[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: cmake-build-system
From: |
Ludovic Courtès |
Subject: |
Re: cmake-build-system |
Date: |
Mon, 06 Jul 2015 00:44:31 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/24.5 (gnu/linux) |
Ricardo Wurmus <address@hidden> skribis:
> Attached is a patch for the cmake-build-system. I have encountered a
> couple of applications that failed the validate-runpath phase because
> the RUNPATH was not set properly.
>
> I found that setting BUILD_WITH_INSTALL_RPATH=TRUE helped when the
> INSTALL_RPATH was also set to contain all inputs, like this:
>
> (string-append "-DCMAKE_INSTALL_RPATH="
> (string-join (map (match-lambda
> ((name . directory)
> (string-append directory "/lib")))
> %build-inputs) ";"))
Ideally we would never have to do this because ld-wrapper is supposed to
take care of that.
This is in powertabeditor, right? What happens if you build it without
this and with VERBOSE=1 (as a ‘make’ variable) and
GUIX_LD_WRAPPER_DEBUG=1 (environment variable)? I would like to see
where the -rpath get lost or overwritten.
> The attached patch only adds BUILD_WITH_INSTALL_RPATH=TRUE. I’m not
> sure if this is sufficient, but I think it is actually a requirement.
Unfortunately CMake has complex RPATH handling and it seems to resist
us. It would be good to see, taking powertabeditor as an example, how
we can achieve better results. I don’t know whether
BUILD_WITH_INSTALL_RPATH=TRUE is necessary, but I think we must first
better understand what’s going on.
Thank you,
Ludo’.