[lld] r247475 - Implement -rpath.

[Joerg Sonnenberger via llvm-commits]

On Sun, Sep 13, 2015 at 10:09:45PM -0400, Ed Maste wrote:
> On 13 September 2015 at 19:49, Hal Finkel via llvm-commits
> <llvm-commits at lists.llvm.org> wrote:
> > ----- Original Message -----
> >> From: "Joerg Sonnenberger via llvm-commits" <llvm-commits at lists.llvm.org>
> >> To: "llvm-commits" <llvm-commits at lists.llvm.org>
> >> Sent: Sunday, September 13, 2015 4:25:26 PM
> >> Subject: Re: [lld] r247475 - Implement -rpath.
> >>
> >> On Sun, Sep 13, 2015 at 02:56:44PM -0500, Hal Finkel wrote:
> >> > Just to confirm, generating both DT_RPATH and DT_RUNPATH is a
> >> > reasonable option for NetBSD and older FreeBSD systems?
> >>
> >> NetBSD completely ignores DT_RUNPATH. But providing both wouldn't be
> >> an
> >> option for systems that moved to the broken DT_RPATH behavior.
> >
> > What do you mean? On the systems that support both, DT_RPATH is ignored when DT_RUNPATH is present, right? Or is this not always true?
> 
> As far as I know this is the case. The default toolchain on FreeBSD
> currently emits binaries with both, and DT_RPATH is indeed ignored
> when DT_RUNPATH is present.
> 
> Joerg, are you aware of systems that "moved to" the silly (but
> documented) DT_RPATH behaviour? My understanding is that a number of
> systems implemented the silly behaviour from the beginning, and later
> introduced DT_RUNPATH to fix that.

I don't see anything in the standards that would recomment ignoring the
DT_RPATH, but I might be wrong. 

Joerg