Andre,
We have seen that on all our cygwin machines (32 and 64 bit) running on
Windows 7 as well. This was on the latest update to cygwin and cmake 2.8.9.
We had 2.8.6 running before and did not experience these BAD COMMAND errors.
(Did get some random timeout errors though).
I've tried remo
The regex check for version of java does not find the version.
My system reports java -version as:
java version "1.7.0_b147-icedtea"
I changed line 110-113:
IF(var MATCHES "java version \"[0-9]+\\.[0-9]+\\.[0-9_.]+[oem-]*\".*")
# This is most likely Sun / OpenJDK, or maybe
need to disect the INSTALL
command?
I like the;
set(CMAKE_JAVA_MANIFEST mymanifest.mf)
option for manifests since they are more specific to java jar files.
>
> On Thursday 20 October 2011 11:59:05 Allen D Byrne wrote:
> > I've upgraded my Java project to the latest 2.8.6 releas
I needed to change the UseJava.cmake script to allow INSTALL components and to
have more flexible locations for JAR resource files.
In addition I needed to change the java_copy_file function to use the
execute_process command.
Attached is the diff for my changes, hoping the correct folks will re
I have finally successfully compiled our library and tests on Win7 with mpich2.
However I had to make the following changes in FindMPI.cmake starting at line
409. I added the fortran block and removed the quotes from the
set(MPI_LIBRARIES_WORK ...) commands. The qoutes really messed up VS linkin
I've upgraded my Java project to the latest 2.8.6 release. The only thing
missing for me is to allow components on the install commands.
I just did the following:
In UseJava.cmake (same for INSTALL_JNI_SYMLINK):
function(INSTALL_JAR _TARGET_NAME _DESTINATION)
get_property(__FILES
TA
Anyone have a clue how we can get CDash to display the valgrind report
properly. Last month I noted that with cmake 2.8.4 we could not get the
valgrind reports displayed properly, the report shows the raw compressed (or
otherwise encrypted data). We have the same problem using CMake 2.8.3, 2.8.2
t drive it
> with 2.8.4. It may take me a few days to get around to this, but I'll
> keep you posted.
>
>
> Thanks,
> David
>
>
> On Wed, Feb 23, 2011 at 4:31 PM, Allen D Byrne wrote:
> > Further results:
> >
> > 2.8.2 works with
Further results:
2.8.2 works with MemoryCheck
2.8.4 does not display even with MemoryCheck.
Here is out CTestConfig.cmake file:
#
SET (CTEST_PROJECT_NAME "HDF5.1.8")
SET (CTEST_NIGHTLY_START_TIME "20:00:00 CST")
SET (CTEST_DROP_METHOD "http")
SET (CTEST
David,
Results so far:
MemCheck with 2.8.4 - CDash does not display DynamicAnalysis data
MemoryCheck with 2.8.3 displayed same issue
The DynamicAnalysis_xxx.log files have the data.
The DynamicAnalysis_xxx.xml files have the test data, but the tags data
looks encyrpted.
We will rerun with 2.8
memory checking step executes,
> producing the dynamic analysis xml file, but not doing any other build
> steps, and not submitting it.
>
> I will investigate when this changed and why, but for now, hopefully
> that can get you going without completely revamping your scripts.
>
&
The question is: Has anyone experienced problems with Dynamic Analysis reports
on CDash with the new cmake 2.8.4?
Our execution seemed to have produced the correct logs, yet on CDash the report
is 0 (ZERO). Looking at the output of the test it just shows one long line of
characters (no words)
One more detail, this seems to be a problem only on non windows (non-make) -
windows VS machines ran the tests in correct declaration order.
Allen
> Bug reported:
> http://public.kitware.com/Bug/view.php?id=11877
>
>
> On Fri, Feb 18, 2011 at 10:30 AM, David Cole wrote:
> > Yup. ctest -N and
Just using standard gcc/gfortran on the fedora distro. It is the 64bit version?
You can grab one of our recent src tarballs from
"http://www.hdfgroup.org/ftp/HDF5/hdf5-1.8.6/";
Allen
PS. I will be out of touch for the next 36 hours
> On Thu, Feb 17, 2011 at 4:20 PM, Allen D
17, 2011 at 3:57 PM, Allen D Byrne wrote:
> >> I just installed 2.8.4 and on my fedora 14 linux box the ctest -D
> >> Experimental stage decided to run the tests in some unknown to me order.
> >> The
> >> previous versions always ran them in order of definit
I just installed 2.8.4 and on my fedora 14 linux box the ctest -D Experimental
stage decided to run the tests in some unknown to me order. The previous
versions always ran them in order of definition. Any thing I can do to get the
ordering back?
Allen
__
AL and not
CMAKE_JAVA_CLASS_OUTPUT_PATH.
