Re: [CMake] Disabling INSTALL target for subdirectory

2017-02-15 Thread Craig Scott
You can use the COMPONENT option to install() for your own project and then select a component-based package if you are packaging with CPack. We currently do this with builds that incorporate other projects with add_subdirectory() similar to your case. If you are using install as a build target (i

Re: [CMake] Disabling INSTALL target for subdirectory

2017-02-15 Thread Konstantin Podsvirov
Hello all!10:00, 16 February 2017 г., Milan Ziegler :Hi,we are using the CMake install mechanism "install(...)" to copy allrelevant files and target binaries of the product into the CMake installprefix folder, for further packaging. So far so good.We are also using several thirdparty libraries, hos

[CMake] Disabling INSTALL target for subdirectory

2017-02-15 Thread Milan Ziegler
Hi, we are using the CMake install mechanism "install(...)" to copy all relevant files and target binaries of the product into the CMake install prefix folder, for further packaging. So far so good. We are also using several thirdparty libraries, hosting them inside our own repository and buildin

[CMake] Errors from CDash

2017-02-15 Thread Thompson, KT
When I run php5 autoRemoveBuilds.php all for my cdash 2.2.3 installation, I see the following errors (and old builds are not removed): removing builds for all -- removing old buildids for projectid: 2 PHP Warning: mysql_fetch_array() expects parameter 1 to be resource, boolean g8 PHP Warning

Re: [CMake] cmake 3.7.2 centos 7 :: kwsys.testSystemTools (Failed)

2017-02-15 Thread Chuck Atkins
Hi Adrian, So, would this imply that cmake is not supported on EL7 because gcc does > not support cxx11 ? > That's not really what's happening. CMake is testing the C compiler to determine whether or not it supports C11 features and the C++ compiler to see if it supports C++11. The error messag

Re: [CMake] set_if_not_set

2017-02-15 Thread Konstantin Tokarev
15.02.2017, 20:08, "Bill Newcomb" : > My top-level CMakeLists.txt files always end up with a lot of > > if(NOT_DEFINED foo) >  set(foo bar) > endif() > > things to allow people to pass interesting things on the command line or > to write their own "set this and that" CMakeLists.txt and includ

Re: [CMake] Problem with override file ...

2017-02-15 Thread Michael Burns via CMake
I found that changing the add_definitions lines to set(CMAKE_C_FLAGS_INIT ...) or set(CMAKE_CXX_FLAGS_INIT ...) resolved the issue. I have not found documentation anywhere that says "this is how it must be done". So, is this the expected behavior when using add_defin

[CMake] set_if_not_set

2017-02-15 Thread Bill Newcomb
My top-level CMakeLists.txt files always end up with a lot of if(NOT_DEFINED foo) set(foo bar) endif() things to allow people to pass interesting things on the command line or to write their own "set this and that" CMakeLists.txt and include the top-level one. It would improve readability

Re: [CMake] Intel C Generator for Cmake?

2017-02-15 Thread J Decker
Toolchain is probably what you want http://stackoverflow.com/questions/9129233/recommended-ways-to-use-cmake-with-icc-via-configuration-options https://cmake.org/Wiki/CMake_FAQ#How_do_I_use_a_different_compiler.3F "cmake intel toochain" search. On Wed, Feb 15, 2017 at 6:42 AM, Thompson, KT

Re: [CMake] import external library

2017-02-15 Thread Florent Castelli
You can either unpack the target at generation time (recommended). Or you could create an INTERFACE target instead. The first solution is recommended as I believe it's important to have access to all the sources (for IDEs for example) before actually starting the build process. /Florent On 1

Re: [CMake] Intel C Generator for Cmake?

2017-02-15 Thread Thompson, KT
Tony, I don’t have access to the Intel compiler on Windows to try this out. However, in an old post I found, it was suggested that you can use cmake’s “-T” option to select the Intel compiler when the generator is “Visual Studio”. For example: -T"Intel C++ Compiler XE 13.0" Ref: https://softw

Re: [CMake] Potential bug: Having executables named "test" causes Ninja generator to generate ambiguous rules...

