On Apr 21, 2010, at 3:09 AM, IOhannes m zmoelnig wrote:
On 2010-04-20 21:04, Hans-Christoph Steiner wrote:
is there a reason why /sw is not in the PATH?
Some build systems are not happy with the tools provided in /sw, so
its
good to make them optional. You can add /sw to the PATH within the
configure.in, then people will inherit that setting when building on
their own machine. I think its just a matter of adding this to the
top:
PATH="/sw:${PATH}"
.hc
configure.in is probably a bad choice, as this file has to be
processed
by binaries found in /sw/bin, rather than has to call binaries in /
sw/bin.
i guess you meant autogen.sh
anyhow, the problem with that is, that autogen.sh is a generic, and i
don't feel like adding every single path anyone might happen to have
on
their machine :-)
i guess it should be set more on a per-machine basis: after all the
sysad of a certain host will always know best, which paths on the
machine are relevant, whereas the choices of an upstream developer
like
me will always be bad.
i will probably add /sw/bin to the path in the master build
script/Makefile, if this is ok for you.
I think it should be targeted to where its needed. This sounds too
broad, tho I am not sure which Makefile you mean. Why not just do a
test against uname to set it in autogen.sh. Something like:
if [ `uname -s` -eq "Darwin" ]; then
PATH="/sw:${PATH}"
fi
.hc
----------------------------------------------------------------------------
If you are not part of the solution, you are part of the problem.
_______________________________________________
GEM-dev mailing list
GEM-dev@iem.at
http://lists.puredata.info/listinfo/gem-dev
