option). And one to 
execute executables from the command line (with /B option).

David

On Mar 31, 2010, at 7:34 PM, David Cole wrote:

 Try something like this instead:
 SETLOCAL
 SET PATH=...;%PATH%
 execute_something.exe
 set my_error=%ERRORLEVEL%
 IF %my_error% neq 0 (
  EXIT /B %my_error%
 )
 ENDLOCAL
 
 I think the problem is probably that ERRORLEVEL is changing to 0 after you 
 execute the batch script IF command... Saving it in your own variable 
 prevents that sort of sole last error madness...
 
 Let me know if that does not work. This should be working as you expect.
 
 
 On Wed, Mar 31, 2010 at 3:59 AM, David Graf david.g...@28msec.com wrote:
 Hello Everybody
 
 I want to execute batch files in CTest tests (to set the correct paths to the 
 PATH environment variables before executing an executable). Unfortunately, 
 that does not work smoothly. The reason is the error handling.
 If an error happens in the batch file, the batch file has to be killed with 
 an exit command. Otherwise, ctest doesn't receive an error code. My script 
 has to look like this:
 
 SETLOCAL
 SET PATH=...;%PATH%
 execute_something.exe
 IF %ERRORLEVEL% neq 0 (
  EXIT %ERRORLEVEL%
 )
 ENDLOCAL
 
 
 Unfortunately, this is not handy because the batch file cannot be executed 
 anymore from the cmd. Actually, it can. But it kills that cmd every time an 
 error happens. Not really nice. In my opinion, the batch file should look 
 like this:
 
 SETLOCAL
 SET PATH=...;%PATH%
 execute_something.exe
 IF %ERRORLEVEL% neq 0 (
  EXIT /B %ERRORLEVEL%
 )
 ENDLOCAL
 
 Because of the argument '/B', the cmd is not killed anymore, but the error 
 code is correctly set! Unfortunately, ctest doesn't behave correctly anymore 
 after this change. After this change, every test passes. No matter what is 
 going on. CTest doesn't seem to check the %ERRORLEVEL%.
 
 Is this a missing feature in ctest? Or am I missing something?
 
 Cheers,
 
 David Graf
 
 
 ___
 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 at: 
 http://www.cmake.org/Wiki/CMake_FAQ
 
 Follow this link to subscribe/unsubscribe:
 http://www.cmake.org/mailman/listinfo/cmake
 

___
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 at: 
http://www.cmake.org/Wiki/CMake_FAQ

Follow this link to subscribe/unsubscribe:
http://www.cmake.org/mailman/listinfo/cmake

Re: [CMake] Batch Files in CTest (instead executables)

2010-04-01 Thread David Graf
Hello Michael

I fixed it differently. I just two different scripts. One for ctest and one to 
execute executables from the command line directly.

David 

On Mar 31, 2010, at 3:16 PM, Michael Wild wrote:

 Why on earth is it so bad to have a trailing line with --- YAY TEST PASSED 
 --- (or something like this) in the output of ctest -V?
 
 On 31. Mar, 2010, at 15:04 , David Graf wrote:
 
 Hello
 
 But this is also not a solution for me. Our testdriver prints some debugging 
 information that can be seen if ctest is executed with the -V option. 
 Therefore, it is impossible to echo SUCCESS or FAILED.
 
 David
 
 On Mar 31, 2010, at 1:23 PM, Michael Wild wrote:
 
 
 On 31. Mar, 2010, at 11:47 , David Graf wrote:
 
 Hello Michael
 
 Instead of checking the exit status, you could have CTest check the 
 output for success or failure. Have a look at SET_TESTS_PROPERTIES.
 
 Unfortunately, that doesn't work in our case. We are running a testdriver 
 for the result checking. There reason is: Our tests produce xml and XML 
 cannot be compared like a string. It must be translated into canonical 
 form before it can be compared. We wrote this part in C++. Therefore, we 
 cannot do the comparison in CMake.
 