2017-02-15 Thread Nils Gladitz
On 15.02.2017 15:16, Malfettone, Kris wrote: Just noticed a problem in my example. In my attempt to simplify my example I moved both outputs into the same folder which in my opinion is invalid because both output file names are the same. However, the same basic setup does reproduce the prob

Re: [CMake] Potential bug: Having executables named "test" causes Ninja generator to generate ambiguous rules...

2017-02-15 Thread Malfettone, Kris
Just noticed a problem in my example. In my attempt to simplify my example I moved both outputs into the same folder which in my opinion is invalid because both output file names are the same. However, the same basic setup does reproduce the problem if you move the two targets into their own s

Re: [CMake] Potential bug: Having executables named "test" causes Ninja generator to generate ambiguous rules...

2017-02-15 Thread Malfettone, Kris
So I was able to reproduce the problem with this very simple CMakeLists.txt setup. I cannot attach source files due to my company's mail server automatically stripping them out attachments with code but here is all you would need to reproduce the problem: Single CMakeLists.txt with the followi

Re: [CMake] Potential bug: Having executables named "test" causes Ninja generator to generate ambiguous rules...

2017-02-15 Thread Malfettone, Kris
Good to know. Thanks for the quick response. Though I do not believe I am violating that particular rule here. That is why I do not receive either a warning or error from CMake in my output. The only warning I see is from ninja: ninja: warning: multiple rules generate MD/MDF/Formats/SIG/BCP/v

Re: [CMake] Intel C Generator for Cmake?

2017-02-15 Thread Tony Garratt
Re: my previous email. To use the Intel C compilers I just can modify the VS settings. So I have a way forward. Regards, Tony On Wed, Feb 15, 2017 at 12:25 PM, Petr Kmoch wrote: > Hi Tony, > > generators are for different *buildsystems*: a generator for Makefiles, a > generator for Visual Studi

[CMake] import external library

2017-02-15 Thread Wagner Martin
Hi @all, I’ve tried importing an external library like that: stackoverflow.com/questions/31274577 I have the same use case, except that my custom target does not call make but does unpack a tar archive containing headers and .a. My problem is that the step “ # specify where the library is and w

Re: [CMake] Intel C Generator for Cmake?

2017-02-15 Thread Tony Garratt
I tried setting the FC and CC variables and yet cmake still wants to use the Microsoft compiler (BTW what I am building is the CGNS - the CFD common file format). Is there something I am missing? CGNS_Build>set CC=icc CGNS_Build>set FC=ifort CGNS_Build>set CC CC=icc CGNS_Build>set FC FC=ifort

Re: [CMake] Intel C Generator for Cmake?

2017-02-15 Thread Tony Garratt
Thank you so much Petr. I am new to cmake so did quite understand how everything fits together. I will give your suggestion a try. Regards, Tony On Wed, Feb 15, 2017 at 12:25 PM, Petr Kmoch wrote: > Hi Tony, > > generators are for different *buildsystems*: a generator for Makefiles, a > gen

Re: [CMake] Intel C Generator for Cmake?

2017-02-15 Thread Petr Kmoch
Hi Tony, generators are for different *buildsystems*: a generator for Makefiles, a generator for Visual Studio solutions, a generator for Ninja files, a generator for Eclipse projects etc. Intel C and Intel Fortran are compilers, not buildsystems. You should be able to use them with any generator

[CMake] Intel C Generator for Cmake?

2017-02-15 Thread Tony Garratt
I am having to switch to cmake from autotools for a third party tool I want to build. We use Intel C and Intel Fortran on Windows. I see no mention of an Intel C generator for cmake. I presume this means there is not one? If so, can I create my own Intel C generator please? Rgeards, Tony -- *Dr

Re: [CMake] Potential bug: Having executables named "test" causes Ninja generator to generate ambiguous rules...

2017-02-15 Thread Nils Gladitz
On 14.02.2017 22:53, Malfettone, Kris wrote: If so, are executables named “test” no longer supported? Am I missing something or wrong in some way? Issues existed before 3.0 but since 3.0 "test" and similarly problematic target names have been explicitly reserved. Any CMake Version >= 3.0 w