Re: [PATCH] Fixing the subtle python PLIST breakage on installing with an older version present

["Dr. Thomas Orgis" <thomas.orgis%uni-hamburg.de@localhost>]

Am Fri, 5 Apr 2019 13:07:57 -0400
schrieb Greg Troxel <gdt%lexort.com@localhost>: 

> Thanks for explaining; that makes sense.  I generally prefer to have an
> upstream bug filed 

To clarify/repeat: This is not just an upstream bug. Upstream has
machinery in place to force --enable-new-dtags (RUNPATH usage, hence
LD_LIBRARY_PATH effective during install phase) regardless of the
default linker setup on the sytem. Pkgsrc carries a patch that
_disables_ that. Hence my patch is needed to fix the install in the
case of RPATH semantics (LD_LIBRARY_PATH not in effect) on the host
system.

So this really fixes a bug in pkgsrc with the incomplete patching.

The question about an upstream bug lies with the initial decision to
avoid python overriding the linker setup. We might discuss with them if
they want to replace the LD_LIBRARY_PATH usage with LD_PRELOAD and get
rid of the --enable-new-dtags, so, integrating both patches from pkgsrc.

But since pkgsrc breaks things itself currently, the fix for that
doesn't have to wait, IMHO.


Alrighty then,

Thomas

-- 
Dr. Thomas Orgis
HPC @ Universität Hamburg