[CMake] Target usage requirements for shared libraries

2014-07-17 Thread Walter Gray
Hi there! I've got a question I can't seem to find a good answer to: What is the idiomatic way in CMake 3.0 to handle bundling shared libraries? I've exhausted my google-fu and all of the examples seem to be from before CMake 3 was released. None of them take any advantage of 3.0's new and IMO

[CMake] find_package and INTERFACE targets in 3.0

2014-07-17 Thread Walter Gray
Hi list! I'm a big fan of the new INTERFACE targets & target usage requirements, but none of the provided Find.cmake files seem to take advantage of the new paradigm. Checking the wiki, the old Find.cmake seem to be deprecated, with most of the information on authoring new packages assuming

Re: [CMake] RPM packaging dependencies expressed in subdirectories

2014-07-17 Thread Thibault Hild
Thank you again Nils for the guidelines, now its time to experiment :). Thibault On Wed, Jul 16, 2014 at 5:40 PM, Nils Gladitz wrote: > On 16.07.2014 16:37, Thibault Hild wrote: > >> Thanks Nils for the hint. >> >> I guess you had to use Cmake constructs like: >> if("new dep" MATCHES "a regex

[CMake] About the FindCUDA.cmake module ande Separate Compilation

2014-07-17 Thread Notargiacomo Thibault
Dear All I am a user of cmake build system and its differents modules, that were very helpful in the past. But I met some specific issues with the FindCUDA.cmake module, for about a year now. Especially about the Separate Compilation feature, that never worked for me, I previously had to bypass

[CMake] Adding custom environment variables to generated Makefile

2014-07-17 Thread Carlchristian Eckert
Hello, We are migrating our project from standard Unix Makefiles to CMake. The old Makefile contained a line that cleared an environment variable like this: export CPLUS_INCLUDE_PATH= This step is necessary due to some library/path confusion in combination with 3rd party libraries on some system

Re: [CMake] Fails for Solaris

2014-07-17 Thread Chuck Atkins
Hi Amos, Although I'm ure it's *possible*, I'd be surprised if changing the compiler made any difference. A few questions would help narrow down the issue: - Which source are you building from? Is it the 3.0.0 release tgz or is it a particular git hash? - Are you doing an in-source bui