Re: pkg/46983: openoffice3 fails to start properly

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

The following reply was made to PR pkg/46983; it has been noted by GNATS.

From: Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: tsutsui%ceres.dti.ne.jp@localhost
Subject: Re: pkg/46983: openoffice3 fails to start properly
Date: Tue, 12 Mar 2013 23:59:13 +0900

 joerg@ wrote:
 
 >  On Mon, Mar 11, 2013 at 06:10:04PM +0000, Izumi Tsutsui wrote:
 >  >  joerg@ wrote:
 >  >  
 >  >  >  On Mon, Mar 11, 2013 at 03:40:13PM +0000, Izumi Tsutsui wrote:
 >  >  >  >  - revert patch-ch rev 1.3 and patch-cs rev 1.4 to pass $ORIGIN for 
 > -rpath
 >  >  >  
 >  >  >  I disagree with this part. We don't support $ORIGIN e.g. on NetBSD and
 >  >  >  it shouldn't be used in pkgsrc.
 >  >  
 >  >  If you can't build working binaries, such policy without analysys
 >  >  doesn't make sense.
 >  
 >  $ORIGIN simply doesn't work, as such it can't be correct. I haven't done
 >  a full analysis to understand why the patch doesn't work or changing it
 >  back seems to fix the issue. What does end up in the final binary in
 >  both cases (with and without the patches)?
 
 You can simply test it:
 # pkg_add 
http://ftp.netbsd.org/pub/pkgsrc/packages/NetBSD/i386/6.0_2012Q4/All/openoffice3
 % soffice
 
 With the original patches (i.e. package binaries currently available),
 soffice hangs (or dumps core, seems to depend on environments)
 right after startup splash window shows all modules are loaded.
 
 With the reverted patches (without -Wl,-rpath,${PREFIX}/lib),
 soffice just works and I could create slides for event last weekend.
 
 Oshima-san also reports that removing whole
 "LINKFLAGSRUNPATH*=-Wl,-rpath,${PREFIX}/lib"
 line also makes soffice binary work.
 Does it make sense for you?
 
 Probably unconditional "-rpath,/usr/pkg/lib" makes some openoffice3
 internal binaries loading wrong libs.
 
 ---
 Izumi Tsutsui