 David
 
 What I meant is this to output a string from your batch script:
 
 SETLOCAL
 SET PATH=...;%PATH%
 execute_something.exe
 IF %ERRORLEVEL% neq 0 ( 
 ECHO FAILED
 ) ELSE (
 ECHO SUCCESS
 )
 ENDLOCAL
 
 
 Michael
 
 

___
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 at: 
http://www.cmake.org/Wiki/CMake_FAQ

Follow this link to subscribe/unsubscribe:
http://www.cmake.org/mailman/listinfo/cmake


[CMake] Batch Files in CTest (instead executables)

2010-03-31 Thread David Graf
Hello Everybody

I want to execute batch files in CTest tests (to set the correct paths to the 
PATH environment variables before executing an executable). Unfortunately, that 
does not work smoothly. The reason is the error handling.
If an error happens in the batch file, the batch file has to be killed with an 
exit command. Otherwise, ctest doesn't receive an error code. My script has to 
look like this:

SETLOCAL
SET PATH=...;%PATH%
execute_something.exe
IF %ERRORLEVEL% neq 0 ( 
  EXIT %ERRORLEVEL%
)
ENDLOCAL


Unfortunately, this is not handy because the batch file cannot be executed 
anymore from the cmd. Actually, it can. But it kills that cmd every time an 
error happens. Not really nice. In my opinion, the batch file should look like 
this:

SETLOCAL
SET PATH=...;%PATH%
execute_something.exe
IF %ERRORLEVEL% neq 0 ( 
  EXIT /B %ERRORLEVEL%
)
ENDLOCAL

Because of the argument '/B', the cmd is not killed anymore, but the error code 
is correctly set! Unfortunately, ctest doesn't behave correctly anymore after 
this change. After this change, every test passes. No matter what is going on. 
CTest doesn't seem to check the %ERRORLEVEL%.

Is this a missing feature in ctest? Or am I missing something?

Cheers,

David Graf


___
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 at: 
http://www.cmake.org/Wiki/CMake_FAQ

Follow this link to subscribe/unsubscribe:
http://www.cmake.org/mailman/listinfo/cmake


Re: [CMake] Batch Files in CTest (instead executables)

2010-03-31 Thread David Graf
Hello Michael

 Instead of checking the exit status, you could have CTest check the output 
 for success or failure. Have a look at SET_TESTS_PROPERTIES.

Unfortunately, that doesn't work in our case. We are running a testdriver for 
the result checking. There reason is: Our tests produce xml and XML cannot be 
compared like a string. It must be translated into canonical form before it can 
be compared. We wrote this part in C++. Therefore, we cannot do the comparison 
in CMake.

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 at: 
http://www.cmake.org/Wiki/CMake_FAQ

Follow this link to subscribe/unsubscribe:
http://www.cmake.org/mailman/listinfo/cmake


Re: [CMake] Batch Files in CTest (instead executables)

2010-03-31 Thread David Graf
Hello

But this is also not a solution for me. Our testdriver prints some debugging 
information that can be seen if ctest is executed with the -V option. 
Therefore, it is impossible to echo SUCCESS or FAILED.

David

On Mar 31, 2010, at 1:23 PM, Michael Wild wrote:

 
 On 31. Mar, 2010, at 11:47 , David Graf wrote:
 
 Hello Michael
 
 Instead of checking the exit status, you could have CTest check the output 
 for success or failure. Have a look at SET_TESTS_PROPERTIES.
 
 Unfortunately, that doesn't work in our case. We are running a testdriver 
 for the result checking. There reason is: Our tests produce xml and XML 
 cannot be compared like a string. It must be translated into canonical form 
 before it can be compared. We wrote this part in C++. Therefore, we cannot 
 do the comparison in CMake.
 
