[GRASS5] GRASS 6.0.1RC1 released
Markus Neteler
neteler at itc.it
Fri Aug 5 03:34:48 EDT 2005
On Thu, Aug 04, 2005 at 10:05:57PM +0200, Markus Neteler wrote:
> (cc ML for the archive)
>
> On Thu, Aug 04, 2005 at 11:59:09AM -0500, William Kyngesburye wrote:
> > On Aug 4, 2005, at 10:42 AM, Markus Neteler wrote:
> > >On Thu, Aug 04, 2005 at 09:16:03AM -0500, William K wrote:
> > >
> > >>bug #2544
> > >>
> > >
> > >OK.
> > >
> > >Find attached a modified
> > > configure
> > >to test. Please report, then I'll update in CVS.
> > >
> > it works
>
> If there are no objections, I'll update aclocal.m4 and
> configure in CVS tomorrow (for RC2 and 6.1-CVS).
Forgot to attach the patch.
Now attached.
Markus
> > >Are we sure that it doesn't break things for other
> > >Mac versions?
> > >
> > I've used that form of SHLIB_LD in 10.3 and 10.4. And a slightly
> > different form when I was on 10.2, but that was because of changes in
> > the GRASS make variables for version and install dir.
> >
> > It's the standard way of building libraries for all versions of Mac
> > OS X (the install_name and xxx_version parts). The flat_namespace
> > and fno_common are pretty normal for Mac OS X UNIX-ported software
> > and have pretty much the same effect (as far as I can tell) as the
> > single_module option that was there before, but I never see single-
> > module used coming out of libtool-enabled source. (now that I said
> > that, I'll probably find something that uses single-module and find
> > out why ^_^)
> >
> >
> > I'm not sure about the Rhapsody part. Libtool handles rhapsody and
> > darwin1 the same for some stuff, but darwin* is handled alone (Mac OS
> > 10.4 is darwin 8.x, 10.3 is darwin 7.x, 10.2 is 6.x), so there is
> > much that is darwin-specific and only a little that is shared between
> > darwin and rhapsody. This is just from searching for 'rhapsody' and
> > 'darwin' in ltmain.sh, I really don't know much about rhapsody or the
> > early days of darwin.
> >
> >
> > One note: if I could figure out a reliable method for getting a
> > release version as integer, I would add that to the current_version
> > option. But GRASS_VERSION_RELEASE gets RC and CVS at various
> > stages.
>
> Ah, now I understand. Yes, we have "1RC1" currently as RELEASE.
> It doesn't harm for other platforms (apparently).
>
> > Does it matter whether the versions are added indirectly in
> > configure, like the way I have it now (\${GRASS_VERSION_MAJOR}, etc)
> > or directly in configure (remove the backslash so it expands
> > immediately)? The darwin section could have a little extra parsing
> > before setting SHLIB_LD to say: if it's CVS, make the release '0', if
> > it's a RC, extract the target release number (not the RC number),
> > otherwise use the number as is.
>
> Discussion is open!
> My intention was simply to get the RC state into the VERSION to
> avoid confusion. Probably we need a modified scheme for that.
>
> Markus
>
>
-------------- next part --------------
Index: aclocal.m4
===================================================================
RCS file: /grassrepository/grass6/aclocal.m4,v
retrieving revision 1.16
diff -u -r1.16 aclocal.m4
--- aclocal.m4 6 Oct 2004 14:50:55 -0000 1.16
+++ aclocal.m4 5 Aug 2005 07:33:45 -0000
@@ -922,11 +922,8 @@
esac
;;
Rhapsody-*|Darwin-*)
- SHLIB_CFLAGS=""
- SHLIB_LD="cc -dynamiclib -single_module \${LDFLAGS}"
-# GRASS_SHLIB_LD_EXTRAS="-compatibility_version ${LIB_VER} -current_version \${LIB_VER} -install_name \${DYLIB_INSTALL_DIR}/\${GRASS_LIB_FILE} -prebind -seg1addr 0xa000000"
-#check if we need -U,_cuserid:
-# SHLIB_LD="cc -dynamiclib -Wl,-flat_namespace,-U,_cuserid -undefined suppress \${LDFLAGS}"
+ SHLIB_CFLAGS="-fno-common"
+ SHLIB_LD="cc -dynamiclib -flat_namespace -compatibility_version \${GRASS_VERSION_MAJOR}.\${GRASS_VERSION_MINOR} -current_version \${GRASS_VERSION_MAJOR}.\${GRASS_VERSION_MINOR} -install_name \${INST_DIR}/lib/lib\${LIB_NAME}\${SHLIB_SUFFIX}"
GRASS_SHLIB_LD_EXTRAS=""
SHLIB_LD_LIBS='${LIBS}'
SHLIB_SUFFIX=".dylib"
More information about the grass-dev
mailing list