Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

CMake: 2.8.12.2+ #92

Closed
ax3l opened this issue Jul 29, 2015 · 4 comments
Closed

CMake: 2.8.12.2+ #92

ax3l opened this issue Jul 29, 2015 · 4 comments

Comments

@ax3l
Copy link
Member

ax3l commented Jul 29, 2015

Since the default CMake module does now require cmake 2.8.12.2+, one could push the cmake versions for the project itself (cmake_minimum_required) and the documentation of required software itself to the same version to avoid confusion for users.

(Aka "I installed in successfully but my project does not work with it..." -> let's force a CMake version that will hold the full way to a project.)

@ax3l ax3l added the feature label Jul 29, 2015
@ax3l ax3l added this to the 2.2.0crp: More New features milestone Jul 29, 2015
@ax3l ax3l assigned ax3l and unassigned slizzered Jul 29, 2015
ax3l added a commit to ax3l/scatteralloc that referenced this issue Jul 29, 2015
Require CMake 2.8.12.2+ for the whole project to
ensure the full build chain until a project that
is using the FindmallocMC.cmake module will pass.

Dependency increase in
  ComputationalRadiationPhysics/cmake-modules#8
@slizzered
Copy link
Contributor

Sounds like a good idea to improve usability 👍

Just a remark: We could think about using a configFile instead of a FindModule in the future, since mallocMC already uses CMake. It looks to me like the concept is not 100% polished yet for non-Windows platforms (see), but might be a nice alternative because we can ship the find-script directly with the mallocMC package.

slizzered added a commit that referenced this issue Jul 30, 2015
Close #92: Update CMake Version to 2.8.12.2+
@ax3l
Copy link
Member Author

ax3l commented Jul 30, 2015

regarding configFiles: afaik and what I got as feedback in the mailing lists, modules are kind of deprecated: also the mainline does not prefer accepting new modules for software.

I am still not 100% overwhelmed by cmake config files, since they do not solve the chicken-egg problem of finding the install in the first place (at least how I understood them). because in the end, one still needs to provide a hint (e.g., via CMAKE_PREFIX_PATH) to find the already installed config file.
Nevertheless, I guess they make multi-platform support way easier by relying on that hint and taking the system default from the system cmake instead of looking up all system paths for all systems in a module.

that said, I would welcome to see a configFile implementation (and to be enlightened about the deeper concept).

@slizzered
Copy link
Contributor

when you say "modules", are you talking about FindModule stuff like the one we worked on?

The problem with finding the config files for a project appears to be somewhat solved under Windows (canonical place in the windows registry), but there is no good solution available for Linux yet.** So for now, configFiles are not the ideal solution that they might become in the future.


** on a per-user basis, configs can be placed under ~/.cmake/packages, but sadly there is no such default folder on system-level

@ax3l
Copy link
Member Author

ax3l commented Jul 30, 2015

when you say "modules", are you talking about FindModule stuff like the one we worked on?

yes. cmake find_package modules, stored under $CMAKE_SHARED/Modules/. (e.g., /usr/share/cmake-3.0/Modules/)

there is no good solution available for Linux yet.**

yes, that place in $HOME was what I saw, too. nevertheless, I remember that setting up a useful environment for a cluster / module system is kind-of painful/not yet possible.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants