Hello,

On šeštadienis 18 Vasaris 2012 20:27:37 Sam Morris wrote:
> I hate to bikeshed, but I must jump in here. The support_cppflags patch
> changes how CMake works WRT its documentation, its behaviour on other
> platforms and even behaviour on other Linux distributions.
> 
> Changing how CMake works in this way without upstream's agreement to
> follow suit will only cause confusion for end-users, who will not expect
> this undocumented divergence in CMake's behaviour.

I fail to understand why a single environment variable is such a big deal?  
Anyway, I don't have hard feelings about it and I am going to be reasonable. 
If upstream decides that respecting CPPFLAGS is an absolute no-go, let it be, 
I will revert the patch.

> I understand that you want to make the job of packagers easier, but IMHO
> it's their job to understand how their build system works, including how
> to correctly pass various common build flags into the build system.

Btw, my main rationalle was not that I wanted to make job easier for anybody. 
It's just that I believe it's the right thing to do. All this time I thought 
CPPFLAGS was an alias for CXXFLAGS but I was wrong. Apparently, Brad had a 
similar reaction at first [1] so maybe upstream will change their mind.

[1] http://www.cmake.org/Bug/view.php?id=12928#c28452

> Some
> less intrusive ways to make this easier for them might be:
> 
>       * Improve CMake's documentation
>       * Providing a quick "cheat sheet for Debian package maintainers"
>         in a README.Debian file
>       * Patch CMake to output a warning message when CPPFLAGS is set in
>         the environment
> 
> Hence please consider reverting this patch, and working with upstream to
> find a solution that works for them as well as us. :)

Thank you for suggestions. They may be useful if CPPFLAGS plan does not work 
out. Until then, I will push towards proper CPPFLAGS support. Unfortunately, 
it does not help that support_cppflags.diff patch is not upstreamable as it 
is, but I don't have time to develop anything better at the moment.

-- 
Modestas Vainius <mo...@debian.org>

Attachment: signature.asc
Description: This is a digitally signed message part.

Reply via email to