 David
 
 What I meant is this to output a string from your batch script:
 
 SETLOCAL
 SET PATH=...;%PATH%
 execute_something.exe
 IF %ERRORLEVEL% neq 0 ( 
  ECHO FAILED
 ) ELSE (
  ECHO SUCCESS
 )
 ENDLOCAL
 
 
 Michael

___
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 at: 
http://www.cmake.org/Wiki/CMake_FAQ

Follow this link to subscribe/unsubscribe:
http://www.cmake.org/mailman/listinfo/cmake


[CMake] Weired behavior with CTest Fedora 10 g++ 4.3.2

2008-12-17 Thread David Graf

Hello

We have a weired problem when we compile and test our software on Fedora 
10 with g++ 4.3.2.
We have around 600 tests. When we execute these tests in one ctest run, 
two random tests always fail (without any regularity). We tested it on 
two different fedora installation with ctest 2.4 and 2.6. This strange 
behavior occurs on both machines.
After some debugging, we observed that the failing tests do not produce 
any log output. Even when adding the std::cout out statements into the 
code, we do not get some output. We assume there is some bad interaction 
between Fedora 10 and ctest.


Has someone noticed the same strange behavior?

David
___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake


[CMake] post build add_test

2008-12-06 Thread David Graf
Hello

Is it somehow possible to make the ADD_TEST command dependent on a target?
Is it possible to add tests to a project after a certain target has been
built? Currently, I have the impression that tests can only be added to a
project when the cmake code is executed, before the build process starts not
later. But maybe, there is a trick to add tests after the buid process.

I have the situation that I wanna use an executable of my own project to
compute the test cases (the executable is built out of external sources that
are not tested in my project). The result of the executable is a list of
strings that I wanna use to add tests to my project.

David
___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake

[CMake] CPack DEB Packaging: Automate dependency resolution

2008-11-01 Thread David Graf

Hello

Currently in my project, I provide all dependencies for DEB packaging in 
CPack by setting the variable CPACK_DEBIAN_PACKAGE_DEPENDS. Is there a 
possibility to automate this process (similar to the behavior of the 
CPack RPM packaging mechanism)?


As far as I understood it correctly, such a mechanism must be 
implemented with the Linux tool 'objdump' to get the dependencies of 
shared libraries. I already thought about implementing my own script to 
get these dependencies of my shared libraries automatically. But first, 
I wanted to make sure that I do not reinvent the wheel. Especially 
because I found the following entry in 'CPackDeb.cmake' next to the 
dependency section:  ' TODO: automate 'objdump -p | grep NEEDED'.


David
___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake


Re: [CMake] RPM Package Dependencies

2008-10-31 Thread David Graf




Hello Eric

Thanks a lot! It works!

I made a feature request (http://public.kitware.com/Bug/view.php?id=7904).
Please attach your updated CPackRPM.cmake to it.

David



Eric Noulard wrote:

  2008/10/30 David Graf [EMAIL PROTECTED]:
  
  
Hello

I would like to add a dependency to a RPM packages that is generated by
CPack 2.6 manually. Is this possible (similar to the DEB solution that is
setting the variable CPACK_DEBIAN_PACKAGE_DEPENDS)?


  
  
No currently, there is no such feature.
However Would you try to replace the CPackRPM.cmake file
currently installed on your system with the one attached to this mail.

Then try to define
CPACK_RPM_PACKAGE_REQUIRES in your CMakeLists.txt
with the appropriate value.

The rebuild your RPM with cpack.
Then verify that the dependency is properly set using:

rpm -qpR yourrpm.rpm

If all this did go smooth, would you be kind enough to
file a feature request in the bug tracker and I'll try to make the update
if the CMake dev team is interested.

If it does not work for you,
please file a feature request :=)

I'll try harder.

  




___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake

[CMake] RPM Package Dependencies

2008-10-30 Thread David Graf

Hello

I would like to add a dependency to a RPM packages that is generated by 
CPack 2.6 manually. Is this possible (similar to the DEB solution that 
is setting the variable CPACK_DEBIAN_PACKAGE_DEPENDS)?


David
___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake


[CMake] set env vars with nsis installer created by cpack

2008-08-07 Thread David Graf
Hello

I am creating an nsis package from my cmake project with cpack and have the
following question:
Is there a possibility in cpack to force the nsis installer to add an entry
to the windows environment variables which contains the path to the
installed package as value? The environment vars should contain after
installation an entry like PROJECTNAME_HOME=C:\Program Files\projectname.

David
___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake

Re: [CMake] Ruby Python Binding with SWIG in CMake

2008-05-13 Thread David Graf




I managed to solve the problem by generating each SWIG module in a
separate directory. But it is not possible to use one SWIG interface
file for every module. Thus, I added in every directory a '.i' file
that just includes the original '.i' file from the upper dir (%include
../test.i).

David


Alan W. Irwin wrote:
On 2008-05-12 11:00-0700 David Graf wrote:
  
  
  Hello


I have problems to generate Python and Ruby bindings with SWIG in CMake

out of one SWIG interface file. In the following example, I have a
small

interface file and a CMake file that should generate Ruby and Python

bindings. But the Ruby binding is not generated at all and the Python

binding is broken because the two generation processes produce a
library

file with the same name (_test.so).


Does someone know how to produce bindings from one interface for
several

languages? I need bindings for Ruby, Python, PHP, Java, C#.


  
  
Multiple bindings should not be a problem for swig and CMake. For
example,
  
PLplot implements python and java bindings using swig. If you want to
see
  
how this example works in detail, look at
  
http://plplot.svn.sourceforge.net/viewvc/plplot/trunk. In particular,
you
  
should look at bindings/swig-support/plplotcapi.i (which defines our
API for
  
swig), and CMakeLists.txt files and *.i files in bindings/python and
  
bindings/java.
  
  
Hope that working example helps you to figure out what is wrong for
your own
  
attempt to build multiple bindings using swig and CMake.
  
  
Alan
  
__
  
Alan W. Irwin
  
  
Astronomical research affiliation with Department of Physics and
Astronomy,
  
University of Victoria (astrowww.phys.uvic.ca).
  
  
Programming affiliations with the FreeEOS equation-of-state
implementation
  
for stellar interiors (freeeos.sf.net); PLplot scientific plotting
software
  
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads
of
  
Linux Links project (loll.sf.net); and the Linux Brochure Project
  
(lbproject.sf.net).
  
__
  
  
Linux-powered Science
  
__
  




___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake

[CMake] Deb Package Dependencies

2008-05-12 Thread David Graf

Hello

In CPack 2.6, is it possible to add package dependencies to the 
generated deb package?


David


___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake


[CMake] Ruby Python Binding with SWIG in CMake

2008-05-12 Thread David Graf




Hello

I have problems to generate Python and Ruby bindings with SWIG in CMake
out of one SWIG interface file. In the following example, I have a
small interface file and a CMake file that should generate Ruby and
Python bindings. But the Ruby binding is not generated at all and the
Python binding is broken because the two generation processes produce a
library file with the same name (_test.so).

Does someone know how to produce bindings from one interface for
several languages? I need bindings for Ruby, Python, PHP, Java, C#.

David

Example:

SWIG Interface File:
%module test
%{
 int get() { return 5; }
%}

int get();

CMake File:
PROJECT(swig_test)

FIND_PACKAGE(SWIG)
INCLUDE(${SWIG_USE_FILE})

SET_SOURCE_FILES_PROPERTIES(test.i PROPERTIES CPLUSPLUS ON)

FIND_PACKAGE(PythonLibs)
INCLUDE_DIRECTORIES(${PYTHON_INCLUDE_PATH})
SWIG_ADD_MODULE(test python test.i)
SWIG_LINK_LIBRARIES(test ${PYTHON_LIBRARIES})

FIND_PACKAGE(Ruby)
INCLUDE_DIRECTORIES(${RUBY_INCLUDE_PATH})
SWIG_ADD_MODULE(test ruby test.i)
SWIG_LINK_LIBRARIES(test ${RUBY_LIBRARY})




___
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake