The discussion about the libraries is getting more and more confusing …
With these settings all works fine for me without bothering
To export DYLD_LIBRARY_PATH or LD_LIBRARY_PATH
# do not skip the full RPATH for the build tree
SET( CMAKE_SKIP_BUILD_RPATH FALSE)
# when building, don't use the install RPATH
# only later on when installing
SET( CMAKE_BUILD_WITH_INSTALL_RPATH FALSE)
if( APPLE )
SET( CMAKE_INSTALL_RPATH "@executable_path/../${_lib_dir}")
else()
SET( CMAKE_INSTALL_RPATH "$ORIGIN/../${_lib_dir}")
endif()
# add the automatically determined parts of the RPATH
# which point to directories outside the build tree to the install RPATH
SET( CMAKE_INSTALL_RPATH_USE_LINK_PATH TRUE)
Without the RPATH setting the exports are required
( tested on El Capitan, high Sierra, Mojave and Fedora 29 )
I usually install to /opt/ooRexx and add /opt/ooRexx/bin to the PATH
Until a few revision ago I could move the the installed tree anywhere with any
name
And the test suite would run successfully
Now something has changed, my scripts run
But the test suite fails with an abort trap 6 in the bif/DATE.testGroup
E
> On 2 Jan 2019, at 23:13, René Jansen <rvjan...@xs4all.nl> wrote:
>
> Hi P.O.,
>
> what Erich meant is that for some time now the *environment variable*
> containing the dyld library path is not inherited in subshells. We have to
> use dyld as that is the way shared libraries on macOS work. We have to work
> at making sure this environment variable is not needed anymore - and that it
> will run on systems with SIP enabled, which would be most user systems. This
> is possible using the install_name_tool, although for librexxutil a solution
> must be found, for which, if I am not mistaken, Rick just opened a sandbox
> directory. Did you try to run without either LD_LIBRARY_PATH or
> DYLD_LIBRARY_PATH?
>
> René.
>
>> On 2 Jan 2019, at 18:00, P.O. Jonsson <oor...@jonases.se
>> <mailto:oor...@jonases.se>> wrote:
>>
>> A small remark:
>>
>> Erich has found out that the dyld stuff will not work since it is not
>> inherited. A rexx program calling another rexx program (example: test cases)
>> fail. He can put it in more technical terms I am sure, but my take on it is
>> that this is a dead end on Darwin/macOS
>>
>> Hälsningar/Regards/Grüsse,
>> P.O. Jonsson
>> oor...@jonases.se <mailto:oor...@jonases.se>
>>
>>
>
> _______________________________________________
> Oorexx-devel mailing list
> Oorexx-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/oorexx-devel
_______________________________________________
Oorexx-devel mailing list
Oorexx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/oorexx-devel