For the resurce files, since the add_jar already copies the files I just added
a prefix parameter for resource files. A bit brute force but effective.
Allen
> On Saturday 05 February 2011 16:04:15 Allen D Byrne wrote:
> > Yes! That is where I
Yes! That is where I meant! (dang email doesn't read minds properly:)
BTW, I have a problem I couldn't fix. When doing an add_jar with source in the
current dir and resource files (image files) in a sub-directory, the resource
files in the jar are at the class base instead of below.
Example;
W
Hendrik,
Thanks! I put '"' around the classpath on line 233 of UseJava.cmake file and
windows was happy to compile the sources.
Andreas,
So that makes only two changes to UseJava.cmake, both in the add_jar
function;
Quotes around the ${CMAKE_JAVA_INCLUDE_PATH_FINAL} in line 233 and
ad
Andreas,
The problem is a CMake issue with how ';' is interpreted. In the
USE_JAVA.CMAKE file (lines 227-243):
> if (_JAVA_COMPILE_FILES)
> # Compile the java files and create a list of class files
> add_custom_command(
> TARGET ${_TARGET_NAME}
> COMMAN
OK, I got past my JNI issue (must set java version number) - except now I have
a problem with the classpath for building the java files only on windows! On
linux everyone's happy. On windows, using VS2008, the classpath in the project
files have been converted from ';' to spaces?
Any ideas?
A
While Marcus states what is likely the preferred way, I just added an
ADD_DEPENDENCIES(internal_target external_target) statement that seems to
always work?
Allen
> Date: Wed, 12 Jan 2011 13:23:52 -0500
> From: "Marcus D. Hanwell"
> Subject: Re: [CMake] How to have a target depend on an Exter
Andreas,
Thanks for providing these files! I have finally been able to progress on
converting our java product to cmake.
I have one suggestion:
Add to the add_java function the classpath for the target;
set(${_TARGET_NAME}_CLASSPATH
".${CMAKE_JAVA_INCLUDE_PATH_FINAL}/${_TARGET_NAME}.jar" PA
I will try it tomorrow, when the machines are idle.
Thanks,
Allen
> On 11/09/2010 08:03 AM, Brad King wrote:
> > I think this regression was an existing bug that was exposed
> > by the other fix.
>
> It was, and here is the fix to the exposed bug:
>
> http://cmake.org/gitweb?p=cmake.git;a=
stead of ""
and
DEBUG:
RELEASE:
SuppressStartupBanner="true" Preprocess="preprocessYes" instead of ""
Taking out the new AdditionalOptions=" /W1 /libs:dll /threads" allowed the
project to succeed.
Allen
> On 11/8/2010 1:15 PM, Allen D Byrne wrote:
> > This
08) generator, then rerun cmake and
> regenerate.
>
> Now you are a happy bunny and can compile using the IDE if desired - or skip
> C and use nmake
>
> JB
>
>
> From: cmake-boun...@cmake.org [mailto:cmake-boun...@cmake.org] On Behalf Of
> Allen D Byrne
> Sent: 0
Friday I upgraded my windows VS2008 and IVF111 machines to cmake 2.8.3. Over
the weekend they all failed to configure with fortran :
>-- Check for working Fortran compiler using: Visual Studio 9 2008 -- broken
>CMake Error at C:/Program Files/CMake
>2.8/share/cmake-2.8/Modules/CMakeTestFortranCo
I found the answer to my problem in bug 0009214 : -D option can't use value
with equals sign
The trick is to qualify the tag with STRING;
"TEST_ARGS:STRING=v;-param1;-parm2=5;--param3"
Everything is working for now.
Allen
> I have a test script where I pass a list of arguments in throu
I have a test script where I pass a list of arguments in through a variable -
TEST_ARGS. No problems until I need to pass in a parameter with an equal sign:
"TEST_ARGS=v;-param1;-parm2=5;--param3"
An equal sign anywhere in the list wipes the entire list of arguments and
TEST_ARGS is then
Well it turns out the problem was in the execute_process script. I had the :
OUTPUT_FILE ${TEST_OUTPUT}
ERROR_FILE ${TEST_OUTPUT}
both going to one file - Windows couldn't handle that. Changing that allowed
the command to actually execute,
NOW, I have to deal with the CRLF issue i
As stated in my previous post, I have a script, named runTest.cmake, which
works on linux but fails on Windows with:
"The process cannot access the file because it is being used by another
process"
This script calls execute_process() with an executable to capture the output
for a comparison
I have a test script that works on linux but fails to work properly on Windows.
It runs a command, captures the standard out and compares it to a reference
file.
On linux everything works 99% of the time (issue with '=' or '--' in a
parameter argument - later question). On windows, the executio
32 matches
Mail list logo
