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
make zsh and bash completion installation more consistent #832
Comments
The current hack is using the It could be possible to honor I think there will be no easy way to honor both Elektras Honoring the cmake and/or pkg-config paths means we will get the appropriate paths that where generated on package generation and that will get sourced or loaded automatically. e.g. homebrews pkg-config file of bash-completion will expose the correct folder for it to be sourced. |
With hack I mainly mean setting Yes, and there is no guarantee that a modified path from pkg-config will work. So you are right, its better to take it as is. Maybe we can honor What I do not like about this fallback is that in one case it would be a |
In wheezy the pkg-config variable bash-completion completionsdir could not be found. Because we use the same debian installation rules, the different fallback caused wheezy packaging to fail: http://community.markus-raab.org:8080/job/elektra-git-buildpackage-wheezy/lastFailedBuild/consoleFull I now changed the fallback to Still there are left-overs: for zsh no pkg-config path is searched for. Does one exist? I am not really in favour of |
CMAKE_INSTALL_PREFIX
Do you have the
As far as i can tell there is no pkgconfig file installed with zsh. We will have to suffice with a hardcoded path or find out if the path is exposed elsewhere somehow.
I will look into refactoring the cmake instructions into making the variable INTERNAL. |
And just to be clear. The variable you see in the cmake cache is not |
And making this statement. It is just plain wrong for variables we do not expose tough cache, as they are not meant to be user settable. We are not talking about opinions here, but plain and clearly about technical facts. In addition for modern systems with up to date packages An it is not a hack to just use the same variable name for older systems that to not set this variable themselves but a clean way to provide backward compatibility for 4 year old software. I already explained at the Trough doing this we are repeating the same discourse again and again. And the fault is not on my side but on yours in this case. As like i said just ignore my arguments and repeat yours without adding any facts or arguments for your statement which is pretty much like a personal insult. Your implicit statement by doing so is that you do not even see worth in answering to them. |
Yes, CACHE and INTERNAL also seems overkill for me, but i did not want to annoy you for such irrelevant stuff. Unfortunately CMake does not define a clear interface for the outside world, every variable can be manipulated. Thus basically every variable (that has an effect which can be exploited) is a question of documentation, legacy and complications. A usual compromise is to not care about what is not documented. So we simply remove the information that Another issue is the inconsistency between bash and zsh, so if the user can manipulate one, the other should be changable, too. So what about having no |
This has been done #841 . The idea to check for |
For me I had to use:
|
@edusantana thank you for your report! I assume without @sanssecours is it possible with homebrew to move some files to different locations if a user runs brew? |
Since Homebrew uses a Ruby DSL I would assume the answer is yes. I think the problem is that the CMake variable libelektra/cmake/ElektraCache.cmake Lines 238 to 242 in 09ef913
. I just pushed a commit that should hopefully fix the problem. |
Is it possible that pkg-config respects
CMAKE_INSTALL_PREFIX
? Then we could avoidBASH_COMPLETION_COMPLETIONSDIR
hack.The text was updated successfully, but these errors were encountered: